Article

Key clinical point: Increased cardiovascular (CV) risk, as assessed by the Systematic Coronary Risk Evaluation 2 system, was associated with a lower risk for prevalent and incident migraine, notably in women.

Major findings: Compared with individuals with the lowest CV risk scores (< 1%), those with CV risk scores of 1% to < 2.5% had a higher risk for prevalent migraine (odds ratio [OR], 0.93; 95% CI, 0.90-0.96) and incident migraine (OR, 0.63; 95% CI, 0.57-0.69), whereas individuals with CV risk scores ≥ 10% had a lower risk for prevalent migraine (OR, 0.43; 95% CI, 0.39-0.48) and incident migraine (OR, 0.17; 95% CI, 0.10-0.27). The association between migraine and CV risk was stronger in women than in men within the same CV risk categories.

Study details: This ~13-year prospective population-based study involved 140,915 adults (58.5% women), with 25,915 having prevalent migraine and 2224 having incident migraine.

Disclosure: The study was supported by the Dutch Research Council and others. Two authors reported receiving research grants, advisory board fees, or speaking fees from various sources.

Source: Al-Hassany L, MaassenVanDenBrink A, Kurth T. Cardiovascular risk scores and migraine status. JAMA Netw Open. Published online October 22, 2024. Source

Key clinical point: Increased cardiovascular (CV) risk, as assessed by the Systematic Coronary Risk Evaluation 2 system, was associated with a lower risk for prevalent and incident migraine, notably in women.

Major findings: Compared with individuals with the lowest CV risk scores (< 1%), those with CV risk scores of 1% to < 2.5% had a higher risk for prevalent migraine (odds ratio [OR], 0.93; 95% CI, 0.90-0.96) and incident migraine (OR, 0.63; 95% CI, 0.57-0.69), whereas individuals with CV risk scores ≥ 10% had a lower risk for prevalent migraine (OR, 0.43; 95% CI, 0.39-0.48) and incident migraine (OR, 0.17; 95% CI, 0.10-0.27). The association between migraine and CV risk was stronger in women than in men within the same CV risk categories.

Study details: This ~13-year prospective population-based study involved 140,915 adults (58.5% women), with 25,915 having prevalent migraine and 2224 having incident migraine.

Disclosure: The study was supported by the Dutch Research Council and others. Two authors reported receiving research grants, advisory board fees, or speaking fees from various sources.

Source: Al-Hassany L, MaassenVanDenBrink A, Kurth T. Cardiovascular risk scores and migraine status. JAMA Netw Open. Published online October 22, 2024. Source

Key clinical point: Increased cardiovascular (CV) risk, as assessed by the Systematic Coronary Risk Evaluation 2 system, was associated with a lower risk for prevalent and incident migraine, notably in women.

Major findings: Compared with individuals with the lowest CV risk scores (< 1%), those with CV risk scores of 1% to < 2.5% had a higher risk for prevalent migraine (odds ratio [OR], 0.93; 95% CI, 0.90-0.96) and incident migraine (OR, 0.63; 95% CI, 0.57-0.69), whereas individuals with CV risk scores ≥ 10% had a lower risk for prevalent migraine (OR, 0.43; 95% CI, 0.39-0.48) and incident migraine (OR, 0.17; 95% CI, 0.10-0.27). The association between migraine and CV risk was stronger in women than in men within the same CV risk categories.

Study details: This ~13-year prospective population-based study involved 140,915 adults (58.5% women), with 25,915 having prevalent migraine and 2224 having incident migraine.

Disclosure: The study was supported by the Dutch Research Council and others. Two authors reported receiving research grants, advisory board fees, or speaking fees from various sources.

Source: Al-Hassany L, MaassenVanDenBrink A, Kurth T. Cardiovascular risk scores and migraine status. JAMA Netw Open. Published online October 22, 2024. Source

Key clinical point: In patients with migraine and those with prior treatment failure and medication overuse (MO), eptinezumab significantly reduced acute headache medication (AHM), with sustained reductions for up to 18 months.

Major findings: Eptinezumab significantly reduced AHM compared with placebo in patients with migraine and those with MO (P < .001 for all). These reductions in AHM were sustained through the 18-month treatment period, with continued reductions in the extension period. 

Study details: This DELIVER trial post-hoc analysis of 890 patients, 49.2% had MO who failed 2-4 preventive treatments, among these 865 patients continued in the 48-week extension period. 

Disclosure: This study was funded by H. Lundbeck A/S. Two authors reported receiving speaking fees from various sources. Five authors declared being full-time employees of H. Lundbeck A/S. Other authors declared multiple ties with various sources.

Source: Gryglas-Dworak A, Schim J, Ettrup A, et al. Long-term reductions in acute headache medication use after eptinezumab treatment in patients with migraine and prior preventive treatment failures: Post hoc analysis of the DELIVER randomized trial. Headache. Published online November 5, 2024. Source

Key clinical point: In patients with migraine and those with prior treatment failure and medication overuse (MO), eptinezumab significantly reduced acute headache medication (AHM), with sustained reductions for up to 18 months.

Major findings: Eptinezumab significantly reduced AHM compared with placebo in patients with migraine and those with MO (P < .001 for all). These reductions in AHM were sustained through the 18-month treatment period, with continued reductions in the extension period. 

Study details: This DELIVER trial post-hoc analysis of 890 patients, 49.2% had MO who failed 2-4 preventive treatments, among these 865 patients continued in the 48-week extension period. 

Disclosure: This study was funded by H. Lundbeck A/S. Two authors reported receiving speaking fees from various sources. Five authors declared being full-time employees of H. Lundbeck A/S. Other authors declared multiple ties with various sources.

Source: Gryglas-Dworak A, Schim J, Ettrup A, et al. Long-term reductions in acute headache medication use after eptinezumab treatment in patients with migraine and prior preventive treatment failures: Post hoc analysis of the DELIVER randomized trial. Headache. Published online November 5, 2024. Source

Key clinical point: In patients with migraine and those with prior treatment failure and medication overuse (MO), eptinezumab significantly reduced acute headache medication (AHM), with sustained reductions for up to 18 months.

Major findings: Eptinezumab significantly reduced AHM compared with placebo in patients with migraine and those with MO (P < .001 for all). These reductions in AHM were sustained through the 18-month treatment period, with continued reductions in the extension period. 

Study details: This DELIVER trial post-hoc analysis of 890 patients, 49.2% had MO who failed 2-4 preventive treatments, among these 865 patients continued in the 48-week extension period. 

Disclosure: This study was funded by H. Lundbeck A/S. Two authors reported receiving speaking fees from various sources. Five authors declared being full-time employees of H. Lundbeck A/S. Other authors declared multiple ties with various sources.

Source: Gryglas-Dworak A, Schim J, Ettrup A, et al. Long-term reductions in acute headache medication use after eptinezumab treatment in patients with migraine and prior preventive treatment failures: Post hoc analysis of the DELIVER randomized trial. Headache. Published online November 5, 2024. Source

Key clinical point: Patients with migraine, particularly those with severe symptoms, were at a higher risk for occupational burnout than those without migraine.

Major findings: Patients with vs without migraine experienced increased occupational burnout (mean burnout score, 3.46 vs 2.82), had higher rates of depression (mean Depression, Anxiety and Stress Scale–21 score, 0.864 vs 0.664), worked longer hours (median weekly hours, 40.0 vs 36.0), and preferred remote work (20.3% vs 10.3%). Migraine severity, as measured by the Migraine Disability Assessment, was significantly associated with increased occupational burnout (P < .001).

Study details: This observational cohort study included 675 patients with migraine and 232 patients without migraine, matched according to age, gender, and primary clinic. 

Disclosure: The study was funded by Teva Pharmaceuticals. Gal Ifergane received consulting fees and honoraria from various sources, including Teva Pharmaceuticals, and others declared no conflicts of interest.

Source: Peles I, Sharvit S, Zlotnik Y, et al. Migraine and work — beyond absenteeism: Migraine severity and occupational burnout — A cohort study. Cephalalgia. Published online October 18, 2024. Source

Key clinical point: Patients with migraine, particularly those with severe symptoms, were at a higher risk for occupational burnout than those without migraine.

Major findings: Patients with vs without migraine experienced increased occupational burnout (mean burnout score, 3.46 vs 2.82), had higher rates of depression (mean Depression, Anxiety and Stress Scale–21 score, 0.864 vs 0.664), worked longer hours (median weekly hours, 40.0 vs 36.0), and preferred remote work (20.3% vs 10.3%). Migraine severity, as measured by the Migraine Disability Assessment, was significantly associated with increased occupational burnout (P < .001).

Study details: This observational cohort study included 675 patients with migraine and 232 patients without migraine, matched according to age, gender, and primary clinic. 

Disclosure: The study was funded by Teva Pharmaceuticals. Gal Ifergane received consulting fees and honoraria from various sources, including Teva Pharmaceuticals, and others declared no conflicts of interest.

Source: Peles I, Sharvit S, Zlotnik Y, et al. Migraine and work — beyond absenteeism: Migraine severity and occupational burnout — A cohort study. Cephalalgia. Published online October 18, 2024. Source

Key clinical point: Patients with migraine, particularly those with severe symptoms, were at a higher risk for occupational burnout than those without migraine.

Major findings: Patients with vs without migraine experienced increased occupational burnout (mean burnout score, 3.46 vs 2.82), had higher rates of depression (mean Depression, Anxiety and Stress Scale–21 score, 0.864 vs 0.664), worked longer hours (median weekly hours, 40.0 vs 36.0), and preferred remote work (20.3% vs 10.3%). Migraine severity, as measured by the Migraine Disability Assessment, was significantly associated with increased occupational burnout (P < .001).

Study details: This observational cohort study included 675 patients with migraine and 232 patients without migraine, matched according to age, gender, and primary clinic. 

Disclosure: The study was funded by Teva Pharmaceuticals. Gal Ifergane received consulting fees and honoraria from various sources, including Teva Pharmaceuticals, and others declared no conflicts of interest.

Source: Peles I, Sharvit S, Zlotnik Y, et al. Migraine and work — beyond absenteeism: Migraine severity and occupational burnout — A cohort study. Cephalalgia. Published online October 18, 2024. Source

Key clinical point: Migraine, including migraine with aura (MWA) and migraine without aura (MWoA), was not a significant factor for stroke caused by cerebral small-vessel disease (CSVD) in adult patients hospitalized with first-ever ischemic stroke. 

Major findings: Patients with migraine (P = .003), those with MWA (P = .02), and those with MWoA (P = .04) were found to have a significantly lower prevalence of CSVD lesions than those without migraine. There was no significant association between CSVD lesions of any grade and migraine (P = .34), MWA (P = .60), and MWoA (P = .51).  

Study details: This retrospective case-control study included 646 patients who were hospitalized for their first-ever ischemic stroke, of whom 225 (34.82%) had migraine. Among those with migraine, 115 (17.8%) had MWA and 110 (17.02%) had MWoA. 

Disclosure: This study received no specific funding. One author declared receiving speaker fees from various pharmaceutical companies. 

Source: Cloet F, Gueyraud G, Lerebours F, Munio M, Larrue V, Gollion C. Stroke due to small-vessel disease and migraine: A case-control study of a young adult with ischemic stroke population. Cephalalgia. 2024;44:1-8. Source 

Key clinical point: Migraine, including migraine with aura (MWA) and migraine without aura (MWoA), was not a significant factor for stroke caused by cerebral small-vessel disease (CSVD) in adult patients hospitalized with first-ever ischemic stroke. 

Major findings: Patients with migraine (P = .003), those with MWA (P = .02), and those with MWoA (P = .04) were found to have a significantly lower prevalence of CSVD lesions than those without migraine. There was no significant association between CSVD lesions of any grade and migraine (P = .34), MWA (P = .60), and MWoA (P = .51).  

Study details: This retrospective case-control study included 646 patients who were hospitalized for their first-ever ischemic stroke, of whom 225 (34.82%) had migraine. Among those with migraine, 115 (17.8%) had MWA and 110 (17.02%) had MWoA. 

Disclosure: This study received no specific funding. One author declared receiving speaker fees from various pharmaceutical companies. 

Source: Cloet F, Gueyraud G, Lerebours F, Munio M, Larrue V, Gollion C. Stroke due to small-vessel disease and migraine: A case-control study of a young adult with ischemic stroke population. Cephalalgia. 2024;44:1-8. Source 

Key clinical point: Migraine, including migraine with aura (MWA) and migraine without aura (MWoA), was not a significant factor for stroke caused by cerebral small-vessel disease (CSVD) in adult patients hospitalized with first-ever ischemic stroke. 

Major findings: Patients with migraine (P = .003), those with MWA (P = .02), and those with MWoA (P = .04) were found to have a significantly lower prevalence of CSVD lesions than those without migraine. There was no significant association between CSVD lesions of any grade and migraine (P = .34), MWA (P = .60), and MWoA (P = .51).  

Study details: This retrospective case-control study included 646 patients who were hospitalized for their first-ever ischemic stroke, of whom 225 (34.82%) had migraine. Among those with migraine, 115 (17.8%) had MWA and 110 (17.02%) had MWoA. 

Disclosure: This study received no specific funding. One author declared receiving speaker fees from various pharmaceutical companies. 

Source: Cloet F, Gueyraud G, Lerebours F, Munio M, Larrue V, Gollion C. Stroke due to small-vessel disease and migraine: A case-control study of a young adult with ischemic stroke population. Cephalalgia. 2024;44:1-8. Source 

Key clinical point: Switching to intravenous eptinezumab may benefit patients with treatment-refractory migraine who have previously failed subcutaneous calcitonin gene-related peptide-receptor (CGRP-R) monoclonal antibodies (mAbs).

Major findings: At 12 and 24 weeks of eptinezumab treatment, 23.1% and 29.7% of patients had ≥30% reduction in monthly migraine days, whereas 15.4% and 21.4% had ≥30% reduction in monthly headache days, respectively. At 21-24 weeks, 38.5% and 52.4% of patients showed significant reductions in the Headache Impact Test and Migraine Disability Assessment scores, respectively. No adverse events were reported during the 24-week treatment period.

Study details: This monocentric retrospective longitudinal cohort study included 41 patients with migraine unresponsive to ≥1 subcutaneous CGRP mAb, who received an initial 100 mg dose of intravenous eptinezumab, followed by 100 or 300 mg after 12 weeks.

Disclosure: This study was supported by the Lundbeck Foundation. Several authors declared receiving personal fees from various sources. 

Source: Triller P, Blessing VN, Overeem LH, et al. Efficacy of eptinezumab in non-responders to subcutaneous monoclonal antibodies against CGRP and the CGRP receptor: A retrospective cohort study. Cephalalgia. Published online October 29, 2024. Source

Key clinical point: Switching to intravenous eptinezumab may benefit patients with treatment-refractory migraine who have previously failed subcutaneous calcitonin gene-related peptide-receptor (CGRP-R) monoclonal antibodies (mAbs).

Major findings: At 12 and 24 weeks of eptinezumab treatment, 23.1% and 29.7% of patients had ≥30% reduction in monthly migraine days, whereas 15.4% and 21.4% had ≥30% reduction in monthly headache days, respectively. At 21-24 weeks, 38.5% and 52.4% of patients showed significant reductions in the Headache Impact Test and Migraine Disability Assessment scores, respectively. No adverse events were reported during the 24-week treatment period.

Study details: This monocentric retrospective longitudinal cohort study included 41 patients with migraine unresponsive to ≥1 subcutaneous CGRP mAb, who received an initial 100 mg dose of intravenous eptinezumab, followed by 100 or 300 mg after 12 weeks.

Disclosure: This study was supported by the Lundbeck Foundation. Several authors declared receiving personal fees from various sources. 

Source: Triller P, Blessing VN, Overeem LH, et al. Efficacy of eptinezumab in non-responders to subcutaneous monoclonal antibodies against CGRP and the CGRP receptor: A retrospective cohort study. Cephalalgia. Published online October 29, 2024. Source

Key clinical point: Switching to intravenous eptinezumab may benefit patients with treatment-refractory migraine who have previously failed subcutaneous calcitonin gene-related peptide-receptor (CGRP-R) monoclonal antibodies (mAbs).

Major findings: At 12 and 24 weeks of eptinezumab treatment, 23.1% and 29.7% of patients had ≥30% reduction in monthly migraine days, whereas 15.4% and 21.4% had ≥30% reduction in monthly headache days, respectively. At 21-24 weeks, 38.5% and 52.4% of patients showed significant reductions in the Headache Impact Test and Migraine Disability Assessment scores, respectively. No adverse events were reported during the 24-week treatment period.

