Original Research

About 1 in 10 Americans have diabetes mellitus (DM), of which about 90% to 95% are diagnosed with type 2 DM (T2DM) and veterans are disproportionately affected.^1,2 About 25% enrolled in the Veterans Health Administration (VHA) have T2DM, which has been attributed to exposure to herbicides (eg, Agent Orange), decreased physical activity resulting from past physical strain, chronic pain, and other physical limitations resulting from military service.^3-5

Pharmacologic management of DM is guided by the effectiveness of lifestyle interventions and comorbid diagnoses. Current DM management guidelines recommend patients with comorbid atherosclerotic cardiovascular disease, chronic kidney disease, or congestive heart failure receive first-line diabetes therapy with a sodium-glucose cotransporter-2 (SGLT-2) inhibitor or glucagon-like peptide-1 receptor (GLP-1) agonist.

Metformin remains a first-line pharmacologic option for the treatment of T2DM with the goal of achieving glycemic management when lifestyle interventions are insufficient.^6,7 Newer antihyperglycemic therapies have been studied as adjunct therapy to metformin. However, there is limited literature comparing metformin directly to other medication classes for the treatment of T2DM.^8-13 A systematic review of treatment-naive patients found HbA_1c reductions were similar whether patients received metformin vs an SGLT-2 inhibitor, GLP-1 agonist, sulfonylurea, or thiazolidinedione monotherapy.¹⁰ The analysis found dipeptidyl-peptidase-4 (DPP-4) inhibitors had inferior HbA_1c reduction compared to metformin.¹⁰ A Japanese systematic review compared metformin to thiazolidinediones, sulfonylureas, glinides, DPP-4 inhibitors, α-glucosidase inhibitors, or SGLT-2 inhibitors for ≥ 12 weeks but found no statistically significant differences in HbA_1c reduction.¹¹ The AWARD-3 trial compared once-weekly dulaglutide to metformin in treatment-experienced patients and found greater improvement in HbA_1c and achievement of HbA_1c goal with dulaglutide.¹³ While these studies show some comparisons of metformin to alternative pharmacologic therapy, researchers have not looked at what happens to patients’ HbA_1c levels when an event, such as a recall, prompts a rapid change to a different antihyperglycemic agent.

On May 28, 2020, the US Food and Drug Administration (FDA) asked 5 pharmaceutical companies to voluntarily recall certain formulations of metformin. This action was taken when FDA testing revealed unacceptably high levels of N-Nitrosodimethylamine, a probable carcinogen.¹⁴ This FDA recall of metformin extended-release, referred to as metformin sustained-action (SA) within the VHA electronic medication file but the same type of formulation, prompted clinicians to revisit and revise the pharmacologic regimens of patients taking the drug. Because of the paucity of head-to-head trials comparing metformin with newer alternative antihyperglycemic therapies, the effect of treatment change was unknown. In response, we aimed to establish a data registry within Veterans Integrated Service Network (VISN) 6.

Registry Development

The VISN 6 registry was established to gather long-term, observational, head-to-head data that would allow review of HbA_1c levels before and after the recall, as well as HbA_1c levels broken down by the agent that patients were switched to after the recall. Another goal was to explore prescribing trends following the recall.

Data Access Request Tracker approval was obtained and a US Department of Veterans Affairs (VA) Information and Computing Infrastructure workspace was developed to host the registry data. The research cohort was established from this data, and the registry framework was finalized using Structured Query Language (SQL). The SQL coding allows for recurring data updates for all individuals within the cohort including date of birth, race, sex, ethnicity, VHA facility visited, weight, body mass index, HbA_1c level, creatinine clearance, serum creatinine, antihyperglycemic medication prescriptions, adverse drug reactions, medication adherence (as defined by ≥ 80% refill history), and hospitalizations related to diabetes. For the purposes of this initial analysis, registry data included demographics, diabetes medications, and HbA_1c results.

METHODS

This study was a concurrent, observational, multicenter, registry-based study conducted at the Western North Carolina VA Health Care System (WNCVAHCS). The study was approved by the WNCVAHCS institutional review board and research and development committees.

All patients aged ≥ 18 years with T2DM and receiving health care from VISN 6 facilities who had an active metformin SA prescription on, and 1 year prior to, June 1, 2020 (the initial date VHA began implementing the FDA metformin recall) were entered into the registry. Data from 1 year prior were collected to provide a baseline. Veterans were excluded if they received metformin SA for any indication other than T2DM, there was no pre- or postrecall HbA_1c measurement, or death. We included 15,594 VISN 6 veterans.

Registry data were analyzed to determine whether a significant change in HbA_1c level occurred after the metformin recall and in response to alternative agents being prescribed. Data from veterans who met all inclusion criteria were assessed during the year before and after June 1, 2020. Demographic data were analyzed using frequency and descriptive statistics. The Shapiro Wilkes test was performed, and data were found to be nonparametric; therefore the Wilcoxon signed-rank test was used to evaluate the hypothesis that HbA_1c levels were not impacted by the recall.

Our sample size allowed us to create exact matched pairs of 9130 individuals and utilize rank-biserial correlation to establish effect size. Following this initial population-level test, we constructed 2 models. The first, a linear mixed-effects model, focused solely on the interaction effects between the pre- and postrecall periods and various medication classes on HbA_1c levels. Second, we constructed a random-effects within-between model (REWB) to evaluate the impact ofmedication classes and demographic variables. Statistical significance was measured at P < .05 with conservative power at .90. The effect size was set to 1.0, reflecting a minimum clinically important difference. Literature establishes 0.5 as a modest level of HbA_1c improvement and 1.0 as a clinically significant improvement.

RESULTS

Preliminary results included 15,594 veterans who received a metformin SA prescription as of June 1, 2020 from VISN 6 facilities; 15,392 veterans had a drug exposure end on June 1, 2020, indicating their standard therapy of metformin SA was discontinued following the FDA recall. Two hundred and two veterans were excluded from the registry because they continued to receive metformin SA from existing stock at a VISN6 facility. After identifying veterans with data for 1 year prior (June 1, 2019) to the index date and 1 year after (June 1, 2021) the study population was adjusted to 9130. The population was predominantly males aged> 60 years. Roughly 55% of the registry identified as White and nearly 40% as Black, and 2% indentified as Hispanic (Table 1).

Wilcoxon Signed-Rank Test

We created exact pairs by iterating the data and finding the closest measurements for each patient before and after the recall. This has the advantage over averaging a patient’s pre- and post-HbA_1c levels, as it allows for a rank-biserial correlation. Using the nonparametric Wilcoxon signed-rank test, V was 20,100,707 (P < .001), indicating a significant effect. The –0.29 rank-biserial correlation, which was computed to assess the effect size of the recall, suggests that the median HbA_1c level was lower postrecall vs prerecall. The magnitude of the correlation suggests a moderate effect size, and while the recall had a noticeable impact at a population level, it was not extreme (Table 2).

Linear Mixed-Effects Model

The binary variable for medication class exposure suggests the use of a logit link function for binary outcomes within the multilevel modeling framework.¹⁵ We employed a linear mixed-effects model to investigate the impact that switching from metformin SA to other T2DM medications had on HbA_1c levels. The model was adjusted for patient-specific random effects and included interaction terms between the recall period (before and after) and the usage of different T2DM medications.

Model Fit and Random Effects

The model demonstrated a residual maximum likelihood criterion of 100,219.7, indicating its fit to the data. Notably, the random effects analysis revealed a substantial variability in baseline HbA_1c levels across patients (SD, 0.94), highlighting the importance of individual differences in DM management. Medication classes with zero or near-zero exposure rate were removed. Due to demographic homogeneity, the model did not converge on demographic variables. Veterans were taking a mean of 1.8 T2DM medications and metformin SA was most common (Table 3).

During the postrecall period, metformin SA remained the most frequently prescribed medication class. This may be attributed to the existence of multiple manufacturers of metformin SA, some of which may not have been impacted by the recall. VISN 6 medical centers could have sought metformin SA outside of the usual procurement path following the recall.

Complex Random Effects Model

We employed a complex REWB model that evaluated the impact of medication classes on HbA_1c levels, accounting for both within and between subject effects of these medications, along with demographic variables (sex, race, and ethnicity) (eAppendix). This model accounts for individual-level changes over time (within-patient effects) and between groups of patients (between-patient effects). This is a more comprehensive model aimed at understanding the broader impact of medications on HbA_1c levels across diverse patient groups.

Most demographic categories did not demonstrate significant effects in this model. Black individuals experienced a slight increase in HbA_1c levels compared with other racial categories that was not statistically significant. However, this model confirms the findings from the linear mixed-effects model that GLP-1 agonists showed a substantial decrease in HbA_1c levels within patients (coefficient –0.5; 95% CI, –0.56 to –0.44; P < .001) and a moderate increase between patients (coefficient, 0.21; 95% CI, 0.12-0.31; P < .001). Additionally, SGLT-2 inhibitors had a notable decrease within patients (coefficient, –0.27; 95% CI, –0.32 to –0.22; P < .001).Another notable finding with our REWB model is insulin usage was associated with high HbA_1c levels, but only between subjects. Long-acting insulin (coefficient, 0.96; 95% CI, 0.90-1.01; P <. 001) and mixed insulin (coefficient, 1.09; 95% CI, 0.94-1.24; P < .001) both displayed marked increases between patients, suggesting future analysis may benefit from stratifying across insulin users and nonusers.

Fixed Effect Analysis

The fixed effects analysis yielded several notable findings. The intercept, representing the mean baseline HbA_1c level, was estimated at 7.8% (58 mmol/mol). The coefficient for the period (postrecall) was not statistically significant, indicating no overall change in HbA_1c levels from before to after the recall when specific medication classes were not considered (Table 4). Among medication classes examined, several showed significant associations with HbA_1c levels. DPP-4 inhibitors and GLP-1 agonists were associated with a decrease in HbA_1c levels, with coefficients of −0.08 and −0.24, respectively. Long-acting insulin and metformin immediate-release (IR) were associated with an increase in HbA_1c levels, as indicated by their positive coefficients of 0.38 and 0.16, respectively. Mixed insulin formulations and sulfonylureas showed an association with decreased HbA_1c levels.

Interaction Effects

The interaction terms between the recall period and the medication classes provided insights into the differential impact of the medication switch postrecall. Notably, the interaction term for long-acting insulin (coefficient, −0.10) was significant, suggesting a differential effect on HbA_1c levels postrecall. Other medications, like metformin IR, also exhibited significant interaction effects, indicating changes in the impact on HbA_1c levels in the postrecall period. The binary variable for medication class exposure suggests the use of a logit link function for binary outcomes within the multilevel modeling framework.¹⁵ We did not address the potential for cross cluster heterogeneity due to different medication classes.

DISCUSSION

This study is an ongoing, concurrent, observational, multicenter, registry-based study consisting of VISN 6 veterans who have T2DM and were prescribed metformin SA on June 1, 2020. This initial aim was to evaluate change in HbA_1c levels following the FDA metformin recall. While there was substantial variability in baseline HbA_1c levels across the patients, the mean baseline HbA_1c level at 7.5% (58 mmol/mol). Patients taking GLP-1 agonists showed substantial decrease in HbA_1c levels (coefficient; –0.5; 95% CI, –0.56 to –0.44; P <. 001). Patients taking SGLT-2 inhibitors had a notable decrease in HbA_1c (coefficient, –0.27; 95% CI, –0.32 to –0.22; P < .001). Despite this, the coefficient for the postrecall period was not statistically significant, indicating no overall change in HbA_1c levels from pre- to postrecall when specific medication classes were not considered.

Further analysis included assessment of prescribing trends postrecall. There was an increase in SGLT-2 inhibitor, GLP-1 agonist, and DPP-4 inhibitor prescribing. Considering the growing evidence of the cardiovascular and renal benefits of these medication classes, specifically the GLP-1 agonists and SGLT-2 inhibitors, this trend would be expected.

Limitations

This study cohort did not capture veterans with T2DM who transferred their health care to VISN 6 after June 1, 2020, and continued to receive metformin SA from the prior facility. Inclusion of these veterans would have increased the registry population. Additionally, the cohort did not identify veterans who continued to receive metformin SA through a source other than the VA. Without that information, the registry cohort may include veterans thought to have either transitioned to a different therapy or to no other T2DM therapy after the recall.

Given that DM can progress over time, it is possible the transition to a new medication after the recall was the result of suboptimal management, or in response to an adverse effect from a previous medication, and not solely due to the metformin SA recall. In addition, there are several factors that could impact HbA_1c level over time that were not accounted for in this study, such as medication adherence and lifestyle modifications.

The notable level of metformin SA prescriptions, despite the recall, may be attributed to several factors. First, not all patients stopped metformin completely. Review of the prescription data indicated that some veterans were provided with limited refills at select VA medical centers that had supplies (medication lots not recalled). Access to a safe supply of metformin SA after the recall may have varied among VISN 6 facilities. It is also possible that as new supplies of metformin SA became available, veterans restarted metformin SA. This may have been resumed while continuing a new medication prescribed at the beginning of the recall. As the year progressed after the recall, an increase in metformin SA prescriptions likely occurred as supplies became available and clinicians/veterans chose to resume this medication therapy.

Conclusions

Results of this initial registry study found no difference in HbA_1c levels across the study population after the metformin SA recall. However, there was clinical difference in the HbA_1c within veterans prescribed SGLT-2 inhibitors and GLP-1 agonists. As expected, prescribing trends showed an increase in these agents after the recall. With the known benefits of these medications beyond glucose lowering, it is anticipated the cohort of veterans prescribed these medications will continue to grow.

The VISN 6 research registry allowed this study to gain an important snapshot in time following the metformin SA recall, and will serve as an important resource for future DM research endeavors. It will allow for ongoing evaluation of the impact of the transition to alternative T2DM medications after the metformin SA recall. Future exploration will include evaluation of adverse drug reactions, DM-related hospitalizations, emergency department visits related to T2DM, changes in renal function, and cardiovascular events among all diabetes medication classes.

Acknowledgments

The study team thanks the Veterans Affairs Informatics and Computing Infrastructure for their help and expertise throughout this project. The authors acknowledge the contributions of Philip Nelson, PharmD, and Brian Peek, PharmD.

About 1 in 10 Americans have diabetes mellitus (DM), of which about 90% to 95% are diagnosed with type 2 DM (T2DM) and veterans are disproportionately affected.^1,2 About 25% enrolled in the Veterans Health Administration (VHA) have T2DM, which has been attributed to exposure to herbicides (eg, Agent Orange), decreased physical activity resulting from past physical strain, chronic pain, and other physical limitations resulting from military service.^3-5

Pharmacologic management of DM is guided by the effectiveness of lifestyle interventions and comorbid diagnoses. Current DM management guidelines recommend patients with comorbid atherosclerotic cardiovascular disease, chronic kidney disease, or congestive heart failure receive first-line diabetes therapy with a sodium-glucose cotransporter-2 (SGLT-2) inhibitor or glucagon-like peptide-1 receptor (GLP-1) agonist.

Metformin remains a first-line pharmacologic option for the treatment of T2DM with the goal of achieving glycemic management when lifestyle interventions are insufficient.^6,7 Newer antihyperglycemic therapies have been studied as adjunct therapy to metformin. However, there is limited literature comparing metformin directly to other medication classes for the treatment of T2DM.^8-13 A systematic review of treatment-naive patients found HbA_1c reductions were similar whether patients received metformin vs an SGLT-2 inhibitor, GLP-1 agonist, sulfonylurea, or thiazolidinedione monotherapy.¹⁰ The analysis found dipeptidyl-peptidase-4 (DPP-4) inhibitors had inferior HbA_1c reduction compared to metformin.¹⁰ A Japanese systematic review compared metformin to thiazolidinediones, sulfonylureas, glinides, DPP-4 inhibitors, α-glucosidase inhibitors, or SGLT-2 inhibitors for ≥ 12 weeks but found no statistically significant differences in HbA_1c reduction.¹¹ The AWARD-3 trial compared once-weekly dulaglutide to metformin in treatment-experienced patients and found greater improvement in HbA_1c and achievement of HbA_1c goal with dulaglutide.¹³ While these studies show some comparisons of metformin to alternative pharmacologic therapy, researchers have not looked at what happens to patients’ HbA_1c levels when an event, such as a recall, prompts a rapid change to a different antihyperglycemic agent.

On May 28, 2020, the US Food and Drug Administration (FDA) asked 5 pharmaceutical companies to voluntarily recall certain formulations of metformin. This action was taken when FDA testing revealed unacceptably high levels of N-Nitrosodimethylamine, a probable carcinogen.¹⁴ This FDA recall of metformin extended-release, referred to as metformin sustained-action (SA) within the VHA electronic medication file but the same type of formulation, prompted clinicians to revisit and revise the pharmacologic regimens of patients taking the drug. Because of the paucity of head-to-head trials comparing metformin with newer alternative antihyperglycemic therapies, the effect of treatment change was unknown. In response, we aimed to establish a data registry within Veterans Integrated Service Network (VISN) 6.

Registry Development

The VISN 6 registry was established to gather long-term, observational, head-to-head data that would allow review of HbA_1c levels before and after the recall, as well as HbA_1c levels broken down by the agent that patients were switched to after the recall. Another goal was to explore prescribing trends following the recall.

Data Access Request Tracker approval was obtained and a US Department of Veterans Affairs (VA) Information and Computing Infrastructure workspace was developed to host the registry data. The research cohort was established from this data, and the registry framework was finalized using Structured Query Language (SQL). The SQL coding allows for recurring data updates for all individuals within the cohort including date of birth, race, sex, ethnicity, VHA facility visited, weight, body mass index, HbA_1c level, creatinine clearance, serum creatinine, antihyperglycemic medication prescriptions, adverse drug reactions, medication adherence (as defined by ≥ 80% refill history), and hospitalizations related to diabetes. For the purposes of this initial analysis, registry data included demographics, diabetes medications, and HbA_1c results.

METHODS

This study was a concurrent, observational, multicenter, registry-based study conducted at the Western North Carolina VA Health Care System (WNCVAHCS). The study was approved by the WNCVAHCS institutional review board and research and development committees.

All patients aged ≥ 18 years with T2DM and receiving health care from VISN 6 facilities who had an active metformin SA prescription on, and 1 year prior to, June 1, 2020 (the initial date VHA began implementing the FDA metformin recall) were entered into the registry. Data from 1 year prior were collected to provide a baseline. Veterans were excluded if they received metformin SA for any indication other than T2DM, there was no pre- or postrecall HbA_1c measurement, or death. We included 15,594 VISN 6 veterans.

Registry data were analyzed to determine whether a significant change in HbA_1c level occurred after the metformin recall and in response to alternative agents being prescribed. Data from veterans who met all inclusion criteria were assessed during the year before and after June 1, 2020. Demographic data were analyzed using frequency and descriptive statistics. The Shapiro Wilkes test was performed, and data were found to be nonparametric; therefore the Wilcoxon signed-rank test was used to evaluate the hypothesis that HbA_1c levels were not impacted by the recall.

Our sample size allowed us to create exact matched pairs of 9130 individuals and utilize rank-biserial correlation to establish effect size. Following this initial population-level test, we constructed 2 models. The first, a linear mixed-effects model, focused solely on the interaction effects between the pre- and postrecall periods and various medication classes on HbA_1c levels. Second, we constructed a random-effects within-between model (REWB) to evaluate the impact ofmedication classes and demographic variables. Statistical significance was measured at P < .05 with conservative power at .90. The effect size was set to 1.0, reflecting a minimum clinically important difference. Literature establishes 0.5 as a modest level of HbA_1c improvement and 1.0 as a clinically significant improvement.

RESULTS

Preliminary results included 15,594 veterans who received a metformin SA prescription as of June 1, 2020 from VISN 6 facilities; 15,392 veterans had a drug exposure end on June 1, 2020, indicating their standard therapy of metformin SA was discontinued following the FDA recall. Two hundred and two veterans were excluded from the registry because they continued to receive metformin SA from existing stock at a VISN6 facility. After identifying veterans with data for 1 year prior (June 1, 2019) to the index date and 1 year after (June 1, 2021) the study population was adjusted to 9130. The population was predominantly males aged> 60 years. Roughly 55% of the registry identified as White and nearly 40% as Black, and 2% indentified as Hispanic (Table 1).

Wilcoxon Signed-Rank Test

We created exact pairs by iterating the data and finding the closest measurements for each patient before and after the recall. This has the advantage over averaging a patient’s pre- and post-HbA_1c levels, as it allows for a rank-biserial correlation. Using the nonparametric Wilcoxon signed-rank test, V was 20,100,707 (P < .001), indicating a significant effect. The –0.29 rank-biserial correlation, which was computed to assess the effect size of the recall, suggests that the median HbA_1c level was lower postrecall vs prerecall. The magnitude of the correlation suggests a moderate effect size, and while the recall had a noticeable impact at a population level, it was not extreme (Table 2).

Linear Mixed-Effects Model

The binary variable for medication class exposure suggests the use of a logit link function for binary outcomes within the multilevel modeling framework.¹⁵ We employed a linear mixed-effects model to investigate the impact that switching from metformin SA to other T2DM medications had on HbA_1c levels. The model was adjusted for patient-specific random effects and included interaction terms between the recall period (before and after) and the usage of different T2DM medications.

Model Fit and Random Effects

The model demonstrated a residual maximum likelihood criterion of 100,219.7, indicating its fit to the data. Notably, the random effects analysis revealed a substantial variability in baseline HbA_1c levels across patients (SD, 0.94), highlighting the importance of individual differences in DM management. Medication classes with zero or near-zero exposure rate were removed. Due to demographic homogeneity, the model did not converge on demographic variables. Veterans were taking a mean of 1.8 T2DM medications and metformin SA was most common (Table 3).

During the postrecall period, metformin SA remained the most frequently prescribed medication class. This may be attributed to the existence of multiple manufacturers of metformin SA, some of which may not have been impacted by the recall. VISN 6 medical centers could have sought metformin SA outside of the usual procurement path following the recall.

Complex Random Effects Model

We employed a complex REWB model that evaluated the impact of medication classes on HbA_1c levels, accounting for both within and between subject effects of these medications, along with demographic variables (sex, race, and ethnicity) (eAppendix). This model accounts for individual-level changes over time (within-patient effects) and between groups of patients (between-patient effects). This is a more comprehensive model aimed at understanding the broader impact of medications on HbA_1c levels across diverse patient groups.

Most demographic categories did not demonstrate significant effects in this model. Black individuals experienced a slight increase in HbA_1c levels compared with other racial categories that was not statistically significant. However, this model confirms the findings from the linear mixed-effects model that GLP-1 agonists showed a substantial decrease in HbA_1c levels within patients (coefficient –0.5; 95% CI, –0.56 to –0.44; P < .001) and a moderate increase between patients (coefficient, 0.21; 95% CI, 0.12-0.31; P < .001). Additionally, SGLT-2 inhibitors had a notable decrease within patients (coefficient, –0.27; 95% CI, –0.32 to –0.22; P < .001).Another notable finding with our REWB model is insulin usage was associated with high HbA_1c levels, but only between subjects. Long-acting insulin (coefficient, 0.96; 95% CI, 0.90-1.01; P <. 001) and mixed insulin (coefficient, 1.09; 95% CI, 0.94-1.24; P < .001) both displayed marked increases between patients, suggesting future analysis may benefit from stratifying across insulin users and nonusers.

Fixed Effect Analysis

The fixed effects analysis yielded several notable findings. The intercept, representing the mean baseline HbA_1c level, was estimated at 7.8% (58 mmol/mol). The coefficient for the period (postrecall) was not statistically significant, indicating no overall change in HbA_1c levels from before to after the recall when specific medication classes were not considered (Table 4). Among medication classes examined, several showed significant associations with HbA_1c levels. DPP-4 inhibitors and GLP-1 agonists were associated with a decrease in HbA_1c levels, with coefficients of −0.08 and −0.24, respectively. Long-acting insulin and metformin immediate-release (IR) were associated with an increase in HbA_1c levels, as indicated by their positive coefficients of 0.38 and 0.16, respectively. Mixed insulin formulations and sulfonylureas showed an association with decreased HbA_1c levels.

Interaction Effects

The interaction terms between the recall period and the medication classes provided insights into the differential impact of the medication switch postrecall. Notably, the interaction term for long-acting insulin (coefficient, −0.10) was significant, suggesting a differential effect on HbA_1c levels postrecall. Other medications, like metformin IR, also exhibited significant interaction effects, indicating changes in the impact on HbA_1c levels in the postrecall period. The binary variable for medication class exposure suggests the use of a logit link function for binary outcomes within the multilevel modeling framework.¹⁵ We did not address the potential for cross cluster heterogeneity due to different medication classes.

DISCUSSION

This study is an ongoing, concurrent, observational, multicenter, registry-based study consisting of VISN 6 veterans who have T2DM and were prescribed metformin SA on June 1, 2020. This initial aim was to evaluate change in HbA_1c levels following the FDA metformin recall. While there was substantial variability in baseline HbA_1c levels across the patients, the mean baseline HbA_1c level at 7.5% (58 mmol/mol). Patients taking GLP-1 agonists showed substantial decrease in HbA_1c levels (coefficient; –0.5; 95% CI, –0.56 to –0.44; P <. 001). Patients taking SGLT-2 inhibitors had a notable decrease in HbA_1c (coefficient, –0.27; 95% CI, –0.32 to –0.22; P < .001). Despite this, the coefficient for the postrecall period was not statistically significant, indicating no overall change in HbA_1c levels from pre- to postrecall when specific medication classes were not considered.