Study details: This monocentric retrospective longitudinal cohort study included 41 patients with migraine unresponsive to ≥1 subcutaneous CGRP mAb, who received an initial 100 mg dose of intravenous eptinezumab, followed by 100 or 300 mg after 12 weeks.

Disclosure: This study was supported by the Lundbeck Foundation. Several authors declared receiving personal fees from various sources. 

Source: Triller P, Blessing VN, Overeem LH, et al. Efficacy of eptinezumab in non-responders to subcutaneous monoclonal antibodies against CGRP and the CGRP receptor: A retrospective cohort study. Cephalalgia. Published online October 29, 2024. Source

Key clinical point: An increased waist-to-height ratio (WHtR), indicating central obesity, was associated with an increased incidence of migraine, particularly in individuals aged <60 years. 

Major findings: Each unit increase in WHtR was associated with a 70% increase in the incidence of migraines (odds ratio [OR], 1.70; 95% CI, 1.04-2.78). Individuals in the highest WHtR quartile (WHtR values, 0.64-1.01) had a 13% greater incidence of migraines than those in the lowest quartile (WHtR values, 0.35-0.52; OR, 1.13; 95% CI, 0.99-1.28). For individuals aged <60 years, each unit increase in WHtR was associated with an 82% increased risk for migraine (P < .01); however, WHtR was negatively associated with migraine risk among those aged ≥60 years. 

Study details: This cross-sectional study analyzed data from the National Health and Nutrition Examination Survey, including 13,344 participants, of whom 2764 had migraines. Disclosure: The study did not receive any funding. The authors declared no conflicts of interest. 

Source: Jin J, Zheng Y, Gao T, Lin X, Li S, Huang C. Associations between the waist-to-height ratio index and migraine: A cross-section study of the NHANES 1999–2004. PLoS ONE. Published online October 23, 2024. Source

Key clinical point: An increased waist-to-height ratio (WHtR), indicating central obesity, was associated with an increased incidence of migraine, particularly in individuals aged <60 years. 

Major findings: Each unit increase in WHtR was associated with a 70% increase in the incidence of migraines (odds ratio [OR], 1.70; 95% CI, 1.04-2.78). Individuals in the highest WHtR quartile (WHtR values, 0.64-1.01) had a 13% greater incidence of migraines than those in the lowest quartile (WHtR values, 0.35-0.52; OR, 1.13; 95% CI, 0.99-1.28). For individuals aged <60 years, each unit increase in WHtR was associated with an 82% increased risk for migraine (P < .01); however, WHtR was negatively associated with migraine risk among those aged ≥60 years. 

Study details: This cross-sectional study analyzed data from the National Health and Nutrition Examination Survey, including 13,344 participants, of whom 2764 had migraines. Disclosure: The study did not receive any funding. The authors declared no conflicts of interest. 

Source: Jin J, Zheng Y, Gao T, Lin X, Li S, Huang C. Associations between the waist-to-height ratio index and migraine: A cross-section study of the NHANES 1999–2004. PLoS ONE. Published online October 23, 2024. Source

Key clinical point: An increased waist-to-height ratio (WHtR), indicating central obesity, was associated with an increased incidence of migraine, particularly in individuals aged <60 years. 

Major findings: Each unit increase in WHtR was associated with a 70% increase in the incidence of migraines (odds ratio [OR], 1.70; 95% CI, 1.04-2.78). Individuals in the highest WHtR quartile (WHtR values, 0.64-1.01) had a 13% greater incidence of migraines than those in the lowest quartile (WHtR values, 0.35-0.52; OR, 1.13; 95% CI, 0.99-1.28). For individuals aged <60 years, each unit increase in WHtR was associated with an 82% increased risk for migraine (P < .01); however, WHtR was negatively associated with migraine risk among those aged ≥60 years. 

Study details: This cross-sectional study analyzed data from the National Health and Nutrition Examination Survey, including 13,344 participants, of whom 2764 had migraines. Disclosure: The study did not receive any funding. The authors declared no conflicts of interest. 

Source: Jin J, Zheng Y, Gao T, Lin X, Li S, Huang C. Associations between the waist-to-height ratio index and migraine: A cross-section study of the NHANES 1999–2004. PLoS ONE. Published online October 23, 2024. Source

Supporting meaningful research that has a positive impact on the health and quality of life of veterans is a priority of the US Department of Veterans Affairs Office of Research and Development.¹ For nearly a century, VA researchers have been conducting high quality studies. To continue this trajectory, it is imperative to attract, train, and retain exceptional investigators while nurturing their development throughout their careers.²

Mentorship is defined as guidance provided by an experienced and trusted party to another (usually junior) individual with the intent of helping the person succeed. It benefits the mentee, mentor, and their institutions.³ Mentorship is crucial for personal and professional development as well as productivity, which may help reduce clinician burnout.^4-7 Conversely, a lack of mentorship could have negative effects on work satisfaction and stagnate career progression.⁸

Mentorship is vital for developing and advancing a VA investigator’s research agenda. Funding, grant writing, and research design were among the most discussed topics in a large comprehensive mentorship program for academic faculty.⁹ However, there are several known barriers to effective research mentorship; among them include a lack of resources, time constraints, and competing clinical priorities.^10,11

Finding time for effective one-on-one research mentoring is difficult within the time constraints of clinical duties; a group mentorship model may help overcome this barrier. Group mentorship can aid in personal and professional development because no single mentor can effectively meet every mentoring need of an individual.¹² Group mentorship also allows for the exchange of ideas among individuals with different backgrounds and the ability to utilize the strengths of each member of the group. For example, a member may have methodological expertise, while another may be skilled in grantsmanship. A team of mentors may be more beneficial for both the mentors (eg, establish a more manageable workload) and the mentee (eg, gains a broader perspective of expertise) when compared to having a single mentor.³

Peer mentorship within the group setting may also yield additional benefits. For example, having a supportive peer group may help reduce stress levels and burnout, while also improving overall well-being.^3,13 Formal mentorship programs do not frequently discuss concerns such as work-life balance, so including peers as mentors may help fill this void.⁹ Peer mentorship has also been found to be beneficial in providing mentees with pooled resources and shared learning.^12,13 This article describes the components, benefits, impacts, and challenges of a group research mentorship program for VA clinicians interested in conducting VArelevant research.

Program Description

The VA Clinical Research Mentorship Program was initiated at the VA Ann Arbor Healthcare System (VAAAHS) in October 2015 by the Chief of Medicine to assist VA clinician investigators with developing and submitting VA clinical science and health services research grant applications. The program offers group and one-on-one consultation services through the expertise of 2 experienced investigators/faculty mentors who also serve as program directors, each of whom devote about 3 to 5 hours per month to activities associated with the mentorship program (eg, attending the meeting, reviewing materials sent by mentees, and one-on-one discussions with mentees).

The program also fostered peer-led mentorship. This encourages all attendees to provide feedback during group sessions and communication by mentees outside the group sessions. An experienced project manager serves as program coordinator and contributes about 4 hours per month for activities such as attending, scheduling, and sending reminders for each meeting, distributing handouts, reviewing materials, and answering mentee’s questions via email. A statistician and additional research staff (ie, an epidemiologist and research assistant) do not attend the recurring meetings, but are available for offline consultation as needed. The program runs on a 12-month cycle with regular meetings occurring twice monthly during the 9-month academic period. Resources to support the program, primarily program director(s) and project coordinator effort, are provided by the Chief of Medicine and through the VAAAHS affiliated VA Health Systems Research (formerly Health Services Research & Development) Center of Innovation.

Invitations for new mentees are sent annually. Mentees expressing interest in the program outside of its annual recruitment period are evaluated for inclusion on a rolling basis. Recruitment begins with the program coordinator sending email notifications to all VAAAHS Medicine Service faculty, section chiefs, and division chiefs at the VAAAHS academic affiliate. Recipients are encouraged to distribute the announcement to eligible applicants and refer them to the application materials for entry consideration into the program. The application consists of the applicant’s curriculum vitae and a 1-page summary that includes a description of their research area of interest, how it is relevant to the VA, in addition to an idea for a research study, its potential significance, and proposed methodology. Applicant materials are reviewed by the program coordinator and program directors. The applicants are evaluated using a simple scoring approach that focuses on the applicant’s research area and agenda, past research training, past research productivity, potential for obtaining VA funding, and whether they have sufficient research time.

Program eligibility initially required being a physician with ≥ 1/8 VA appointment from the Medicine Service. However, clinicians with clinical appointments from other VA services are also accepted for participation as needed. Applicants must have previous research experience and have a career goal to obtain external funding for conducting and publishing original research. Those who have previously served as a principal investigator on a funded VA grant proposal are not eligible as new applicants but can remain in the program as peer mentors. The number of annual applicants varies and ranges from 1 to 11; on average, about 90% of applicants receive invitations to join the program.

Sessions

The program holds recurring meetings twice monthly for 1 hour during the 9-month academic year. However, program directors are available year-round, and mentees are encouraged to communicate questions or concerns via email during nonacademic months. Prior to the COVID-19 pandemic, all meetings were held in-person. However, the group pivoted to virtual meetings and continues to utilize this format. The dedicated program coordinator is responsible for coordinating meetings and distributing meeting materials.

Each session is informal, flexible, and supportive. Attendance is not enforced, and mentees are allowed to join meetings as their schedules permit; however, program directors and program coordinator attend each meeting. In advance of each session, the program coordinator sends out a call for agenda items to all active members invited to discuss any research related items. Each mentee presents their ideas to lead the discussion for their portion of the meeting with no defined format required.

A variety of topics are covered including, but not limited to: (1) grant-specific concerns (eg, questions related to specific aim pages, grantsmanship, postsubmission comments from reviewers, or postaward logistics); (2) research procedures (eg, questions related to methodological practices or institutional review board concerns); (3) manuscript or presentation preparation; and (4) careerrelated issues. The program coordinator distributes handouts prior to meetings and mentees may record their presentations. These handouts may include, but are not limited to, specific aims pages, analytical plans, grant solicitations, and PowerPoint presentations. If a resource that can benefit the entire group is mentioned during the meeting, the program coordinator is responsible for distribution.

The program follows a group facilitated discussion format. Program directors facilitate each meeting, but input is encouraged from all attendees. This model allows for mentees to learn from the faculty mentors as well as peer mentees in a simultaneous and efficient fashion. Group discussions foster collective problem solving, peer support, and resource sharing that would not be possible through individualized mentorship. Participants have access to varied expertise during each session which reduces the need to seek specialized help elsewhere. Participants are also encouraged to contact the program directors or research staff for consultation as needed. Some one-on-one consultations have transitioned to a more sustained and ongoing mentorship relationship between a program director and mentee, but most are often brief email exchanges or a single meeting.

Participants

Since its inception in 2015, 35 clinicians have enrolled in the program. The mentees are equally distributed by sex and practice in a variety of disciplines including gastroenterology, hematology/oncology, cardiology, and general medicine (Table 1). Mentees have submitted 33 grant proposals addressing a variety of health care issues to a diverse group of federal and nonfederal funding agencies (Table 2). As of May 15, 2024, 19 (58%) of the submitted applications have been funded.

Many factors contribute to a successfully funded grant application, and several mentees report that participating in the mentorship program was helpful. For example, a mentee became the first lead investigator for a VA Cooperative Studies Program funded at VAAAHS. The VA Cooperative Studies Program, a division of the Office of Research and Development, plans and conducts large multicenter clinical trials and epidemiological studies within the VA via a vast network of clinician investigators, statisticians, and other key research experts.¹⁴

Several program mentees have also received VA Clinical Science Research and Development Career Development Awards. The VA Career Development program supports investigators during their early research careers with a goal of retaining talented researchers committed to improving the health and care of veterans.¹⁵

Survey Responses

Mentee productivity and updates are tracked through direct mentee input, as requested by the program coordinator. Since 2022, participants could complete an end-of-year survey based on an assessment tool used in a VAAAHS nonresearch mentorship program.¹⁶ The survey, distributed to mentees and program directors, requests feedback on logistics (eg, if the meeting was a good use of time and barriers to attendance); perceptions of effectiveness (eg, ability to discuss agenda items, helpfulness with setting and reaching research goals, and quality of mentors’ feedback); and the impact of the mentoring program on work satisfaction and clinician burnout. Respondents are also encouraged to leave open-ended qualitative feedback.

To date the survey has elicited 19 responses. Seventeen (89%) indicated that they agree or strongly agree the meetings were an effective use of their time and 11 (58%) indicated that they were able to discuss all or most of the items they wanted to during the meeting. Sixteen respondents (84%) agreed the program helped them set and achieve their research goals and 14 respondents (74%) agreed the feedback they received during the meeting was specific, actionable, and focused on how to improve their research agenda. Seventeen respondents (89%) agreed the program increased their work satisfaction, while 13 respondents (68%) felt the program reduced levels of clinician burnout.

As attendance was not mandatory, the survey asked participants how often they attended meetings during the past year. Responses were mixed: 4 (21%) respondents attended regularly (12 to 16 times per year) and 8 (42%) attended most sessions (8 to 11 times per year). Noted barriers to attendance included conflicts with patient care activities and conflicts with other high priority meetings.

Mentees also provided qualitive feedback regarding the program. They highlighted the supportive environment, valuable expertise of the mentors, and usefulness of obtaining tailored feedback from the group. “This group is an amazing resource to anyone developing a research career,” a mentee noted, adding that the program directors “fostered an incredibly supportive group where research ideas and methodology can be explored in a nonthreatening and creative environment.”

Conclusions

This mentorship program aims to help aspiring VA clinician investigators develop and submit competitive research grant applications. The addition of the program to the existing robust research environments at VAAAHS and its academic affiliate appears to have contributed to this success, with 58% of applications submitted by program mentees receiving funding.

In addition to funding success, we also found that most participants have a favorable impression of the program. Of the participants who responded to the program evaluation survey, nearly all indicated the program was an effective use of their time. The program also appeared to increase work satisfaction and reduce levels of clinician burnout. Barriers to attendance were also noted, with the most frequent being scheduling conflicts.

This program’s format includes facilitated group discussion as well as peer mentorship. This collaborative structure allows for an efficient and rich learning experience. Feedback from multiple perspectives encourages natural networking and relationship building. Incorporating the collective wisdom of the faculty mentors and peer mentees is beneficial; it not only empowers the mentees but also enriches the experience for the mentors. This program can serve as a model for other VA facilities—or non-VA academic medical centers—to enhance their research programs.

Supporting meaningful research that has a positive impact on the health and quality of life of veterans is a priority of the US Department of Veterans Affairs Office of Research and Development.¹ For nearly a century, VA researchers have been conducting high quality studies. To continue this trajectory, it is imperative to attract, train, and retain exceptional investigators while nurturing their development throughout their careers.²

Mentorship is defined as guidance provided by an experienced and trusted party to another (usually junior) individual with the intent of helping the person succeed. It benefits the mentee, mentor, and their institutions.³ Mentorship is crucial for personal and professional development as well as productivity, which may help reduce clinician burnout.^4-7 Conversely, a lack of mentorship could have negative effects on work satisfaction and stagnate career progression.⁸

Mentorship is vital for developing and advancing a VA investigator’s research agenda. Funding, grant writing, and research design were among the most discussed topics in a large comprehensive mentorship program for academic faculty.⁹ However, there are several known barriers to effective research mentorship; among them include a lack of resources, time constraints, and competing clinical priorities.^10,11

Finding time for effective one-on-one research mentoring is difficult within the time constraints of clinical duties; a group mentorship model may help overcome this barrier. Group mentorship can aid in personal and professional development because no single mentor can effectively meet every mentoring need of an individual.¹² Group mentorship also allows for the exchange of ideas among individuals with different backgrounds and the ability to utilize the strengths of each member of the group. For example, a member may have methodological expertise, while another may be skilled in grantsmanship. A team of mentors may be more beneficial for both the mentors (eg, establish a more manageable workload) and the mentee (eg, gains a broader perspective of expertise) when compared to having a single mentor.³

Peer mentorship within the group setting may also yield additional benefits. For example, having a supportive peer group may help reduce stress levels and burnout, while also improving overall well-being.^3,13 Formal mentorship programs do not frequently discuss concerns such as work-life balance, so including peers as mentors may help fill this void.⁹ Peer mentorship has also been found to be beneficial in providing mentees with pooled resources and shared learning.^12,13 This article describes the components, benefits, impacts, and challenges of a group research mentorship program for VA clinicians interested in conducting VArelevant research.