Further analysis included assessment of prescribing trends postrecall. There was an increase in SGLT-2 inhibitor, GLP-1 agonist, and DPP-4 inhibitor prescribing. Considering the growing evidence of the cardiovascular and renal benefits of these medication classes, specifically the GLP-1 agonists and SGLT-2 inhibitors, this trend would be expected.

Limitations

This study cohort did not capture veterans with T2DM who transferred their health care to VISN 6 after June 1, 2020, and continued to receive metformin SA from the prior facility. Inclusion of these veterans would have increased the registry population. Additionally, the cohort did not identify veterans who continued to receive metformin SA through a source other than the VA. Without that information, the registry cohort may include veterans thought to have either transitioned to a different therapy or to no other T2DM therapy after the recall.

Given that DM can progress over time, it is possible the transition to a new medication after the recall was the result of suboptimal management, or in response to an adverse effect from a previous medication, and not solely due to the metformin SA recall. In addition, there are several factors that could impact HbA_1c level over time that were not accounted for in this study, such as medication adherence and lifestyle modifications.

The notable level of metformin SA prescriptions, despite the recall, may be attributed to several factors. First, not all patients stopped metformin completely. Review of the prescription data indicated that some veterans were provided with limited refills at select VA medical centers that had supplies (medication lots not recalled). Access to a safe supply of metformin SA after the recall may have varied among VISN 6 facilities. It is also possible that as new supplies of metformin SA became available, veterans restarted metformin SA. This may have been resumed while continuing a new medication prescribed at the beginning of the recall. As the year progressed after the recall, an increase in metformin SA prescriptions likely occurred as supplies became available and clinicians/veterans chose to resume this medication therapy.

Conclusions

Results of this initial registry study found no difference in HbA_1c levels across the study population after the metformin SA recall. However, there was clinical difference in the HbA_1c within veterans prescribed SGLT-2 inhibitors and GLP-1 agonists. As expected, prescribing trends showed an increase in these agents after the recall. With the known benefits of these medications beyond glucose lowering, it is anticipated the cohort of veterans prescribed these medications will continue to grow.

The VISN 6 research registry allowed this study to gain an important snapshot in time following the metformin SA recall, and will serve as an important resource for future DM research endeavors. It will allow for ongoing evaluation of the impact of the transition to alternative T2DM medications after the metformin SA recall. Future exploration will include evaluation of adverse drug reactions, DM-related hospitalizations, emergency department visits related to T2DM, changes in renal function, and cardiovascular events among all diabetes medication classes.

Acknowledgments

The study team thanks the Veterans Affairs Informatics and Computing Infrastructure for their help and expertise throughout this project. The authors acknowledge the contributions of Philip Nelson, PharmD, and Brian Peek, PharmD.

About 1 in 10 Americans have diabetes mellitus (DM), of which about 90% to 95% are diagnosed with type 2 DM (T2DM) and veterans are disproportionately affected.^1,2 About 25% enrolled in the Veterans Health Administration (VHA) have T2DM, which has been attributed to exposure to herbicides (eg, Agent Orange), decreased physical activity resulting from past physical strain, chronic pain, and other physical limitations resulting from military service.^3-5

Pharmacologic management of DM is guided by the effectiveness of lifestyle interventions and comorbid diagnoses. Current DM management guidelines recommend patients with comorbid atherosclerotic cardiovascular disease, chronic kidney disease, or congestive heart failure receive first-line diabetes therapy with a sodium-glucose cotransporter-2 (SGLT-2) inhibitor or glucagon-like peptide-1 receptor (GLP-1) agonist.

Metformin remains a first-line pharmacologic option for the treatment of T2DM with the goal of achieving glycemic management when lifestyle interventions are insufficient.^6,7 Newer antihyperglycemic therapies have been studied as adjunct therapy to metformin. However, there is limited literature comparing metformin directly to other medication classes for the treatment of T2DM.^8-13 A systematic review of treatment-naive patients found HbA_1c reductions were similar whether patients received metformin vs an SGLT-2 inhibitor, GLP-1 agonist, sulfonylurea, or thiazolidinedione monotherapy.¹⁰ The analysis found dipeptidyl-peptidase-4 (DPP-4) inhibitors had inferior HbA_1c reduction compared to metformin.¹⁰ A Japanese systematic review compared metformin to thiazolidinediones, sulfonylureas, glinides, DPP-4 inhibitors, α-glucosidase inhibitors, or SGLT-2 inhibitors for ≥ 12 weeks but found no statistically significant differences in HbA_1c reduction.¹¹ The AWARD-3 trial compared once-weekly dulaglutide to metformin in treatment-experienced patients and found greater improvement in HbA_1c and achievement of HbA_1c goal with dulaglutide.¹³ While these studies show some comparisons of metformin to alternative pharmacologic therapy, researchers have not looked at what happens to patients’ HbA_1c levels when an event, such as a recall, prompts a rapid change to a different antihyperglycemic agent.

On May 28, 2020, the US Food and Drug Administration (FDA) asked 5 pharmaceutical companies to voluntarily recall certain formulations of metformin. This action was taken when FDA testing revealed unacceptably high levels of N-Nitrosodimethylamine, a probable carcinogen.¹⁴ This FDA recall of metformin extended-release, referred to as metformin sustained-action (SA) within the VHA electronic medication file but the same type of formulation, prompted clinicians to revisit and revise the pharmacologic regimens of patients taking the drug. Because of the paucity of head-to-head trials comparing metformin with newer alternative antihyperglycemic therapies, the effect of treatment change was unknown. In response, we aimed to establish a data registry within Veterans Integrated Service Network (VISN) 6.

Registry Development

The VISN 6 registry was established to gather long-term, observational, head-to-head data that would allow review of HbA_1c levels before and after the recall, as well as HbA_1c levels broken down by the agent that patients were switched to after the recall. Another goal was to explore prescribing trends following the recall.

Data Access Request Tracker approval was obtained and a US Department of Veterans Affairs (VA) Information and Computing Infrastructure workspace was developed to host the registry data. The research cohort was established from this data, and the registry framework was finalized using Structured Query Language (SQL). The SQL coding allows for recurring data updates for all individuals within the cohort including date of birth, race, sex, ethnicity, VHA facility visited, weight, body mass index, HbA_1c level, creatinine clearance, serum creatinine, antihyperglycemic medication prescriptions, adverse drug reactions, medication adherence (as defined by ≥ 80% refill history), and hospitalizations related to diabetes. For the purposes of this initial analysis, registry data included demographics, diabetes medications, and HbA_1c results.

METHODS

This study was a concurrent, observational, multicenter, registry-based study conducted at the Western North Carolina VA Health Care System (WNCVAHCS). The study was approved by the WNCVAHCS institutional review board and research and development committees.

All patients aged ≥ 18 years with T2DM and receiving health care from VISN 6 facilities who had an active metformin SA prescription on, and 1 year prior to, June 1, 2020 (the initial date VHA began implementing the FDA metformin recall) were entered into the registry. Data from 1 year prior were collected to provide a baseline. Veterans were excluded if they received metformin SA for any indication other than T2DM, there was no pre- or postrecall HbA_1c measurement, or death. We included 15,594 VISN 6 veterans.

Registry data were analyzed to determine whether a significant change in HbA_1c level occurred after the metformin recall and in response to alternative agents being prescribed. Data from veterans who met all inclusion criteria were assessed during the year before and after June 1, 2020. Demographic data were analyzed using frequency and descriptive statistics. The Shapiro Wilkes test was performed, and data were found to be nonparametric; therefore the Wilcoxon signed-rank test was used to evaluate the hypothesis that HbA_1c levels were not impacted by the recall.

Our sample size allowed us to create exact matched pairs of 9130 individuals and utilize rank-biserial correlation to establish effect size. Following this initial population-level test, we constructed 2 models. The first, a linear mixed-effects model, focused solely on the interaction effects between the pre- and postrecall periods and various medication classes on HbA_1c levels. Second, we constructed a random-effects within-between model (REWB) to evaluate the impact ofmedication classes and demographic variables. Statistical significance was measured at P < .05 with conservative power at .90. The effect size was set to 1.0, reflecting a minimum clinically important difference. Literature establishes 0.5 as a modest level of HbA_1c improvement and 1.0 as a clinically significant improvement.

RESULTS

Preliminary results included 15,594 veterans who received a metformin SA prescription as of June 1, 2020 from VISN 6 facilities; 15,392 veterans had a drug exposure end on June 1, 2020, indicating their standard therapy of metformin SA was discontinued following the FDA recall. Two hundred and two veterans were excluded from the registry because they continued to receive metformin SA from existing stock at a VISN6 facility. After identifying veterans with data for 1 year prior (June 1, 2019) to the index date and 1 year after (June 1, 2021) the study population was adjusted to 9130. The population was predominantly males aged> 60 years. Roughly 55% of the registry identified as White and nearly 40% as Black, and 2% indentified as Hispanic (Table 1).

Wilcoxon Signed-Rank Test

We created exact pairs by iterating the data and finding the closest measurements for each patient before and after the recall. This has the advantage over averaging a patient’s pre- and post-HbA_1c levels, as it allows for a rank-biserial correlation. Using the nonparametric Wilcoxon signed-rank test, V was 20,100,707 (P < .001), indicating a significant effect. The –0.29 rank-biserial correlation, which was computed to assess the effect size of the recall, suggests that the median HbA_1c level was lower postrecall vs prerecall. The magnitude of the correlation suggests a moderate effect size, and while the recall had a noticeable impact at a population level, it was not extreme (Table 2).

Linear Mixed-Effects Model

The binary variable for medication class exposure suggests the use of a logit link function for binary outcomes within the multilevel modeling framework.¹⁵ We employed a linear mixed-effects model to investigate the impact that switching from metformin SA to other T2DM medications had on HbA_1c levels. The model was adjusted for patient-specific random effects and included interaction terms between the recall period (before and after) and the usage of different T2DM medications.

Model Fit and Random Effects

The model demonstrated a residual maximum likelihood criterion of 100,219.7, indicating its fit to the data. Notably, the random effects analysis revealed a substantial variability in baseline HbA_1c levels across patients (SD, 0.94), highlighting the importance of individual differences in DM management. Medication classes with zero or near-zero exposure rate were removed. Due to demographic homogeneity, the model did not converge on demographic variables. Veterans were taking a mean of 1.8 T2DM medications and metformin SA was most common (Table 3).

During the postrecall period, metformin SA remained the most frequently prescribed medication class. This may be attributed to the existence of multiple manufacturers of metformin SA, some of which may not have been impacted by the recall. VISN 6 medical centers could have sought metformin SA outside of the usual procurement path following the recall.

Complex Random Effects Model

We employed a complex REWB model that evaluated the impact of medication classes on HbA_1c levels, accounting for both within and between subject effects of these medications, along with demographic variables (sex, race, and ethnicity) (eAppendix). This model accounts for individual-level changes over time (within-patient effects) and between groups of patients (between-patient effects). This is a more comprehensive model aimed at understanding the broader impact of medications on HbA_1c levels across diverse patient groups.

Most demographic categories did not demonstrate significant effects in this model. Black individuals experienced a slight increase in HbA_1c levels compared with other racial categories that was not statistically significant. However, this model confirms the findings from the linear mixed-effects model that GLP-1 agonists showed a substantial decrease in HbA_1c levels within patients (coefficient –0.5; 95% CI, –0.56 to –0.44; P < .001) and a moderate increase between patients (coefficient, 0.21; 95% CI, 0.12-0.31; P < .001). Additionally, SGLT-2 inhibitors had a notable decrease within patients (coefficient, –0.27; 95% CI, –0.32 to –0.22; P < .001).Another notable finding with our REWB model is insulin usage was associated with high HbA_1c levels, but only between subjects. Long-acting insulin (coefficient, 0.96; 95% CI, 0.90-1.01; P <. 001) and mixed insulin (coefficient, 1.09; 95% CI, 0.94-1.24; P < .001) both displayed marked increases between patients, suggesting future analysis may benefit from stratifying across insulin users and nonusers.

Fixed Effect Analysis

The fixed effects analysis yielded several notable findings. The intercept, representing the mean baseline HbA_1c level, was estimated at 7.8% (58 mmol/mol). The coefficient for the period (postrecall) was not statistically significant, indicating no overall change in HbA_1c levels from before to after the recall when specific medication classes were not considered (Table 4). Among medication classes examined, several showed significant associations with HbA_1c levels. DPP-4 inhibitors and GLP-1 agonists were associated with a decrease in HbA_1c levels, with coefficients of −0.08 and −0.24, respectively. Long-acting insulin and metformin immediate-release (IR) were associated with an increase in HbA_1c levels, as indicated by their positive coefficients of 0.38 and 0.16, respectively. Mixed insulin formulations and sulfonylureas showed an association with decreased HbA_1c levels.

Interaction Effects

The interaction terms between the recall period and the medication classes provided insights into the differential impact of the medication switch postrecall. Notably, the interaction term for long-acting insulin (coefficient, −0.10) was significant, suggesting a differential effect on HbA_1c levels postrecall. Other medications, like metformin IR, also exhibited significant interaction effects, indicating changes in the impact on HbA_1c levels in the postrecall period. The binary variable for medication class exposure suggests the use of a logit link function for binary outcomes within the multilevel modeling framework.¹⁵ We did not address the potential for cross cluster heterogeneity due to different medication classes.

DISCUSSION

This study is an ongoing, concurrent, observational, multicenter, registry-based study consisting of VISN 6 veterans who have T2DM and were prescribed metformin SA on June 1, 2020. This initial aim was to evaluate change in HbA_1c levels following the FDA metformin recall. While there was substantial variability in baseline HbA_1c levels across the patients, the mean baseline HbA_1c level at 7.5% (58 mmol/mol). Patients taking GLP-1 agonists showed substantial decrease in HbA_1c levels (coefficient; –0.5; 95% CI, –0.56 to –0.44; P <. 001). Patients taking SGLT-2 inhibitors had a notable decrease in HbA_1c (coefficient, –0.27; 95% CI, –0.32 to –0.22; P < .001). Despite this, the coefficient for the postrecall period was not statistically significant, indicating no overall change in HbA_1c levels from pre- to postrecall when specific medication classes were not considered.

Further analysis included assessment of prescribing trends postrecall. There was an increase in SGLT-2 inhibitor, GLP-1 agonist, and DPP-4 inhibitor prescribing. Considering the growing evidence of the cardiovascular and renal benefits of these medication classes, specifically the GLP-1 agonists and SGLT-2 inhibitors, this trend would be expected.

Limitations

This study cohort did not capture veterans with T2DM who transferred their health care to VISN 6 after June 1, 2020, and continued to receive metformin SA from the prior facility. Inclusion of these veterans would have increased the registry population. Additionally, the cohort did not identify veterans who continued to receive metformin SA through a source other than the VA. Without that information, the registry cohort may include veterans thought to have either transitioned to a different therapy or to no other T2DM therapy after the recall.

Given that DM can progress over time, it is possible the transition to a new medication after the recall was the result of suboptimal management, or in response to an adverse effect from a previous medication, and not solely due to the metformin SA recall. In addition, there are several factors that could impact HbA_1c level over time that were not accounted for in this study, such as medication adherence and lifestyle modifications.

The notable level of metformin SA prescriptions, despite the recall, may be attributed to several factors. First, not all patients stopped metformin completely. Review of the prescription data indicated that some veterans were provided with limited refills at select VA medical centers that had supplies (medication lots not recalled). Access to a safe supply of metformin SA after the recall may have varied among VISN 6 facilities. It is also possible that as new supplies of metformin SA became available, veterans restarted metformin SA. This may have been resumed while continuing a new medication prescribed at the beginning of the recall. As the year progressed after the recall, an increase in metformin SA prescriptions likely occurred as supplies became available and clinicians/veterans chose to resume this medication therapy.

Conclusions

Results of this initial registry study found no difference in HbA_1c levels across the study population after the metformin SA recall. However, there was clinical difference in the HbA_1c within veterans prescribed SGLT-2 inhibitors and GLP-1 agonists. As expected, prescribing trends showed an increase in these agents after the recall. With the known benefits of these medications beyond glucose lowering, it is anticipated the cohort of veterans prescribed these medications will continue to grow.

The VISN 6 research registry allowed this study to gain an important snapshot in time following the metformin SA recall, and will serve as an important resource for future DM research endeavors. It will allow for ongoing evaluation of the impact of the transition to alternative T2DM medications after the metformin SA recall. Future exploration will include evaluation of adverse drug reactions, DM-related hospitalizations, emergency department visits related to T2DM, changes in renal function, and cardiovascular events among all diabetes medication classes.

Acknowledgments

The study team thanks the Veterans Affairs Informatics and Computing Infrastructure for their help and expertise throughout this project. The authors acknowledge the contributions of Philip Nelson, PharmD, and Brian Peek, PharmD.

To the Editor:

Although biologics provide major therapeutic benefits for dermatologic conditions, they also come with a substantial cost, making them among the most expensive medications available. Medicare and Medicaid spending on biologics for dermatologic conditions increased by 320% from 2012 to 2018, reaching a staggering $10.6 billion in 2018 alone.¹ Biosimilars show promise in reducing health care spending for dermatologic conditions; however, their utilization has been limited due to multiple factors, including delayed market entry from patent thickets, exclusionary formulary contracts, and prescriber skepticism regarding their safety and efficacy.² For instance, a national survey of 1201 US physicians in specialties that are high prescribers of biologics reported that 55% doubted the safety and appropriateness of biosimilars.³

US Food and Drug Administration approval of biosimilars for adalimumab and etanercept offers the potential to reduce health care spending for dermatologic conditions. However, this cost reduction is dependent on utilization rates among dermatologists. In this national cross-sectional review of Medicare data, we predicted the impact of these biosimilars on dermatologic Medicare costs and demonstrated how differing utilization rates among dermatologists can influence potential savings.

To model 2023 utilization and cost reduction from biosimilars, we analyzed Medicare Part D data from 2020 on existing biosimilars, including granulocyte colony–stimulating factors, erythropoiesis-stimulating agents, and tumor necrosis factor α inhibitors.⁴ Methods in line with a 2021 report from the US Department of Health and Human Services⁵ as well as those of Yazdany et al⁶ were used. For each class, we calculated the 2020 distribution of biosimilar and originator drug claims as well as biosimilar cost reduction per 30-day claim. We utilized 2018-2021 annual growth rates for branded adalimumab and etanercept to estimate 30-day claims for 2023 and the cost of these branded agents in the absence of biosimilars. The hypothetical 2023 cost reduction from adalimumab and etanercept biosimilars was estimated by assuming 2020 biosimilar utilization rates and mean cost reduction per claim. This study utilized publicly available or aggregate summary data (not attributable to specific patients) and did not qualify as human subject research; therefore, institutional review board approval was not required.

In 2020, biosimilar utilization proportions ranged from 6.4% (tumor necrosis factor α inhibitors) to 82.7% (granulocyte colony–stimulating factors), with a mean across all classes of 35.7%. On average, the cost per 30-day claim of biosimilars was 66.8% of originator agents (Table 1). In 2021, we identified 57,868 30-day claims for branded adalimumab and etanercept submitted by dermatologists. From 2018 to 2021, 30-day branded adalimumab claims increased by 1.27% annually (cost + 10.62% annually), while claims for branded etanercept decreased by 13.0% annually (cost + 5.68% annually). Assuming these trends, the cost of branded adalimumab and etanercept was estimated to be $539 million in 2023. Applying the aforementioned 35.7% utilization, the introduction of biosimilars in dermatology would yield a cost reduction of approximately $118 million (21.9%). A high utilization rate (82.7%) of biosimilars among dermatologists would increase cost savings to $199 million (36.9%)(Table 2).

We found that the extent of savings from biosimilars was dependent on the utilization rates among dermatologists, with the highest utilization rate almost doubling the total savings of average utilization rates. Given the impact of high utilization and the wide variation observed, understanding the factors that have influenced uptake of biosimilars is important to increasing utilization as these medications become integrated into dermatology. For instance, limited uptake of infliximab initially may have been influenced by concerns about efficacy and increased adverse events.^8,9 In contrast, the high utilization of filgrastim biosimilars (82.7%) may be attributed to its longevity in the market and familiarity to prescribers, as filgrastim was the first biosimilar to be approved in the United States.¹⁰

Promoting reasonable utilization of biosimilars may require prescriber education on their safety and approval processes, which could foster increased utilization and reduce skepticism.⁴ Under the Biologics Price Competition and Innovation Act, the US Food and Drug Administration approves biosimilars only when they exhibit “high similarity” and show no “clinically meaningful differences” compared to the reference biologic, with no added safety risks or reduced efficacy.¹¹ Moreover, a 2023 systematic review of 17 studies found no major difference in efficacy and safety between biosimilars and originators of etanercept, infliximab, and other biologics.¹² Understanding these findings may reassure dermatologists and patients about the reliability and safety of biosimilars.

A limitation of our study is that it solely assesses Medicare data and estimates derived from existing (separate) biologic classes. It also does not account for potential expenditure shifts to newer biologic agents (eg, IL-12/17/23 inhibitors) or changes in manufacturer behavior or promotions. Nevertheless, it indicates notable financial savings from new biosimilar agents in dermatology; along with their compelling efficacy and safety profiles, this could represent a substantial benefit to patients and the health care system.

To the Editor:

Although biologics provide major therapeutic benefits for dermatologic conditions, they also come with a substantial cost, making them among the most expensive medications available. Medicare and Medicaid spending on biologics for dermatologic conditions increased by 320% from 2012 to 2018, reaching a staggering $10.6 billion in 2018 alone.¹ Biosimilars show promise in reducing health care spending for dermatologic conditions; however, their utilization has been limited due to multiple factors, including delayed market entry from patent thickets, exclusionary formulary contracts, and prescriber skepticism regarding their safety and efficacy.² For instance, a national survey of 1201 US physicians in specialties that are high prescribers of biologics reported that 55% doubted the safety and appropriateness of biosimilars.³

US Food and Drug Administration approval of biosimilars for adalimumab and etanercept offers the potential to reduce health care spending for dermatologic conditions. However, this cost reduction is dependent on utilization rates among dermatologists. In this national cross-sectional review of Medicare data, we predicted the impact of these biosimilars on dermatologic Medicare costs and demonstrated how differing utilization rates among dermatologists can influence potential savings.

To model 2023 utilization and cost reduction from biosimilars, we analyzed Medicare Part D data from 2020 on existing biosimilars, including granulocyte colony–stimulating factors, erythropoiesis-stimulating agents, and tumor necrosis factor α inhibitors.⁴ Methods in line with a 2021 report from the US Department of Health and Human Services⁵ as well as those of Yazdany et al⁶ were used. For each class, we calculated the 2020 distribution of biosimilar and originator drug claims as well as biosimilar cost reduction per 30-day claim. We utilized 2018-2021 annual growth rates for branded adalimumab and etanercept to estimate 30-day claims for 2023 and the cost of these branded agents in the absence of biosimilars. The hypothetical 2023 cost reduction from adalimumab and etanercept biosimilars was estimated by assuming 2020 biosimilar utilization rates and mean cost reduction per claim. This study utilized publicly available or aggregate summary data (not attributable to specific patients) and did not qualify as human subject research; therefore, institutional review board approval was not required.

In 2020, biosimilar utilization proportions ranged from 6.4% (tumor necrosis factor α inhibitors) to 82.7% (granulocyte colony–stimulating factors), with a mean across all classes of 35.7%. On average, the cost per 30-day claim of biosimilars was 66.8% of originator agents (Table 1). In 2021, we identified 57,868 30-day claims for branded adalimumab and etanercept submitted by dermatologists. From 2018 to 2021, 30-day branded adalimumab claims increased by 1.27% annually (cost + 10.62% annually), while claims for branded etanercept decreased by 13.0% annually (cost + 5.68% annually). Assuming these trends, the cost of branded adalimumab and etanercept was estimated to be $539 million in 2023. Applying the aforementioned 35.7% utilization, the introduction of biosimilars in dermatology would yield a cost reduction of approximately $118 million (21.9%). A high utilization rate (82.7%) of biosimilars among dermatologists would increase cost savings to $199 million (36.9%)(Table 2).

We found that the extent of savings from biosimilars was dependent on the utilization rates among dermatologists, with the highest utilization rate almost doubling the total savings of average utilization rates. Given the impact of high utilization and the wide variation observed, understanding the factors that have influenced uptake of biosimilars is important to increasing utilization as these medications become integrated into dermatology. For instance, limited uptake of infliximab initially may have been influenced by concerns about efficacy and increased adverse events.^8,9 In contrast, the high utilization of filgrastim biosimilars (82.7%) may be attributed to its longevity in the market and familiarity to prescribers, as filgrastim was the first biosimilar to be approved in the United States.¹⁰

Promoting reasonable utilization of biosimilars may require prescriber education on their safety and approval processes, which could foster increased utilization and reduce skepticism.⁴ Under the Biologics Price Competition and Innovation Act, the US Food and Drug Administration approves biosimilars only when they exhibit “high similarity” and show no “clinically meaningful differences” compared to the reference biologic, with no added safety risks or reduced efficacy.¹¹ Moreover, a 2023 systematic review of 17 studies found no major difference in efficacy and safety between biosimilars and originators of etanercept, infliximab, and other biologics.¹² Understanding these findings may reassure dermatologists and patients about the reliability and safety of biosimilars.

A limitation of our study is that it solely assesses Medicare data and estimates derived from existing (separate) biologic classes. It also does not account for potential expenditure shifts to newer biologic agents (eg, IL-12/17/23 inhibitors) or changes in manufacturer behavior or promotions. Nevertheless, it indicates notable financial savings from new biosimilar agents in dermatology; along with their compelling efficacy and safety profiles, this could represent a substantial benefit to patients and the health care system.

To the Editor:

Although biologics provide major therapeutic benefits for dermatologic conditions, they also come with a substantial cost, making them among the most expensive medications available. Medicare and Medicaid spending on biologics for dermatologic conditions increased by 320% from 2012 to 2018, reaching a staggering $10.6 billion in 2018 alone.¹ Biosimilars show promise in reducing health care spending for dermatologic conditions; however, their utilization has been limited due to multiple factors, including delayed market entry from patent thickets, exclusionary formulary contracts, and prescriber skepticism regarding their safety and efficacy.² For instance, a national survey of 1201 US physicians in specialties that are high prescribers of biologics reported that 55% doubted the safety and appropriateness of biosimilars.³

US Food and Drug Administration approval of biosimilars for adalimumab and etanercept offers the potential to reduce health care spending for dermatologic conditions. However, this cost reduction is dependent on utilization rates among dermatologists. In this national cross-sectional review of Medicare data, we predicted the impact of these biosimilars on dermatologic Medicare costs and demonstrated how differing utilization rates among dermatologists can influence potential savings.

To model 2023 utilization and cost reduction from biosimilars, we analyzed Medicare Part D data from 2020 on existing biosimilars, including granulocyte colony–stimulating factors, erythropoiesis-stimulating agents, and tumor necrosis factor α inhibitors.⁴ Methods in line with a 2021 report from the US Department of Health and Human Services⁵ as well as those of Yazdany et al⁶ were used. For each class, we calculated the 2020 distribution of biosimilar and originator drug claims as well as biosimilar cost reduction per 30-day claim. We utilized 2018-2021 annual growth rates for branded adalimumab and etanercept to estimate 30-day claims for 2023 and the cost of these branded agents in the absence of biosimilars. The hypothetical 2023 cost reduction from adalimumab and etanercept biosimilars was estimated by assuming 2020 biosimilar utilization rates and mean cost reduction per claim. This study utilized publicly available or aggregate summary data (not attributable to specific patients) and did not qualify as human subject research; therefore, institutional review board approval was not required.

In 2020, biosimilar utilization proportions ranged from 6.4% (tumor necrosis factor α inhibitors) to 82.7% (granulocyte colony–stimulating factors), with a mean across all classes of 35.7%. On average, the cost per 30-day claim of biosimilars was 66.8% of originator agents (Table 1). In 2021, we identified 57,868 30-day claims for branded adalimumab and etanercept submitted by dermatologists. From 2018 to 2021, 30-day branded adalimumab claims increased by 1.27% annually (cost + 10.62% annually), while claims for branded etanercept decreased by 13.0% annually (cost + 5.68% annually). Assuming these trends, the cost of branded adalimumab and etanercept was estimated to be $539 million in 2023. Applying the aforementioned 35.7% utilization, the introduction of biosimilars in dermatology would yield a cost reduction of approximately $118 million (21.9%). A high utilization rate (82.7%) of biosimilars among dermatologists would increase cost savings to $199 million (36.9%)(Table 2).

We found that the extent of savings from biosimilars was dependent on the utilization rates among dermatologists, with the highest utilization rate almost doubling the total savings of average utilization rates. Given the impact of high utilization and the wide variation observed, understanding the factors that have influenced uptake of biosimilars is important to increasing utilization as these medications become integrated into dermatology. For instance, limited uptake of infliximab initially may have been influenced by concerns about efficacy and increased adverse events.^8,9 In contrast, the high utilization of filgrastim biosimilars (82.7%) may be attributed to its longevity in the market and familiarity to prescribers, as filgrastim was the first biosimilar to be approved in the United States.¹⁰

Promoting reasonable utilization of biosimilars may require prescriber education on their safety and approval processes, which could foster increased utilization and reduce skepticism.⁴ Under the Biologics Price Competition and Innovation Act, the US Food and Drug Administration approves biosimilars only when they exhibit “high similarity” and show no “clinically meaningful differences” compared to the reference biologic, with no added safety risks or reduced efficacy.¹¹ Moreover, a 2023 systematic review of 17 studies found no major difference in efficacy and safety between biosimilars and originators of etanercept, infliximab, and other biologics.¹² Understanding these findings may reassure dermatologists and patients about the reliability and safety of biosimilars.

A limitation of our study is that it solely assesses Medicare data and estimates derived from existing (separate) biologic classes. It also does not account for potential expenditure shifts to newer biologic agents (eg, IL-12/17/23 inhibitors) or changes in manufacturer behavior or promotions. Nevertheless, it indicates notable financial savings from new biosimilar agents in dermatology; along with their compelling efficacy and safety profiles, this could represent a substantial benefit to patients and the health care system.

To the Editor:

Patients with primary psychiatric disorders with dermatologic manifestations often seek treatment from dermatologists instead of psychiatrists.¹ For example, patients with delusions of parasitosis may lack insight into the underlying etiology of their disease and instead fixate on establishing an organic cause for their symptoms. As a result, it is an increasingly common practice for dermatologists to diagnose and treat psychiatric conditions.¹ The goal of this study was to evaluate trends for the top 5 antipsychotics most frequently prescribed by dermatologists in the Medicare Part D database.

In this retrospective analysis, we consulted the Medicare Provider Utilization and Payment Data for January 2013 through December 2020, which is provided to the public by the Centers for Medicare & Medicaid Services.² Only prescribing data from dermatologists were included in this study by using the built-in filter on the website to select “dermatology” as the prescriber type. All other provider types were excluded. We chose the top 5 most prescribed antipsychotics based on the number of supply days reported. Supply days—defined by Medicare as the number of days’ worth of medication that is prescribed—were used as a metric for ­utilization; therefore, each drug’s total supply days prescribed by dermatologists were calculated using this combined filter of drug name and total supply days using the database.

To analyze utilization over time, the annual average growth rate (AAGR) was calculated by determining the growth rate in total supply days annually from 2013 to 2020 and then averaging those rates to determine the overall AAGR. For greater clinical relevance, we calculated the average growth in supply days for the entire study period by determining the difference in the number of supply days for each year and then averaging these values. This was done to consider overall trends across dermatology rather than individual dermatologist prescribing patterns.

Based on our analysis, the antipsychotics most frequently prescribed by dermatologists for Medicare patients from January 2013 to December 2020 were pimozide, quetiapine, risperidone, olanzapine, and aripiprazole. The AAGR for each drug was 2.35%, 4.89%, 5.59%, 9.48%, and 20.72%, respectively, which is consistent with increased utilization over the study period for all 5 drugs (Table 1). The change in cost per supply day for the same period was 1.3%, –66.1%, –60.2%, –81.7%, and –84.3%, respectively. The net difference in cost per supply day over this entire period was $0.02, –$2.79, –$1.06, –$5.37, and –$21.22, respectively (Table 2).

To the Editor:

Patients with primary psychiatric disorders with dermatologic manifestations often seek treatment from dermatologists instead of psychiatrists.¹ For example, patients with delusions of parasitosis may lack insight into the underlying etiology of their disease and instead fixate on establishing an organic cause for their symptoms. As a result, it is an increasingly common practice for dermatologists to diagnose and treat psychiatric conditions.¹ The goal of this study was to evaluate trends for the top 5 antipsychotics most frequently prescribed by dermatologists in the Medicare Part D database.

In this retrospective analysis, we consulted the Medicare Provider Utilization and Payment Data for January 2013 through December 2020, which is provided to the public by the Centers for Medicare & Medicaid Services.² Only prescribing data from dermatologists were included in this study by using the built-in filter on the website to select “dermatology” as the prescriber type. All other provider types were excluded. We chose the top 5 most prescribed antipsychotics based on the number of supply days reported. Supply days—defined by Medicare as the number of days’ worth of medication that is prescribed—were used as a metric for ­utilization; therefore, each drug’s total supply days prescribed by dermatologists were calculated using this combined filter of drug name and total supply days using the database.

To analyze utilization over time, the annual average growth rate (AAGR) was calculated by determining the growth rate in total supply days annually from 2013 to 2020 and then averaging those rates to determine the overall AAGR. For greater clinical relevance, we calculated the average growth in supply days for the entire study period by determining the difference in the number of supply days for each year and then averaging these values. This was done to consider overall trends across dermatology rather than individual dermatologist prescribing patterns.

Based on our analysis, the antipsychotics most frequently prescribed by dermatologists for Medicare patients from January 2013 to December 2020 were pimozide, quetiapine, risperidone, olanzapine, and aripiprazole. The AAGR for each drug was 2.35%, 4.89%, 5.59%, 9.48%, and 20.72%, respectively, which is consistent with increased utilization over the study period for all 5 drugs (Table 1). The change in cost per supply day for the same period was 1.3%, –66.1%, –60.2%, –81.7%, and –84.3%, respectively. The net difference in cost per supply day over this entire period was $0.02, –$2.79, –$1.06, –$5.37, and –$21.22, respectively (Table 2).

To the Editor:

Patients with primary psychiatric disorders with dermatologic manifestations often seek treatment from dermatologists instead of psychiatrists.¹ For example, patients with delusions of parasitosis may lack insight into the underlying etiology of their disease and instead fixate on establishing an organic cause for their symptoms. As a result, it is an increasingly common practice for dermatologists to diagnose and treat psychiatric conditions.¹ The goal of this study was to evaluate trends for the top 5 antipsychotics most frequently prescribed by dermatologists in the Medicare Part D database.

In this retrospective analysis, we consulted the Medicare Provider Utilization and Payment Data for January 2013 through December 2020, which is provided to the public by the Centers for Medicare & Medicaid Services.² Only prescribing data from dermatologists were included in this study by using the built-in filter on the website to select “dermatology” as the prescriber type. All other provider types were excluded. We chose the top 5 most prescribed antipsychotics based on the number of supply days reported. Supply days—defined by Medicare as the number of days’ worth of medication that is prescribed—were used as a metric for ­utilization; therefore, each drug’s total supply days prescribed by dermatologists were calculated using this combined filter of drug name and total supply days using the database.

To analyze utilization over time, the annual average growth rate (AAGR) was calculated by determining the growth rate in total supply days annually from 2013 to 2020 and then averaging those rates to determine the overall AAGR. For greater clinical relevance, we calculated the average growth in supply days for the entire study period by determining the difference in the number of supply days for each year and then averaging these values. This was done to consider overall trends across dermatology rather than individual dermatologist prescribing patterns.

Based on our analysis, the antipsychotics most frequently prescribed by dermatologists for Medicare patients from January 2013 to December 2020 were pimozide, quetiapine, risperidone, olanzapine, and aripiprazole. The AAGR for each drug was 2.35%, 4.89%, 5.59%, 9.48%, and 20.72%, respectively, which is consistent with increased utilization over the study period for all 5 drugs (Table 1). The change in cost per supply day for the same period was 1.3%, –66.1%, –60.2%, –81.7%, and –84.3%, respectively. The net difference in cost per supply day over this entire period was $0.02, –$2.79, –$1.06, –$5.37, and –$21.22, respectively (Table 2).

Psoriasis is a chronic inflammatory disease affecting approximately 8 million adults in the United States and 2% of the global population.^1,2 Psoriasis causes pain, itching, and disfigurement and is associated with a physical, psychological, and economic burden that substantially affects health-related quality of life.^3-5

Setting treatment goals and treating to target are evidence-based approaches that have been successfully applied to several chronic diseases to improve patient outcomes, including diabetes, hypertension, and rheumatoid arthritis.^6-9 Treat-to-target strategies generally set low disease activity (or remission) as an overall goal and seek to achieve this using available therapeutic options as necessary. Introduced following the availability of biologics and targeted systemic therapies, treat-to-target strategies generally provide guidance on expectations of treatment but not specific treatments, as personalized treatment decisions depend on an assessment of individual patients and consider clinical and demographic features as well as preferences for available therapeutic options. If targets are not achieved in the assigned time span, adjustments can be made to the treatment approach in close consultation with the patient. If the target is reached, follow-up visits can be scheduled to ensure improvement is maintained or to establish if more aggressive goals could be selected.

Treat-to-target strategies for the management of psoriasis developed by the National Psoriasis Foundation (NPF) Medical Board include reducing the extent of psoriasis to 1% or lower total body surface area (BSA) after 3 months of treatment.¹⁰ Treatment targets endorsed by the European Academy of Dermatology and Venereology (EADV) in guidelines on the use of systemic therapies in psoriasis include achieving a 75% or greater reduction in Psoriasis Area and Severity Index (PASI) score within 3 to 4 months of treatment.¹¹

In clinical practice, many patients do not achieve these treatment targets, and topical treatments alone generally are insufficient in achieving treatment goals for psoriasis.^12,13 Moreover, conventional topical treatments (eg, topical corticosteroids) used by most patients with psoriasis regardless of disease severity are associated with adverse events that can limit their use. Most topical corticosteroids have US Food and Drug Administration label restrictions relating to sites of application, duration and extent of use, and frequency of administration.^14,15

Tapinarof cream 1% (VTAMA [Dermavant Sciences, Inc]) is a first-in-class topical nonsteroidal aryl hydrocarbon receptor agonist that was approved by the US Food and Drug Administration for the treatment of plaque psoriasis in adults¹⁶ and is being studied for the treatment of plaque psoriasis in children 2 years and older as well as for atopic dermatitis in adults and children 2 years and older. In PSOARING 1 (ClinicalTrials .gov identifier NCT03956355) and PSOARING 2 (NCT03983980)—identical 12-week pivotal phase 3 trials—monotherapy with tapinarof cream 1% once daily (QD) demonstrated statistically significant efficacy vs vehicle cream and was well tolerated in adults with mild to severe plaque psoriasis (Supplementary Figure S1).¹⁷ Lebwohl et al¹⁷ reported that significantly higher PASI75 responses were observed at week 12 with tapinarof cream vs vehicle in PSOARING 1 and PSOARING 2 (36% and 48% vs 10% and 7%, respectively; both P<.0001). A significantly higher PASI90 response of 19% and 21% at week 12 also was observed with tapinarof cream vs 2% and 3% with vehicle in PSOARING 1 and PSOARING 2, respectively (P=.0005 and P<.0001).¹⁷

In PSOARING 3 (NCT04053387)—the long-term extension trial (Supplementary Figure S1)—efficacy continued to improve or was maintained beyond the two 12-week trials, with improvements in total BSA affected and PASI scores for up to 52 weeks.¹⁸ Tapinarof cream 1% QD demonstrated positive, rapid, and durable outcomes in PSOARING 3, including high rates of complete disease clearance (Physician Global Assessment [PGA] score=0 [clear])(40.9% [312/763]), durability of response on treatment with no evidence of tachyphylaxis, and a remittive effect of approximately 4 months when off therapy (defined as maintenance of a PGA score of 0 [clear] or 1 [almost clear] after first achieving a PGA score of 0).¹⁸

Herein, we report absolute treatment targets for patients with plaque psoriasis who received tapinarof cream 1% QD in the PSOARING trials that are at least as stringent as the corresponding NPF and EADV targets of achieving a total BSA affected of 1% or lower or a PASI75 response within 3 to 4 months, respectively.

METHODS

Study Design

The pooled efficacy analyses included all patients with a baseline PGA score of 2 or higher (mild or worse) before treatment with tapinarof cream 1% QD in the PSOARING trials. This included patients who received tapinarof cream 1% in PSOARING 1 and PSOARING 2 who may or may not have continued into PSOARING 3, as well as those who received the vehicle in PSOARING 1 and PSOARING 2 who enrolled in PSOARING 3 and had a PGA score of 2 or higher before receiving tapinarof cream 1%.

Trial Participants

Full methods, including inclusion and exclusion criteria, for the PSOARING trials have been previously reported.^17,18 Patients were aged 18 to 75 years and had chronic plaque psoriasis that was stable for at least 6 months before randomization; 3% to 20% total BSA affected (excluding the scalp, palms, fingernails, toenails, and soles); and a PGA score of 2 (mild), 3 (moderate), or 4 (severe) at baseline.

The clinical trials were conducted in compliance with the guidelines for Good Clinical Practice and the Declaration of Helsinki. Approval was obtained from local ethics committees or institutional review boards at each center. All patients provided written informed consent.

Trial Treatment

In PSOARING 1 and PSOARING 2, patients were randomized (2:1) to receive tapinarof cream 1% or vehicle QD for 12 weeks. In PSOARING 3 (the long-term extension trial), patients received up to 40 weeks of open-label tapinarof, followed by 4 weeks of follow-up off treatment. Patients received intermittent or continuous treatment with tapinarof cream 1% in PSOARING 3 based on PGA score: those entering the trial with a PGA score of 1 or higher received tapinarof cream 1% until complete disease clearance was achieved (defined as a PGA score of 0 [clear]). Those entering PSOARING 3 with or achieving a PGA score of 0 (clear) discontinued treatment and were observed for the duration of maintenance of a PGA score of 0 (clear) or 1 (almost clear) while off therapy (the protocol-defined “duration of remittive effect”). If disease worsening (defined as a PGA score 2 or higher) occurred, tapinarof cream 1% was restarted and continued until a PGA score of 0 (clear) was achieved. This pattern of treatment, discontinuation on achieving a PGA score of 0 (clear), and retreatment on disease worsening continued until the end of the trial. As a result, patients in PSOARING 3 could receive tapinarof cream 1% continuously or intermittently for 40 weeks.

Outcome Measures and Statistical Analyses

The assessment of total BSA affected by plaque psoriasis is an estimate of the total extent of disease as a percentage of total skin area. In the PSOARING trials, the skin surface of one hand (palm and digits) was assumed to be approximately equivalent to 1% BSA. The total BSA affected by psoriasis was evaluated from 0% to 100%, with greater total BSA affected being an indication of more extensive disease. The BSA efficacy outcomes used in these analyses were based post hoc on the proportion of patients who achieved a 1% or lower or 0.5% or lower total BSA affected. The smallest BSA affected increment that investigators were trained to measure and could record was 0.1%.

Psoriasis Area and Severity Index scores assess both the severity and extent of psoriasis. A PASI score lower than 5 often is considered indicative of mild psoriasis, a score of 5 to 10 indicates moderate disease, and a score higher than 10 indicates severe disease.¹⁹ The maximum PASI score is 72. The PASI efficacy outcomes used in these analyses were based post hoc on the proportion of patients who achieved an absolute total PASI score of 3 or lower, 2 or lower, and 1 or lower.

Efficacy analyses were based on pooled data for all patients in the PSOARING trials who had a PGA score of 2 to 4 (mild to severe) before treatment with tapinarof cream 1% in the intention-to-treat population using observed cases. Time-to-target analyses were based on Kaplan-Meier (KM) estimates using observed cases.

Safety analyses included the incidence and frequency of adverse events and were based on all patients who received tapinarof cream 1% in the PSOARING trials.

RESULTS

Baseline Patient Demographics and Disease Characteristics

The pooled efficacy analyses included 915 eligible patients (Table). At baseline, the mean (SD) age was 50.2 (13.25) years, 58.7% were male, the mean (SD) weight was 92.2 (23.67) kg, and the mean (SD) body mass index was 31.6 (7.53) kg/m². The percentage of patients with a PGA score of 2 (mild), 3 (moderate), or 4 (severe) was 13.9%, 78.1%, and 8.0%, respectively. The mean (SD) PASI score was 8.7 (4.23) and mean (SD) total BSA affected was 7.8% (4.98).

Efficacy

Achievement of BSA-Affected Targets—The NPF-recommended target of 1% or lower total BSA affected within 3 months was achieved by 40% of patients (KM estimate [95% CI, 37%-43%])(Figure 1). Across the total trial period of up to 52 weeks, a total BSA affected of 1% or lower was achieved by 61% of patients (561/915), with the median time to target of approximately 4 months (KM estimate: 120 days [95% CI, 113-141])(Supplementary Figure S2a). Approximately 50% of patients (455/915) achieved a total BSA affected of 0.5% or lower, with a median time to target of 199 days (KM estimate [95% CI, 172-228)(Figure 1; Supplementary Figure S2b).

Achievement of Absolute PASI Targets—Across the total trial period (up to 52 weeks), an absolute total PASI score of 3 or lower was achieved by 75% of patients (686/915), with a median time to achieve this of 2 months (KM estimate: 58 days [95% CI, 57-63]); approximately 67% of patients (612/915) achieved a total PASI score of 2 or lower, with a median time to achieve of 3 months (KM estimate: 87 days [95% CI, 85-110])(Figure 2; Supplementary Figures S3a and S3b). A PASI score of 1 or lower was achieved by approximately 50% of patients (460/915), with a median time to achieve of approximately 6 months (KM estimate: 185 days [95% CI, 169-218])(Figure 2, Supplementary Figure S3c).

Illustrative Case—Case photography showing the clinical response in a 63-year-old man with moderate plaque psoriasis in PSOARING 2 is shown in Figure 3. After 12 weeks of treatment with tapinarof cream 1% QD, the patient achieved all primary and secondary efficacy end points. In addition to achieving the regulatory end point of a PGA score of 0 (clear) or 1 (almost clear) and a decrease from baseline of at least 2 points, achievement of 0% total BSA affected and a total PASI score of 0 at week 12 exceeded the NPF and EADV consensus treatment targets.^10,11 Targets were achieved as early as week 4, with a total BSA affected of 0.5% or lower and a total PASI score of 1 or lower, illustrated by marked skin clearing and only faint residual erythema that completely resolved at week 12, with the absence of postinflammatory hyperpigmentation.

Safety

Safety data for the PSOARING trials have been previously reported.^17,18 The most common treatment-emergent adverse events were folliculitis, contact dermatitis, upper respiratory tract infection, and nasopharyngitis. Treatment-emergent adverse events generally were mild or moderate in severity and did not lead to trial discontinuation.^17,18

COMMENT

Treat-to-target management approaches aim to improve patient outcomes by striving to achieve optimal goals. The treat-to-target approach supports shared decision-making between clinicians and patients based on common expectations of what constitutes treatment success.

The findings of this analysis based on pooled data from a large cohort of patients demonstrate that a high proportion of patients can achieve or exceed recommended treatment targets with tapinarof cream 1% QD and maintain improvements long-term. The NPF-recommended treatment target of 1% or lower BSA affected within approximately 3 months (90 days) of treatment was achieved by 40% of tapinarof-treated patients. In addition, 1% or lower BSA affected at any time during the trials was achieved by 61% of patients (median, approximately 4 months). The analyses also indicated that PASI total scores of 3 or lower and 2 or lower were achieved by 75% and 67% of tapinarof-treated patients, respectively, within 2 to 3 months.

These findings support the previously reported efficacy of tapinarof cream, including high rates of complete disease clearance (40.9% [312/763]), durable response following treatment interruption, an off-therapy remittive effect of approximately 4 months, and good disease control on therapy with no evidence of tachyphylaxis.^17,18

CONCLUSION

Taken together with previously reported tapinarof efficacy and safety results, our findings demonstrate that a high proportion of patients treated with tapinarof cream as monotherapy can achieve aggressive treatment targets set by both US and European guidelines developed for systemic and biologic therapies. Tapinarof cream 1% QD is an effective topical treatment option for patients with plaque psoriasis that has been approved without restrictions relating to severity or extent of disease treated, duration of use, or application sites, including application to sensitive and intertriginous skin.

Acknowledgments—Editorial and medical writing support under the guidance of the authors was provided by Melanie Govender, MSc (Med), ApotheCom (United Kingdom), and was funded by Dermavant Sciences, Inc, in accordance with Good Publication Practice (GPP) guidelines.

Psoriasis is a chronic inflammatory disease affecting approximately 8 million adults in the United States and 2% of the global population.^1,2 Psoriasis causes pain, itching, and disfigurement and is associated with a physical, psychological, and economic burden that substantially affects health-related quality of life.^3-5

Setting treatment goals and treating to target are evidence-based approaches that have been successfully applied to several chronic diseases to improve patient outcomes, including diabetes, hypertension, and rheumatoid arthritis.^6-9 Treat-to-target strategies generally set low disease activity (or remission) as an overall goal and seek to achieve this using available therapeutic options as necessary. Introduced following the availability of biologics and targeted systemic therapies, treat-to-target strategies generally provide guidance on expectations of treatment but not specific treatments, as personalized treatment decisions depend on an assessment of individual patients and consider clinical and demographic features as well as preferences for available therapeutic options. If targets are not achieved in the assigned time span, adjustments can be made to the treatment approach in close consultation with the patient. If the target is reached, follow-up visits can be scheduled to ensure improvement is maintained or to establish if more aggressive goals could be selected.