Program Description

The VA Clinical Research Mentorship Program was initiated at the VA Ann Arbor Healthcare System (VAAAHS) in October 2015 by the Chief of Medicine to assist VA clinician investigators with developing and submitting VA clinical science and health services research grant applications. The program offers group and one-on-one consultation services through the expertise of 2 experienced investigators/faculty mentors who also serve as program directors, each of whom devote about 3 to 5 hours per month to activities associated with the mentorship program (eg, attending the meeting, reviewing materials sent by mentees, and one-on-one discussions with mentees).

The program also fostered peer-led mentorship. This encourages all attendees to provide feedback during group sessions and communication by mentees outside the group sessions. An experienced project manager serves as program coordinator and contributes about 4 hours per month for activities such as attending, scheduling, and sending reminders for each meeting, distributing handouts, reviewing materials, and answering mentee’s questions via email. A statistician and additional research staff (ie, an epidemiologist and research assistant) do not attend the recurring meetings, but are available for offline consultation as needed. The program runs on a 12-month cycle with regular meetings occurring twice monthly during the 9-month academic period. Resources to support the program, primarily program director(s) and project coordinator effort, are provided by the Chief of Medicine and through the VAAAHS affiliated VA Health Systems Research (formerly Health Services Research & Development) Center of Innovation.

Invitations for new mentees are sent annually. Mentees expressing interest in the program outside of its annual recruitment period are evaluated for inclusion on a rolling basis. Recruitment begins with the program coordinator sending email notifications to all VAAAHS Medicine Service faculty, section chiefs, and division chiefs at the VAAAHS academic affiliate. Recipients are encouraged to distribute the announcement to eligible applicants and refer them to the application materials for entry consideration into the program. The application consists of the applicant’s curriculum vitae and a 1-page summary that includes a description of their research area of interest, how it is relevant to the VA, in addition to an idea for a research study, its potential significance, and proposed methodology. Applicant materials are reviewed by the program coordinator and program directors. The applicants are evaluated using a simple scoring approach that focuses on the applicant’s research area and agenda, past research training, past research productivity, potential for obtaining VA funding, and whether they have sufficient research time.

Program eligibility initially required being a physician with ≥ 1/8 VA appointment from the Medicine Service. However, clinicians with clinical appointments from other VA services are also accepted for participation as needed. Applicants must have previous research experience and have a career goal to obtain external funding for conducting and publishing original research. Those who have previously served as a principal investigator on a funded VA grant proposal are not eligible as new applicants but can remain in the program as peer mentors. The number of annual applicants varies and ranges from 1 to 11; on average, about 90% of applicants receive invitations to join the program.

Sessions

The program holds recurring meetings twice monthly for 1 hour during the 9-month academic year. However, program directors are available year-round, and mentees are encouraged to communicate questions or concerns via email during nonacademic months. Prior to the COVID-19 pandemic, all meetings were held in-person. However, the group pivoted to virtual meetings and continues to utilize this format. The dedicated program coordinator is responsible for coordinating meetings and distributing meeting materials.

Each session is informal, flexible, and supportive. Attendance is not enforced, and mentees are allowed to join meetings as their schedules permit; however, program directors and program coordinator attend each meeting. In advance of each session, the program coordinator sends out a call for agenda items to all active members invited to discuss any research related items. Each mentee presents their ideas to lead the discussion for their portion of the meeting with no defined format required.

A variety of topics are covered including, but not limited to: (1) grant-specific concerns (eg, questions related to specific aim pages, grantsmanship, postsubmission comments from reviewers, or postaward logistics); (2) research procedures (eg, questions related to methodological practices or institutional review board concerns); (3) manuscript or presentation preparation; and (4) careerrelated issues. The program coordinator distributes handouts prior to meetings and mentees may record their presentations. These handouts may include, but are not limited to, specific aims pages, analytical plans, grant solicitations, and PowerPoint presentations. If a resource that can benefit the entire group is mentioned during the meeting, the program coordinator is responsible for distribution.

The program follows a group facilitated discussion format. Program directors facilitate each meeting, but input is encouraged from all attendees. This model allows for mentees to learn from the faculty mentors as well as peer mentees in a simultaneous and efficient fashion. Group discussions foster collective problem solving, peer support, and resource sharing that would not be possible through individualized mentorship. Participants have access to varied expertise during each session which reduces the need to seek specialized help elsewhere. Participants are also encouraged to contact the program directors or research staff for consultation as needed. Some one-on-one consultations have transitioned to a more sustained and ongoing mentorship relationship between a program director and mentee, but most are often brief email exchanges or a single meeting.

Participants

Since its inception in 2015, 35 clinicians have enrolled in the program. The mentees are equally distributed by sex and practice in a variety of disciplines including gastroenterology, hematology/oncology, cardiology, and general medicine (Table 1). Mentees have submitted 33 grant proposals addressing a variety of health care issues to a diverse group of federal and nonfederal funding agencies (Table 2). As of May 15, 2024, 19 (58%) of the submitted applications have been funded.

Many factors contribute to a successfully funded grant application, and several mentees report that participating in the mentorship program was helpful. For example, a mentee became the first lead investigator for a VA Cooperative Studies Program funded at VAAAHS. The VA Cooperative Studies Program, a division of the Office of Research and Development, plans and conducts large multicenter clinical trials and epidemiological studies within the VA via a vast network of clinician investigators, statisticians, and other key research experts.¹⁴

Several program mentees have also received VA Clinical Science Research and Development Career Development Awards. The VA Career Development program supports investigators during their early research careers with a goal of retaining talented researchers committed to improving the health and care of veterans.¹⁵

Survey Responses

Mentee productivity and updates are tracked through direct mentee input, as requested by the program coordinator. Since 2022, participants could complete an end-of-year survey based on an assessment tool used in a VAAAHS nonresearch mentorship program.¹⁶ The survey, distributed to mentees and program directors, requests feedback on logistics (eg, if the meeting was a good use of time and barriers to attendance); perceptions of effectiveness (eg, ability to discuss agenda items, helpfulness with setting and reaching research goals, and quality of mentors’ feedback); and the impact of the mentoring program on work satisfaction and clinician burnout. Respondents are also encouraged to leave open-ended qualitative feedback.

To date the survey has elicited 19 responses. Seventeen (89%) indicated that they agree or strongly agree the meetings were an effective use of their time and 11 (58%) indicated that they were able to discuss all or most of the items they wanted to during the meeting. Sixteen respondents (84%) agreed the program helped them set and achieve their research goals and 14 respondents (74%) agreed the feedback they received during the meeting was specific, actionable, and focused on how to improve their research agenda. Seventeen respondents (89%) agreed the program increased their work satisfaction, while 13 respondents (68%) felt the program reduced levels of clinician burnout.

As attendance was not mandatory, the survey asked participants how often they attended meetings during the past year. Responses were mixed: 4 (21%) respondents attended regularly (12 to 16 times per year) and 8 (42%) attended most sessions (8 to 11 times per year). Noted barriers to attendance included conflicts with patient care activities and conflicts with other high priority meetings.

Mentees also provided qualitive feedback regarding the program. They highlighted the supportive environment, valuable expertise of the mentors, and usefulness of obtaining tailored feedback from the group. “This group is an amazing resource to anyone developing a research career,” a mentee noted, adding that the program directors “fostered an incredibly supportive group where research ideas and methodology can be explored in a nonthreatening and creative environment.”

Conclusions

This mentorship program aims to help aspiring VA clinician investigators develop and submit competitive research grant applications. The addition of the program to the existing robust research environments at VAAAHS and its academic affiliate appears to have contributed to this success, with 58% of applications submitted by program mentees receiving funding.

In addition to funding success, we also found that most participants have a favorable impression of the program. Of the participants who responded to the program evaluation survey, nearly all indicated the program was an effective use of their time. The program also appeared to increase work satisfaction and reduce levels of clinician burnout. Barriers to attendance were also noted, with the most frequent being scheduling conflicts.

This program’s format includes facilitated group discussion as well as peer mentorship. This collaborative structure allows for an efficient and rich learning experience. Feedback from multiple perspectives encourages natural networking and relationship building. Incorporating the collective wisdom of the faculty mentors and peer mentees is beneficial; it not only empowers the mentees but also enriches the experience for the mentors. This program can serve as a model for other VA facilities—or non-VA academic medical centers—to enhance their research programs.

Supporting meaningful research that has a positive impact on the health and quality of life of veterans is a priority of the US Department of Veterans Affairs Office of Research and Development.¹ For nearly a century, VA researchers have been conducting high quality studies. To continue this trajectory, it is imperative to attract, train, and retain exceptional investigators while nurturing their development throughout their careers.²

Mentorship is defined as guidance provided by an experienced and trusted party to another (usually junior) individual with the intent of helping the person succeed. It benefits the mentee, mentor, and their institutions.³ Mentorship is crucial for personal and professional development as well as productivity, which may help reduce clinician burnout.^4-7 Conversely, a lack of mentorship could have negative effects on work satisfaction and stagnate career progression.⁸

Mentorship is vital for developing and advancing a VA investigator’s research agenda. Funding, grant writing, and research design were among the most discussed topics in a large comprehensive mentorship program for academic faculty.⁹ However, there are several known barriers to effective research mentorship; among them include a lack of resources, time constraints, and competing clinical priorities.^10,11

Finding time for effective one-on-one research mentoring is difficult within the time constraints of clinical duties; a group mentorship model may help overcome this barrier. Group mentorship can aid in personal and professional development because no single mentor can effectively meet every mentoring need of an individual.¹² Group mentorship also allows for the exchange of ideas among individuals with different backgrounds and the ability to utilize the strengths of each member of the group. For example, a member may have methodological expertise, while another may be skilled in grantsmanship. A team of mentors may be more beneficial for both the mentors (eg, establish a more manageable workload) and the mentee (eg, gains a broader perspective of expertise) when compared to having a single mentor.³

Peer mentorship within the group setting may also yield additional benefits. For example, having a supportive peer group may help reduce stress levels and burnout, while also improving overall well-being.^3,13 Formal mentorship programs do not frequently discuss concerns such as work-life balance, so including peers as mentors may help fill this void.⁹ Peer mentorship has also been found to be beneficial in providing mentees with pooled resources and shared learning.^12,13 This article describes the components, benefits, impacts, and challenges of a group research mentorship program for VA clinicians interested in conducting VArelevant research.

Program Description

The VA Clinical Research Mentorship Program was initiated at the VA Ann Arbor Healthcare System (VAAAHS) in October 2015 by the Chief of Medicine to assist VA clinician investigators with developing and submitting VA clinical science and health services research grant applications. The program offers group and one-on-one consultation services through the expertise of 2 experienced investigators/faculty mentors who also serve as program directors, each of whom devote about 3 to 5 hours per month to activities associated with the mentorship program (eg, attending the meeting, reviewing materials sent by mentees, and one-on-one discussions with mentees).

The program also fostered peer-led mentorship. This encourages all attendees to provide feedback during group sessions and communication by mentees outside the group sessions. An experienced project manager serves as program coordinator and contributes about 4 hours per month for activities such as attending, scheduling, and sending reminders for each meeting, distributing handouts, reviewing materials, and answering mentee’s questions via email. A statistician and additional research staff (ie, an epidemiologist and research assistant) do not attend the recurring meetings, but are available for offline consultation as needed. The program runs on a 12-month cycle with regular meetings occurring twice monthly during the 9-month academic period. Resources to support the program, primarily program director(s) and project coordinator effort, are provided by the Chief of Medicine and through the VAAAHS affiliated VA Health Systems Research (formerly Health Services Research & Development) Center of Innovation.

Invitations for new mentees are sent annually. Mentees expressing interest in the program outside of its annual recruitment period are evaluated for inclusion on a rolling basis. Recruitment begins with the program coordinator sending email notifications to all VAAAHS Medicine Service faculty, section chiefs, and division chiefs at the VAAAHS academic affiliate. Recipients are encouraged to distribute the announcement to eligible applicants and refer them to the application materials for entry consideration into the program. The application consists of the applicant’s curriculum vitae and a 1-page summary that includes a description of their research area of interest, how it is relevant to the VA, in addition to an idea for a research study, its potential significance, and proposed methodology. Applicant materials are reviewed by the program coordinator and program directors. The applicants are evaluated using a simple scoring approach that focuses on the applicant’s research area and agenda, past research training, past research productivity, potential for obtaining VA funding, and whether they have sufficient research time.

Program eligibility initially required being a physician with ≥ 1/8 VA appointment from the Medicine Service. However, clinicians with clinical appointments from other VA services are also accepted for participation as needed. Applicants must have previous research experience and have a career goal to obtain external funding for conducting and publishing original research. Those who have previously served as a principal investigator on a funded VA grant proposal are not eligible as new applicants but can remain in the program as peer mentors. The number of annual applicants varies and ranges from 1 to 11; on average, about 90% of applicants receive invitations to join the program.

Sessions

The program holds recurring meetings twice monthly for 1 hour during the 9-month academic year. However, program directors are available year-round, and mentees are encouraged to communicate questions or concerns via email during nonacademic months. Prior to the COVID-19 pandemic, all meetings were held in-person. However, the group pivoted to virtual meetings and continues to utilize this format. The dedicated program coordinator is responsible for coordinating meetings and distributing meeting materials.

Each session is informal, flexible, and supportive. Attendance is not enforced, and mentees are allowed to join meetings as their schedules permit; however, program directors and program coordinator attend each meeting. In advance of each session, the program coordinator sends out a call for agenda items to all active members invited to discuss any research related items. Each mentee presents their ideas to lead the discussion for their portion of the meeting with no defined format required.

A variety of topics are covered including, but not limited to: (1) grant-specific concerns (eg, questions related to specific aim pages, grantsmanship, postsubmission comments from reviewers, or postaward logistics); (2) research procedures (eg, questions related to methodological practices or institutional review board concerns); (3) manuscript or presentation preparation; and (4) careerrelated issues. The program coordinator distributes handouts prior to meetings and mentees may record their presentations. These handouts may include, but are not limited to, specific aims pages, analytical plans, grant solicitations, and PowerPoint presentations. If a resource that can benefit the entire group is mentioned during the meeting, the program coordinator is responsible for distribution.

The program follows a group facilitated discussion format. Program directors facilitate each meeting, but input is encouraged from all attendees. This model allows for mentees to learn from the faculty mentors as well as peer mentees in a simultaneous and efficient fashion. Group discussions foster collective problem solving, peer support, and resource sharing that would not be possible through individualized mentorship. Participants have access to varied expertise during each session which reduces the need to seek specialized help elsewhere. Participants are also encouraged to contact the program directors or research staff for consultation as needed. Some one-on-one consultations have transitioned to a more sustained and ongoing mentorship relationship between a program director and mentee, but most are often brief email exchanges or a single meeting.

Participants

Since its inception in 2015, 35 clinicians have enrolled in the program. The mentees are equally distributed by sex and practice in a variety of disciplines including gastroenterology, hematology/oncology, cardiology, and general medicine (Table 1). Mentees have submitted 33 grant proposals addressing a variety of health care issues to a diverse group of federal and nonfederal funding agencies (Table 2). As of May 15, 2024, 19 (58%) of the submitted applications have been funded.

Many factors contribute to a successfully funded grant application, and several mentees report that participating in the mentorship program was helpful. For example, a mentee became the first lead investigator for a VA Cooperative Studies Program funded at VAAAHS. The VA Cooperative Studies Program, a division of the Office of Research and Development, plans and conducts large multicenter clinical trials and epidemiological studies within the VA via a vast network of clinician investigators, statisticians, and other key research experts.¹⁴

Several program mentees have also received VA Clinical Science Research and Development Career Development Awards. The VA Career Development program supports investigators during their early research careers with a goal of retaining talented researchers committed to improving the health and care of veterans.¹⁵

Survey Responses

Mentee productivity and updates are tracked through direct mentee input, as requested by the program coordinator. Since 2022, participants could complete an end-of-year survey based on an assessment tool used in a VAAAHS nonresearch mentorship program.¹⁶ The survey, distributed to mentees and program directors, requests feedback on logistics (eg, if the meeting was a good use of time and barriers to attendance); perceptions of effectiveness (eg, ability to discuss agenda items, helpfulness with setting and reaching research goals, and quality of mentors’ feedback); and the impact of the mentoring program on work satisfaction and clinician burnout. Respondents are also encouraged to leave open-ended qualitative feedback.

To date the survey has elicited 19 responses. Seventeen (89%) indicated that they agree or strongly agree the meetings were an effective use of their time and 11 (58%) indicated that they were able to discuss all or most of the items they wanted to during the meeting. Sixteen respondents (84%) agreed the program helped them set and achieve their research goals and 14 respondents (74%) agreed the feedback they received during the meeting was specific, actionable, and focused on how to improve their research agenda. Seventeen respondents (89%) agreed the program increased their work satisfaction, while 13 respondents (68%) felt the program reduced levels of clinician burnout.

As attendance was not mandatory, the survey asked participants how often they attended meetings during the past year. Responses were mixed: 4 (21%) respondents attended regularly (12 to 16 times per year) and 8 (42%) attended most sessions (8 to 11 times per year). Noted barriers to attendance included conflicts with patient care activities and conflicts with other high priority meetings.

Mentees also provided qualitive feedback regarding the program. They highlighted the supportive environment, valuable expertise of the mentors, and usefulness of obtaining tailored feedback from the group. “This group is an amazing resource to anyone developing a research career,” a mentee noted, adding that the program directors “fostered an incredibly supportive group where research ideas and methodology can be explored in a nonthreatening and creative environment.”

Conclusions

This mentorship program aims to help aspiring VA clinician investigators develop and submit competitive research grant applications. The addition of the program to the existing robust research environments at VAAAHS and its academic affiliate appears to have contributed to this success, with 58% of applications submitted by program mentees receiving funding.

In addition to funding success, we also found that most participants have a favorable impression of the program. Of the participants who responded to the program evaluation survey, nearly all indicated the program was an effective use of their time. The program also appeared to increase work satisfaction and reduce levels of clinician burnout. Barriers to attendance were also noted, with the most frequent being scheduling conflicts.

This program’s format includes facilitated group discussion as well as peer mentorship. This collaborative structure allows for an efficient and rich learning experience. Feedback from multiple perspectives encourages natural networking and relationship building. Incorporating the collective wisdom of the faculty mentors and peer mentees is beneficial; it not only empowers the mentees but also enriches the experience for the mentors. This program can serve as a model for other VA facilities—or non-VA academic medical centers—to enhance their research programs.

Limb loss is a significant and growing concern in the United States. Nearly 2 million Americans are living with limb loss, and up to 185,000 people undergo amputations annually.^1-4 Of these patients, about 35% are women.⁵ The Veterans Health Administration (VHA) provides about 10% of US amputations.^6-8 Between 2015 and 2019, the number of prosthetic devices provided to female veterans increased from 3.3 million to 4.6 million.^5,9,10

Previous research identified disparities in prosthetic care between men and women, both within and outside the VHA. These disparities include slower prosthesis prescription and receipt among women, in addition to differences in self-reported mobility, satisfaction, rates of prosthesis rejection, and challenges related to prosthesis appearance and fit.^5,10,11 Recent studies suggest women tend to have worse outcomes following amputation, and are underrepresented in amputation research.^12,13 However, these disparities are poorly described in a large, national sample. Because women represent a growing portion of patients with limb loss in the VHA, understanding their needs is critical.¹⁴

The Johnny Isakson and David P. Roe, MD Veterans Health Care and Benefits Improvement Act of 2020 was enacted, in part, to improve the care provided to women veterans.¹⁵ The law required the VHA to conduct a survey of ≥ 50,000 veterans to assess the satisfaction of women veterans with prostheses provided by the VHA. To comply with this legislation and understand how women veterans rate their prostheses and related care in the VHA, the US Department of Veterans Affairs (VA) Center for Collaborative Evaluation (VACE) conducted a large national survey of veterans with limb loss that oversampled women veterans. This article describes the survey results, including characteristics of female veterans with limb loss receiving care from the VHA, assesses their satisfaction with prostheses and prosthetic care, and highlights where their responses differ from those of male veterans.

Methods

We conducted a cross-sectional, mixedmode survey of eligible amputees in the VHA Support Service Capital Assets Amputee Data Cube. We identified a cohort of veterans with any major amputation (above the ankle or wrist) or partial hand or foot amputation who received VHA care between October 1, 2019, and September 30, 2020. The final cohort yielded 46,646 potentially eligible veterans. Thirty-three had invalid contact information, leaving 46,613 veterans who were asked to participate, including 1356 women.

Survey

We created a survey instrument de novo that included questions from validated instruments, including the Trinity Amputation Prosthesis and Experience Scales to assess prosthetic device satisfaction, the Prosthesis Evaluation Questionnaire to assess quality of life (QOL) satisfaction, and the Orthotics Prosthetics Users Survey to assess prosthesis-related care satisfaction. ^16-18 Additional questions were incorporated from a survey of veterans with upper limb amputation to assess the importance of cosmetic considerations related to the prosthesis and comfort with prosthesis use in intimate relationships.¹⁹ Questions were also included to assess amputation type, year of amputation, if a prosthesis was currently used, reasons for ceasing use of a prosthesis, reasons for never using a prosthesis, the types of prostheses used, intensity of prosthesis use, satisfaction with time required to receive a prosthetic limb, and if the prosthesis reflected the veteran’s selfidentified gender. Veterans were asked to answer questions based on their most recent amputation.

We tested the survey using cognitive interviews with 6 veterans to refine the survey and better understand how veterans interpreted the questions. Pilot testers completed the survey and participated in individual interviews with experienced interviewers (CL and RRK) to describe how they selected their responses.²⁰ This feedback was used to refine the survey. The online survey was programmed using Qualtrics Software and manually translated into Spanish.

Given the multimodal design, surveys were distributed by email, text message, and US Postal Service (USPS). Surveys were emailed to all veterans for whom a valid email address was available. If emails were undeliverable, veterans were contacted via text message or the USPS. Surveys were distributed by text message to all veterans without an email address but with a cellphone number. We were unable to consistently identify invalid numbers among all text message recipients. Invitations with a survey URL and QR code were sent via USPS to veterans who had no valid email address or cellphone number. Targeted efforts were made to increase the response rate for women. A random sample of 200 women who had not completed the survey 2 weeks prior to the closing date (15% of women in sample) was selected to receive personal phone calls. Another random sample of 400 women was selected to receive personalized outreach emails. The survey data were confidential, and responses could not be traced to identifying information.

Data Analyses

We conducted a descriptive analysis, including percentages and means for responses to variables focused on describing amputation characteristics, prosthesis characteristics, and QOL. All data, including missing values, were used to document the percentage of respondents for each question. Removing missing data from the denominator when calculating percentages could introduce bias to the analysis because we cannot be certain data are missing at random. Missing variables were removed to avoid underinflation of mean scores.

We compared responses across 2 groups: individuals who self-identified as men and individuals who self-identified as women. For each question, we assessed whether each of these groups differed significantly from the remaining sample. For example, we examined whether the percentage of men who answered affirmatively to a question was significantly higher or lower than that of individuals not identifying as male, and whether the percentage of women who answered affirmatively was significantly higher or lower than that of individuals not identifying as female. We utilized x² tests to determine significant differences for percentage calculations and t tests to determine significant differences in means across gender.

Since conducting multiple comparisons within a dataset may result in inflating statistical significance (type 1 errors), we used a more conservative estimate of statistical significance (α = 0.01) and high significance (α = 0.001). This study was deemed quality improvement by the VHA Rehabilitation and Prosthetic Services (12RPS) and acknowledged by the VA Research Office at Eastern Colorado Health Care System and was not subject to institutional review board review.

Results

Surveys were distributed to 46,613 veterans and were completed by 4981 respondents for a 10.7% overall response rate. Survey respondents were generally similar to the eligible population invited to participate, but the proportion of women who completed the survey was higher than the proportion of women eligible to participate (2.0% of eligible population vs 16.7% of respondents), likely due to specific efforts to target women. Survey respondents were slightly younger than the general population (67.3 years vs 68.7 years), less likely to be male (97.1% vs 83.3%), showed similar representation of Operation Enduring Freedom/Operation Iraqi Freedom (OEF/OIF) veterans (4.4% vs 4.1%), and were less likely to have diabetes (58.0% vs 52.7% had diabetes) (Table 1).

The mean age of male respondents was 67.3 years, while the mean age of female respondents was 58.3 years. The majority of respondents were male (83.3%) and White (77.2%). Female respondents were less likely to have diabetes (35.4% of women vs 53.5% of men) and less likely to report that their most recent amputation resulted from diabetes (10.1% of women vs 22.2% of men). Women respondents were more likely to report an amputation due to other causes, such as adverse results of surgery, neurologic disease, suicide attempt, blood clots, tumors, rheumatoid arthritis, and revisions of previous amputations. Most women respondents did not serve during the OEF or OIF eras. The most common amputation site for women respondents was lower limb, either below the knee and above the ankle or above the knee.

Most participants use an everyday prosthesis, but women were more likely to report using a sports-specific prosthesis (Table 2). Overall, most respondents report using a prosthesis (87.7%); however, women were more likely to report not using a prosthesis (19.4% of women vs 11.1% of men; P ≤ .01). Additionally, a lower proportion of women report using a prosthesis for < 12 hours per day (30.6% of women vs 46.4% of men; P ≤ .01) or using a prosthesis every day (54.8% of women vs 74.6% of men; P ≤ .001).

In the overall sample, the mean satisfaction score with a prosthesis was 2.7 on a 5-point scale, and women had slightly lower overall satisfaction scores (2.6 for women vs 2.7 for men; P ≤ .001) (Table 3). Women also had lower satisfaction scores related to appearance, usefulness, reliability, and comfort. Women were more likely to indicate that it was very important to be able to wear jewelry and accessories (20.2% of women vs 11.6% of men; P ≤ .01), while men were less likely to indicate that it was somewhat or very important that the prosthesis not restrict clothing or shoes (95.2% of women vs 82.9% of men; P ≤ .001). Men were more likely than women to report being comfortable or very comfortable using their prosthesis in intimate contact: 40.5% vs 29.0%, respectively (P ≤ .001).

Overall, participants reported high satisfaction with appointment times, wait times, courteous treatment, opportunities to express concerns, and staff responsiveness. Men were slightly more likely than women to be satisfied with training (P ≤ 0.001) and problem discussion (P ≤ 0.01) (Table 4). There were no statistically significant differences in satisfaction or QOL ratings between women and men. The overall sample rated both QOL and satisfaction with QOL 6.7 on a 10-point scale.

Discussion

The goal of this study was to characterize the experience of veterans with limb loss receiving care in the VHA and assess their satisfaction with prostheses and prosthetic care. We received responses from nearly 5000 veterans, 158 of whom were women. Women veteran respondents were slightly younger and less likely to have an amputation due to diabetes. We did not observe significant differences in amputation level between men and women but women were less likely to use a prosthesis, reported lower intensity of prosthesis use, and were less satisfied with certain aspects of their prostheses. Women may also be less satisfied with prosthesis training and problem discussion. However, we found no differences in QOL ratings between men and women.

Findings indicating women were more likely to report not using a prosthesis and that a lower proportion of women report using a prosthesis for > 12 hours a day or every day are consistent with previous research. 21,22 Interestingly, women were more likely to report using a sports-specific prosthesis. This is notable because prior research suggests that individuals with amputations may avoid participating in sports and exercise, and a lack of access to sports-specific prostheses may inhibit physical activity.^23,24 Women in this sample were slightly less satisfied with their prostheses overall and reported lower satisfaction scores regarding appearance, usefulness, reliability, and comfort, consistent with previous findings.²⁵

A lower percentage of women in this sample reported being comfortable or very comfortable using their prosthesis during intimate contact. Previous research on prosthesis satisfaction suggests individuals who rate prosthesis satisfaction lower also report lower body image across genders. ²⁶ While women in this sample did not rate their prosthesis satisfaction lower than men, they did report lower intensity of prosthesis use, suggesting potential issues with their prostheses this survey did not evaluate. Women indicated the importance of prostheses not restricting jewelry, accessories, clothing, or shoes. These results have significant clinical and social implications. A recent qualitative study emphasizes that women veterans feel prostheses are primarily designed for men and may not work well with their physiological needs.⁹ Research focused on limbs better suited to women’s bodies could result in better fitting sockets, lightweight limbs, or less bulky designs. Additional research has also explored the difficulties in accommodating a range of footwear for patients with lower limb amputation. One study found that varying footwear heights affect the function of adjustable prosthetic feet in ways that may not be optimal.²⁷

Ratings of satisfaction with prosthesisrelated services between men and women in this sample are consistent with a recent study showing that women veterans do not have significant differences in satisfaction with prosthesis-related services.²⁸ However, this study focused specifically on lower limb amputations, while the respondents of this study include those with both upper and lower limb amputations. Importantly, our findings that women are less likely to be satisfied with prosthesis training and problem discussions support recent qualitative findings in which women expressed a desire to work with prosthetists who listen to them, take their concerns seriously, and seek solutions that fit their needs. We did not observe a difference in QOL ratings between men and women in the sample despite lower satisfaction among women with some elements of prosthesis-related services. Previous research suggests many factors impact QOL after amputation, most notably time since amputation.^16,29

Limitations

This survey was deployed in a short timeline that did not allow for careful sample selection or implementing strategies to increase response rate. Additionally, the study was conducted among veterans receiving care in the VHA, and findings may not be generalizable to limb loss in other settings. Finally, the discrepancy in number of respondents who identified as men vs women made it difficult to compare differences between the 2 groups.

Conclusions

This is the largest sample of survey respondents of veterans with limb loss to date. While the findings suggest veterans are generally satisfied with prosthetic-related services overall, they also highlight several areas for improvement with services or prostheses. Given that most veterans with limb loss are men, there is a significant discrepancy between the number of women and men respondents. Additional studies with more comparable numbers of men and women have found similar ratings of satisfaction with prostheses and services.²⁸ Further research specifically focused on improving the experiences of women should focus on better characterizing their experiences and identifying how they differ from those of male veterans. For example, understanding how to engage female veterans with limb loss in prosthesis training and problem discussions may improve their experience with their care teams and improve their use of prostheses. Understanding experiences and needs that are specific to women could lead to the development of processes, resources, or devices that are tailored to the unique requirements of women with limb loss.

Limb loss is a significant and growing concern in the United States. Nearly 2 million Americans are living with limb loss, and up to 185,000 people undergo amputations annually.^1-4 Of these patients, about 35% are women.⁵ The Veterans Health Administration (VHA) provides about 10% of US amputations.^6-8 Between 2015 and 2019, the number of prosthetic devices provided to female veterans increased from 3.3 million to 4.6 million.^5,9,10

Previous research identified disparities in prosthetic care between men and women, both within and outside the VHA. These disparities include slower prosthesis prescription and receipt among women, in addition to differences in self-reported mobility, satisfaction, rates of prosthesis rejection, and challenges related to prosthesis appearance and fit.^5,10,11 Recent studies suggest women tend to have worse outcomes following amputation, and are underrepresented in amputation research.^12,13 However, these disparities are poorly described in a large, national sample. Because women represent a growing portion of patients with limb loss in the VHA, understanding their needs is critical.¹⁴

The Johnny Isakson and David P. Roe, MD Veterans Health Care and Benefits Improvement Act of 2020 was enacted, in part, to improve the care provided to women veterans.¹⁵ The law required the VHA to conduct a survey of ≥ 50,000 veterans to assess the satisfaction of women veterans with prostheses provided by the VHA. To comply with this legislation and understand how women veterans rate their prostheses and related care in the VHA, the US Department of Veterans Affairs (VA) Center for Collaborative Evaluation (VACE) conducted a large national survey of veterans with limb loss that oversampled women veterans. This article describes the survey results, including characteristics of female veterans with limb loss receiving care from the VHA, assesses their satisfaction with prostheses and prosthetic care, and highlights where their responses differ from those of male veterans.

Methods

We conducted a cross-sectional, mixedmode survey of eligible amputees in the VHA Support Service Capital Assets Amputee Data Cube. We identified a cohort of veterans with any major amputation (above the ankle or wrist) or partial hand or foot amputation who received VHA care between October 1, 2019, and September 30, 2020. The final cohort yielded 46,646 potentially eligible veterans. Thirty-three had invalid contact information, leaving 46,613 veterans who were asked to participate, including 1356 women.