Treat-to-target strategies for the management of psoriasis developed by the National Psoriasis Foundation (NPF) Medical Board include reducing the extent of psoriasis to 1% or lower total body surface area (BSA) after 3 months of treatment.¹⁰ Treatment targets endorsed by the European Academy of Dermatology and Venereology (EADV) in guidelines on the use of systemic therapies in psoriasis include achieving a 75% or greater reduction in Psoriasis Area and Severity Index (PASI) score within 3 to 4 months of treatment.¹¹

In clinical practice, many patients do not achieve these treatment targets, and topical treatments alone generally are insufficient in achieving treatment goals for psoriasis.^12,13 Moreover, conventional topical treatments (eg, topical corticosteroids) used by most patients with psoriasis regardless of disease severity are associated with adverse events that can limit their use. Most topical corticosteroids have US Food and Drug Administration label restrictions relating to sites of application, duration and extent of use, and frequency of administration.^14,15

Tapinarof cream 1% (VTAMA [Dermavant Sciences, Inc]) is a first-in-class topical nonsteroidal aryl hydrocarbon receptor agonist that was approved by the US Food and Drug Administration for the treatment of plaque psoriasis in adults¹⁶ and is being studied for the treatment of plaque psoriasis in children 2 years and older as well as for atopic dermatitis in adults and children 2 years and older. In PSOARING 1 (ClinicalTrials .gov identifier NCT03956355) and PSOARING 2 (NCT03983980)—identical 12-week pivotal phase 3 trials—monotherapy with tapinarof cream 1% once daily (QD) demonstrated statistically significant efficacy vs vehicle cream and was well tolerated in adults with mild to severe plaque psoriasis (Supplementary Figure S1).¹⁷ Lebwohl et al¹⁷ reported that significantly higher PASI75 responses were observed at week 12 with tapinarof cream vs vehicle in PSOARING 1 and PSOARING 2 (36% and 48% vs 10% and 7%, respectively; both P<.0001). A significantly higher PASI90 response of 19% and 21% at week 12 also was observed with tapinarof cream vs 2% and 3% with vehicle in PSOARING 1 and PSOARING 2, respectively (P=.0005 and P<.0001).¹⁷

In PSOARING 3 (NCT04053387)—the long-term extension trial (Supplementary Figure S1)—efficacy continued to improve or was maintained beyond the two 12-week trials, with improvements in total BSA affected and PASI scores for up to 52 weeks.¹⁸ Tapinarof cream 1% QD demonstrated positive, rapid, and durable outcomes in PSOARING 3, including high rates of complete disease clearance (Physician Global Assessment [PGA] score=0 [clear])(40.9% [312/763]), durability of response on treatment with no evidence of tachyphylaxis, and a remittive effect of approximately 4 months when off therapy (defined as maintenance of a PGA score of 0 [clear] or 1 [almost clear] after first achieving a PGA score of 0).¹⁸

Herein, we report absolute treatment targets for patients with plaque psoriasis who received tapinarof cream 1% QD in the PSOARING trials that are at least as stringent as the corresponding NPF and EADV targets of achieving a total BSA affected of 1% or lower or a PASI75 response within 3 to 4 months, respectively.

METHODS

Study Design

The pooled efficacy analyses included all patients with a baseline PGA score of 2 or higher (mild or worse) before treatment with tapinarof cream 1% QD in the PSOARING trials. This included patients who received tapinarof cream 1% in PSOARING 1 and PSOARING 2 who may or may not have continued into PSOARING 3, as well as those who received the vehicle in PSOARING 1 and PSOARING 2 who enrolled in PSOARING 3 and had a PGA score of 2 or higher before receiving tapinarof cream 1%.

Trial Participants

Full methods, including inclusion and exclusion criteria, for the PSOARING trials have been previously reported.^17,18 Patients were aged 18 to 75 years and had chronic plaque psoriasis that was stable for at least 6 months before randomization; 3% to 20% total BSA affected (excluding the scalp, palms, fingernails, toenails, and soles); and a PGA score of 2 (mild), 3 (moderate), or 4 (severe) at baseline.

The clinical trials were conducted in compliance with the guidelines for Good Clinical Practice and the Declaration of Helsinki. Approval was obtained from local ethics committees or institutional review boards at each center. All patients provided written informed consent.

Trial Treatment

In PSOARING 1 and PSOARING 2, patients were randomized (2:1) to receive tapinarof cream 1% or vehicle QD for 12 weeks. In PSOARING 3 (the long-term extension trial), patients received up to 40 weeks of open-label tapinarof, followed by 4 weeks of follow-up off treatment. Patients received intermittent or continuous treatment with tapinarof cream 1% in PSOARING 3 based on PGA score: those entering the trial with a PGA score of 1 or higher received tapinarof cream 1% until complete disease clearance was achieved (defined as a PGA score of 0 [clear]). Those entering PSOARING 3 with or achieving a PGA score of 0 (clear) discontinued treatment and were observed for the duration of maintenance of a PGA score of 0 (clear) or 1 (almost clear) while off therapy (the protocol-defined “duration of remittive effect”). If disease worsening (defined as a PGA score 2 or higher) occurred, tapinarof cream 1% was restarted and continued until a PGA score of 0 (clear) was achieved. This pattern of treatment, discontinuation on achieving a PGA score of 0 (clear), and retreatment on disease worsening continued until the end of the trial. As a result, patients in PSOARING 3 could receive tapinarof cream 1% continuously or intermittently for 40 weeks.

Outcome Measures and Statistical Analyses

The assessment of total BSA affected by plaque psoriasis is an estimate of the total extent of disease as a percentage of total skin area. In the PSOARING trials, the skin surface of one hand (palm and digits) was assumed to be approximately equivalent to 1% BSA. The total BSA affected by psoriasis was evaluated from 0% to 100%, with greater total BSA affected being an indication of more extensive disease. The BSA efficacy outcomes used in these analyses were based post hoc on the proportion of patients who achieved a 1% or lower or 0.5% or lower total BSA affected. The smallest BSA affected increment that investigators were trained to measure and could record was 0.1%.

Psoriasis Area and Severity Index scores assess both the severity and extent of psoriasis. A PASI score lower than 5 often is considered indicative of mild psoriasis, a score of 5 to 10 indicates moderate disease, and a score higher than 10 indicates severe disease.¹⁹ The maximum PASI score is 72. The PASI efficacy outcomes used in these analyses were based post hoc on the proportion of patients who achieved an absolute total PASI score of 3 or lower, 2 or lower, and 1 or lower.

Efficacy analyses were based on pooled data for all patients in the PSOARING trials who had a PGA score of 2 to 4 (mild to severe) before treatment with tapinarof cream 1% in the intention-to-treat population using observed cases. Time-to-target analyses were based on Kaplan-Meier (KM) estimates using observed cases.

Safety analyses included the incidence and frequency of adverse events and were based on all patients who received tapinarof cream 1% in the PSOARING trials.

RESULTS

Baseline Patient Demographics and Disease Characteristics

The pooled efficacy analyses included 915 eligible patients (Table). At baseline, the mean (SD) age was 50.2 (13.25) years, 58.7% were male, the mean (SD) weight was 92.2 (23.67) kg, and the mean (SD) body mass index was 31.6 (7.53) kg/m². The percentage of patients with a PGA score of 2 (mild), 3 (moderate), or 4 (severe) was 13.9%, 78.1%, and 8.0%, respectively. The mean (SD) PASI score was 8.7 (4.23) and mean (SD) total BSA affected was 7.8% (4.98).

Efficacy

Achievement of BSA-Affected Targets—The NPF-recommended target of 1% or lower total BSA affected within 3 months was achieved by 40% of patients (KM estimate [95% CI, 37%-43%])(Figure 1). Across the total trial period of up to 52 weeks, a total BSA affected of 1% or lower was achieved by 61% of patients (561/915), with the median time to target of approximately 4 months (KM estimate: 120 days [95% CI, 113-141])(Supplementary Figure S2a). Approximately 50% of patients (455/915) achieved a total BSA affected of 0.5% or lower, with a median time to target of 199 days (KM estimate [95% CI, 172-228)(Figure 1; Supplementary Figure S2b).

Achievement of Absolute PASI Targets—Across the total trial period (up to 52 weeks), an absolute total PASI score of 3 or lower was achieved by 75% of patients (686/915), with a median time to achieve this of 2 months (KM estimate: 58 days [95% CI, 57-63]); approximately 67% of patients (612/915) achieved a total PASI score of 2 or lower, with a median time to achieve of 3 months (KM estimate: 87 days [95% CI, 85-110])(Figure 2; Supplementary Figures S3a and S3b). A PASI score of 1 or lower was achieved by approximately 50% of patients (460/915), with a median time to achieve of approximately 6 months (KM estimate: 185 days [95% CI, 169-218])(Figure 2, Supplementary Figure S3c).

Illustrative Case—Case photography showing the clinical response in a 63-year-old man with moderate plaque psoriasis in PSOARING 2 is shown in Figure 3. After 12 weeks of treatment with tapinarof cream 1% QD, the patient achieved all primary and secondary efficacy end points. In addition to achieving the regulatory end point of a PGA score of 0 (clear) or 1 (almost clear) and a decrease from baseline of at least 2 points, achievement of 0% total BSA affected and a total PASI score of 0 at week 12 exceeded the NPF and EADV consensus treatment targets.^10,11 Targets were achieved as early as week 4, with a total BSA affected of 0.5% or lower and a total PASI score of 1 or lower, illustrated by marked skin clearing and only faint residual erythema that completely resolved at week 12, with the absence of postinflammatory hyperpigmentation.

Safety

Safety data for the PSOARING trials have been previously reported.^17,18 The most common treatment-emergent adverse events were folliculitis, contact dermatitis, upper respiratory tract infection, and nasopharyngitis. Treatment-emergent adverse events generally were mild or moderate in severity and did not lead to trial discontinuation.^17,18

COMMENT

Treat-to-target management approaches aim to improve patient outcomes by striving to achieve optimal goals. The treat-to-target approach supports shared decision-making between clinicians and patients based on common expectations of what constitutes treatment success.

The findings of this analysis based on pooled data from a large cohort of patients demonstrate that a high proportion of patients can achieve or exceed recommended treatment targets with tapinarof cream 1% QD and maintain improvements long-term. The NPF-recommended treatment target of 1% or lower BSA affected within approximately 3 months (90 days) of treatment was achieved by 40% of tapinarof-treated patients. In addition, 1% or lower BSA affected at any time during the trials was achieved by 61% of patients (median, approximately 4 months). The analyses also indicated that PASI total scores of 3 or lower and 2 or lower were achieved by 75% and 67% of tapinarof-treated patients, respectively, within 2 to 3 months.

These findings support the previously reported efficacy of tapinarof cream, including high rates of complete disease clearance (40.9% [312/763]), durable response following treatment interruption, an off-therapy remittive effect of approximately 4 months, and good disease control on therapy with no evidence of tachyphylaxis.^17,18

CONCLUSION

Taken together with previously reported tapinarof efficacy and safety results, our findings demonstrate that a high proportion of patients treated with tapinarof cream as monotherapy can achieve aggressive treatment targets set by both US and European guidelines developed for systemic and biologic therapies. Tapinarof cream 1% QD is an effective topical treatment option for patients with plaque psoriasis that has been approved without restrictions relating to severity or extent of disease treated, duration of use, or application sites, including application to sensitive and intertriginous skin.

Acknowledgments—Editorial and medical writing support under the guidance of the authors was provided by Melanie Govender, MSc (Med), ApotheCom (United Kingdom), and was funded by Dermavant Sciences, Inc, in accordance with Good Publication Practice (GPP) guidelines.

Psoriasis is a chronic inflammatory disease affecting approximately 8 million adults in the United States and 2% of the global population.^1,2 Psoriasis causes pain, itching, and disfigurement and is associated with a physical, psychological, and economic burden that substantially affects health-related quality of life.^3-5

Setting treatment goals and treating to target are evidence-based approaches that have been successfully applied to several chronic diseases to improve patient outcomes, including diabetes, hypertension, and rheumatoid arthritis.^6-9 Treat-to-target strategies generally set low disease activity (or remission) as an overall goal and seek to achieve this using available therapeutic options as necessary. Introduced following the availability of biologics and targeted systemic therapies, treat-to-target strategies generally provide guidance on expectations of treatment but not specific treatments, as personalized treatment decisions depend on an assessment of individual patients and consider clinical and demographic features as well as preferences for available therapeutic options. If targets are not achieved in the assigned time span, adjustments can be made to the treatment approach in close consultation with the patient. If the target is reached, follow-up visits can be scheduled to ensure improvement is maintained or to establish if more aggressive goals could be selected.

Treat-to-target strategies for the management of psoriasis developed by the National Psoriasis Foundation (NPF) Medical Board include reducing the extent of psoriasis to 1% or lower total body surface area (BSA) after 3 months of treatment.¹⁰ Treatment targets endorsed by the European Academy of Dermatology and Venereology (EADV) in guidelines on the use of systemic therapies in psoriasis include achieving a 75% or greater reduction in Psoriasis Area and Severity Index (PASI) score within 3 to 4 months of treatment.¹¹

In clinical practice, many patients do not achieve these treatment targets, and topical treatments alone generally are insufficient in achieving treatment goals for psoriasis.^12,13 Moreover, conventional topical treatments (eg, topical corticosteroids) used by most patients with psoriasis regardless of disease severity are associated with adverse events that can limit their use. Most topical corticosteroids have US Food and Drug Administration label restrictions relating to sites of application, duration and extent of use, and frequency of administration.^14,15

Tapinarof cream 1% (VTAMA [Dermavant Sciences, Inc]) is a first-in-class topical nonsteroidal aryl hydrocarbon receptor agonist that was approved by the US Food and Drug Administration for the treatment of plaque psoriasis in adults¹⁶ and is being studied for the treatment of plaque psoriasis in children 2 years and older as well as for atopic dermatitis in adults and children 2 years and older. In PSOARING 1 (ClinicalTrials .gov identifier NCT03956355) and PSOARING 2 (NCT03983980)—identical 12-week pivotal phase 3 trials—monotherapy with tapinarof cream 1% once daily (QD) demonstrated statistically significant efficacy vs vehicle cream and was well tolerated in adults with mild to severe plaque psoriasis (Supplementary Figure S1).¹⁷ Lebwohl et al¹⁷ reported that significantly higher PASI75 responses were observed at week 12 with tapinarof cream vs vehicle in PSOARING 1 and PSOARING 2 (36% and 48% vs 10% and 7%, respectively; both P<.0001). A significantly higher PASI90 response of 19% and 21% at week 12 also was observed with tapinarof cream vs 2% and 3% with vehicle in PSOARING 1 and PSOARING 2, respectively (P=.0005 and P<.0001).¹⁷

In PSOARING 3 (NCT04053387)—the long-term extension trial (Supplementary Figure S1)—efficacy continued to improve or was maintained beyond the two 12-week trials, with improvements in total BSA affected and PASI scores for up to 52 weeks.¹⁸ Tapinarof cream 1% QD demonstrated positive, rapid, and durable outcomes in PSOARING 3, including high rates of complete disease clearance (Physician Global Assessment [PGA] score=0 [clear])(40.9% [312/763]), durability of response on treatment with no evidence of tachyphylaxis, and a remittive effect of approximately 4 months when off therapy (defined as maintenance of a PGA score of 0 [clear] or 1 [almost clear] after first achieving a PGA score of 0).¹⁸

Herein, we report absolute treatment targets for patients with plaque psoriasis who received tapinarof cream 1% QD in the PSOARING trials that are at least as stringent as the corresponding NPF and EADV targets of achieving a total BSA affected of 1% or lower or a PASI75 response within 3 to 4 months, respectively.

METHODS

Study Design

The pooled efficacy analyses included all patients with a baseline PGA score of 2 or higher (mild or worse) before treatment with tapinarof cream 1% QD in the PSOARING trials. This included patients who received tapinarof cream 1% in PSOARING 1 and PSOARING 2 who may or may not have continued into PSOARING 3, as well as those who received the vehicle in PSOARING 1 and PSOARING 2 who enrolled in PSOARING 3 and had a PGA score of 2 or higher before receiving tapinarof cream 1%.

Trial Participants

Full methods, including inclusion and exclusion criteria, for the PSOARING trials have been previously reported.^17,18 Patients were aged 18 to 75 years and had chronic plaque psoriasis that was stable for at least 6 months before randomization; 3% to 20% total BSA affected (excluding the scalp, palms, fingernails, toenails, and soles); and a PGA score of 2 (mild), 3 (moderate), or 4 (severe) at baseline.

The clinical trials were conducted in compliance with the guidelines for Good Clinical Practice and the Declaration of Helsinki. Approval was obtained from local ethics committees or institutional review boards at each center. All patients provided written informed consent.

Trial Treatment

In PSOARING 1 and PSOARING 2, patients were randomized (2:1) to receive tapinarof cream 1% or vehicle QD for 12 weeks. In PSOARING 3 (the long-term extension trial), patients received up to 40 weeks of open-label tapinarof, followed by 4 weeks of follow-up off treatment. Patients received intermittent or continuous treatment with tapinarof cream 1% in PSOARING 3 based on PGA score: those entering the trial with a PGA score of 1 or higher received tapinarof cream 1% until complete disease clearance was achieved (defined as a PGA score of 0 [clear]). Those entering PSOARING 3 with or achieving a PGA score of 0 (clear) discontinued treatment and were observed for the duration of maintenance of a PGA score of 0 (clear) or 1 (almost clear) while off therapy (the protocol-defined “duration of remittive effect”). If disease worsening (defined as a PGA score 2 or higher) occurred, tapinarof cream 1% was restarted and continued until a PGA score of 0 (clear) was achieved. This pattern of treatment, discontinuation on achieving a PGA score of 0 (clear), and retreatment on disease worsening continued until the end of the trial. As a result, patients in PSOARING 3 could receive tapinarof cream 1% continuously or intermittently for 40 weeks.

Outcome Measures and Statistical Analyses

The assessment of total BSA affected by plaque psoriasis is an estimate of the total extent of disease as a percentage of total skin area. In the PSOARING trials, the skin surface of one hand (palm and digits) was assumed to be approximately equivalent to 1% BSA. The total BSA affected by psoriasis was evaluated from 0% to 100%, with greater total BSA affected being an indication of more extensive disease. The BSA efficacy outcomes used in these analyses were based post hoc on the proportion of patients who achieved a 1% or lower or 0.5% or lower total BSA affected. The smallest BSA affected increment that investigators were trained to measure and could record was 0.1%.

Psoriasis Area and Severity Index scores assess both the severity and extent of psoriasis. A PASI score lower than 5 often is considered indicative of mild psoriasis, a score of 5 to 10 indicates moderate disease, and a score higher than 10 indicates severe disease.¹⁹ The maximum PASI score is 72. The PASI efficacy outcomes used in these analyses were based post hoc on the proportion of patients who achieved an absolute total PASI score of 3 or lower, 2 or lower, and 1 or lower.

Efficacy analyses were based on pooled data for all patients in the PSOARING trials who had a PGA score of 2 to 4 (mild to severe) before treatment with tapinarof cream 1% in the intention-to-treat population using observed cases. Time-to-target analyses were based on Kaplan-Meier (KM) estimates using observed cases.

Safety analyses included the incidence and frequency of adverse events and were based on all patients who received tapinarof cream 1% in the PSOARING trials.

RESULTS

Baseline Patient Demographics and Disease Characteristics

The pooled efficacy analyses included 915 eligible patients (Table). At baseline, the mean (SD) age was 50.2 (13.25) years, 58.7% were male, the mean (SD) weight was 92.2 (23.67) kg, and the mean (SD) body mass index was 31.6 (7.53) kg/m². The percentage of patients with a PGA score of 2 (mild), 3 (moderate), or 4 (severe) was 13.9%, 78.1%, and 8.0%, respectively. The mean (SD) PASI score was 8.7 (4.23) and mean (SD) total BSA affected was 7.8% (4.98).

Efficacy

Achievement of BSA-Affected Targets—The NPF-recommended target of 1% or lower total BSA affected within 3 months was achieved by 40% of patients (KM estimate [95% CI, 37%-43%])(Figure 1). Across the total trial period of up to 52 weeks, a total BSA affected of 1% or lower was achieved by 61% of patients (561/915), with the median time to target of approximately 4 months (KM estimate: 120 days [95% CI, 113-141])(Supplementary Figure S2a). Approximately 50% of patients (455/915) achieved a total BSA affected of 0.5% or lower, with a median time to target of 199 days (KM estimate [95% CI, 172-228)(Figure 1; Supplementary Figure S2b).

Achievement of Absolute PASI Targets—Across the total trial period (up to 52 weeks), an absolute total PASI score of 3 or lower was achieved by 75% of patients (686/915), with a median time to achieve this of 2 months (KM estimate: 58 days [95% CI, 57-63]); approximately 67% of patients (612/915) achieved a total PASI score of 2 or lower, with a median time to achieve of 3 months (KM estimate: 87 days [95% CI, 85-110])(Figure 2; Supplementary Figures S3a and S3b). A PASI score of 1 or lower was achieved by approximately 50% of patients (460/915), with a median time to achieve of approximately 6 months (KM estimate: 185 days [95% CI, 169-218])(Figure 2, Supplementary Figure S3c).

Illustrative Case—Case photography showing the clinical response in a 63-year-old man with moderate plaque psoriasis in PSOARING 2 is shown in Figure 3. After 12 weeks of treatment with tapinarof cream 1% QD, the patient achieved all primary and secondary efficacy end points. In addition to achieving the regulatory end point of a PGA score of 0 (clear) or 1 (almost clear) and a decrease from baseline of at least 2 points, achievement of 0% total BSA affected and a total PASI score of 0 at week 12 exceeded the NPF and EADV consensus treatment targets.^10,11 Targets were achieved as early as week 4, with a total BSA affected of 0.5% or lower and a total PASI score of 1 or lower, illustrated by marked skin clearing and only faint residual erythema that completely resolved at week 12, with the absence of postinflammatory hyperpigmentation.

Safety

Safety data for the PSOARING trials have been previously reported.^17,18 The most common treatment-emergent adverse events were folliculitis, contact dermatitis, upper respiratory tract infection, and nasopharyngitis. Treatment-emergent adverse events generally were mild or moderate in severity and did not lead to trial discontinuation.^17,18

COMMENT

Treat-to-target management approaches aim to improve patient outcomes by striving to achieve optimal goals. The treat-to-target approach supports shared decision-making between clinicians and patients based on common expectations of what constitutes treatment success.

The findings of this analysis based on pooled data from a large cohort of patients demonstrate that a high proportion of patients can achieve or exceed recommended treatment targets with tapinarof cream 1% QD and maintain improvements long-term. The NPF-recommended treatment target of 1% or lower BSA affected within approximately 3 months (90 days) of treatment was achieved by 40% of tapinarof-treated patients. In addition, 1% or lower BSA affected at any time during the trials was achieved by 61% of patients (median, approximately 4 months). The analyses also indicated that PASI total scores of 3 or lower and 2 or lower were achieved by 75% and 67% of tapinarof-treated patients, respectively, within 2 to 3 months.

These findings support the previously reported efficacy of tapinarof cream, including high rates of complete disease clearance (40.9% [312/763]), durable response following treatment interruption, an off-therapy remittive effect of approximately 4 months, and good disease control on therapy with no evidence of tachyphylaxis.^17,18

CONCLUSION

Taken together with previously reported tapinarof efficacy and safety results, our findings demonstrate that a high proportion of patients treated with tapinarof cream as monotherapy can achieve aggressive treatment targets set by both US and European guidelines developed for systemic and biologic therapies. Tapinarof cream 1% QD is an effective topical treatment option for patients with plaque psoriasis that has been approved without restrictions relating to severity or extent of disease treated, duration of use, or application sites, including application to sensitive and intertriginous skin.

Acknowledgments—Editorial and medical writing support under the guidance of the authors was provided by Melanie Govender, MSc (Med), ApotheCom (United Kingdom), and was funded by Dermavant Sciences, Inc, in accordance with Good Publication Practice (GPP) guidelines.

To the Editor:

Skin cancers are extremely common worldwide. Malignant melanomas comprise approximately 1 in 5 of these cancers. Exposure to UV radiation is postulated to be responsible for a global rise in melanoma cases over the past 50 years.¹ Pediatric melanoma is a particularly rare condition that affects approximately 6 in every 1 million children.² Melanoma incidence in children ranges by age, increasing by approximately 10-fold from age 1 to 4 years to age 15 to 19 years. Tumor ulceration is a feature more commonly seen among children younger than 10 years and is associated with worse outcomes. Tumor thickness and ulceration strongly predict sentinel lymph node metastases among children, which also is associated with a poor prognosis.³

A recent study evaluating stage IV melanoma survival rates in adolescents and young adults (AYAs) vs older adults found that survival is much worse among AYAs. Thicker tumors and public health insurance also were associated with worse survival rates for AYAs, while early detection was associated with better survival rates.⁴

Health disparities and their role in the prognosis of pediatric melanoma is another important factor. One study analyzed this relationship at the state level using Texas Cancer Registry data (1995-2009).⁵ Patients’ socioeconomic status (SES) and driving distance to the nearest pediatric cancer care center were included in the analysis. Hispanic children were found to be 3 times more likely to present with advanced disease than non-Hispanic White children. Although SES and distance to the nearest treatment center were not found to affect the melanoma stage at presentation, Hispanic ethnicity or being in the lowest SES quartile were correlated with a higher mortality risk.⁵

When considering specific subtypes of melanoma, acral lentiginous melanoma (ALM) is known to develop in patients with skin of color. A 2023 study by Holman et al⁶ reported that the percentage of melanomas that were ALMs ranged from 0.8% in non-Hispanic White individuals to 19.1% in Hispanic Black, American Indian/Alaska Native, and Asian/Pacific Islander individuals. However, ALM is rare in children. In a pooled cohort study with patient information retrieved from the nationwide Dutch Pathology Registry, only 1 child and 1 adolescent were found to have ALM across a total of 514 patients.⁷ We sought to analyze pediatric melanoma outcomes based on race and other barriers to appropriate care.