Survey

We created a survey instrument de novo that included questions from validated instruments, including the Trinity Amputation Prosthesis and Experience Scales to assess prosthetic device satisfaction, the Prosthesis Evaluation Questionnaire to assess quality of life (QOL) satisfaction, and the Orthotics Prosthetics Users Survey to assess prosthesis-related care satisfaction. ^16-18 Additional questions were incorporated from a survey of veterans with upper limb amputation to assess the importance of cosmetic considerations related to the prosthesis and comfort with prosthesis use in intimate relationships.¹⁹ Questions were also included to assess amputation type, year of amputation, if a prosthesis was currently used, reasons for ceasing use of a prosthesis, reasons for never using a prosthesis, the types of prostheses used, intensity of prosthesis use, satisfaction with time required to receive a prosthetic limb, and if the prosthesis reflected the veteran’s selfidentified gender. Veterans were asked to answer questions based on their most recent amputation.

We tested the survey using cognitive interviews with 6 veterans to refine the survey and better understand how veterans interpreted the questions. Pilot testers completed the survey and participated in individual interviews with experienced interviewers (CL and RRK) to describe how they selected their responses.²⁰ This feedback was used to refine the survey. The online survey was programmed using Qualtrics Software and manually translated into Spanish.

Given the multimodal design, surveys were distributed by email, text message, and US Postal Service (USPS). Surveys were emailed to all veterans for whom a valid email address was available. If emails were undeliverable, veterans were contacted via text message or the USPS. Surveys were distributed by text message to all veterans without an email address but with a cellphone number. We were unable to consistently identify invalid numbers among all text message recipients. Invitations with a survey URL and QR code were sent via USPS to veterans who had no valid email address or cellphone number. Targeted efforts were made to increase the response rate for women. A random sample of 200 women who had not completed the survey 2 weeks prior to the closing date (15% of women in sample) was selected to receive personal phone calls. Another random sample of 400 women was selected to receive personalized outreach emails. The survey data were confidential, and responses could not be traced to identifying information.

Data Analyses

We conducted a descriptive analysis, including percentages and means for responses to variables focused on describing amputation characteristics, prosthesis characteristics, and QOL. All data, including missing values, were used to document the percentage of respondents for each question. Removing missing data from the denominator when calculating percentages could introduce bias to the analysis because we cannot be certain data are missing at random. Missing variables were removed to avoid underinflation of mean scores.

We compared responses across 2 groups: individuals who self-identified as men and individuals who self-identified as women. For each question, we assessed whether each of these groups differed significantly from the remaining sample. For example, we examined whether the percentage of men who answered affirmatively to a question was significantly higher or lower than that of individuals not identifying as male, and whether the percentage of women who answered affirmatively was significantly higher or lower than that of individuals not identifying as female. We utilized x² tests to determine significant differences for percentage calculations and t tests to determine significant differences in means across gender.

Since conducting multiple comparisons within a dataset may result in inflating statistical significance (type 1 errors), we used a more conservative estimate of statistical significance (α = 0.01) and high significance (α = 0.001). This study was deemed quality improvement by the VHA Rehabilitation and Prosthetic Services (12RPS) and acknowledged by the VA Research Office at Eastern Colorado Health Care System and was not subject to institutional review board review.

Results

Surveys were distributed to 46,613 veterans and were completed by 4981 respondents for a 10.7% overall response rate. Survey respondents were generally similar to the eligible population invited to participate, but the proportion of women who completed the survey was higher than the proportion of women eligible to participate (2.0% of eligible population vs 16.7% of respondents), likely due to specific efforts to target women. Survey respondents were slightly younger than the general population (67.3 years vs 68.7 years), less likely to be male (97.1% vs 83.3%), showed similar representation of Operation Enduring Freedom/Operation Iraqi Freedom (OEF/OIF) veterans (4.4% vs 4.1%), and were less likely to have diabetes (58.0% vs 52.7% had diabetes) (Table 1).

The mean age of male respondents was 67.3 years, while the mean age of female respondents was 58.3 years. The majority of respondents were male (83.3%) and White (77.2%). Female respondents were less likely to have diabetes (35.4% of women vs 53.5% of men) and less likely to report that their most recent amputation resulted from diabetes (10.1% of women vs 22.2% of men). Women respondents were more likely to report an amputation due to other causes, such as adverse results of surgery, neurologic disease, suicide attempt, blood clots, tumors, rheumatoid arthritis, and revisions of previous amputations. Most women respondents did not serve during the OEF or OIF eras. The most common amputation site for women respondents was lower limb, either below the knee and above the ankle or above the knee.

Most participants use an everyday prosthesis, but women were more likely to report using a sports-specific prosthesis (Table 2). Overall, most respondents report using a prosthesis (87.7%); however, women were more likely to report not using a prosthesis (19.4% of women vs 11.1% of men; P ≤ .01). Additionally, a lower proportion of women report using a prosthesis for < 12 hours per day (30.6% of women vs 46.4% of men; P ≤ .01) or using a prosthesis every day (54.8% of women vs 74.6% of men; P ≤ .001).

In the overall sample, the mean satisfaction score with a prosthesis was 2.7 on a 5-point scale, and women had slightly lower overall satisfaction scores (2.6 for women vs 2.7 for men; P ≤ .001) (Table 3). Women also had lower satisfaction scores related to appearance, usefulness, reliability, and comfort. Women were more likely to indicate that it was very important to be able to wear jewelry and accessories (20.2% of women vs 11.6% of men; P ≤ .01), while men were less likely to indicate that it was somewhat or very important that the prosthesis not restrict clothing or shoes (95.2% of women vs 82.9% of men; P ≤ .001). Men were more likely than women to report being comfortable or very comfortable using their prosthesis in intimate contact: 40.5% vs 29.0%, respectively (P ≤ .001).

Overall, participants reported high satisfaction with appointment times, wait times, courteous treatment, opportunities to express concerns, and staff responsiveness. Men were slightly more likely than women to be satisfied with training (P ≤ 0.001) and problem discussion (P ≤ 0.01) (Table 4). There were no statistically significant differences in satisfaction or QOL ratings between women and men. The overall sample rated both QOL and satisfaction with QOL 6.7 on a 10-point scale.

Discussion

The goal of this study was to characterize the experience of veterans with limb loss receiving care in the VHA and assess their satisfaction with prostheses and prosthetic care. We received responses from nearly 5000 veterans, 158 of whom were women. Women veteran respondents were slightly younger and less likely to have an amputation due to diabetes. We did not observe significant differences in amputation level between men and women but women were less likely to use a prosthesis, reported lower intensity of prosthesis use, and were less satisfied with certain aspects of their prostheses. Women may also be less satisfied with prosthesis training and problem discussion. However, we found no differences in QOL ratings between men and women.

Findings indicating women were more likely to report not using a prosthesis and that a lower proportion of women report using a prosthesis for > 12 hours a day or every day are consistent with previous research. 21,22 Interestingly, women were more likely to report using a sports-specific prosthesis. This is notable because prior research suggests that individuals with amputations may avoid participating in sports and exercise, and a lack of access to sports-specific prostheses may inhibit physical activity.^23,24 Women in this sample were slightly less satisfied with their prostheses overall and reported lower satisfaction scores regarding appearance, usefulness, reliability, and comfort, consistent with previous findings.²⁵

A lower percentage of women in this sample reported being comfortable or very comfortable using their prosthesis during intimate contact. Previous research on prosthesis satisfaction suggests individuals who rate prosthesis satisfaction lower also report lower body image across genders. ²⁶ While women in this sample did not rate their prosthesis satisfaction lower than men, they did report lower intensity of prosthesis use, suggesting potential issues with their prostheses this survey did not evaluate. Women indicated the importance of prostheses not restricting jewelry, accessories, clothing, or shoes. These results have significant clinical and social implications. A recent qualitative study emphasizes that women veterans feel prostheses are primarily designed for men and may not work well with their physiological needs.⁹ Research focused on limbs better suited to women’s bodies could result in better fitting sockets, lightweight limbs, or less bulky designs. Additional research has also explored the difficulties in accommodating a range of footwear for patients with lower limb amputation. One study found that varying footwear heights affect the function of adjustable prosthetic feet in ways that may not be optimal.²⁷

Ratings of satisfaction with prosthesisrelated services between men and women in this sample are consistent with a recent study showing that women veterans do not have significant differences in satisfaction with prosthesis-related services.²⁸ However, this study focused specifically on lower limb amputations, while the respondents of this study include those with both upper and lower limb amputations. Importantly, our findings that women are less likely to be satisfied with prosthesis training and problem discussions support recent qualitative findings in which women expressed a desire to work with prosthetists who listen to them, take their concerns seriously, and seek solutions that fit their needs. We did not observe a difference in QOL ratings between men and women in the sample despite lower satisfaction among women with some elements of prosthesis-related services. Previous research suggests many factors impact QOL after amputation, most notably time since amputation.^16,29

Limitations

This survey was deployed in a short timeline that did not allow for careful sample selection or implementing strategies to increase response rate. Additionally, the study was conducted among veterans receiving care in the VHA, and findings may not be generalizable to limb loss in other settings. Finally, the discrepancy in number of respondents who identified as men vs women made it difficult to compare differences between the 2 groups.

Conclusions

This is the largest sample of survey respondents of veterans with limb loss to date. While the findings suggest veterans are generally satisfied with prosthetic-related services overall, they also highlight several areas for improvement with services or prostheses. Given that most veterans with limb loss are men, there is a significant discrepancy between the number of women and men respondents. Additional studies with more comparable numbers of men and women have found similar ratings of satisfaction with prostheses and services.²⁸ Further research specifically focused on improving the experiences of women should focus on better characterizing their experiences and identifying how they differ from those of male veterans. For example, understanding how to engage female veterans with limb loss in prosthesis training and problem discussions may improve their experience with their care teams and improve their use of prostheses. Understanding experiences and needs that are specific to women could lead to the development of processes, resources, or devices that are tailored to the unique requirements of women with limb loss.

Limb loss is a significant and growing concern in the United States. Nearly 2 million Americans are living with limb loss, and up to 185,000 people undergo amputations annually.^1-4 Of these patients, about 35% are women.⁵ The Veterans Health Administration (VHA) provides about 10% of US amputations.^6-8 Between 2015 and 2019, the number of prosthetic devices provided to female veterans increased from 3.3 million to 4.6 million.^5,9,10

Previous research identified disparities in prosthetic care between men and women, both within and outside the VHA. These disparities include slower prosthesis prescription and receipt among women, in addition to differences in self-reported mobility, satisfaction, rates of prosthesis rejection, and challenges related to prosthesis appearance and fit.^5,10,11 Recent studies suggest women tend to have worse outcomes following amputation, and are underrepresented in amputation research.^12,13 However, these disparities are poorly described in a large, national sample. Because women represent a growing portion of patients with limb loss in the VHA, understanding their needs is critical.¹⁴

The Johnny Isakson and David P. Roe, MD Veterans Health Care and Benefits Improvement Act of 2020 was enacted, in part, to improve the care provided to women veterans.¹⁵ The law required the VHA to conduct a survey of ≥ 50,000 veterans to assess the satisfaction of women veterans with prostheses provided by the VHA. To comply with this legislation and understand how women veterans rate their prostheses and related care in the VHA, the US Department of Veterans Affairs (VA) Center for Collaborative Evaluation (VACE) conducted a large national survey of veterans with limb loss that oversampled women veterans. This article describes the survey results, including characteristics of female veterans with limb loss receiving care from the VHA, assesses their satisfaction with prostheses and prosthetic care, and highlights where their responses differ from those of male veterans.

Methods

We conducted a cross-sectional, mixedmode survey of eligible amputees in the VHA Support Service Capital Assets Amputee Data Cube. We identified a cohort of veterans with any major amputation (above the ankle or wrist) or partial hand or foot amputation who received VHA care between October 1, 2019, and September 30, 2020. The final cohort yielded 46,646 potentially eligible veterans. Thirty-three had invalid contact information, leaving 46,613 veterans who were asked to participate, including 1356 women.

Survey

We created a survey instrument de novo that included questions from validated instruments, including the Trinity Amputation Prosthesis and Experience Scales to assess prosthetic device satisfaction, the Prosthesis Evaluation Questionnaire to assess quality of life (QOL) satisfaction, and the Orthotics Prosthetics Users Survey to assess prosthesis-related care satisfaction. ^16-18 Additional questions were incorporated from a survey of veterans with upper limb amputation to assess the importance of cosmetic considerations related to the prosthesis and comfort with prosthesis use in intimate relationships.¹⁹ Questions were also included to assess amputation type, year of amputation, if a prosthesis was currently used, reasons for ceasing use of a prosthesis, reasons for never using a prosthesis, the types of prostheses used, intensity of prosthesis use, satisfaction with time required to receive a prosthetic limb, and if the prosthesis reflected the veteran’s selfidentified gender. Veterans were asked to answer questions based on their most recent amputation.

We tested the survey using cognitive interviews with 6 veterans to refine the survey and better understand how veterans interpreted the questions. Pilot testers completed the survey and participated in individual interviews with experienced interviewers (CL and RRK) to describe how they selected their responses.²⁰ This feedback was used to refine the survey. The online survey was programmed using Qualtrics Software and manually translated into Spanish.

Given the multimodal design, surveys were distributed by email, text message, and US Postal Service (USPS). Surveys were emailed to all veterans for whom a valid email address was available. If emails were undeliverable, veterans were contacted via text message or the USPS. Surveys were distributed by text message to all veterans without an email address but with a cellphone number. We were unable to consistently identify invalid numbers among all text message recipients. Invitations with a survey URL and QR code were sent via USPS to veterans who had no valid email address or cellphone number. Targeted efforts were made to increase the response rate for women. A random sample of 200 women who had not completed the survey 2 weeks prior to the closing date (15% of women in sample) was selected to receive personal phone calls. Another random sample of 400 women was selected to receive personalized outreach emails. The survey data were confidential, and responses could not be traced to identifying information.

Data Analyses

We conducted a descriptive analysis, including percentages and means for responses to variables focused on describing amputation characteristics, prosthesis characteristics, and QOL. All data, including missing values, were used to document the percentage of respondents for each question. Removing missing data from the denominator when calculating percentages could introduce bias to the analysis because we cannot be certain data are missing at random. Missing variables were removed to avoid underinflation of mean scores.

We compared responses across 2 groups: individuals who self-identified as men and individuals who self-identified as women. For each question, we assessed whether each of these groups differed significantly from the remaining sample. For example, we examined whether the percentage of men who answered affirmatively to a question was significantly higher or lower than that of individuals not identifying as male, and whether the percentage of women who answered affirmatively was significantly higher or lower than that of individuals not identifying as female. We utilized x² tests to determine significant differences for percentage calculations and t tests to determine significant differences in means across gender.

Since conducting multiple comparisons within a dataset may result in inflating statistical significance (type 1 errors), we used a more conservative estimate of statistical significance (α = 0.01) and high significance (α = 0.001). This study was deemed quality improvement by the VHA Rehabilitation and Prosthetic Services (12RPS) and acknowledged by the VA Research Office at Eastern Colorado Health Care System and was not subject to institutional review board review.

Results

Surveys were distributed to 46,613 veterans and were completed by 4981 respondents for a 10.7% overall response rate. Survey respondents were generally similar to the eligible population invited to participate, but the proportion of women who completed the survey was higher than the proportion of women eligible to participate (2.0% of eligible population vs 16.7% of respondents), likely due to specific efforts to target women. Survey respondents were slightly younger than the general population (67.3 years vs 68.7 years), less likely to be male (97.1% vs 83.3%), showed similar representation of Operation Enduring Freedom/Operation Iraqi Freedom (OEF/OIF) veterans (4.4% vs 4.1%), and were less likely to have diabetes (58.0% vs 52.7% had diabetes) (Table 1).

The mean age of male respondents was 67.3 years, while the mean age of female respondents was 58.3 years. The majority of respondents were male (83.3%) and White (77.2%). Female respondents were less likely to have diabetes (35.4% of women vs 53.5% of men) and less likely to report that their most recent amputation resulted from diabetes (10.1% of women vs 22.2% of men). Women respondents were more likely to report an amputation due to other causes, such as adverse results of surgery, neurologic disease, suicide attempt, blood clots, tumors, rheumatoid arthritis, and revisions of previous amputations. Most women respondents did not serve during the OEF or OIF eras. The most common amputation site for women respondents was lower limb, either below the knee and above the ankle or above the knee.

Most participants use an everyday prosthesis, but women were more likely to report using a sports-specific prosthesis (Table 2). Overall, most respondents report using a prosthesis (87.7%); however, women were more likely to report not using a prosthesis (19.4% of women vs 11.1% of men; P ≤ .01). Additionally, a lower proportion of women report using a prosthesis for < 12 hours per day (30.6% of women vs 46.4% of men; P ≤ .01) or using a prosthesis every day (54.8% of women vs 74.6% of men; P ≤ .001).

In the overall sample, the mean satisfaction score with a prosthesis was 2.7 on a 5-point scale, and women had slightly lower overall satisfaction scores (2.6 for women vs 2.7 for men; P ≤ .001) (Table 3). Women also had lower satisfaction scores related to appearance, usefulness, reliability, and comfort. Women were more likely to indicate that it was very important to be able to wear jewelry and accessories (20.2% of women vs 11.6% of men; P ≤ .01), while men were less likely to indicate that it was somewhat or very important that the prosthesis not restrict clothing or shoes (95.2% of women vs 82.9% of men; P ≤ .001). Men were more likely than women to report being comfortable or very comfortable using their prosthesis in intimate contact: 40.5% vs 29.0%, respectively (P ≤ .001).

Overall, participants reported high satisfaction with appointment times, wait times, courteous treatment, opportunities to express concerns, and staff responsiveness. Men were slightly more likely than women to be satisfied with training (P ≤ 0.001) and problem discussion (P ≤ 0.01) (Table 4). There were no statistically significant differences in satisfaction or QOL ratings between women and men. The overall sample rated both QOL and satisfaction with QOL 6.7 on a 10-point scale.

Discussion

The goal of this study was to characterize the experience of veterans with limb loss receiving care in the VHA and assess their satisfaction with prostheses and prosthetic care. We received responses from nearly 5000 veterans, 158 of whom were women. Women veteran respondents were slightly younger and less likely to have an amputation due to diabetes. We did not observe significant differences in amputation level between men and women but women were less likely to use a prosthesis, reported lower intensity of prosthesis use, and were less satisfied with certain aspects of their prostheses. Women may also be less satisfied with prosthesis training and problem discussion. However, we found no differences in QOL ratings between men and women.

Findings indicating women were more likely to report not using a prosthesis and that a lower proportion of women report using a prosthesis for > 12 hours a day or every day are consistent with previous research. 21,22 Interestingly, women were more likely to report using a sports-specific prosthesis. This is notable because prior research suggests that individuals with amputations may avoid participating in sports and exercise, and a lack of access to sports-specific prostheses may inhibit physical activity.^23,24 Women in this sample were slightly less satisfied with their prostheses overall and reported lower satisfaction scores regarding appearance, usefulness, reliability, and comfort, consistent with previous findings.²⁵

A lower percentage of women in this sample reported being comfortable or very comfortable using their prosthesis during intimate contact. Previous research on prosthesis satisfaction suggests individuals who rate prosthesis satisfaction lower also report lower body image across genders. ²⁶ While women in this sample did not rate their prosthesis satisfaction lower than men, they did report lower intensity of prosthesis use, suggesting potential issues with their prostheses this survey did not evaluate. Women indicated the importance of prostheses not restricting jewelry, accessories, clothing, or shoes. These results have significant clinical and social implications. A recent qualitative study emphasizes that women veterans feel prostheses are primarily designed for men and may not work well with their physiological needs.⁹ Research focused on limbs better suited to women’s bodies could result in better fitting sockets, lightweight limbs, or less bulky designs. Additional research has also explored the difficulties in accommodating a range of footwear for patients with lower limb amputation. One study found that varying footwear heights affect the function of adjustable prosthetic feet in ways that may not be optimal.²⁷

Ratings of satisfaction with prosthesisrelated services between men and women in this sample are consistent with a recent study showing that women veterans do not have significant differences in satisfaction with prosthesis-related services.²⁸ However, this study focused specifically on lower limb amputations, while the respondents of this study include those with both upper and lower limb amputations. Importantly, our findings that women are less likely to be satisfied with prosthesis training and problem discussions support recent qualitative findings in which women expressed a desire to work with prosthetists who listen to them, take their concerns seriously, and seek solutions that fit their needs. We did not observe a difference in QOL ratings between men and women in the sample despite lower satisfaction among women with some elements of prosthesis-related services. Previous research suggests many factors impact QOL after amputation, most notably time since amputation.^16,29

Limitations

This survey was deployed in a short timeline that did not allow for careful sample selection or implementing strategies to increase response rate. Additionally, the study was conducted among veterans receiving care in the VHA, and findings may not be generalizable to limb loss in other settings. Finally, the discrepancy in number of respondents who identified as men vs women made it difficult to compare differences between the 2 groups.

Conclusions

This is the largest sample of survey respondents of veterans with limb loss to date. While the findings suggest veterans are generally satisfied with prosthetic-related services overall, they also highlight several areas for improvement with services or prostheses. Given that most veterans with limb loss are men, there is a significant discrepancy between the number of women and men respondents. Additional studies with more comparable numbers of men and women have found similar ratings of satisfaction with prostheses and services.²⁸ Further research specifically focused on improving the experiences of women should focus on better characterizing their experiences and identifying how they differ from those of male veterans. For example, understanding how to engage female veterans with limb loss in prosthesis training and problem discussions may improve their experience with their care teams and improve their use of prostheses. Understanding experiences and needs that are specific to women could lead to the development of processes, resources, or devices that are tailored to the unique requirements of women with limb loss.

As Veterans Day approaches, stores and restaurants will offer discounts and free meals to veterans. Children will write thank you letters, and citizens nationwide will raise flags to honor and thank veterans. We can never repay those who lost their life, health, or livelihood in defense of the nation. Since the American Revolution, and in gratitude for that incalculable debt, the US government, on behalf of the American public, has seen fit to grant a host of benefits and services to those who wore the uniform.^2,3 Among the best known are health care, burial services, compensation and pensions, home loans, and the GI Bill.

Less recognized yet arguably essential for the fair and consistent provision of these entitlements is a legal principle: the veteran’s canon. A canon is a system of rules or maxims used to interpret legal instruments, such as statutes. They are not rules but serve as a “principle that guides the interpretation of the text.”⁴ Since I am not a lawyer, I will undoubtedly oversimplify this legal principle, but I hope to get enough right to explain why the veteran’s canon should matter to federal health care professionals.

At its core, the veteran’s canon means that when the US Department of Veterans Affairs (VA) and a veteran have a legal dispute about VA benefits, the courts will give deference to the veteran. Underscoring that any ambiguity in the statute is resolved in the veteran’s favor, the canon is known in legal circles as the Gardner deference. This is a reference to a 1994 case in which a Korean War veteran underwent surgery in a VA facility for a herniated disc he alleged caused pain and weakness in his left lower extremity.⁵ Gardner argued that federal statutes 38 USC § 1151 underlying corresponding VA regulation 38 CFR § 3.358(c)(3) granted disability benefits to veterans injured during VA treatment. The VA denied the disability claim, contending the regulation restricted compensation to veterans whose injury was the fault of the VA; thus, the disability had to have been the result of negligent treatment or an unforeseen therapeutic accident.⁵

The case wound its way through various appeals boards and courts until the Supreme Court of the United States (SCOTUS) ruled that the statute’s context left no ambiguity, and that any care provided under VA auspices was covered under the statute. What is important for this column is that the justices opined that had ambiguity been present, it would have legally necessitated, “applying the rule that interpretive doubt is to be resolved in the veteran’s favor.”⁵ In Gardner’s case, the courts reaffirmed nearly 80 years of judicial precedent upholding the veteran’s canon.

Thirty years later, Rudisill v McDonough again questioned the veteran’s canon.⁶ Educational benefits, namely the GI Bill, were the issue in this case. Rudisill served during 3 different periods in the US Army, totaling 8 years. Two educational programs overlapped during Rudisill’s tenure in the military: the Montgomery GI Bill and Post-9/11 Veterans Educational Assistance Act. Rudisill had used a portion of his Montgomery benefits to fund his undergraduate education and now wished to use the more extensive Post-9/11 assistance to finance his graduate degree. Rudisill and the VA disagreed about when his combined benefits would be capped, either at 36 or 48 months. After working its way through appeals courts, SCOTUS was again called upon for judgment.

The justices found that Rudisill qualified under both programs and could use them in any order he wished up to the cap. The majority found no ambiguity in the statute; however, if interpretation was required, the majority of justices indicated that the veteran’s canon would have supported Rudisill. While this sounds like good news for veterans, 2 justices authored a dissenting opinion that questioned the constitutional grounding of the veteran’s canon, noting that the “canon appears to have developed almost by accident.”⁶ The minority opinion suggested that when the veteran’s canon allocates resources to pay for specific veteran benefits, other interests and groups are deprived of those same resources, resulting in potential inequity.⁷

The potential ethical import and clinical impact of striking down the veteran’s canon is serious. It is especially concerning given that in a recent case, the SCOTUS ruling struck down another legal interpretation that also benefited the VA and ultimately veterans: the Chevron deference.⁸ This precedent held that when a legal dispute arises about the meaning of a specific federal agency regulation or policy, the courts should defer to the federal agency’s presumably superior understanding of the matter. The principle places the locus of decision-making with the subject-matter experts of the respective agency rather than the courts.

Ironically, given the legislative purposes of both interpretive principles, their overturning would likely introduce much more uncertainty, variation, and unpredictability in cases involving veteran benefits. This is bad news for both veterans and the VA. Veterans might not prevail as often in court when they have a reasonable claim, leading to more aggressive challenges. In response, the VA would have a heavier and more costly burden of administrative proof to defend sound decisions.⁹ Recently, the VA has tried to reduce the backlog of claims. The inability to have legal recourse to Chevron or Gardener could result in even more delay in adjudicating veterans’ claims that enable them to access benefits and services, already an object of congressional pressure.¹⁰

Courts will continue to debate the issue with another judicial test of the canon on the current SCOTUS docket (Bufkin v McDonough).¹¹ The veteran’s canon was put in place to equalize the power differential between the VA and the veteran: in administrative language, to make it more likely than not that the veteran would prevail when regulations were ambiguous. There are many legal and political rationales for veteran’s canon, including enabling veterans to file claims for service-connected illnesses. The veteran’s cannon helped Vietnam War-era veterans receive VA care while researchers were still studying the sequela of Agent Orange exposure. ¹² The legislative purpose of the veteran’s canon is the same as that of all VA benefits and services commemorated on Veterans Day. As expressed by SCOTUS justices in the wake of World War II, the benefit statutes should be “liberally construed for the benefit of those who left private life to serve their country in its hour of greatest need.”¹³

As Veterans Day approaches, stores and restaurants will offer discounts and free meals to veterans. Children will write thank you letters, and citizens nationwide will raise flags to honor and thank veterans. We can never repay those who lost their life, health, or livelihood in defense of the nation. Since the American Revolution, and in gratitude for that incalculable debt, the US government, on behalf of the American public, has seen fit to grant a host of benefits and services to those who wore the uniform.^2,3 Among the best known are health care, burial services, compensation and pensions, home loans, and the GI Bill.

Less recognized yet arguably essential for the fair and consistent provision of these entitlements is a legal principle: the veteran’s canon. A canon is a system of rules or maxims used to interpret legal instruments, such as statutes. They are not rules but serve as a “principle that guides the interpretation of the text.”⁴ Since I am not a lawyer, I will undoubtedly oversimplify this legal principle, but I hope to get enough right to explain why the veteran’s canon should matter to federal health care professionals.

At its core, the veteran’s canon means that when the US Department of Veterans Affairs (VA) and a veteran have a legal dispute about VA benefits, the courts will give deference to the veteran. Underscoring that any ambiguity in the statute is resolved in the veteran’s favor, the canon is known in legal circles as the Gardner deference. This is a reference to a 1994 case in which a Korean War veteran underwent surgery in a VA facility for a herniated disc he alleged caused pain and weakness in his left lower extremity.⁵ Gardner argued that federal statutes 38 USC § 1151 underlying corresponding VA regulation 38 CFR § 3.358(c)(3) granted disability benefits to veterans injured during VA treatment. The VA denied the disability claim, contending the regulation restricted compensation to veterans whose injury was the fault of the VA; thus, the disability had to have been the result of negligent treatment or an unforeseen therapeutic accident.⁵

The case wound its way through various appeals boards and courts until the Supreme Court of the United States (SCOTUS) ruled that the statute’s context left no ambiguity, and that any care provided under VA auspices was covered under the statute. What is important for this column is that the justices opined that had ambiguity been present, it would have legally necessitated, “applying the rule that interpretive doubt is to be resolved in the veteran’s favor.”⁵ In Gardner’s case, the courts reaffirmed nearly 80 years of judicial precedent upholding the veteran’s canon.

Thirty years later, Rudisill v McDonough again questioned the veteran’s canon.⁶ Educational benefits, namely the GI Bill, were the issue in this case. Rudisill served during 3 different periods in the US Army, totaling 8 years. Two educational programs overlapped during Rudisill’s tenure in the military: the Montgomery GI Bill and Post-9/11 Veterans Educational Assistance Act. Rudisill had used a portion of his Montgomery benefits to fund his undergraduate education and now wished to use the more extensive Post-9/11 assistance to finance his graduate degree. Rudisill and the VA disagreed about when his combined benefits would be capped, either at 36 or 48 months. After working its way through appeals courts, SCOTUS was again called upon for judgment.

The justices found that Rudisill qualified under both programs and could use them in any order he wished up to the cap. The majority found no ambiguity in the statute; however, if interpretation was required, the majority of justices indicated that the veteran’s canon would have supported Rudisill. While this sounds like good news for veterans, 2 justices authored a dissenting opinion that questioned the constitutional grounding of the veteran’s canon, noting that the “canon appears to have developed almost by accident.”⁶ The minority opinion suggested that when the veteran’s canon allocates resources to pay for specific veteran benefits, other interests and groups are deprived of those same resources, resulting in potential inequity.⁷

The potential ethical import and clinical impact of striking down the veteran’s canon is serious. It is especially concerning given that in a recent case, the SCOTUS ruling struck down another legal interpretation that also benefited the VA and ultimately veterans: the Chevron deference.⁸ This precedent held that when a legal dispute arises about the meaning of a specific federal agency regulation or policy, the courts should defer to the federal agency’s presumably superior understanding of the matter. The principle places the locus of decision-making with the subject-matter experts of the respective agency rather than the courts.

Ironically, given the legislative purposes of both interpretive principles, their overturning would likely introduce much more uncertainty, variation, and unpredictability in cases involving veteran benefits. This is bad news for both veterans and the VA. Veterans might not prevail as often in court when they have a reasonable claim, leading to more aggressive challenges. In response, the VA would have a heavier and more costly burden of administrative proof to defend sound decisions.⁹ Recently, the VA has tried to reduce the backlog of claims. The inability to have legal recourse to Chevron or Gardener could result in even more delay in adjudicating veterans’ claims that enable them to access benefits and services, already an object of congressional pressure.¹⁰

Courts will continue to debate the issue with another judicial test of the canon on the current SCOTUS docket (Bufkin v McDonough).¹¹ The veteran’s canon was put in place to equalize the power differential between the VA and the veteran: in administrative language, to make it more likely than not that the veteran would prevail when regulations were ambiguous. There are many legal and political rationales for veteran’s canon, including enabling veterans to file claims for service-connected illnesses. The veteran’s cannon helped Vietnam War-era veterans receive VA care while researchers were still studying the sequela of Agent Orange exposure. ¹² The legislative purpose of the veteran’s canon is the same as that of all VA benefits and services commemorated on Veterans Day. As expressed by SCOTUS justices in the wake of World War II, the benefit statutes should be “liberally construed for the benefit of those who left private life to serve their country in its hour of greatest need.”¹³

As Veterans Day approaches, stores and restaurants will offer discounts and free meals to veterans. Children will write thank you letters, and citizens nationwide will raise flags to honor and thank veterans. We can never repay those who lost their life, health, or livelihood in defense of the nation. Since the American Revolution, and in gratitude for that incalculable debt, the US government, on behalf of the American public, has seen fit to grant a host of benefits and services to those who wore the uniform.^2,3 Among the best known are health care, burial services, compensation and pensions, home loans, and the GI Bill.