We conducted a search of the Surveillance, Epidemiology, and End Results (SEER) database from January 1995 to December 2016 for patients aged 21 years and younger with a primary melanoma diagnosis. The primary outcome was the 5-year survival rate. County-level SES variables were used to calculate a prosperity index. Kaplan-Meier analysis and Cox proportional hazards model were used to compare 5-year survival rates among the different racial/ethnic groups.

A sample of 2742 patients was identified during the study period and followed for 5 years. Eighty-two percent were White, 6% Hispanic, 2% Asian, 1% Black, and 5% classified as other/unknown race (data were missing for 4%). The cohort was predominantly female (61%). White patients were more likely to present with localized disease than any other race/ethnicity (83% vs 65% in Hispanic, 60% in Asian/Pacific Islander, and 45% in Black patients [P<.05]).

Black and Hispanic patients had the worst 5-year survival rates on bivariate analysis. On multivariate analysis, this finding remained significant for Hispanic patients when compared with White patients (hazard ratio, 2.37 [P<.05]). Increasing age, male sex, advanced stage at diagnosis, and failure to receive surgery were associated with increased odds of mortality.

Patients with regionalized and disseminated disease had increased odds of mortality (6.16 and 64.45, respectively; P<.05) compared with patients with localized disease. Socioeconomic status and urbanization were not found to influence 5-year survival rates.

Pediatric melanoma often presents a clinical challenge with special considerations. Pediatric-specific predisposing risk factors for melanoma and an atypical clinical presentation are some of the major concerns that necessitate a tailored approach to this malignancy, especially among different age groups, skin types, and racial and socioeconomic groups.⁵

Standard ABCDE criteria often are inadequate for accurate detection of pediatric melanomas. Initial lesions often manifest as raised, red, amelanotic lesions mimicking pyogenic granulomas. Lesions tend to be very small (<6 mm in diameter) and can be uniform in color, thereby making the melanoma more difficult to detect compared to the characteristic findings in adults.⁵ Bleeding or ulceration often can be a warning sign during physical examination.

With regard to incidence, pediatric melanoma is relatively rare. Since the 1970s, the incidence of pediatric melanoma has been increasing; however, a recent analysis of the SEER database showed a decreasing trend from 2000 to 2010.⁴

Our analysis of the SEER data showed an increased risk for pediatric melanoma in older adolescents. In addition, the incidence of pediatric melanoma was higher in females of all racial groups except Asian/Pacific Islander individuals. However, SES was not found to significantly influence the 5-year survival rate in pediatric melanoma.

White pediatric patients were more likely to present with localized disease compared with other races. Pediatric melanoma patients with regional disease had a 6-fold increase in mortality rate vs those with localized disease; those with disseminated disease had a 65-fold higher risk. Consistent with this, Black and Hispanic patients had the worst 5-year survival rates on bivariate analysis.

These findings suggest a relationship between race, melanoma spread, and disease severity. Patient education programs need to be directed specifically to minority groups to improve their knowledge on evolving skin lesions and sun protection practices. Physicians also need to have heightened suspicion and better knowledge of the unique traits of pediatric melanoma.⁵

Given the considerable influence these disparities can have on melanoma outcomes, further research is needed to characterize outcomes based on race and determine obstacles to appropriate care. Improved public outreach initiatives that accommodate specific cultural barriers (eg, language, traditional patterns of behavior) also are required to improve current circumstances.

To the Editor:

Skin cancers are extremely common worldwide. Malignant melanomas comprise approximately 1 in 5 of these cancers. Exposure to UV radiation is postulated to be responsible for a global rise in melanoma cases over the past 50 years.¹ Pediatric melanoma is a particularly rare condition that affects approximately 6 in every 1 million children.² Melanoma incidence in children ranges by age, increasing by approximately 10-fold from age 1 to 4 years to age 15 to 19 years. Tumor ulceration is a feature more commonly seen among children younger than 10 years and is associated with worse outcomes. Tumor thickness and ulceration strongly predict sentinel lymph node metastases among children, which also is associated with a poor prognosis.³

A recent study evaluating stage IV melanoma survival rates in adolescents and young adults (AYAs) vs older adults found that survival is much worse among AYAs. Thicker tumors and public health insurance also were associated with worse survival rates for AYAs, while early detection was associated with better survival rates.⁴

Health disparities and their role in the prognosis of pediatric melanoma is another important factor. One study analyzed this relationship at the state level using Texas Cancer Registry data (1995-2009).⁵ Patients’ socioeconomic status (SES) and driving distance to the nearest pediatric cancer care center were included in the analysis. Hispanic children were found to be 3 times more likely to present with advanced disease than non-Hispanic White children. Although SES and distance to the nearest treatment center were not found to affect the melanoma stage at presentation, Hispanic ethnicity or being in the lowest SES quartile were correlated with a higher mortality risk.⁵

When considering specific subtypes of melanoma, acral lentiginous melanoma (ALM) is known to develop in patients with skin of color. A 2023 study by Holman et al⁶ reported that the percentage of melanomas that were ALMs ranged from 0.8% in non-Hispanic White individuals to 19.1% in Hispanic Black, American Indian/Alaska Native, and Asian/Pacific Islander individuals. However, ALM is rare in children. In a pooled cohort study with patient information retrieved from the nationwide Dutch Pathology Registry, only 1 child and 1 adolescent were found to have ALM across a total of 514 patients.⁷ We sought to analyze pediatric melanoma outcomes based on race and other barriers to appropriate care.

We conducted a search of the Surveillance, Epidemiology, and End Results (SEER) database from January 1995 to December 2016 for patients aged 21 years and younger with a primary melanoma diagnosis. The primary outcome was the 5-year survival rate. County-level SES variables were used to calculate a prosperity index. Kaplan-Meier analysis and Cox proportional hazards model were used to compare 5-year survival rates among the different racial/ethnic groups.

A sample of 2742 patients was identified during the study period and followed for 5 years. Eighty-two percent were White, 6% Hispanic, 2% Asian, 1% Black, and 5% classified as other/unknown race (data were missing for 4%). The cohort was predominantly female (61%). White patients were more likely to present with localized disease than any other race/ethnicity (83% vs 65% in Hispanic, 60% in Asian/Pacific Islander, and 45% in Black patients [P<.05]).

Black and Hispanic patients had the worst 5-year survival rates on bivariate analysis. On multivariate analysis, this finding remained significant for Hispanic patients when compared with White patients (hazard ratio, 2.37 [P<.05]). Increasing age, male sex, advanced stage at diagnosis, and failure to receive surgery were associated with increased odds of mortality.

Patients with regionalized and disseminated disease had increased odds of mortality (6.16 and 64.45, respectively; P<.05) compared with patients with localized disease. Socioeconomic status and urbanization were not found to influence 5-year survival rates.

Pediatric melanoma often presents a clinical challenge with special considerations. Pediatric-specific predisposing risk factors for melanoma and an atypical clinical presentation are some of the major concerns that necessitate a tailored approach to this malignancy, especially among different age groups, skin types, and racial and socioeconomic groups.⁵

Standard ABCDE criteria often are inadequate for accurate detection of pediatric melanomas. Initial lesions often manifest as raised, red, amelanotic lesions mimicking pyogenic granulomas. Lesions tend to be very small (<6 mm in diameter) and can be uniform in color, thereby making the melanoma more difficult to detect compared to the characteristic findings in adults.⁵ Bleeding or ulceration often can be a warning sign during physical examination.

With regard to incidence, pediatric melanoma is relatively rare. Since the 1970s, the incidence of pediatric melanoma has been increasing; however, a recent analysis of the SEER database showed a decreasing trend from 2000 to 2010.⁴

Our analysis of the SEER data showed an increased risk for pediatric melanoma in older adolescents. In addition, the incidence of pediatric melanoma was higher in females of all racial groups except Asian/Pacific Islander individuals. However, SES was not found to significantly influence the 5-year survival rate in pediatric melanoma.

White pediatric patients were more likely to present with localized disease compared with other races. Pediatric melanoma patients with regional disease had a 6-fold increase in mortality rate vs those with localized disease; those with disseminated disease had a 65-fold higher risk. Consistent with this, Black and Hispanic patients had the worst 5-year survival rates on bivariate analysis.

These findings suggest a relationship between race, melanoma spread, and disease severity. Patient education programs need to be directed specifically to minority groups to improve their knowledge on evolving skin lesions and sun protection practices. Physicians also need to have heightened suspicion and better knowledge of the unique traits of pediatric melanoma.⁵

Given the considerable influence these disparities can have on melanoma outcomes, further research is needed to characterize outcomes based on race and determine obstacles to appropriate care. Improved public outreach initiatives that accommodate specific cultural barriers (eg, language, traditional patterns of behavior) also are required to improve current circumstances.

To the Editor:

Skin cancers are extremely common worldwide. Malignant melanomas comprise approximately 1 in 5 of these cancers. Exposure to UV radiation is postulated to be responsible for a global rise in melanoma cases over the past 50 years.¹ Pediatric melanoma is a particularly rare condition that affects approximately 6 in every 1 million children.² Melanoma incidence in children ranges by age, increasing by approximately 10-fold from age 1 to 4 years to age 15 to 19 years. Tumor ulceration is a feature more commonly seen among children younger than 10 years and is associated with worse outcomes. Tumor thickness and ulceration strongly predict sentinel lymph node metastases among children, which also is associated with a poor prognosis.³

A recent study evaluating stage IV melanoma survival rates in adolescents and young adults (AYAs) vs older adults found that survival is much worse among AYAs. Thicker tumors and public health insurance also were associated with worse survival rates for AYAs, while early detection was associated with better survival rates.⁴

Health disparities and their role in the prognosis of pediatric melanoma is another important factor. One study analyzed this relationship at the state level using Texas Cancer Registry data (1995-2009).⁵ Patients’ socioeconomic status (SES) and driving distance to the nearest pediatric cancer care center were included in the analysis. Hispanic children were found to be 3 times more likely to present with advanced disease than non-Hispanic White children. Although SES and distance to the nearest treatment center were not found to affect the melanoma stage at presentation, Hispanic ethnicity or being in the lowest SES quartile were correlated with a higher mortality risk.⁵

When considering specific subtypes of melanoma, acral lentiginous melanoma (ALM) is known to develop in patients with skin of color. A 2023 study by Holman et al⁶ reported that the percentage of melanomas that were ALMs ranged from 0.8% in non-Hispanic White individuals to 19.1% in Hispanic Black, American Indian/Alaska Native, and Asian/Pacific Islander individuals. However, ALM is rare in children. In a pooled cohort study with patient information retrieved from the nationwide Dutch Pathology Registry, only 1 child and 1 adolescent were found to have ALM across a total of 514 patients.⁷ We sought to analyze pediatric melanoma outcomes based on race and other barriers to appropriate care.

We conducted a search of the Surveillance, Epidemiology, and End Results (SEER) database from January 1995 to December 2016 for patients aged 21 years and younger with a primary melanoma diagnosis. The primary outcome was the 5-year survival rate. County-level SES variables were used to calculate a prosperity index. Kaplan-Meier analysis and Cox proportional hazards model were used to compare 5-year survival rates among the different racial/ethnic groups.

A sample of 2742 patients was identified during the study period and followed for 5 years. Eighty-two percent were White, 6% Hispanic, 2% Asian, 1% Black, and 5% classified as other/unknown race (data were missing for 4%). The cohort was predominantly female (61%). White patients were more likely to present with localized disease than any other race/ethnicity (83% vs 65% in Hispanic, 60% in Asian/Pacific Islander, and 45% in Black patients [P<.05]).

Black and Hispanic patients had the worst 5-year survival rates on bivariate analysis. On multivariate analysis, this finding remained significant for Hispanic patients when compared with White patients (hazard ratio, 2.37 [P<.05]). Increasing age, male sex, advanced stage at diagnosis, and failure to receive surgery were associated with increased odds of mortality.

Patients with regionalized and disseminated disease had increased odds of mortality (6.16 and 64.45, respectively; P<.05) compared with patients with localized disease. Socioeconomic status and urbanization were not found to influence 5-year survival rates.

Pediatric melanoma often presents a clinical challenge with special considerations. Pediatric-specific predisposing risk factors for melanoma and an atypical clinical presentation are some of the major concerns that necessitate a tailored approach to this malignancy, especially among different age groups, skin types, and racial and socioeconomic groups.⁵

Standard ABCDE criteria often are inadequate for accurate detection of pediatric melanomas. Initial lesions often manifest as raised, red, amelanotic lesions mimicking pyogenic granulomas. Lesions tend to be very small (<6 mm in diameter) and can be uniform in color, thereby making the melanoma more difficult to detect compared to the characteristic findings in adults.⁵ Bleeding or ulceration often can be a warning sign during physical examination.

With regard to incidence, pediatric melanoma is relatively rare. Since the 1970s, the incidence of pediatric melanoma has been increasing; however, a recent analysis of the SEER database showed a decreasing trend from 2000 to 2010.⁴

Our analysis of the SEER data showed an increased risk for pediatric melanoma in older adolescents. In addition, the incidence of pediatric melanoma was higher in females of all racial groups except Asian/Pacific Islander individuals. However, SES was not found to significantly influence the 5-year survival rate in pediatric melanoma.

White pediatric patients were more likely to present with localized disease compared with other races. Pediatric melanoma patients with regional disease had a 6-fold increase in mortality rate vs those with localized disease; those with disseminated disease had a 65-fold higher risk. Consistent with this, Black and Hispanic patients had the worst 5-year survival rates on bivariate analysis.

These findings suggest a relationship between race, melanoma spread, and disease severity. Patient education programs need to be directed specifically to minority groups to improve their knowledge on evolving skin lesions and sun protection practices. Physicians also need to have heightened suspicion and better knowledge of the unique traits of pediatric melanoma.⁵

Given the considerable influence these disparities can have on melanoma outcomes, further research is needed to characterize outcomes based on race and determine obstacles to appropriate care. Improved public outreach initiatives that accommodate specific cultural barriers (eg, language, traditional patterns of behavior) also are required to improve current circumstances.

Colorectal cancer (CRC) is the third-most diagnosed cancer after breast and lung cancer, and is the second leading cause of global cancer related deaths.¹ In 2023 in the United States, > 150,000 individuals were diagnosed with CRC and 52,000 died.²

Colonoscopy is an effective CRC screening method and the lone method recommended for polyp surveillance. Inadequate bowel preparation (IBP) has been estimated to occur in about 6% to 26% of colonoscopies. ^3,4 The prevalence varies based on a variety of comorbidities, including immobility, diabetes mellitus, neurologic disorders, and use of opioids, with more occurrences of IBP noted in older adult, non-English speaking, and male individuals.^4-6

The quality of bowel preparation is integral to the effectiveness of screening and surveillance colonoscopies. IBP has been associated with missed adenomas and significantly lower adenoma detection rates.^7-9 In particular, IBP is independently associated with an increased risk of CRC in the future.³ Accordingly, the US Multisociety Task Force recommends repeat colonoscopies for individuals with IBP within 1 year.¹⁰ Ensuring that these individuals receive repeat colonoscopies is an essential part of CRC prevention. The benefit of repeat colonoscopy after IBP is highlighted by a retrospective analysis from Fung and colleagues that showed 81% of repeat colonoscopies had adequate bowel preparation, with higher numbers of adenomas detected on repeat compared to initial colonoscopies.¹¹

Given the impact of bowel preparation quality on the diagnostic capability of the colonoscopy, adherence to guidelines for repeat colonoscopies in cases of IBP is paramount for effective CRC prevention. This study aims to measure the frequency of repeat colonoscopy after IBP and the factors associated with adherence to recommendations.

METHODS

Individuals who underwent colonoscopy at the Minneapolis Veterans Affairs Medical Center (MVAMC) from January 1, 2016, to October 19, 2021, were identified to allow for 400 days of follow-up from the index colonoscopy to the data collection date. During the COVID-19 pandemic, the colonoscopy procedure capacity was reduced by 50% from June 1, 2020, to December 1, 2020, delaying nonurgent procedures, including screening and surveillance colonoscopies.

Individuals who underwent colonoscopy for CRC screening or polyp surveillance, or following a positive fecal immunohistochemistry test (FIT) or virtual computed tomography colonoscopy were included. Patients with colonoscopy indications for iron deficiency anemia, gastrointestinal bleeding, disease activity assessment of inflammatory bowel disease, abdominal pain, or changes in bowel movement pattern were excluded. IBP was defined as recording a Boston Bowel Preparation Scale (BBPS) score of < 6, or < 2 in any segment, or described as poor or inadequate using the Aronchick scale.

Age, sex, race, marital status, distance to MVAMC, smoking status, comorbidities, and concurrent medication use, including antiplatelet, anticoagulation, and prescription opiates at the time of index colonoscopy were obtained from the Veterans Health Administration (VHA) Corporate Data Warehouse (CDW) using structured query language processing of colonoscopy procedure notes to extract preparation scores and other procedure information. The CDW contains extracts from VHA clinical and administrative systems that contain complete clinical data from October 1999.¹² Current smoking status was defined as any smoking activity at the time the questionnaire was administered during a routine clinic visit within 400 days from the index colonoscopy.

Only individuals who were recommended to have repeat colonoscopy within 1 year were included. The intervals of 365 days and 400 days (1 year + about 1 additional month) were used in the event that the individual had a delay in scheduling their 1-year repeat colonoscopy. For individuals who did not undergo a colonoscopy at MVAMC within 400 days, a manual chart review of all available records was performed to determine whether a colonoscopy was performed at a non-VA facility.

Patients received written instructions for bowel preparation 2 weeks prior to the procedure. The preparation included magnesium citrate and a split dose of 4 liters of polyethylene glycol. Patients were also advised to start a low-fiber diet 3 days prior to the procedure and a clear liquid diet the day before the procedure. Patients with a history of IBP or those undergoing procedures with anesthesia received an additional 2 liters for a total of 6 liters of polyethylene glycol.

Statistical analysis

Baseline characteristics were reported as mean (SD) or median and IQR for continuous variables and percentage for categorical variables. Individuals who returned for colonoscopy within 400 days were compared to those who did not identify factors associated with adherence to recommendations. The data on individuals who returned for colonoscopy within 400 days were also analyzed for additional minor delays in the timing of the repeat colonoscopy. Continuous data were compared using Mann-Whitney U tests. Categorical data were compared using X² or Fisher exact tests. Missing data were imputed from the analyses. All analyses were performed using SAS JMP Pro version 16. P < .05 was considered statistically significant.

RESULTS

There were 18,241 total colonoscopies performed between January 1, 2016, to October 19, 2021, and 13,818 colonoscopies had indications for screening for colon cancer, positive FIT, virtual colonoscopy, or surveillance. Of the 10,466 unique patients there were 5369 patients for polyp surveillance, 4054 patients for CRC screening, and 1043 patients for positive FIT or virtual colonoscopy. Of these, 571 individuals (5.5%) had IBP. Repeat colonoscopy within 1 year was recommended for 485 individuals (84.9%) who were included in this study (153 CRC screenings and 46 positive FITs) but not for 86 individuals (15.1%) (Figure 1). Among included patients, the mean (SD) age was 66.6 (7.2) years, and the majority were male (460 [94.8%]) and White (435 [89.7%]) (Table). Two hundred and forty-three (50.1%) were married.

Adherence to Recommended Interval Colonoscopy

Of the 485 patients with IBP who were recommended for follow-up colonoscopy, 287 (59.2%) had a colonoscopy within 1 year, and 198 (40.8%) did not; 17 patients (13.5%) had repeat colonoscopy within 366 to 400 days. Five (1.0%) individuals had a repeat colonoscopy the next day, and 77 (15.9%) had a repeat colonoscopy within 7 days. One hundred and twentysix (26.0%) individuals underwent no repeat colonoscopy during the study period (Figure 2).

To account for the COVID-19 pandemic, the adherence rate of repeat colonoscopy within 1 year prepandemic (January 1, 2016, to December 1, 2018) was calculated along with the adherence rate postpandemic (January 1, 2019 to the end of the study). The rates were similar: 199 of 330 (60.3%) individuals prepandemic vs 88 of 155 (56.8%) individuals postpandemic (Figure 3).

Significant Associations

Age, sex, and race were not associated with adherence to repeat colonoscopy within 1 year. Individuals living ≤ 40 miles from the endoscopy center were more likely to undergo a repeat colonoscopy within 1 year compared with those who lived > 40 miles away (61.7% vs 51.0%, P = .02). Current smoking status was associated with a lower rate of repeat colonoscopy within 1 year (25.8% vs 35.9%; P = .02). There were no differences with respect to inflammatory bowel disease diagnosis, mental health diagnosis, diabetes mellitus, cirrhosis, or medications used, including opioids, anticoagulation, and antiplatelet therapy.

Outcomes

Among individuals who had a repeat colonoscopy the day after the index colonoscopy, 53 of 56 individuals (94.6%) had adequate bowel preparation. Among individuals who had a repeat colonoscopy within 7 days, 70 of 77 (90.9%) had adequate bowel preparation. Of 287 individuals with a repeat colonoscopy within 1 year, 251 (87.5%) had adequate bowel preparation on the repeat colonoscopy. By 400 days after the index colonoscopy, 268 of 304 individuals (88.2%) had adequate bowel preparation.

In this study conducted at a large VA medical center, we found that 5.6% of individuals undergoing colonoscopies had IBP, a rate comparable to prior studies (6% to 26%).^3,4 Only 59.2% of individuals underwent repeat colonoscopies within 1 year, as recommended after an index colonoscopy with IBP. Smoking and living longer distances (> 40 miles) from the endoscopy center were associated with a decreased adherence to the repeat colonoscopy recommendation.

Current guidelines recommend repeat colonoscopy for individuals with IBP within 1 year.¹⁰ In cases of IBP, the advanced adenoma miss rate is 36% upon repeat colonoscopy within 1 year.¹³ Despite the importance of a follow-up colonoscopy, clinician adherence with this recommendation remains low.^10,14,15 However, in this study cohort, 485 of 571 individuals with IBP (84.9%) received recommendations for a repeat colonoscopy within 1 year. In the US, only 31.9% of 260,314 colonoscopies with IBP included recommendations for a follow-up colonoscopy within 1 year.¹⁴ This could be related to variations in endoscopist practice as well as patient risk factors for developing polyps, including family history of cancer and personal history of prior polyps. The findings of multiple polyps, high-risk adenomas, and cancer on the index colonoscopy also influences the endoscopist for repeat colonoscopy within 1 year.¹⁴

The timing for repeat colonoscopies within 1 year will be determined by the patients, clinicians, and available scheduling. In this study, the earlier repeat colonoscopies, especially those occurring the day after the index colonoscopy, had the highest success rate of adequate bowel preparation. In a prior study, repeating colonoscopies within the same day or the next day was also found to have a higher rate of adequate bowel preparation than repeat colonoscopies within 1 year (88.9% vs 83.5%).¹⁶

Ensuring the return of individuals with IBP for repeat colonoscopy is a challenging task. We identified that individuals who live further away from MVAMC and current smokers had a decreased probability of returning for a repeat colonoscopy. Toro and colleagues found a 68.7% return rate for a repeat colonoscopy within 1 year with individuals age ≥ 60 years, and patients who were White were less likely to proceed with a repeat colonoscopy within 1 year.¹⁷ The study did not provide data regarding smoking status or distance to the endoscopy center.¹⁷ In a prior study of veterans, the dual diagnosis of psychiatric disorders and substance abuse was associated with missed and canceled colonoscopy appointments.¹⁸ The distance to the endoscopy center has also been previously identified as a barrier to a colonoscopy following an abnormal FIT.¹⁹ Although not identified in this study due to the homogenous demographic profile, social determinants of health such as socioeconomic status, education, and insurance coverage are known barriers to cancer screening but were not evaluated in this study.²⁰

Based on the identified risk factors, we have created a model for utilizing those risk factors to identify individuals at higher risk for noncompliance (ie, those who live further away from the endoscopy center or currently smoke). These individuals are proactively offered to use an intraprocedural bowel cleansing device to achieve adequate bowel preparation or priority rescheduling for a next-day colonoscopy.

Limitations

This study was a single-center study of the veteran population, which is predominantly White and male, thus limiting generalizability. The study is also limited by minimal available data on adenoma detection and colon cancer incidence on subsequent colonoscopies.

CONCLUSIONS

The rate of IBP was 5.5% in individuals undergoing colonoscopy for colon cancer screening, surveillance, positive FIT, or computed tomography colonography. Only 59.2% of those with IBP underwent the recommended repeat colonoscopy within 1 year. Smoking and distance to the endoscopy center were associated with a decreased adherence to the repeat colonoscopy recommendation. Additional efforts are needed to ensure that individuals with IBP return for timely repeat colonoscopy.