Less recognized yet arguably essential for the fair and consistent provision of these entitlements is a legal principle: the veteran’s canon. A canon is a system of rules or maxims used to interpret legal instruments, such as statutes. They are not rules but serve as a “principle that guides the interpretation of the text.”⁴ Since I am not a lawyer, I will undoubtedly oversimplify this legal principle, but I hope to get enough right to explain why the veteran’s canon should matter to federal health care professionals.

At its core, the veteran’s canon means that when the US Department of Veterans Affairs (VA) and a veteran have a legal dispute about VA benefits, the courts will give deference to the veteran. Underscoring that any ambiguity in the statute is resolved in the veteran’s favor, the canon is known in legal circles as the Gardner deference. This is a reference to a 1994 case in which a Korean War veteran underwent surgery in a VA facility for a herniated disc he alleged caused pain and weakness in his left lower extremity.⁵ Gardner argued that federal statutes 38 USC § 1151 underlying corresponding VA regulation 38 CFR § 3.358(c)(3) granted disability benefits to veterans injured during VA treatment. The VA denied the disability claim, contending the regulation restricted compensation to veterans whose injury was the fault of the VA; thus, the disability had to have been the result of negligent treatment or an unforeseen therapeutic accident.⁵

The case wound its way through various appeals boards and courts until the Supreme Court of the United States (SCOTUS) ruled that the statute’s context left no ambiguity, and that any care provided under VA auspices was covered under the statute. What is important for this column is that the justices opined that had ambiguity been present, it would have legally necessitated, “applying the rule that interpretive doubt is to be resolved in the veteran’s favor.”⁵ In Gardner’s case, the courts reaffirmed nearly 80 years of judicial precedent upholding the veteran’s canon.

Thirty years later, Rudisill v McDonough again questioned the veteran’s canon.⁶ Educational benefits, namely the GI Bill, were the issue in this case. Rudisill served during 3 different periods in the US Army, totaling 8 years. Two educational programs overlapped during Rudisill’s tenure in the military: the Montgomery GI Bill and Post-9/11 Veterans Educational Assistance Act. Rudisill had used a portion of his Montgomery benefits to fund his undergraduate education and now wished to use the more extensive Post-9/11 assistance to finance his graduate degree. Rudisill and the VA disagreed about when his combined benefits would be capped, either at 36 or 48 months. After working its way through appeals courts, SCOTUS was again called upon for judgment.

The justices found that Rudisill qualified under both programs and could use them in any order he wished up to the cap. The majority found no ambiguity in the statute; however, if interpretation was required, the majority of justices indicated that the veteran’s canon would have supported Rudisill. While this sounds like good news for veterans, 2 justices authored a dissenting opinion that questioned the constitutional grounding of the veteran’s canon, noting that the “canon appears to have developed almost by accident.”⁶ The minority opinion suggested that when the veteran’s canon allocates resources to pay for specific veteran benefits, other interests and groups are deprived of those same resources, resulting in potential inequity.⁷

The potential ethical import and clinical impact of striking down the veteran’s canon is serious. It is especially concerning given that in a recent case, the SCOTUS ruling struck down another legal interpretation that also benefited the VA and ultimately veterans: the Chevron deference.⁸ This precedent held that when a legal dispute arises about the meaning of a specific federal agency regulation or policy, the courts should defer to the federal agency’s presumably superior understanding of the matter. The principle places the locus of decision-making with the subject-matter experts of the respective agency rather than the courts.

Ironically, given the legislative purposes of both interpretive principles, their overturning would likely introduce much more uncertainty, variation, and unpredictability in cases involving veteran benefits. This is bad news for both veterans and the VA. Veterans might not prevail as often in court when they have a reasonable claim, leading to more aggressive challenges. In response, the VA would have a heavier and more costly burden of administrative proof to defend sound decisions.⁹ Recently, the VA has tried to reduce the backlog of claims. The inability to have legal recourse to Chevron or Gardener could result in even more delay in adjudicating veterans’ claims that enable them to access benefits and services, already an object of congressional pressure.¹⁰

Courts will continue to debate the issue with another judicial test of the canon on the current SCOTUS docket (Bufkin v McDonough).¹¹ The veteran’s canon was put in place to equalize the power differential between the VA and the veteran: in administrative language, to make it more likely than not that the veteran would prevail when regulations were ambiguous. There are many legal and political rationales for veteran’s canon, including enabling veterans to file claims for service-connected illnesses. The veteran’s cannon helped Vietnam War-era veterans receive VA care while researchers were still studying the sequela of Agent Orange exposure. ¹² The legislative purpose of the veteran’s canon is the same as that of all VA benefits and services commemorated on Veterans Day. As expressed by SCOTUS justices in the wake of World War II, the benefit statutes should be “liberally construed for the benefit of those who left private life to serve their country in its hour of greatest need.”¹³

Atherosclerotic cardiovascular disease (ASCVD) is a significant cause of morbidity and mortality in the United States. ASCVD involves the buildup of cholesterol plaque in arteries and includes acute coronary syndrome, peripheral arterial disease, and events such as myocardial infarction and stroke.¹ Cardiovascular disease (CVD) risk factors include high cholesterol levels, elevated blood pressure, insulin resistance, elevated blood glucose levels, smoking, poor dietary habits, and a sedentary lifestyle.²

According to the Centers for Disease Control and Prevention, about 86 million adults aged ≥ 20 years have total cholesterol levels > 200 mg/dL. More than half (54.5%) who could benefit are currently taking cholesterol-lowering medications.³ Controlling high cholesterol in American adults, especially veterans, is essential for reducing CVD morbidity and mortality.

The 2018 American College of Cardiology/American Heart Association (ACC/AHA) guideline recommends a low-density lipoprotein cholesterol (LDL-C) target goal of < 70 mg/dL for patients at high risk for ASCVD. Very high-risk ASCVD includes a history of multiple major ASCVD events or 1 major ASCVD event and multiple high-risk conditions (eg, age ≥ 65 years, smoking, or diabetes).⁴ Major ASCVD events include recent acute coronary syndrome (within the past 12 months), a history of myocardial infarction or ischemic stroke, and symptomatic peripheral artery disease. 

The ACC/AHA guideline suggests that if the LDL-C level remains ≥ 70 mg/dL, adding ezetimibe (a dietary cholesterol absorption inhibitor) to maximally tolerated statin therapy is reasonable. If LDL-C levels remain ≥ 70 mg/dL, adding a proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitor, such as alirocumab, is reasonable.⁴ The US Departments of Veterans Affairs/US Department of Defense guidelines recommend using maximally tolerated statins and ezetimibe before PCSK9 inhibitors due to established long-term safety and reduction in CVD events. 

Generic statins and ezetimibe are administered orally and widely available. In contrast, PCSK9 inhibitors have unknown long-term safety profiles, require subcutaneous injection once or twice monthly, and are significantly more expensive. They also require patient education on proper use while providing comparable or lesser relative risk reductions.²

These 3 classes of medication vary in their mechanisms of action to reduce LDL.^5,6 Ezetimibe and several statin medications are included on the Veterans Affairs Sioux Falls Health Care System (VASFHCS) formulary and do not require review prior to prescribing. Alirocumab is available at VASFHCS but is restricted to patients with a history of ASCVD or a diagnosis of familial hypercholesterolemia, and who are receiving maximally tolerated statin and ezetimibe therapy but require further LDL-C lowering to reduce their ASCVD risk. 

Studies have found ezetimibe monotherapy reduces LDL-C in patients with dyslipidemia by 18% after 12 weeks.⁷ One found that the percentage reduction in LDL-C was significantly greater (P < .001) with all doses of ezetimibe plus simvastatin (46% to 59%) compared with either atorvastatin 10 mg (37%) or simvastatin 20 mg (38%) monotherapy after 6 weeks.⁸

Although alirocumab can be added to other lipid therapies, most VASFHCS patients are prescribed alirocumab monotherapy. In the ODYSSEY CHOICE II study, patients were randomly assigned to receive either a placebo or alirocumab 150 mg every 4 weeks or alirocumab 75 mg every 2 weeks. The primary efficacy endpoint was LDL-C percentage change from baseline to week 24. In the alirocumab 150 mg every 4 weeks and 75 mg every 2 weeks groups, the least-squares mean LDL-C changes from baseline to week 24 were 51.7% and 53.5%, respectively, compared to a 4.7% increase in the placebo group (both groups P < .001 vs placebo). The authors also reported that alirocumab 150 mg every 4 weeks as monotherapy demonstrated a 47.4% reduction in LDL-C levels from baseline in a phase 1 study.⁹Although alirocumab monotherapy and ezetimibe plus statin therapy have been shown to effectively decrease LDL-C independently, a direct comparison of alirocumab monotherapy vs ezetimibe plus statin therapy has not been assessed, to our knowledge. Understanding the differences in effectiveness and safety between these 2 regimens will be valuable for clinicians when selecting a medication regimen for veterans with a history of ASCVD.

METHODS

This retrospective, single-center chart review used VASFHCS Computerized Patient Record System (CPRS) and Joint Longitudinal Viewer (JLV) records to compare patients with a history of ASCVD events who were treated with alirocumab monotherapy or ezetimibe plus statin. The 2 groups were randomized in a 1:3 ratio. The primary endpoint was achieving LDL-C < 70 mg/dL after 4 to 12 weeks, 13 to 24 weeks, and 25 to 52 weeks. Secondary endpoints included the mean percentage change from baseline in total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), LDL-C, and triglycerides (TG) over 52 weeks. The incidence of ASCVD events during this period was also assessed. If LDL-C < 70 mg/dL was achieved > 1 time during each time frame, only 1 incident was counted for analysis. Safety was assessed based on the incidence of any adverse event (AE) that led to treatment discontinuation.

Patients were identified by screening the prescription fill history between October 1, 2019, and December 31, 2022. The 52-week data collection period was counted from the first available fill date. Additionally, the prior authorization drug request file from January 1, 2017, to December 31, 2022, was used to obtain a list of patients prescribed alirocumab. Patients were included if they were veterans aged ≥ 18 years and had a history of an ASCVD event, had a alirocumab monotherapy or ezetimibe plus statin prescription between October 1, 2019, and December 31, 2022, or had an approved prior authorization drug request for alirocumab between January 1, 2017, and December 31, 2022. Patients missing a baseline or follow-up lipid panel and those with concurrent use of alirocumab and ezetimibe and/or statin were excluded. 

Baseline characteristics collected for patients included age, sex, race, weight, body mass index, lipid parameters (LDL-C, TC, HDL-C, and TG), dosing of each type of statin before adding ezetimibe, and use of any other antihyperlipidemic medication. We also collected histories of hypertension, hyperlipidemia, diabetes, chronic kidney disease, congestive heart failure, and smoking or tobacco use status. The baseline lipid panel was the most recent lipid panel documented before starting alirocumab or ezetimibe plus statin therapy. Follow-up lipid panel values were gathered at 4 to 12 weeks, 13 to 24 weeks, and 25 to 52 weeks following initiation of either therapy.

High-, moderate-, and low-intensity dosing of statin therapy and alirocumab dosing (75 mg every 2 weeks, 150 mg every 2 weeks, or 300 mg every 4 weeks) were recorded at the specified intervals. However, no patients in this study received the latter dosing regimen. ASCVD events and safety endpoints were recorded based on a review of clinical notes over the 52 weeks following the first available start date.

Statistical Analysis

The primary endpoint of achieving the LDL-C < 70 mg/dL goal from baseline to 4 to 12 weeks, 13 to 24 weeks, and 25 to 52 weeks after initiation was compared between alirocumab monotherapy and ezetimibe plus statin therapy using the χ² test. Mean percentage change from baseline in TC, HDL-C, LDL-C, and TG were compared using the independent t test. P < .05 was considered statistically significant. Incidence of ASCVD events and the safety endpoint (incidence of AEs leading to treatment discontinuation) were also compared using the χ² test. Continuous baseline characteristics were reported mean (SD) and nominal baseline characteristics were reported as a percentage.

RESULTS

There were 80 participants in this study: 20 in the alirocumab monotherapy group and 60 in the ezetimibe plus statin therapy group. More than 100 patients did not meet the prespecified inclusion criteria and were excluded. Mean (SD) age was 75 (8) years in the alirocumab group and 74 (8) years in the ezetimibe plus statin group. There was no significant differences in mean (SD) weight or mean (SD) body mass index. All study participants identified as White and male except for 2 patients in the ezetimibe plus statin therapy group whose race was not documented. Differences in lipid parameters were observed between groups, with mean baseline LDL-C, HDL-C, and TC higher in the alirocumab monotherapy group than in the ezetimibe plus statin therapy group, with significant differences in LDL-C and TC (Table 1).

Fourteen patients (70%) in the alirocumab monotherapy group had hypertension, compared with 31 (52%) in the ezetimibe plus statin therapy group. In both groups, most patients had previously been diagnosed with hyperlipidemia. More patients (60%) in the alirocumab group had diabetes than in the ezetimibe plus statin therapy group (37%). The alirocumab monotherapy group also had a higher percentage of patients with diagnoses of congestive heart failure and used other antihyperlipidemic medications than in the ezetimibe plus statin therapy group. Five patients (25%) in the alirocumab monotherapy group and 12 patients (20%) in the ezetimibe plus statin therapy group took fish oil. In the ezetimibe plus statin therapy group, 2 patients (3%) took gemfibrozil, and 2 patients (3%) took fenofibrate. Six (30%) patients in the alirocumab monotherapy group and 12 (20%) patients in the ezetimibe plus statin therapy group had chronic kidney disease. Although the majority of patients in each group did not use tobacco products, there were more tobacco users in the ezetimibe plus statin therapy group.

In the alirocumab monotherapy group, 15 patients (75%) were prescribed 75 mg every 2 weeks and 5 patients (25%) were prescribed 150 mg every 2 weeks. In the ezetimibe plus statin therapy group, 59 patients (98%) were prescribed ezetimibe 10 mg/d (Table 2). Forty-three patients (72%) were prescribed a high-intensity statin 10 received moderate-intensity (17%) and 7 received low-intensity statin (12%). Most patients were prescribed rosuvastatin (45%), followed by atorvastatin (42%), pravastatin (10%), and simvastatin (3%).

Primary Endpoint

During the 52-week study, more patients met the LDL-C goal of < 70 mg/dL in the alirocumab monotherapy group (70%) than in the ezetimibe plus statin therapy group (57%); however, the difference was not significant (P = .29). Of the patients prescribed alirocumab monotherapy who achieved LDL-C < 70 mg/dL, 15% achieved this goal in 4 to 12 weeks, 40% in 13 to 24 weeks, and 45% in 25 to 52 weeks. In the ezetimibe plus statin therapy group, 28% of patients achieved LDL-C < 70 mg/dL in 4 to 12 weeks, 31% in 13 to 24 weeks, and 41% in 25 to 52 weeks (Table 3).

Secondary Endpoints

During weeks 4 to 52 of treatment, the mean percentage change decreased in LDL-C (37.7% vs 21.4%; P = .01), TC (24.7% vs 12.5%; P = .01), and TG (0.9% vs 7.0%; P = .28) in the alirocumab monotherapy group and the ezetimibe plus statin therapy group, respectively (Table 4). The mean percentage change increased in HDL-C by 3.6% in the alirocumab monotherapy group and 1.8% in the ezetimibe plus statin therapy group (P = .36). During the study, ASCVD events occurred in 1 patient (5%) in the alirocumab monotherapy group and 3 patients (5%) in the ezetimibe plus statin therapy group (P = .99). The patient in the alirocumab monotherapy group had unstable angina 1 month after taking alirocumab. One patient in the ezetimibe plus statin therapy group had coronary artery disease and 2 patients had coronary heart disease that required stents during the 52-week period. There was 1 patient in each group who reported an AE that led to treatment discontinuation (P = .41). One patient stopped alirocumab after a trial of 2 months due to intolerance, but no specific AE was reported in the CPRS. In the ezetimibe plus statin therapy group, 1 patient requested to discontinue ezetimibe after a trial of 3 months without a specific reason noted in the medical record.

DISCUSSION

This study found no statistically significant difference in the incidence of reaching an LDL-C goal of < 70 mg/dL after alirocumab monotherapy initiation compared with ezetimibe plus statin therapy. This occurred despite baseline LDL-C being lower in the ezetimibe plus statin therapy group, which required a smaller reduction in LDL-C to reach the primary goal. Most patients on alirocumab monotherapy were prescribed a lower initial dose of 75 mg every 2 weeks. Of those patients, 30% did not achieve the LDL-C goal < 70 mg/dL. Thus, a higher dose may have led to more patients achieving the LDL-C goal.