Colorectal cancer (CRC) is the third-most diagnosed cancer after breast and lung cancer, and is the second leading cause of global cancer related deaths.¹ In 2023 in the United States, > 150,000 individuals were diagnosed with CRC and 52,000 died.²

Colonoscopy is an effective CRC screening method and the lone method recommended for polyp surveillance. Inadequate bowel preparation (IBP) has been estimated to occur in about 6% to 26% of colonoscopies. ^3,4 The prevalence varies based on a variety of comorbidities, including immobility, diabetes mellitus, neurologic disorders, and use of opioids, with more occurrences of IBP noted in older adult, non-English speaking, and male individuals.^4-6

The quality of bowel preparation is integral to the effectiveness of screening and surveillance colonoscopies. IBP has been associated with missed adenomas and significantly lower adenoma detection rates.^7-9 In particular, IBP is independently associated with an increased risk of CRC in the future.³ Accordingly, the US Multisociety Task Force recommends repeat colonoscopies for individuals with IBP within 1 year.¹⁰ Ensuring that these individuals receive repeat colonoscopies is an essential part of CRC prevention. The benefit of repeat colonoscopy after IBP is highlighted by a retrospective analysis from Fung and colleagues that showed 81% of repeat colonoscopies had adequate bowel preparation, with higher numbers of adenomas detected on repeat compared to initial colonoscopies.¹¹

Given the impact of bowel preparation quality on the diagnostic capability of the colonoscopy, adherence to guidelines for repeat colonoscopies in cases of IBP is paramount for effective CRC prevention. This study aims to measure the frequency of repeat colonoscopy after IBP and the factors associated with adherence to recommendations.

METHODS

Individuals who underwent colonoscopy at the Minneapolis Veterans Affairs Medical Center (MVAMC) from January 1, 2016, to October 19, 2021, were identified to allow for 400 days of follow-up from the index colonoscopy to the data collection date. During the COVID-19 pandemic, the colonoscopy procedure capacity was reduced by 50% from June 1, 2020, to December 1, 2020, delaying nonurgent procedures, including screening and surveillance colonoscopies.

Individuals who underwent colonoscopy for CRC screening or polyp surveillance, or following a positive fecal immunohistochemistry test (FIT) or virtual computed tomography colonoscopy were included. Patients with colonoscopy indications for iron deficiency anemia, gastrointestinal bleeding, disease activity assessment of inflammatory bowel disease, abdominal pain, or changes in bowel movement pattern were excluded. IBP was defined as recording a Boston Bowel Preparation Scale (BBPS) score of < 6, or < 2 in any segment, or described as poor or inadequate using the Aronchick scale.

Age, sex, race, marital status, distance to MVAMC, smoking status, comorbidities, and concurrent medication use, including antiplatelet, anticoagulation, and prescription opiates at the time of index colonoscopy were obtained from the Veterans Health Administration (VHA) Corporate Data Warehouse (CDW) using structured query language processing of colonoscopy procedure notes to extract preparation scores and other procedure information. The CDW contains extracts from VHA clinical and administrative systems that contain complete clinical data from October 1999.¹² Current smoking status was defined as any smoking activity at the time the questionnaire was administered during a routine clinic visit within 400 days from the index colonoscopy.

Only individuals who were recommended to have repeat colonoscopy within 1 year were included. The intervals of 365 days and 400 days (1 year + about 1 additional month) were used in the event that the individual had a delay in scheduling their 1-year repeat colonoscopy. For individuals who did not undergo a colonoscopy at MVAMC within 400 days, a manual chart review of all available records was performed to determine whether a colonoscopy was performed at a non-VA facility.

Patients received written instructions for bowel preparation 2 weeks prior to the procedure. The preparation included magnesium citrate and a split dose of 4 liters of polyethylene glycol. Patients were also advised to start a low-fiber diet 3 days prior to the procedure and a clear liquid diet the day before the procedure. Patients with a history of IBP or those undergoing procedures with anesthesia received an additional 2 liters for a total of 6 liters of polyethylene glycol.

Statistical analysis

Baseline characteristics were reported as mean (SD) or median and IQR for continuous variables and percentage for categorical variables. Individuals who returned for colonoscopy within 400 days were compared to those who did not identify factors associated with adherence to recommendations. The data on individuals who returned for colonoscopy within 400 days were also analyzed for additional minor delays in the timing of the repeat colonoscopy. Continuous data were compared using Mann-Whitney U tests. Categorical data were compared using X² or Fisher exact tests. Missing data were imputed from the analyses. All analyses were performed using SAS JMP Pro version 16. P < .05 was considered statistically significant.

RESULTS

There were 18,241 total colonoscopies performed between January 1, 2016, to October 19, 2021, and 13,818 colonoscopies had indications for screening for colon cancer, positive FIT, virtual colonoscopy, or surveillance. Of the 10,466 unique patients there were 5369 patients for polyp surveillance, 4054 patients for CRC screening, and 1043 patients for positive FIT or virtual colonoscopy. Of these, 571 individuals (5.5%) had IBP. Repeat colonoscopy within 1 year was recommended for 485 individuals (84.9%) who were included in this study (153 CRC screenings and 46 positive FITs) but not for 86 individuals (15.1%) (Figure 1). Among included patients, the mean (SD) age was 66.6 (7.2) years, and the majority were male (460 [94.8%]) and White (435 [89.7%]) (Table). Two hundred and forty-three (50.1%) were married.

Adherence to Recommended Interval Colonoscopy

Of the 485 patients with IBP who were recommended for follow-up colonoscopy, 287 (59.2%) had a colonoscopy within 1 year, and 198 (40.8%) did not; 17 patients (13.5%) had repeat colonoscopy within 366 to 400 days. Five (1.0%) individuals had a repeat colonoscopy the next day, and 77 (15.9%) had a repeat colonoscopy within 7 days. One hundred and twentysix (26.0%) individuals underwent no repeat colonoscopy during the study period (Figure 2).

To account for the COVID-19 pandemic, the adherence rate of repeat colonoscopy within 1 year prepandemic (January 1, 2016, to December 1, 2018) was calculated along with the adherence rate postpandemic (January 1, 2019 to the end of the study). The rates were similar: 199 of 330 (60.3%) individuals prepandemic vs 88 of 155 (56.8%) individuals postpandemic (Figure 3).

Significant Associations

Age, sex, and race were not associated with adherence to repeat colonoscopy within 1 year. Individuals living ≤ 40 miles from the endoscopy center were more likely to undergo a repeat colonoscopy within 1 year compared with those who lived > 40 miles away (61.7% vs 51.0%, P = .02). Current smoking status was associated with a lower rate of repeat colonoscopy within 1 year (25.8% vs 35.9%; P = .02). There were no differences with respect to inflammatory bowel disease diagnosis, mental health diagnosis, diabetes mellitus, cirrhosis, or medications used, including opioids, anticoagulation, and antiplatelet therapy.

Outcomes

Among individuals who had a repeat colonoscopy the day after the index colonoscopy, 53 of 56 individuals (94.6%) had adequate bowel preparation. Among individuals who had a repeat colonoscopy within 7 days, 70 of 77 (90.9%) had adequate bowel preparation. Of 287 individuals with a repeat colonoscopy within 1 year, 251 (87.5%) had adequate bowel preparation on the repeat colonoscopy. By 400 days after the index colonoscopy, 268 of 304 individuals (88.2%) had adequate bowel preparation.

In this study conducted at a large VA medical center, we found that 5.6% of individuals undergoing colonoscopies had IBP, a rate comparable to prior studies (6% to 26%).^3,4 Only 59.2% of individuals underwent repeat colonoscopies within 1 year, as recommended after an index colonoscopy with IBP. Smoking and living longer distances (> 40 miles) from the endoscopy center were associated with a decreased adherence to the repeat colonoscopy recommendation.

Current guidelines recommend repeat colonoscopy for individuals with IBP within 1 year.¹⁰ In cases of IBP, the advanced adenoma miss rate is 36% upon repeat colonoscopy within 1 year.¹³ Despite the importance of a follow-up colonoscopy, clinician adherence with this recommendation remains low.^10,14,15 However, in this study cohort, 485 of 571 individuals with IBP (84.9%) received recommendations for a repeat colonoscopy within 1 year. In the US, only 31.9% of 260,314 colonoscopies with IBP included recommendations for a follow-up colonoscopy within 1 year.¹⁴ This could be related to variations in endoscopist practice as well as patient risk factors for developing polyps, including family history of cancer and personal history of prior polyps. The findings of multiple polyps, high-risk adenomas, and cancer on the index colonoscopy also influences the endoscopist for repeat colonoscopy within 1 year.¹⁴

The timing for repeat colonoscopies within 1 year will be determined by the patients, clinicians, and available scheduling. In this study, the earlier repeat colonoscopies, especially those occurring the day after the index colonoscopy, had the highest success rate of adequate bowel preparation. In a prior study, repeating colonoscopies within the same day or the next day was also found to have a higher rate of adequate bowel preparation than repeat colonoscopies within 1 year (88.9% vs 83.5%).¹⁶

Ensuring the return of individuals with IBP for repeat colonoscopy is a challenging task. We identified that individuals who live further away from MVAMC and current smokers had a decreased probability of returning for a repeat colonoscopy. Toro and colleagues found a 68.7% return rate for a repeat colonoscopy within 1 year with individuals age ≥ 60 years, and patients who were White were less likely to proceed with a repeat colonoscopy within 1 year.¹⁷ The study did not provide data regarding smoking status or distance to the endoscopy center.¹⁷ In a prior study of veterans, the dual diagnosis of psychiatric disorders and substance abuse was associated with missed and canceled colonoscopy appointments.¹⁸ The distance to the endoscopy center has also been previously identified as a barrier to a colonoscopy following an abnormal FIT.¹⁹ Although not identified in this study due to the homogenous demographic profile, social determinants of health such as socioeconomic status, education, and insurance coverage are known barriers to cancer screening but were not evaluated in this study.²⁰

Based on the identified risk factors, we have created a model for utilizing those risk factors to identify individuals at higher risk for noncompliance (ie, those who live further away from the endoscopy center or currently smoke). These individuals are proactively offered to use an intraprocedural bowel cleansing device to achieve adequate bowel preparation or priority rescheduling for a next-day colonoscopy.

Limitations

This study was a single-center study of the veteran population, which is predominantly White and male, thus limiting generalizability. The study is also limited by minimal available data on adenoma detection and colon cancer incidence on subsequent colonoscopies.

CONCLUSIONS

The rate of IBP was 5.5% in individuals undergoing colonoscopy for colon cancer screening, surveillance, positive FIT, or computed tomography colonography. Only 59.2% of those with IBP underwent the recommended repeat colonoscopy within 1 year. Smoking and distance to the endoscopy center were associated with a decreased adherence to the repeat colonoscopy recommendation. Additional efforts are needed to ensure that individuals with IBP return for timely repeat colonoscopy.

Colorectal cancer (CRC) is the third-most diagnosed cancer after breast and lung cancer, and is the second leading cause of global cancer related deaths.¹ In 2023 in the United States, > 150,000 individuals were diagnosed with CRC and 52,000 died.²

Colonoscopy is an effective CRC screening method and the lone method recommended for polyp surveillance. Inadequate bowel preparation (IBP) has been estimated to occur in about 6% to 26% of colonoscopies. ^3,4 The prevalence varies based on a variety of comorbidities, including immobility, diabetes mellitus, neurologic disorders, and use of opioids, with more occurrences of IBP noted in older adult, non-English speaking, and male individuals.^4-6

The quality of bowel preparation is integral to the effectiveness of screening and surveillance colonoscopies. IBP has been associated with missed adenomas and significantly lower adenoma detection rates.^7-9 In particular, IBP is independently associated with an increased risk of CRC in the future.³ Accordingly, the US Multisociety Task Force recommends repeat colonoscopies for individuals with IBP within 1 year.¹⁰ Ensuring that these individuals receive repeat colonoscopies is an essential part of CRC prevention. The benefit of repeat colonoscopy after IBP is highlighted by a retrospective analysis from Fung and colleagues that showed 81% of repeat colonoscopies had adequate bowel preparation, with higher numbers of adenomas detected on repeat compared to initial colonoscopies.¹¹

Given the impact of bowel preparation quality on the diagnostic capability of the colonoscopy, adherence to guidelines for repeat colonoscopies in cases of IBP is paramount for effective CRC prevention. This study aims to measure the frequency of repeat colonoscopy after IBP and the factors associated with adherence to recommendations.

METHODS

Individuals who underwent colonoscopy at the Minneapolis Veterans Affairs Medical Center (MVAMC) from January 1, 2016, to October 19, 2021, were identified to allow for 400 days of follow-up from the index colonoscopy to the data collection date. During the COVID-19 pandemic, the colonoscopy procedure capacity was reduced by 50% from June 1, 2020, to December 1, 2020, delaying nonurgent procedures, including screening and surveillance colonoscopies.

Individuals who underwent colonoscopy for CRC screening or polyp surveillance, or following a positive fecal immunohistochemistry test (FIT) or virtual computed tomography colonoscopy were included. Patients with colonoscopy indications for iron deficiency anemia, gastrointestinal bleeding, disease activity assessment of inflammatory bowel disease, abdominal pain, or changes in bowel movement pattern were excluded. IBP was defined as recording a Boston Bowel Preparation Scale (BBPS) score of < 6, or < 2 in any segment, or described as poor or inadequate using the Aronchick scale.

Age, sex, race, marital status, distance to MVAMC, smoking status, comorbidities, and concurrent medication use, including antiplatelet, anticoagulation, and prescription opiates at the time of index colonoscopy were obtained from the Veterans Health Administration (VHA) Corporate Data Warehouse (CDW) using structured query language processing of colonoscopy procedure notes to extract preparation scores and other procedure information. The CDW contains extracts from VHA clinical and administrative systems that contain complete clinical data from October 1999.¹² Current smoking status was defined as any smoking activity at the time the questionnaire was administered during a routine clinic visit within 400 days from the index colonoscopy.

Only individuals who were recommended to have repeat colonoscopy within 1 year were included. The intervals of 365 days and 400 days (1 year + about 1 additional month) were used in the event that the individual had a delay in scheduling their 1-year repeat colonoscopy. For individuals who did not undergo a colonoscopy at MVAMC within 400 days, a manual chart review of all available records was performed to determine whether a colonoscopy was performed at a non-VA facility.

Patients received written instructions for bowel preparation 2 weeks prior to the procedure. The preparation included magnesium citrate and a split dose of 4 liters of polyethylene glycol. Patients were also advised to start a low-fiber diet 3 days prior to the procedure and a clear liquid diet the day before the procedure. Patients with a history of IBP or those undergoing procedures with anesthesia received an additional 2 liters for a total of 6 liters of polyethylene glycol.

Statistical analysis

Baseline characteristics were reported as mean (SD) or median and IQR for continuous variables and percentage for categorical variables. Individuals who returned for colonoscopy within 400 days were compared to those who did not identify factors associated with adherence to recommendations. The data on individuals who returned for colonoscopy within 400 days were also analyzed for additional minor delays in the timing of the repeat colonoscopy. Continuous data were compared using Mann-Whitney U tests. Categorical data were compared using X² or Fisher exact tests. Missing data were imputed from the analyses. All analyses were performed using SAS JMP Pro version 16. P < .05 was considered statistically significant.

RESULTS

There were 18,241 total colonoscopies performed between January 1, 2016, to October 19, 2021, and 13,818 colonoscopies had indications for screening for colon cancer, positive FIT, virtual colonoscopy, or surveillance. Of the 10,466 unique patients there were 5369 patients for polyp surveillance, 4054 patients for CRC screening, and 1043 patients for positive FIT or virtual colonoscopy. Of these, 571 individuals (5.5%) had IBP. Repeat colonoscopy within 1 year was recommended for 485 individuals (84.9%) who were included in this study (153 CRC screenings and 46 positive FITs) but not for 86 individuals (15.1%) (Figure 1). Among included patients, the mean (SD) age was 66.6 (7.2) years, and the majority were male (460 [94.8%]) and White (435 [89.7%]) (Table). Two hundred and forty-three (50.1%) were married.

Adherence to Recommended Interval Colonoscopy

Of the 485 patients with IBP who were recommended for follow-up colonoscopy, 287 (59.2%) had a colonoscopy within 1 year, and 198 (40.8%) did not; 17 patients (13.5%) had repeat colonoscopy within 366 to 400 days. Five (1.0%) individuals had a repeat colonoscopy the next day, and 77 (15.9%) had a repeat colonoscopy within 7 days. One hundred and twentysix (26.0%) individuals underwent no repeat colonoscopy during the study period (Figure 2).

To account for the COVID-19 pandemic, the adherence rate of repeat colonoscopy within 1 year prepandemic (January 1, 2016, to December 1, 2018) was calculated along with the adherence rate postpandemic (January 1, 2019 to the end of the study). The rates were similar: 199 of 330 (60.3%) individuals prepandemic vs 88 of 155 (56.8%) individuals postpandemic (Figure 3).

Significant Associations

Age, sex, and race were not associated with adherence to repeat colonoscopy within 1 year. Individuals living ≤ 40 miles from the endoscopy center were more likely to undergo a repeat colonoscopy within 1 year compared with those who lived > 40 miles away (61.7% vs 51.0%, P = .02). Current smoking status was associated with a lower rate of repeat colonoscopy within 1 year (25.8% vs 35.9%; P = .02). There were no differences with respect to inflammatory bowel disease diagnosis, mental health diagnosis, diabetes mellitus, cirrhosis, or medications used, including opioids, anticoagulation, and antiplatelet therapy.

Outcomes

Among individuals who had a repeat colonoscopy the day after the index colonoscopy, 53 of 56 individuals (94.6%) had adequate bowel preparation. Among individuals who had a repeat colonoscopy within 7 days, 70 of 77 (90.9%) had adequate bowel preparation. Of 287 individuals with a repeat colonoscopy within 1 year, 251 (87.5%) had adequate bowel preparation on the repeat colonoscopy. By 400 days after the index colonoscopy, 268 of 304 individuals (88.2%) had adequate bowel preparation.

In this study conducted at a large VA medical center, we found that 5.6% of individuals undergoing colonoscopies had IBP, a rate comparable to prior studies (6% to 26%).^3,4 Only 59.2% of individuals underwent repeat colonoscopies within 1 year, as recommended after an index colonoscopy with IBP. Smoking and living longer distances (> 40 miles) from the endoscopy center were associated with a decreased adherence to the repeat colonoscopy recommendation.

Current guidelines recommend repeat colonoscopy for individuals with IBP within 1 year.¹⁰ In cases of IBP, the advanced adenoma miss rate is 36% upon repeat colonoscopy within 1 year.¹³ Despite the importance of a follow-up colonoscopy, clinician adherence with this recommendation remains low.^10,14,15 However, in this study cohort, 485 of 571 individuals with IBP (84.9%) received recommendations for a repeat colonoscopy within 1 year. In the US, only 31.9% of 260,314 colonoscopies with IBP included recommendations for a follow-up colonoscopy within 1 year.¹⁴ This could be related to variations in endoscopist practice as well as patient risk factors for developing polyps, including family history of cancer and personal history of prior polyps. The findings of multiple polyps, high-risk adenomas, and cancer on the index colonoscopy also influences the endoscopist for repeat colonoscopy within 1 year.¹⁴

The timing for repeat colonoscopies within 1 year will be determined by the patients, clinicians, and available scheduling. In this study, the earlier repeat colonoscopies, especially those occurring the day after the index colonoscopy, had the highest success rate of adequate bowel preparation. In a prior study, repeating colonoscopies within the same day or the next day was also found to have a higher rate of adequate bowel preparation than repeat colonoscopies within 1 year (88.9% vs 83.5%).¹⁶

Ensuring the return of individuals with IBP for repeat colonoscopy is a challenging task. We identified that individuals who live further away from MVAMC and current smokers had a decreased probability of returning for a repeat colonoscopy. Toro and colleagues found a 68.7% return rate for a repeat colonoscopy within 1 year with individuals age ≥ 60 years, and patients who were White were less likely to proceed with a repeat colonoscopy within 1 year.¹⁷ The study did not provide data regarding smoking status or distance to the endoscopy center.¹⁷ In a prior study of veterans, the dual diagnosis of psychiatric disorders and substance abuse was associated with missed and canceled colonoscopy appointments.¹⁸ The distance to the endoscopy center has also been previously identified as a barrier to a colonoscopy following an abnormal FIT.¹⁹ Although not identified in this study due to the homogenous demographic profile, social determinants of health such as socioeconomic status, education, and insurance coverage are known barriers to cancer screening but were not evaluated in this study.²⁰

Based on the identified risk factors, we have created a model for utilizing those risk factors to identify individuals at higher risk for noncompliance (ie, those who live further away from the endoscopy center or currently smoke). These individuals are proactively offered to use an intraprocedural bowel cleansing device to achieve adequate bowel preparation or priority rescheduling for a next-day colonoscopy.

Limitations

This study was a single-center study of the veteran population, which is predominantly White and male, thus limiting generalizability. The study is also limited by minimal available data on adenoma detection and colon cancer incidence on subsequent colonoscopies.

CONCLUSIONS

The rate of IBP was 5.5% in individuals undergoing colonoscopy for colon cancer screening, surveillance, positive FIT, or computed tomography colonography. Only 59.2% of those with IBP underwent the recommended repeat colonoscopy within 1 year. Smoking and distance to the endoscopy center were associated with a decreased adherence to the repeat colonoscopy recommendation. Additional efforts are needed to ensure that individuals with IBP return for timely repeat colonoscopy.

Medical errors are a persistent problem and leading cause of preventable death in the United States. There is considerable momentum behind the idea that implementation of a just culture is foundational to detecting and learning from errors in pursuit of zero patient harm.^1-6 Just culture is a framework that fosters an environment of trust within health care organizations, aiming to achieve fair outcomes for those involved in incidents or near misses. It emphasizes openness, accountability, and learning, prioritizing the repair of harm and systemic improvement over assigning blame.⁷

A just culture mindset reflects a significant shift in thinking that moves from the tendency to blame and punish others toward a focus on organizational learning and continued process improvement.^8,9 This systemic shift in fundamental thinking transforms how leaders approach staff errors and how they are addressed.¹⁰ In essence, just culture reflects an ethos centered on openness, a deep appreciation of human fallibility, and shared accountability at both the individual and organizational levels.

Organizational learning and innovation are stifled in the absence of a just culture, and there is a tendency for employees to avoid disclosing their own errors as well as those of their colleagues.¹¹ The transformation to a just culture is often slowed or disrupted by personal, systemic, and cultural barriers.¹² It is imperative that all executive, service line, and frontline managers recognize and execute their distinct responsibilities while adjudicating the appropriate course of action in the aftermath of adverse events or near misses. This requires a nuanced understanding of the factors that contribute to errors at the individual and organizational levels to ensure an appropriate response.

The Veterans Health Administration (VHA) is orchestrating an enterprise transformation to develop into a high reliability organization (HRO). This began with a single-site test in 2016, which demonstrated successful results in patient safety culture, patient safety event reporting, and patient safety outcomes.¹³ In 2019, the VHA formally launched its enterprise-wide HRO journey in 18 hospital facilities, followed by successive waves of 67 and 54 facilities in 2021 and 2022, respectively. The VHA journey to transform into an HRO aligns with 3 pillars, 5 principles, and 7 values. The VHA has emphasized the importance of just culture as a foundational element of the HRO framework, specifically under the pillar of leadership. To promote leadership engagement, the VHA has employed an array of approaches that include education, leader coaching, and change management strategies. Given the diversity among VHA facilities, each with local cultures and histories, some sites have more readily implemented a just culture than others.¹⁴ A deeper exploration into potential obstacles, particularly concerning leadership engagement, could be instrumental for formulating strategies that further establish a just culture across the VHA.¹⁵

There is a paucity of empirical research regarding factors that facilitate and/or impede the implementation of a just culture in health care settings.^16,17 Likert scale surveys, such as the Patient Safety Culture Module for the VHA All Employee Survey and its predecessor, the Patient Safety Culture Survey, have been used to assess culture and climate.¹⁸ However, qualitative evaluations directly assessing the lived experiences of those trying to implement a just culture provide additional depth and context that can help identify specific factors that support or impede becoming an HRO. The purpose of this study was to increase understanding of factors that influence the establishment and sustainment of a just culture and to identify specific methods for improving the implementation of just culture principles and practices aligned with HRO.

METHODS

This qualitative study explored facilitators and barriers to establishing and sustaining a just culture as experienced across a subset of VHA facilities by HRO leads or staff assigned with the primary responsibilities of supporting facility-level HRO transformation. HRO leads are assigned responsibility for supporting executive leadership in planning, coordinating, implementing, and monitoring activities to ensure effective high reliability efforts, including focused efforts to establish a robust patient safety program, a culture of safety, and a culture of continuous process improvement.

Virtual focus group discussions held via Microsoft Teams generated in-depth, diverse perspectives from participants across 16 VHA facilities. Qualitative research and evaluation methods provide an enhanced depth of understanding and allow the emergence of detailed data.¹⁹ A qualitative grounded theory approach elicits complex, multifaceted phenomena that cannot be appreciated solely by numeric data.²⁰ Grounded theory was selected to limit preconceived notions and provide a more systematic analysis, including open, axial, and thematic coding. Such methods afford opportunities to adapt to unplanned follow-up questions and thus provide a flexible approach to generate new ideas and concepts.²¹ Additionally, qualitative methods help overcome the tendencies of respondents to agree rather than disagree when presented with Likert-style scales, which tend to skew responses toward the positive.²²

Participants must have been assigned as an HRO lead for ≥ 6 months at the same facility. Potential participants were identified through purposive sampling, considering their leadership roles in HRO and experience with just culture implementation, the size and complexity of their facility, and geographic distribution. Invitations explaining the study and encouraging voluntary participation to participate were emailed. Of 37 HRO leads invited to participate in the study, 16 agreed to participate and attended 1 of 3 hour-long focus group sessions. One session was rescheduled due to limited attendance. Participants represented a mix of VHA sites in terms of geography, facility size, and complexity.

Focus Group Procedures

Demographic data were collected prior to sessions via an online form to better understand the participant population, including facility complexity level, length of time in HRO lead role, clinical background, and facility level just culture training. Each session was led by an experienced focus group facilitator (CV) who was not directly involved with the overall HRO implementation to establish a neutral perspective. Each session was attended by 4 to 7 participants and 2 observers who took notes. The sessions were recorded and included automated transcriptions, which were edited for accuracy.

Focus group sessions began with a brief introduction and an opportunity for participants to ask questions. Participants were then asked a series of open-ended questions to elicit responses regardingfacilitators, barriers, and leadership support needed for implementing just culture. The questions were part of a facilitator guide that included an introductory script and discussion prompts to ensure consistency across focus groups.

Facilitators were defined as factors that increase the likelihood of establishing or sustaining a just culture. Barriers were defined as factors that decrease or inhibit the likelihood of establishing or sustaining just culture. The focus group facilitator encouraged all participants to share their views and provided clarification if needed, as well as prompts and examples where appropriate, but primarily sought to elicit responses to the questions.

Institutional review board review and approval were not required for this quality improvement initiative. The project adhered to ethical standards of research, including asking participants for verbal consent and preserving their confidentiality. Participation was voluntary, and prior to the focus group sessions, participants were provided information explaining the study’s purpose, procedures, and their rights. Participant identities were kept confidential, and all data were anonymized during the analysis phase. Pseudonyms or identifiers were used during data transcription to protect participant identity. All data, including recordings and transcriptions, were stored on password-protected devices accessible only to the research team. Any identifiable information was removed during data analysis to ensure confidentiality.

Analysis

Focus group recordings were transcribed verbatim, capturing all verbal interactions and nonverbal cues that may contribute to understanding the participants' perspectives. Thematic analysis was used to analyze the qualitative data from the focus group discussions.²³ The transcribed data were organized, coded, and analyzed using ATLAS.ti 23 qualitative data software to identify key themes and patterns.

Results

The themes identified include the 5 facilitators, barriers, and recommendations most frequently mentioned by HRO leads across focus group sessions. The nature of each theme is described, along with commonly mentioned examples and direct quotes from participants that illustrate our understanding of their perspectives.

Facilitators

Training and coaching (26 responses). The availability of training around the Just Culture Decision Support Tool (DST) was cited as a practical aid in guiding leaders through complex just culture decisions to ensure consistency and fairness. Additionally, an executive leadership team that served as champions for just culture principles played a vital role in promoting and sustaining the approach: “Training them on the roll-out of the decision support tool with supervisors at all levels, and education for just culture and making it part of our safety forum has helped for the last 4 months.” “Having some regular training and share-out cadences embedded within the schedule as well as dynamic directors and well-trained executive leadership team (ELT) for support has been a facilitator.”

Increased transparency (16 responses). Participants consistently highlighted the importance of leadership transparency as a key facilitator for implementing just culture. Open and honest communication from top-level executives fostered an environment of trust and accountability. Approachable and physically present leadership was seen as essential for creating a culture where employees felt comfortable reporting errors and concerns without fear of retaliation: “They’re surprisingly honest with themselves about what we can do, what we cannot do, and they set the expectations exactly at that.”

Approachable leadership (15 responses). Participants frequently mentioned the importance of having dynamic leadership spearheading the implementation of just culture and leading by example. Having a leadership team that accepts accountability and reinforces consistency in the manner that near misses or mistakes are addressed is paramount to promoting the principles of just culture and increasing psychological safety: “We do have very approachable leadership, which I know is hard if you’re trying to implement that nationwide, it’s hard to implement approachability. But I do think that people raise their concerns, and they’ll stop them in the hallway and ask them questions. So, in terms of comfort level with the executive leadership, I do think that’s high, which would promote psychological safety.”

Feedback loops and follow through (13 responses). Participants emphasized the importance of taking concrete actions to address concerns and improve processes. Regular check-ins with supervisors to discuss matters related to just culture provided a structured opportunity for addressing issues and reinforcing the importance of the approach: “One thing that we’ve really focused on is not only identifying mistakes, but [taking] ownership. We continue to track it until … it’s completed and then a process of how to communicate that back and really using closed loop communication with the staff and letting them know.”

Forums and town halls (10 responses). These platforms created feedback loops that were seen as invaluable tools for sharing near misses, celebrating successes, and promoting open dialogue. Forums and town halls cultivated a culture of continuous improvement and trust: “We’ll celebrate catches, a safety story is inside that catch. So, if we celebrate the change, people feel safer to speak up.” “Truthfully, we’ve had a great relationship since establishing our safety forums and just value open lines of communication.”

Barriers

Inadequate training (30 responses). Insufficient engagement during training—limited bandwidth and availability to attend and actively participate in training—was perceived as detrimental to creating awareness and buy-in from staff, supervisors, and leadership, thereby hindering successful integration of just culture principles. Participants also identified too many conflicting priorities from VHA leadership, which contributes to training and information fatigue among staff and supervisors. “Our biggest barrier is just so many different competing priorities going on. We have so much that we’re asking people to do.” “One hundred percent training is feeling more like a ticked box than actually yielding results, I have a very hard time getting staff engaged.”

Inconsistency between executive leaders and middle managers (28 responses). A lack of consistency in the commitment to and enactment of just culture principles among leaders poses a challenge. Participants gave several examples of inconsistencies in messaging and reinforcement of just culture principles, leading to confusion among staff and hindering adoption. Likewise, the absence of standardized procedures for implementing just culture created variability: “The director coming in and trying to change things, it put a lot of resistance, we struggle with getting the other ELT members on board … some of the messages that come out at times can feel more punitive.”

Middle management resistance (22 responses). In some instances, participants reported middle managers exhibiting attitudes and behaviors that undermined the application of just culture principles and effectiveness. Such attitudes and behaviors were attributed to a lack of adequate training, coaching, and awareness. Other perceived contributions included fear of failure and a desire to distance oneself from staff who have made mistakes: “As soon as someone makes an error, they go straight to suspend them, and that’s the disconnect right there.” “There’s almost a level of working in the opposite direction in some of the mid-management.”

Cultural misalignment (18 responses). The existing culture of distancing oneself from mistakes presented a significant barrier to the adoption of just culture because it clashed with the principles of open reporting and accountability. Staff underreported errors or framed them in a way that minimized personal responsibility, thereby making it more essential to put in the necessary and difficult work to learn from mistakes: “One, you’re going to get in trouble. There’s going to be more work added to you or something of that nature."

Lack of accountability for opposition(17 responses). Participants noted a clear lack of accountability for those who opposed or showed resistance to just culture, which allowed resistance to persist without consequences. In many instances, leaders were described as having overlooked repeated instances of unjust attitudes and behaviors (eg, inappropriate blame or punishment), which allowed those practices to continue. “Executive leadership is standing on the hill and saying we’re a just culture and we do everything correctly, and staff has the expectation that they’re going to be treated with just culture and then the middle management is setting that on fire, then we show them that that’s not just culture, and they continue to have those poor behaviors, but there’s a lack of accountability.”

Limited bandwidth and lack of coordination (14 responses). HRO leads often faced role-specific constraints in having adequate time and authority to coordinate efforts to implement or sustain just culture. This includes challenges with coordination across organizational levels (eg, between the hospital and regional Veterans Integrated Service Network [VISN] management levels) and across departments within the hospital (eg, between human resources and service lines or units). “Our VISN human resources is completely detached. They’ll not cooperate with these efforts, which is hard.” “There’s not enough bandwidth to actually support, I’m just 1 person.” “[There’s] all these mandated trainings of 8 hours when we’re already fatigued, short-staffed, taking 3 other HRO classes.”

Recommendations

Training improvements (24 responses). HRO leads recommended that comprehensive training programs be developed and implemented for staff, supervisors, and leadership to increase awareness and understanding of just culture principles. These training initiatives should focus on fostering a shared understanding of the core tenets of just culture, the importance of error reporting, and the processes involved in fair and consistent decision making (eg, training simulations on use of the Just Culture DST). “We’ve really never had any formal training on the decision support tool. I hope that what’s coming out for next year. We’ll have some more formal training for the tool because I think it would be great to really have our leadership and our supervisors and our managers use that tool.” “We can give a more directed and intentional training to leadership on the 4 foundational practices and what it means to implement those and what it means to utilize that behavioral component of HRO.”

Clear and consistent procedures toincrease accountability (22 responses). To promote a culture of accountability and consistency in the application of just culture principles, organizations should establish clear mechanisms for reporting, investigating, and addressing incidents. Standardized procedures and DSTs can aid in ensuring that responses to errors are equitable and align with just culture principles: “I recommend accountability; if it’s clearly evidenced that you’re not toeing the just culture line, then we need to be able to do something about it and not just finger wag.” “[We need to have] a templated way to approach just culture implementation. The decision support tool is great, I absolutely love having the resources and being able to find a lot of clinical examples and discussion tools like that. But when it comes down to it, not having that kind of official thing to fall back on it can be a little bit rough.”

Additional coaching and consultationsupport (15 responses). To support supervisors in effectively implementing just culture within their teams, participants recommended that organizations provide ongoing coaching and mentorship opportunities. Additionally, third-party consultants with expertise in just culture were described as offering valuable guidance, particularly in cases where internal staff resources or HRO lead bandwidth may be limited. “There are so many consulting agencies with HRO that have been contracted to do different projects, but maybe that can help with an educational program.” “I want to see my executive leadership coach the supervisors up right and then allow them to do one-on-ones and facilitate and empower the frontline staff, and it’s just a good way of transparency and communication.”

Improved leadership sponsorship (15 responses). Participants noted that leadership buy-in is crucial for the successful implementation of just culture. Facilities should actively engage and educate leadership teams on the benefits of just culture and how it aligns with broader patient safety and organizational goals. Leaders should be visible and active champions of its principles, supporting change in their daily engagements with staff. “ELT support is absolutely necessary. Why? Because they will make it important to those in their service lines. They will make it important to those supervisors and managers. If it’s not important to that ELT member, then it’s not going to be important to that manager or that supervisor.”

Improved collaboration with patient safety and human resources (6 responses). Collaborative efforts with patient safety and human resources departments were seen as instrumental in supporting just culture, emphasizing its importance, and effectively addressing issues. Coordinating with these departments specifically contributes to consistent reinforcement and expands the bandwidth of HRO leads. These departments play integral roles in supporting just culture through effective policies, procedures, and communication. “I think it would be really helpful to have common language between what human resources teaches and what is in our decision support tool.”

DISCUSSION

This study sought to collect and synthesize the experiences of leaders across a large, integrated health care system in establishing and sustaining a just culture as part of an enterprise journey to become an HRO.²⁴ The VHA has provided enterprise-wide support (eg, training, leader coaching, and communications) for the implementation of HRO principles and practices with the goal of creating a culture of safety, which includes just culture elements. This support includes enterprise program offices, VISNs, and hospital facilities, though notably, there is variability in how HRO is implemented at the local level. The facilitators, barriers, and recommendations presented in this article are representative of the designated HRO leads at VHA hospital facilities who have direct experience with implementing and sustaining just culture. The themes presented offer specific opportunities for intervention and actionable strategies to enhance just culture initiatives, foster psychological safety and accountability, and ultimately improve the quality of care and patient outcomes.^3,25

Frequently identified facilitators such as providing training and coaching, having leaders who are available and approachable, demonstrating follow-through to address identified issues, and creating venues where errors and successes can be openly discussed.²⁶ These facilitators are aligned with enterprise HRO support strategies orchestrated by the VHA at the enterprise VISN and facility levels to support a culture of safety, continuous process improvement, and leadership commitment.

Frequently identified barriers included inadequate training, inconsistent application of just culture by middle managers vs senior leaders, a lack of accountability or corrective action when unjust corrective actions took place, time and resource constraints, and inadequate coordination across departments (eg, operational departments and human resources) and organizational levels. These factors were identified through focus groups with a limited set of HRO leads. They may reflect challenges to changing culture that may be deeply engrained in individual histories, organizational norms, and systemic practices. Improving upon these just culture initiatives requires multifaceted approaches and working through resistance to change.

VHA HRO leads identified several actionable recommendations that may be used in pursuit of a just culture. First, improvements in training involving how to apply just culture principles and, specifically, the use of the Just Culture DST were identified as an opportunity for improvement. The VHA National Center for Patient Safety developed the DST as an aid for leaders to effectively address errors in line with just culture principles, balancing individual and system accountability.²⁷ The DST specifically addresses human error as well as risky and reckless behavior, and it clarifies the delineation between individual and organizational accountability (Table).³

Limitations

This study employed qualitative methods and sampled a relatively small subset of experienced leaders with specific roles in implementing HRO in the VHA. Thus, it should not be considered representative of the perspectives of all leaders within the VHA or other health care systems. Future studies should assess facilitators and barriers beyond the facility level, including a focus incorporating both the VISN and VHA. More broadly, qualitative methods such as those employed in this study offer great depth and nuance but have limited ability to identify system-wide trends and differences. As such, it may be beneficial to specifically look at sites that are high- or low-performing on measures of patient safety culture to identify differences that may inform implementation strategies based on organizational maturity and readiness for change.

Conclusions

Successful implementation of these recommendations will require ongoing commitment, collaboration, and a sustained effort from all stakeholders involved at multiple levels of the health care system. Monitoring and evaluating progress should be conducted regularly to ensure that recommendations lead to improvements in implementing just culture principles. This quality improvement study adds to the knowledge base on factors that impact the just culture and broader efforts to realize HRO principles and practices in health care systems. The approach of this study may serve as a model for identifying opportunities to improve HRO implementation within the VHA and other settings, especially when paired with ongoing quantitative evaluation of organizational safety culture, just culture behaviors, and patient outcomes.

Medical errors are a persistent problem and leading cause of preventable death in the United States. There is considerable momentum behind the idea that implementation of a just culture is foundational to detecting and learning from errors in pursuit of zero patient harm.^1-6 Just culture is a framework that fosters an environment of trust within health care organizations, aiming to achieve fair outcomes for those involved in incidents or near misses. It emphasizes openness, accountability, and learning, prioritizing the repair of harm and systemic improvement over assigning blame.⁷

A just culture mindset reflects a significant shift in thinking that moves from the tendency to blame and punish others toward a focus on organizational learning and continued process improvement.^8,9 This systemic shift in fundamental thinking transforms how leaders approach staff errors and how they are addressed.¹⁰ In essence, just culture reflects an ethos centered on openness, a deep appreciation of human fallibility, and shared accountability at both the individual and organizational levels.

Organizational learning and innovation are stifled in the absence of a just culture, and there is a tendency for employees to avoid disclosing their own errors as well as those of their colleagues.¹¹ The transformation to a just culture is often slowed or disrupted by personal, systemic, and cultural barriers.¹² It is imperative that all executive, service line, and frontline managers recognize and execute their distinct responsibilities while adjudicating the appropriate course of action in the aftermath of adverse events or near misses. This requires a nuanced understanding of the factors that contribute to errors at the individual and organizational levels to ensure an appropriate response.

The Veterans Health Administration (VHA) is orchestrating an enterprise transformation to develop into a high reliability organization (HRO). This began with a single-site test in 2016, which demonstrated successful results in patient safety culture, patient safety event reporting, and patient safety outcomes.¹³ In 2019, the VHA formally launched its enterprise-wide HRO journey in 18 hospital facilities, followed by successive waves of 67 and 54 facilities in 2021 and 2022, respectively. The VHA journey to transform into an HRO aligns with 3 pillars, 5 principles, and 7 values. The VHA has emphasized the importance of just culture as a foundational element of the HRO framework, specifically under the pillar of leadership. To promote leadership engagement, the VHA has employed an array of approaches that include education, leader coaching, and change management strategies. Given the diversity among VHA facilities, each with local cultures and histories, some sites have more readily implemented a just culture than others.¹⁴ A deeper exploration into potential obstacles, particularly concerning leadership engagement, could be instrumental for formulating strategies that further establish a just culture across the VHA.¹⁵

There is a paucity of empirical research regarding factors that facilitate and/or impede the implementation of a just culture in health care settings.^16,17 Likert scale surveys, such as the Patient Safety Culture Module for the VHA All Employee Survey and its predecessor, the Patient Safety Culture Survey, have been used to assess culture and climate.¹⁸ However, qualitative evaluations directly assessing the lived experiences of those trying to implement a just culture provide additional depth and context that can help identify specific factors that support or impede becoming an HRO. The purpose of this study was to increase understanding of factors that influence the establishment and sustainment of a just culture and to identify specific methods for improving the implementation of just culture principles and practices aligned with HRO.

METHODS

This qualitative study explored facilitators and barriers to establishing and sustaining a just culture as experienced across a subset of VHA facilities by HRO leads or staff assigned with the primary responsibilities of supporting facility-level HRO transformation. HRO leads are assigned responsibility for supporting executive leadership in planning, coordinating, implementing, and monitoring activities to ensure effective high reliability efforts, including focused efforts to establish a robust patient safety program, a culture of safety, and a culture of continuous process improvement.

Virtual focus group discussions held via Microsoft Teams generated in-depth, diverse perspectives from participants across 16 VHA facilities. Qualitative research and evaluation methods provide an enhanced depth of understanding and allow the emergence of detailed data.¹⁹ A qualitative grounded theory approach elicits complex, multifaceted phenomena that cannot be appreciated solely by numeric data.²⁰ Grounded theory was selected to limit preconceived notions and provide a more systematic analysis, including open, axial, and thematic coding. Such methods afford opportunities to adapt to unplanned follow-up questions and thus provide a flexible approach to generate new ideas and concepts.²¹ Additionally, qualitative methods help overcome the tendencies of respondents to agree rather than disagree when presented with Likert-style scales, which tend to skew responses toward the positive.²²

Participants must have been assigned as an HRO lead for ≥ 6 months at the same facility. Potential participants were identified through purposive sampling, considering their leadership roles in HRO and experience with just culture implementation, the size and complexity of their facility, and geographic distribution. Invitations explaining the study and encouraging voluntary participation to participate were emailed. Of 37 HRO leads invited to participate in the study, 16 agreed to participate and attended 1 of 3 hour-long focus group sessions. One session was rescheduled due to limited attendance. Participants represented a mix of VHA sites in terms of geography, facility size, and complexity.

Focus Group Procedures

Demographic data were collected prior to sessions via an online form to better understand the participant population, including facility complexity level, length of time in HRO lead role, clinical background, and facility level just culture training. Each session was led by an experienced focus group facilitator (CV) who was not directly involved with the overall HRO implementation to establish a neutral perspective. Each session was attended by 4 to 7 participants and 2 observers who took notes. The sessions were recorded and included automated transcriptions, which were edited for accuracy.

Focus group sessions began with a brief introduction and an opportunity for participants to ask questions. Participants were then asked a series of open-ended questions to elicit responses regardingfacilitators, barriers, and leadership support needed for implementing just culture. The questions were part of a facilitator guide that included an introductory script and discussion prompts to ensure consistency across focus groups.

Facilitators were defined as factors that increase the likelihood of establishing or sustaining a just culture. Barriers were defined as factors that decrease or inhibit the likelihood of establishing or sustaining just culture. The focus group facilitator encouraged all participants to share their views and provided clarification if needed, as well as prompts and examples where appropriate, but primarily sought to elicit responses to the questions.

Institutional review board review and approval were not required for this quality improvement initiative. The project adhered to ethical standards of research, including asking participants for verbal consent and preserving their confidentiality. Participation was voluntary, and prior to the focus group sessions, participants were provided information explaining the study’s purpose, procedures, and their rights. Participant identities were kept confidential, and all data were anonymized during the analysis phase. Pseudonyms or identifiers were used during data transcription to protect participant identity. All data, including recordings and transcriptions, were stored on password-protected devices accessible only to the research team. Any identifiable information was removed during data analysis to ensure confidentiality.

Analysis

Focus group recordings were transcribed verbatim, capturing all verbal interactions and nonverbal cues that may contribute to understanding the participants' perspectives. Thematic analysis was used to analyze the qualitative data from the focus group discussions.²³ The transcribed data were organized, coded, and analyzed using ATLAS.ti 23 qualitative data software to identify key themes and patterns.

Results

The themes identified include the 5 facilitators, barriers, and recommendations most frequently mentioned by HRO leads across focus group sessions. The nature of each theme is described, along with commonly mentioned examples and direct quotes from participants that illustrate our understanding of their perspectives.

Facilitators

Training and coaching (26 responses). The availability of training around the Just Culture Decision Support Tool (DST) was cited as a practical aid in guiding leaders through complex just culture decisions to ensure consistency and fairness. Additionally, an executive leadership team that served as champions for just culture principles played a vital role in promoting and sustaining the approach: “Training them on the roll-out of the decision support tool with supervisors at all levels, and education for just culture and making it part of our safety forum has helped for the last 4 months.” “Having some regular training and share-out cadences embedded within the schedule as well as dynamic directors and well-trained executive leadership team (ELT) for support has been a facilitator.”

Increased transparency (16 responses). Participants consistently highlighted the importance of leadership transparency as a key facilitator for implementing just culture. Open and honest communication from top-level executives fostered an environment of trust and accountability. Approachable and physically present leadership was seen as essential for creating a culture where employees felt comfortable reporting errors and concerns without fear of retaliation: “They’re surprisingly honest with themselves about what we can do, what we cannot do, and they set the expectations exactly at that.”

Approachable leadership (15 responses). Participants frequently mentioned the importance of having dynamic leadership spearheading the implementation of just culture and leading by example. Having a leadership team that accepts accountability and reinforces consistency in the manner that near misses or mistakes are addressed is paramount to promoting the principles of just culture and increasing psychological safety: “We do have very approachable leadership, which I know is hard if you’re trying to implement that nationwide, it’s hard to implement approachability. But I do think that people raise their concerns, and they’ll stop them in the hallway and ask them questions. So, in terms of comfort level with the executive leadership, I do think that’s high, which would promote psychological safety.”

Feedback loops and follow through (13 responses). Participants emphasized the importance of taking concrete actions to address concerns and improve processes. Regular check-ins with supervisors to discuss matters related to just culture provided a structured opportunity for addressing issues and reinforcing the importance of the approach: “One thing that we’ve really focused on is not only identifying mistakes, but [taking] ownership. We continue to track it until … it’s completed and then a process of how to communicate that back and really using closed loop communication with the staff and letting them know.”

Forums and town halls (10 responses). These platforms created feedback loops that were seen as invaluable tools for sharing near misses, celebrating successes, and promoting open dialogue. Forums and town halls cultivated a culture of continuous improvement and trust: “We’ll celebrate catches, a safety story is inside that catch. So, if we celebrate the change, people feel safer to speak up.” “Truthfully, we’ve had a great relationship since establishing our safety forums and just value open lines of communication.”

Barriers

Inadequate training (30 responses). Insufficient engagement during training—limited bandwidth and availability to attend and actively participate in training—was perceived as detrimental to creating awareness and buy-in from staff, supervisors, and leadership, thereby hindering successful integration of just culture principles. Participants also identified too many conflicting priorities from VHA leadership, which contributes to training and information fatigue among staff and supervisors. “Our biggest barrier is just so many different competing priorities going on. We have so much that we’re asking people to do.” “One hundred percent training is feeling more like a ticked box than actually yielding results, I have a very hard time getting staff engaged.”

Inconsistency between executive leaders and middle managers (28 responses). A lack of consistency in the commitment to and enactment of just culture principles among leaders poses a challenge. Participants gave several examples of inconsistencies in messaging and reinforcement of just culture principles, leading to confusion among staff and hindering adoption. Likewise, the absence of standardized procedures for implementing just culture created variability: “The director coming in and trying to change things, it put a lot of resistance, we struggle with getting the other ELT members on board … some of the messages that come out at times can feel more punitive.”

Middle management resistance (22 responses). In some instances, participants reported middle managers exhibiting attitudes and behaviors that undermined the application of just culture principles and effectiveness. Such attitudes and behaviors were attributed to a lack of adequate training, coaching, and awareness. Other perceived contributions included fear of failure and a desire to distance oneself from staff who have made mistakes: “As soon as someone makes an error, they go straight to suspend them, and that’s the disconnect right there.” “There’s almost a level of working in the opposite direction in some of the mid-management.”

Cultural misalignment (18 responses). The existing culture of distancing oneself from mistakes presented a significant barrier to the adoption of just culture because it clashed with the principles of open reporting and accountability. Staff underreported errors or framed them in a way that minimized personal responsibility, thereby making it more essential to put in the necessary and difficult work to learn from mistakes: “One, you’re going to get in trouble. There’s going to be more work added to you or something of that nature."

Lack of accountability for opposition(17 responses). Participants noted a clear lack of accountability for those who opposed or showed resistance to just culture, which allowed resistance to persist without consequences. In many instances, leaders were described as having overlooked repeated instances of unjust attitudes and behaviors (eg, inappropriate blame or punishment), which allowed those practices to continue. “Executive leadership is standing on the hill and saying we’re a just culture and we do everything correctly, and staff has the expectation that they’re going to be treated with just culture and then the middle management is setting that on fire, then we show them that that’s not just culture, and they continue to have those poor behaviors, but there’s a lack of accountability.”

Limited bandwidth and lack of coordination (14 responses). HRO leads often faced role-specific constraints in having adequate time and authority to coordinate efforts to implement or sustain just culture. This includes challenges with coordination across organizational levels (eg, between the hospital and regional Veterans Integrated Service Network [VISN] management levels) and across departments within the hospital (eg, between human resources and service lines or units). “Our VISN human resources is completely detached. They’ll not cooperate with these efforts, which is hard.” “There’s not enough bandwidth to actually support, I’m just 1 person.” “[There’s] all these mandated trainings of 8 hours when we’re already fatigued, short-staffed, taking 3 other HRO classes.”

Recommendations

Training improvements (24 responses). HRO leads recommended that comprehensive training programs be developed and implemented for staff, supervisors, and leadership to increase awareness and understanding of just culture principles. These training initiatives should focus on fostering a shared understanding of the core tenets of just culture, the importance of error reporting, and the processes involved in fair and consistent decision making (eg, training simulations on use of the Just Culture DST). “We’ve really never had any formal training on the decision support tool. I hope that what’s coming out for next year. We’ll have some more formal training for the tool because I think it would be great to really have our leadership and our supervisors and our managers use that tool.” “We can give a more directed and intentional training to leadership on the 4 foundational practices and what it means to implement those and what it means to utilize that behavioral component of HRO.”

Clear and consistent procedures toincrease accountability (22 responses). To promote a culture of accountability and consistency in the application of just culture principles, organizations should establish clear mechanisms for reporting, investigating, and addressing incidents. Standardized procedures and DSTs can aid in ensuring that responses to errors are equitable and align with just culture principles: “I recommend accountability; if it’s clearly evidenced that you’re not toeing the just culture line, then we need to be able to do something about it and not just finger wag.” “[We need to have] a templated way to approach just culture implementation. The decision support tool is great, I absolutely love having the resources and being able to find a lot of clinical examples and discussion tools like that. But when it comes down to it, not having that kind of official thing to fall back on it can be a little bit rough.”

Additional coaching and consultationsupport (15 responses). To support supervisors in effectively implementing just culture within their teams, participants recommended that organizations provide ongoing coaching and mentorship opportunities. Additionally, third-party consultants with expertise in just culture were described as offering valuable guidance, particularly in cases where internal staff resources or HRO lead bandwidth may be limited. “There are so many consulting agencies with HRO that have been contracted to do different projects, but maybe that can help with an educational program.” “I want to see my executive leadership coach the supervisors up right and then allow them to do one-on-ones and facilitate and empower the frontline staff, and it’s just a good way of transparency and communication.”

Improved leadership sponsorship (15 responses). Participants noted that leadership buy-in is crucial for the successful implementation of just culture. Facilities should actively engage and educate leadership teams on the benefits of just culture and how it aligns with broader patient safety and organizational goals. Leaders should be visible and active champions of its principles, supporting change in their daily engagements with staff. “ELT support is absolutely necessary. Why? Because they will make it important to those in their service lines. They will make it important to those supervisors and managers. If it’s not important to that ELT member, then it’s not going to be important to that manager or that supervisor.”

Improved collaboration with patient safety and human resources (6 responses). Collaborative efforts with patient safety and human resources departments were seen as instrumental in supporting just culture, emphasizing its importance, and effectively addressing issues. Coordinating with these departments specifically contributes to consistent reinforcement and expands the bandwidth of HRO leads. These departments play integral roles in supporting just culture through effective policies, procedures, and communication. “I think it would be really helpful to have common language between what human resources teaches and what is in our decision support tool.”

DISCUSSION

This study sought to collect and synthesize the experiences of leaders across a large, integrated health care system in establishing and sustaining a just culture as part of an enterprise journey to become an HRO.²⁴ The VHA has provided enterprise-wide support (eg, training, leader coaching, and communications) for the implementation of HRO principles and practices with the goal of creating a culture of safety, which includes just culture elements. This support includes enterprise program offices, VISNs, and hospital facilities, though notably, there is variability in how HRO is implemented at the local level. The facilitators, barriers, and recommendations presented in this article are representative of the designated HRO leads at VHA hospital facilities who have direct experience with implementing and sustaining just culture. The themes presented offer specific opportunities for intervention and actionable strategies to enhance just culture initiatives, foster psychological safety and accountability, and ultimately improve the quality of care and patient outcomes.^3,25

Frequently identified facilitators such as providing training and coaching, having leaders who are available and approachable, demonstrating follow-through to address identified issues, and creating venues where errors and successes can be openly discussed.²⁶ These facilitators are aligned with enterprise HRO support strategies orchestrated by the VHA at the enterprise VISN and facility levels to support a culture of safety, continuous process improvement, and leadership commitment.

Frequently identified barriers included inadequate training, inconsistent application of just culture by middle managers vs senior leaders, a lack of accountability or corrective action when unjust corrective actions took place, time and resource constraints, and inadequate coordination across departments (eg, operational departments and human resources) and organizational levels. These factors were identified through focus groups with a limited set of HRO leads. They may reflect challenges to changing culture that may be deeply engrained in individual histories, organizational norms, and systemic practices. Improving upon these just culture initiatives requires multifaceted approaches and working through resistance to change.

VHA HRO leads identified several actionable recommendations that may be used in pursuit of a just culture. First, improvements in training involving how to apply just culture principles and, specifically, the use of the Just Culture DST were identified as an opportunity for improvement. The VHA National Center for Patient Safety developed the DST as an aid for leaders to effectively address errors in line with just culture principles, balancing individual and system accountability.²⁷ The DST specifically addresses human error as well as risky and reckless behavior, and it clarifies the delineation between individual and organizational accountability (Table).³

Limitations

This study employed qualitative methods and sampled a relatively small subset of experienced leaders with specific roles in implementing HRO in the VHA. Thus, it should not be considered representative of the perspectives of all leaders within the VHA or other health care systems. Future studies should assess facilitators and barriers beyond the facility level, including a focus incorporating both the VISN and VHA. More broadly, qualitative methods such as those employed in this study offer great depth and nuance but have limited ability to identify system-wide trends and differences. As such, it may be beneficial to specifically look at sites that are high- or low-performing on measures of patient safety culture to identify differences that may inform implementation strategies based on organizational maturity and readiness for change.

Conclusions

Successful implementation of these recommendations will require ongoing commitment, collaboration, and a sustained effort from all stakeholders involved at multiple levels of the health care system. Monitoring and evaluating progress should be conducted regularly to ensure that recommendations lead to improvements in implementing just culture principles. This quality improvement study adds to the knowledge base on factors that impact the just culture and broader efforts to realize HRO principles and practices in health care systems. The approach of this study may serve as a model for identifying opportunities to improve HRO implementation within the VHA and other settings, especially when paired with ongoing quantitative evaluation of organizational safety culture, just culture behaviors, and patient outcomes.

Medical errors are a persistent problem and leading cause of preventable death in the United States. There is considerable momentum behind the idea that implementation of a just culture is foundational to detecting and learning from errors in pursuit of zero patient harm.^1-6 Just culture is a framework that fosters an environment of trust within health care organizations, aiming to achieve fair outcomes for those involved in incidents or near misses. It emphasizes openness, accountability, and learning, prioritizing the repair of harm and systemic improvement over assigning blame.⁷

A just culture mindset reflects a significant shift in thinking that moves from the tendency to blame and punish others toward a focus on organizational learning and continued process improvement.^8,9 This systemic shift in fundamental thinking transforms how leaders approach staff errors and how they are addressed.¹⁰ In essence, just culture reflects an ethos centered on openness, a deep appreciation of human fallibility, and shared accountability at both the individual and organizational levels.

Organizational learning and innovation are stifled in the absence of a just culture, and there is a tendency for employees to avoid disclosing their own errors as well as those of their colleagues.¹¹ The transformation to a just culture is often slowed or disrupted by personal, systemic, and cultural barriers.¹² It is imperative that all executive, service line, and frontline managers recognize and execute their distinct responsibilities while adjudicating the appropriate course of action in the aftermath of adverse events or near misses. This requires a nuanced understanding of the factors that contribute to errors at the individual and organizational levels to ensure an appropriate response.

The Veterans Health Administration (VHA) is orchestrating an enterprise transformation to develop into a high reliability organization (HRO). This began with a single-site test in 2016, which demonstrated successful results in patient safety culture, patient safety event reporting, and patient safety outcomes.¹³ In 2019, the VHA formally launched its enterprise-wide HRO journey in 18 hospital facilities, followed by successive waves of 67 and 54 facilities in 2021 and 2022, respectively. The VHA journey to transform into an HRO aligns with 3 pillars, 5 principles, and 7 values. The VHA has emphasized the importance of just culture as a foundational element of the HRO framework, specifically under the pillar of leadership. To promote leadership engagement, the VHA has employed an array of approaches that include education, leader coaching, and change management strategies. Given the diversity among VHA facilities, each with local cultures and histories, some sites have more readily implemented a just culture than others.¹⁴ A deeper exploration into potential obstacles, particularly concerning leadership engagement, could be instrumental for formulating strategies that further establish a just culture across the VHA.¹⁵

There is a paucity of empirical research regarding factors that facilitate and/or impede the implementation of a just culture in health care settings.^16,17 Likert scale surveys, such as the Patient Safety Culture Module for the VHA All Employee Survey and its predecessor, the Patient Safety Culture Survey, have been used to assess culture and climate.¹⁸ However, qualitative evaluations directly assessing the lived experiences of those trying to implement a just culture provide additional depth and context that can help identify specific factors that support or impede becoming an HRO. The purpose of this study was to increase understanding of factors that influence the establishment and sustainment of a just culture and to identify specific methods for improving the implementation of just culture principles and practices aligned with HRO.

METHODS

This qualitative study explored facilitators and barriers to establishing and sustaining a just culture as experienced across a subset of VHA facilities by HRO leads or staff assigned with the primary responsibilities of supporting facility-level HRO transformation. HRO leads are assigned responsibility for supporting executive leadership in planning, coordinating, implementing, and monitoring activities to ensure effective high reliability efforts, including focused efforts to establish a robust patient safety program, a culture of safety, and a culture of continuous process improvement.

Virtual focus group discussions held via Microsoft Teams generated in-depth, diverse perspectives from participants across 16 VHA facilities. Qualitative research and evaluation methods provide an enhanced depth of understanding and allow the emergence of detailed data.¹⁹ A qualitative grounded theory approach elicits complex, multifaceted phenomena that cannot be appreciated solely by numeric data.²⁰ Grounded theory was selected to limit preconceived notions and provide a more systematic analysis, including open, axial, and thematic coding. Such methods afford opportunities to adapt to unplanned follow-up questions and thus provide a flexible approach to generate new ideas and concepts.²¹ Additionally, qualitative methods help overcome the tendencies of respondents to agree rather than disagree when presented with Likert-style scales, which tend to skew responses toward the positive.²²

Participants must have been assigned as an HRO lead for ≥ 6 months at the same facility. Potential participants were identified through purposive sampling, considering their leadership roles in HRO and experience with just culture implementation, the size and complexity of their facility, and geographic distribution. Invitations explaining the study and encouraging voluntary participation to participate were emailed. Of 37 HRO leads invited to participate in the study, 16 agreed to participate and attended 1 of 3 hour-long focus group sessions. One session was rescheduled due to limited attendance. Participants represented a mix of VHA sites in terms of geography, facility size, and complexity.

Focus Group Procedures

Demographic data were collected prior to sessions via an online form to better understand the participant population, including facility complexity level, length of time in HRO lead role, clinical background, and facility level just culture training. Each session was led by an experienced focus group facilitator (CV) who was not directly involved with the overall HRO implementation to establish a neutral perspective. Each session was attended by 4 to 7 participants and 2 observers who took notes. The sessions were recorded and included automated transcriptions, which were edited for accuracy.

Focus group sessions began with a brief introduction and an opportunity for participants to ask questions. Participants were then asked a series of open-ended questions to elicit responses regardingfacilitators, barriers, and leadership support needed for implementing just culture. The questions were part of a facilitator guide that included an introductory script and discussion prompts to ensure consistency across focus groups.

Facilitators were defined as factors that increase the likelihood of establishing or sustaining a just culture. Barriers were defined as factors that decrease or inhibit the likelihood of establishing or sustaining just culture. The focus group facilitator encouraged all participants to share their views and provided clarification if needed, as well as prompts and examples where appropriate, but primarily sought to elicit responses to the questions.

Institutional review board review and approval were not required for this quality improvement initiative. The project adhered to ethical standards of research, including asking participants for verbal consent and preserving their confidentiality. Participation was voluntary, and prior to the focus group sessions, participants were provided information explaining the study’s purpose, procedures, and their rights. Participant identities were kept confidential, and all data were anonymized during the analysis phase. Pseudonyms or identifiers were used during data transcription to protect participant identity. All data, including recordings and transcriptions, were stored on password-protected devices accessible only to the research team. Any identifiable information was removed during data analysis to ensure confidentiality.

Analysis

Focus group recordings were transcribed verbatim, capturing all verbal interactions and nonverbal cues that may contribute to understanding the participants' perspectives. Thematic analysis was used to analyze the qualitative data from the focus group discussions.²³ The transcribed data were organized, coded, and analyzed using ATLAS.ti 23 qualitative data software to identify key themes and patterns.

Results

The themes identified include the 5 facilitators, barriers, and recommendations most frequently mentioned by HRO leads across focus group sessions. The nature of each theme is described, along with commonly mentioned examples and direct quotes from participants that illustrate our understanding of their perspectives.

Facilitators

Training and coaching (26 responses). The availability of training around the Just Culture Decision Support Tool (DST) was cited as a practical aid in guiding leaders through complex just culture decisions to ensure consistency and fairness. Additionally, an executive leadership team that served as champions for just culture principles played a vital role in promoting and sustaining the approach: “Training them on the roll-out of the decision support tool with supervisors at all levels, and education for just culture and making it part of our safety forum has helped for the last 4 months.” “Having some regular training and share-out cadences embedded within the schedule as well as dynamic directors and well-trained executive leadership team (ELT) for support has been a facilitator.”

Increased transparency (16 responses). Participants consistently highlighted the importance of leadership transparency as a key facilitator for implementing just culture. Open and honest communication from top-level executives fostered an environment of trust and accountability. Approachable and physically present leadership was seen as essential for creating a culture where employees felt comfortable reporting errors and concerns without fear of retaliation: “They’re surprisingly honest with themselves about what we can do, what we cannot do, and they set the expectations exactly at that.”

Approachable leadership (15 responses). Participants frequently mentioned the importance of having dynamic leadership spearheading the implementation of just culture and leading by example. Having a leadership team that accepts accountability and reinforces consistency in the manner that near misses or mistakes are addressed is paramount to promoting the principles of just culture and increasing psychological safety: “We do have very approachable leadership, which I know is hard if you’re trying to implement that nationwide, it’s hard to implement approachability. But I do think that people raise their concerns, and they’ll stop them in the hallway and ask them questions. So, in terms of comfort level with the executive leadership, I do think that’s high, which would promote psychological safety.”

Feedback loops and follow through (13 responses). Participants emphasized the importance of taking concrete actions to address concerns and improve processes. Regular check-ins with supervisors to discuss matters related to just culture provided a structured opportunity for addressing issues and reinforcing the importance of the approach: “One thing that we’ve really focused on is not only identifying mistakes, but [taking] ownership. We continue to track it until … it’s completed and then a process of how to communicate that back and really using closed loop communication with the staff and letting them know.”

Forums and town halls (10 responses). These platforms created feedback loops that were seen as invaluable tools for sharing near misses, celebrating successes, and promoting open dialogue. Forums and town halls cultivated a culture of continuous improvement and trust: “We’ll celebrate catches, a safety story is inside that catch. So, if we celebrate the change, people feel safer to speak up.” “Truthfully, we’ve had a great relationship since establishing our safety forums and just value open lines of communication.”

Barriers

Inadequate training (30 responses). Insufficient engagement during training—limited bandwidth and availability to attend and actively participate in training—was perceived as detrimental to creating awareness and buy-in from staff, supervisors, and leadership, thereby hindering successful integration of just culture principles. Participants also identified too many conflicting priorities from VHA leadership, which contributes to training and information fatigue among staff and supervisors. “Our biggest barrier is just so many different competing priorities going on. We have so much that we’re asking people to do.” “One hundred percent training is feeling more like a ticked box than actually yielding results, I have a very hard time getting staff engaged.”

Inconsistency between executive leaders and middle managers (28 responses). A lack of consistency in the commitment to and enactment of just culture principles among leaders poses a challenge. Participants gave several examples of inconsistencies in messaging and reinforcement of just culture principles, leading to confusion among staff and hindering adoption. Likewise, the absence of standardized procedures for implementing just culture created variability: “The director coming in and trying to change things, it put a lot of resistance, we struggle with getting the other ELT members on board … some of the messages that come out at times can feel more punitive.”

Middle management resistance (22 responses). In some instances, participants reported middle managers exhibiting attitudes and behaviors that undermined the application of just culture principles and effectiveness. Such attitudes and behaviors were attributed to a lack of adequate training, coaching, and awareness. Other perceived contributions included fear of failure and a desire to distance oneself from staff who have made mistakes: “As soon as someone makes an error, they go straight to suspend them, and that’s the disconnect right there.” “There’s almost a level of working in the opposite direction in some of the mid-management.”

Cultural misalignment (18 responses). The existing culture of distancing oneself from mistakes presented a significant barrier to the adoption of just culture because it clashed with the principles of open reporting and accountability. Staff underreported errors or framed them in a way that minimized personal responsibility, thereby making it more essential to put in the necessary and difficult work to learn from mistakes: “One, you’re going to get in trouble. There’s going to be more work added to you or something of that nature."

Lack of accountability for opposition(17 responses). Participants noted a clear lack of accountability for those who opposed or showed resistance to just culture, which allowed resistance to persist without consequences. In many instances, leaders were described as having overlooked repeated instances of unjust attitudes and behaviors (eg, inappropriate blame or punishment), which allowed those practices to continue. “Executive leadership is standing on the hill and saying we’re a just culture and we do everything correctly, and staff has the expectation that they’re going to be treated with just culture and then the middle management is setting that on fire, then we show them that that’s not just culture, and they continue to have those poor behaviors, but there’s a lack of accountability.”

Limited bandwidth and lack of coordination (14 responses). HRO leads often faced role-specific constraints in having adequate time and authority to coordinate efforts to implement or sustain just culture. This includes challenges with coordination across organizational levels (eg, between the hospital and regional Veterans Integrated Service Network [VISN] management levels) and across departments within the hospital (eg, between human resources and service lines or units). “Our VISN human resources is completely detached. They’ll not cooperate with these efforts, which is hard.” “There’s not enough bandwidth to actually support, I’m just 1 person.” “[There’s] all these mandated trainings of 8 hours when we’re already fatigued, short-staffed, taking 3 other HRO classes.”

Recommendations

Training improvements (24 responses). HRO leads recommended that comprehensive training programs be developed and implemented for staff, supervisors, and leadership to increase awareness and understanding of just culture principles. These training initiatives should focus on fostering a shared understanding of the core tenets of just culture, the importance of error reporting, and the processes involved in fair and consistent decision making (eg, training simulations on use of the Just Culture DST). “We’ve really never had any formal training on the decision support tool. I hope that what’s coming out for next year. We’ll have some more formal training for the tool because I think it would be great to really have our leadership and our supervisors and our managers use that tool.” “We can give a more directed and intentional training to leadership on the 4 foundational practices and what it means to implement those and what it means to utilize that behavioral component of HRO.”

Clear and consistent procedures toincrease accountability (22 responses). To promote a culture of accountability and consistency in the application of just culture principles, organizations should establish clear mechanisms for reporting, investigating, and addressing incidents. Standardized procedures and DSTs can aid in ensuring that responses to errors are equitable and align with just culture principles: “I recommend accountability; if it’s clearly evidenced that you’re not toeing the just culture line, then we need to be able to do something about it and not just finger wag.” “[We need to have] a templated way to approach just culture implementation. The decision support tool is great, I absolutely love having the resources and being able to find a lot of clinical examples and discussion tools like that. But when it comes down to it, not having that kind of official thing to fall back on it can be a little bit rough.”

Additional coaching and consultationsupport (15 responses). To support supervisors in effectively implementing just culture within their teams, participants recommended that organizations provide ongoing coaching and mentorship opportunities. Additionally, third-party consultants with expertise in just culture were described as offering valuable guidance, particularly in cases where internal staff resources or HRO lead bandwidth may be limited. “There are so many consulting agencies with HRO that have been contracted to do different projects, but maybe that can help with an educational program.” “I want to see my executive leadership coach the supervisors up right and then allow them to do one-on-ones and facilitate and empower the frontline staff, and it’s just a good way of transparency and communication.”

Improved leadership sponsorship (15 responses). Participants noted that leadership buy-in is crucial for the successful implementation of just culture. Facilities should actively engage and educate leadership teams on the benefits of just culture and how it aligns with broader patient safety and organizational goals. Leaders should be visible and active champions of its principles, supporting change in their daily engagements with staff. “ELT support is absolutely necessary. Why? Because they will make it important to those in their service lines. They will make it important to those supervisors and managers. If it’s not important to that ELT member, then it’s not going to be important to that manager or that supervisor.”

Improved collaboration with patient safety and human resources (6 responses). Collaborative efforts with patient safety and human resources departments were seen as instrumental in supporting just culture, emphasizing its importance, and effectively addressing issues. Coordinating with these departments specifically contributes to consistent reinforcement and expands the bandwidth of HRO leads. These departments play integral roles in supporting just culture through effective policies, procedures, and communication. “I think it would be really helpful to have common language between what human resources teaches and what is in our decision support tool.”

DISCUSSION

This study sought to collect and synthesize the experiences of leaders across a large, integrated health care system in establishing and sustaining a just culture as part of an enterprise journey to become an HRO.²⁴ The VHA has provided enterprise-wide support (eg, training, leader coaching, and communications) for the implementation of HRO principles and practices with the goal of creating a culture of safety, which includes just culture elements. This support includes enterprise program offices, VISNs, and hospital facilities, though notably, there is variability in how HRO is implemented at the local level. The facilitators, barriers, and recommendations presented in this article are representative of the designated HRO leads at VHA hospital facilities who have direct experience with implementing and sustaining just culture. The themes presented offer specific opportunities for intervention and actionable strategies to enhance just culture initiatives, foster psychological safety and accountability, and ultimately improve the quality of care and patient outcomes.^3,25

Frequently identified facilitators such as providing training and coaching, having leaders who are available and approachable, demonstrating follow-through to address identified issues, and creating venues where errors and successes can be openly discussed.²⁶ These facilitators are aligned with enterprise HRO support strategies orchestrated by the VHA at the enterprise VISN and facility levels to support a culture of safety, continuous process improvement, and leadership commitment.

Frequently identified barriers included inadequate training, inconsistent application of just culture by middle managers vs senior leaders, a lack of accountability or corrective action when unjust corrective actions took place, time and resource constraints, and inadequate coordination across departments (eg, operational departments and human resources) and organizational levels. These factors were identified through focus groups with a limited set of HRO leads. They may reflect challenges to changing culture that may be deeply engrained in individual histories, organizational norms, and systemic practices. Improving upon these just culture initiatives requires multifaceted approaches and working through resistance to change.

VHA HRO leads identified several actionable recommendations that may be used in pursuit of a just culture. First, improvements in training involving how to apply just culture principles and, specifically, the use of the Just Culture DST were identified as an opportunity for improvement. The VHA National Center for Patient Safety developed the DST as an aid for leaders to effectively address errors in line with just culture principles, balancing individual and system accountability.²⁷ The DST specifically addresses human error as well as risky and reckless behavior, and it clarifies the delineation between individual and organizational accountability (Table).³

Limitations

This study employed qualitative methods and sampled a relatively small subset of experienced leaders with specific roles in implementing HRO in the VHA. Thus, it should not be considered representative of the perspectives of all leaders within the VHA or other health care systems. Future studies should assess facilitators and barriers beyond the facility level, including a focus incorporating both the VISN and VHA. More broadly, qualitative methods such as those employed in this study offer great depth and nuance but have limited ability to identify system-wide trends and differences. As such, it may be beneficial to specifically look at sites that are high- or low-performing on measures of patient safety culture to identify differences that may inform implementation strategies based on organizational maturity and readiness for change.

Conclusions

Successful implementation of these recommendations will require ongoing commitment, collaboration, and a sustained effort from all stakeholders involved at multiple levels of the health care system. Monitoring and evaluating progress should be conducted regularly to ensure that recommendations lead to improvements in implementing just culture principles. This quality improvement study adds to the knowledge base on factors that impact the just culture and broader efforts to realize HRO principles and practices in health care systems. The approach of this study may serve as a model for identifying opportunities to improve HRO implementation within the VHA and other settings, especially when paired with ongoing quantitative evaluation of organizational safety culture, just culture behaviors, and patient outcomes.