Secondary endpoints, including mean percentage change in HDL-C and TG and incidence of ASCVD events during 52 weeks of treatment, were not statistically significant. The mean percentage increase in HDL-C was negligible in both groups, while the mean percentage reduction in TG favored the ezetimibe plus statin therapy group. In the ezetimibe plus statin therapy group, patients who also took fenofibrate experienced a significant reduction in TG while none of the patients in the alirocumab group were prescribed fenofibrate. Although the alirocumab monotherapy group had a statistically significant greater reduction in LDL-C and TC compared with those prescribed ezetimibe plus statin, the mean baseline LDL-C and TC were significantly greater in the alirocumab monotherapy group, which could contribute to higher reductions in LDL-C and TC after alirocumab monotherapy.Based on the available literature, we expected greater reductions in LDL-C in both study groups compared with statin therapy alone.^8,9 However, it was unclear whether the LDL-C and TC reductions were clinically significant.

Limitations

The study design did not permit randomization prior to the treatments, restricting our ability to account for some confounding factors, such as diet, exercise, other antihyperlipidemic medication, and medication adherence, which may have affected LDL-C, HDL-C, TG, and TC levels. Differences in baseline characteristics—particularly major risk factors, such as hypertension, diabetes, and tobacco use—also could have confounding affect on lipid levels and ASCVD events. Additionally, patients prescribed alirocumab monotherapy may have switched from statin or ezetimibe therapy, and the washout period was not reviewed or recorded, which could have affected the lipid panel results.

The small sample size of this study also may have limited the ability to detect significant differences between groups. A direct comparison of alirocumab monotherapy vs ezetimibe plus statin therapy has not been performed, making it difficult to prospectively evaluate what sample size would be needed to power this study. A posthoc analysis was used to calculate power, which was found to be only 17%. Many patients were excluded due to a lack of laboratory results within the study period, contributing to the small sample size. 

Another limitation was the reliance on documentation in CPRS and JLV. For example, having documentation of the specific AEs for the 2 patients who discontinued alirocumab or ezetimibe could have helped determine the severity of the AEs. Several patients were followed by non-VA clinicians, which could have contributed to limited documentation in the CPRS and JLV. It is difficult to draw any conclusions regarding ASCVD events and AEs that led to treatment discontinuation between alirocumab monotherapy and ezetimibe plus statin therapy based on the results of this retrospective study due to the limited number of events within the 52-week period.

CONCLUSIONS

This study found that there was no statistically significant difference in LDL-C reduction to < 70 mg/dL between alirocumab monotherapy and ezetimibe plus statin therapy in a small population of veterans with ASCVD, with a higher percentage of participants in both groups achieving that goal in 25 to 52 weeks. There also was no significant difference in percentage change in HDL-C or TG or in incidence of ASCVD events and AEs leading to treatment discontinuation. However, there was a statistically significant difference in percentage reduction for LDL-C and TC during 52 weeks of alirocumab monotherapy vs ezetimibe plus statin therapy.

Although there was no significant difference in LDL-C reduction to < 70 mg/dL, targeting this goal in patients with ASCVD is still clinically warranted. This study does not support a change in current VA criteria for use of alirocumab or a change in current guidelines for secondary prevention of ASCVD. Still, this study does indicate that the efficacy of alirocumab monotherapy is similar to that of ezetimibe plus statin therapy in patients with a history of ASCVD and may be useful in clinical settings when an alternative to ezetimibe plus statin therapy is needed. Alirocumab also may be more effective in lowering LDL-C and TC than ezetimibe plus statin therapy in veterans with ASCVD and could be added to statin therapy or ezetimibe when additional LDL-C or TC reduction is needed.

Atherosclerotic cardiovascular disease (ASCVD) is a significant cause of morbidity and mortality in the United States. ASCVD involves the buildup of cholesterol plaque in arteries and includes acute coronary syndrome, peripheral arterial disease, and events such as myocardial infarction and stroke.¹ Cardiovascular disease (CVD) risk factors include high cholesterol levels, elevated blood pressure, insulin resistance, elevated blood glucose levels, smoking, poor dietary habits, and a sedentary lifestyle.²

According to the Centers for Disease Control and Prevention, about 86 million adults aged ≥ 20 years have total cholesterol levels > 200 mg/dL. More than half (54.5%) who could benefit are currently taking cholesterol-lowering medications.³ Controlling high cholesterol in American adults, especially veterans, is essential for reducing CVD morbidity and mortality.

The 2018 American College of Cardiology/American Heart Association (ACC/AHA) guideline recommends a low-density lipoprotein cholesterol (LDL-C) target goal of < 70 mg/dL for patients at high risk for ASCVD. Very high-risk ASCVD includes a history of multiple major ASCVD events or 1 major ASCVD event and multiple high-risk conditions (eg, age ≥ 65 years, smoking, or diabetes).⁴ Major ASCVD events include recent acute coronary syndrome (within the past 12 months), a history of myocardial infarction or ischemic stroke, and symptomatic peripheral artery disease. 

The ACC/AHA guideline suggests that if the LDL-C level remains ≥ 70 mg/dL, adding ezetimibe (a dietary cholesterol absorption inhibitor) to maximally tolerated statin therapy is reasonable. If LDL-C levels remain ≥ 70 mg/dL, adding a proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitor, such as alirocumab, is reasonable.⁴ The US Departments of Veterans Affairs/US Department of Defense guidelines recommend using maximally tolerated statins and ezetimibe before PCSK9 inhibitors due to established long-term safety and reduction in CVD events. 

Generic statins and ezetimibe are administered orally and widely available. In contrast, PCSK9 inhibitors have unknown long-term safety profiles, require subcutaneous injection once or twice monthly, and are significantly more expensive. They also require patient education on proper use while providing comparable or lesser relative risk reductions.²

These 3 classes of medication vary in their mechanisms of action to reduce LDL.^5,6 Ezetimibe and several statin medications are included on the Veterans Affairs Sioux Falls Health Care System (VASFHCS) formulary and do not require review prior to prescribing. Alirocumab is available at VASFHCS but is restricted to patients with a history of ASCVD or a diagnosis of familial hypercholesterolemia, and who are receiving maximally tolerated statin and ezetimibe therapy but require further LDL-C lowering to reduce their ASCVD risk. 

Studies have found ezetimibe monotherapy reduces LDL-C in patients with dyslipidemia by 18% after 12 weeks.⁷ One found that the percentage reduction in LDL-C was significantly greater (P < .001) with all doses of ezetimibe plus simvastatin (46% to 59%) compared with either atorvastatin 10 mg (37%) or simvastatin 20 mg (38%) monotherapy after 6 weeks.⁸

Although alirocumab can be added to other lipid therapies, most VASFHCS patients are prescribed alirocumab monotherapy. In the ODYSSEY CHOICE II study, patients were randomly assigned to receive either a placebo or alirocumab 150 mg every 4 weeks or alirocumab 75 mg every 2 weeks. The primary efficacy endpoint was LDL-C percentage change from baseline to week 24. In the alirocumab 150 mg every 4 weeks and 75 mg every 2 weeks groups, the least-squares mean LDL-C changes from baseline to week 24 were 51.7% and 53.5%, respectively, compared to a 4.7% increase in the placebo group (both groups P < .001 vs placebo). The authors also reported that alirocumab 150 mg every 4 weeks as monotherapy demonstrated a 47.4% reduction in LDL-C levels from baseline in a phase 1 study.⁹Although alirocumab monotherapy and ezetimibe plus statin therapy have been shown to effectively decrease LDL-C independently, a direct comparison of alirocumab monotherapy vs ezetimibe plus statin therapy has not been assessed, to our knowledge. Understanding the differences in effectiveness and safety between these 2 regimens will be valuable for clinicians when selecting a medication regimen for veterans with a history of ASCVD.

METHODS

This retrospective, single-center chart review used VASFHCS Computerized Patient Record System (CPRS) and Joint Longitudinal Viewer (JLV) records to compare patients with a history of ASCVD events who were treated with alirocumab monotherapy or ezetimibe plus statin. The 2 groups were randomized in a 1:3 ratio. The primary endpoint was achieving LDL-C < 70 mg/dL after 4 to 12 weeks, 13 to 24 weeks, and 25 to 52 weeks. Secondary endpoints included the mean percentage change from baseline in total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), LDL-C, and triglycerides (TG) over 52 weeks. The incidence of ASCVD events during this period was also assessed. If LDL-C < 70 mg/dL was achieved > 1 time during each time frame, only 1 incident was counted for analysis. Safety was assessed based on the incidence of any adverse event (AE) that led to treatment discontinuation.

Patients were identified by screening the prescription fill history between October 1, 2019, and December 31, 2022. The 52-week data collection period was counted from the first available fill date. Additionally, the prior authorization drug request file from January 1, 2017, to December 31, 2022, was used to obtain a list of patients prescribed alirocumab. Patients were included if they were veterans aged ≥ 18 years and had a history of an ASCVD event, had a alirocumab monotherapy or ezetimibe plus statin prescription between October 1, 2019, and December 31, 2022, or had an approved prior authorization drug request for alirocumab between January 1, 2017, and December 31, 2022. Patients missing a baseline or follow-up lipid panel and those with concurrent use of alirocumab and ezetimibe and/or statin were excluded. 

Baseline characteristics collected for patients included age, sex, race, weight, body mass index, lipid parameters (LDL-C, TC, HDL-C, and TG), dosing of each type of statin before adding ezetimibe, and use of any other antihyperlipidemic medication. We also collected histories of hypertension, hyperlipidemia, diabetes, chronic kidney disease, congestive heart failure, and smoking or tobacco use status. The baseline lipid panel was the most recent lipid panel documented before starting alirocumab or ezetimibe plus statin therapy. Follow-up lipid panel values were gathered at 4 to 12 weeks, 13 to 24 weeks, and 25 to 52 weeks following initiation of either therapy.

High-, moderate-, and low-intensity dosing of statin therapy and alirocumab dosing (75 mg every 2 weeks, 150 mg every 2 weeks, or 300 mg every 4 weeks) were recorded at the specified intervals. However, no patients in this study received the latter dosing regimen. ASCVD events and safety endpoints were recorded based on a review of clinical notes over the 52 weeks following the first available start date.

Statistical Analysis

The primary endpoint of achieving the LDL-C < 70 mg/dL goal from baseline to 4 to 12 weeks, 13 to 24 weeks, and 25 to 52 weeks after initiation was compared between alirocumab monotherapy and ezetimibe plus statin therapy using the χ² test. Mean percentage change from baseline in TC, HDL-C, LDL-C, and TG were compared using the independent t test. P < .05 was considered statistically significant. Incidence of ASCVD events and the safety endpoint (incidence of AEs leading to treatment discontinuation) were also compared using the χ² test. Continuous baseline characteristics were reported mean (SD) and nominal baseline characteristics were reported as a percentage.

RESULTS

There were 80 participants in this study: 20 in the alirocumab monotherapy group and 60 in the ezetimibe plus statin therapy group. More than 100 patients did not meet the prespecified inclusion criteria and were excluded. Mean (SD) age was 75 (8) years in the alirocumab group and 74 (8) years in the ezetimibe plus statin group. There was no significant differences in mean (SD) weight or mean (SD) body mass index. All study participants identified as White and male except for 2 patients in the ezetimibe plus statin therapy group whose race was not documented. Differences in lipid parameters were observed between groups, with mean baseline LDL-C, HDL-C, and TC higher in the alirocumab monotherapy group than in the ezetimibe plus statin therapy group, with significant differences in LDL-C and TC (Table 1).

Fourteen patients (70%) in the alirocumab monotherapy group had hypertension, compared with 31 (52%) in the ezetimibe plus statin therapy group. In both groups, most patients had previously been diagnosed with hyperlipidemia. More patients (60%) in the alirocumab group had diabetes than in the ezetimibe plus statin therapy group (37%). The alirocumab monotherapy group also had a higher percentage of patients with diagnoses of congestive heart failure and used other antihyperlipidemic medications than in the ezetimibe plus statin therapy group. Five patients (25%) in the alirocumab monotherapy group and 12 patients (20%) in the ezetimibe plus statin therapy group took fish oil. In the ezetimibe plus statin therapy group, 2 patients (3%) took gemfibrozil, and 2 patients (3%) took fenofibrate. Six (30%) patients in the alirocumab monotherapy group and 12 (20%) patients in the ezetimibe plus statin therapy group had chronic kidney disease. Although the majority of patients in each group did not use tobacco products, there were more tobacco users in the ezetimibe plus statin therapy group.

In the alirocumab monotherapy group, 15 patients (75%) were prescribed 75 mg every 2 weeks and 5 patients (25%) were prescribed 150 mg every 2 weeks. In the ezetimibe plus statin therapy group, 59 patients (98%) were prescribed ezetimibe 10 mg/d (Table 2). Forty-three patients (72%) were prescribed a high-intensity statin 10 received moderate-intensity (17%) and 7 received low-intensity statin (12%). Most patients were prescribed rosuvastatin (45%), followed by atorvastatin (42%), pravastatin (10%), and simvastatin (3%).

Primary Endpoint

During the 52-week study, more patients met the LDL-C goal of < 70 mg/dL in the alirocumab monotherapy group (70%) than in the ezetimibe plus statin therapy group (57%); however, the difference was not significant (P = .29). Of the patients prescribed alirocumab monotherapy who achieved LDL-C < 70 mg/dL, 15% achieved this goal in 4 to 12 weeks, 40% in 13 to 24 weeks, and 45% in 25 to 52 weeks. In the ezetimibe plus statin therapy group, 28% of patients achieved LDL-C < 70 mg/dL in 4 to 12 weeks, 31% in 13 to 24 weeks, and 41% in 25 to 52 weeks (Table 3).

Secondary Endpoints

During weeks 4 to 52 of treatment, the mean percentage change decreased in LDL-C (37.7% vs 21.4%; P = .01), TC (24.7% vs 12.5%; P = .01), and TG (0.9% vs 7.0%; P = .28) in the alirocumab monotherapy group and the ezetimibe plus statin therapy group, respectively (Table 4). The mean percentage change increased in HDL-C by 3.6% in the alirocumab monotherapy group and 1.8% in the ezetimibe plus statin therapy group (P = .36). During the study, ASCVD events occurred in 1 patient (5%) in the alirocumab monotherapy group and 3 patients (5%) in the ezetimibe plus statin therapy group (P = .99). The patient in the alirocumab monotherapy group had unstable angina 1 month after taking alirocumab. One patient in the ezetimibe plus statin therapy group had coronary artery disease and 2 patients had coronary heart disease that required stents during the 52-week period. There was 1 patient in each group who reported an AE that led to treatment discontinuation (P = .41). One patient stopped alirocumab after a trial of 2 months due to intolerance, but no specific AE was reported in the CPRS. In the ezetimibe plus statin therapy group, 1 patient requested to discontinue ezetimibe after a trial of 3 months without a specific reason noted in the medical record.

DISCUSSION

This study found no statistically significant difference in the incidence of reaching an LDL-C goal of < 70 mg/dL after alirocumab monotherapy initiation compared with ezetimibe plus statin therapy. This occurred despite baseline LDL-C being lower in the ezetimibe plus statin therapy group, which required a smaller reduction in LDL-C to reach the primary goal. Most patients on alirocumab monotherapy were prescribed a lower initial dose of 75 mg every 2 weeks. Of those patients, 30% did not achieve the LDL-C goal < 70 mg/dL. Thus, a higher dose may have led to more patients achieving the LDL-C goal.

Secondary endpoints, including mean percentage change in HDL-C and TG and incidence of ASCVD events during 52 weeks of treatment, were not statistically significant. The mean percentage increase in HDL-C was negligible in both groups, while the mean percentage reduction in TG favored the ezetimibe plus statin therapy group. In the ezetimibe plus statin therapy group, patients who also took fenofibrate experienced a significant reduction in TG while none of the patients in the alirocumab group were prescribed fenofibrate. Although the alirocumab monotherapy group had a statistically significant greater reduction in LDL-C and TC compared with those prescribed ezetimibe plus statin, the mean baseline LDL-C and TC were significantly greater in the alirocumab monotherapy group, which could contribute to higher reductions in LDL-C and TC after alirocumab monotherapy.Based on the available literature, we expected greater reductions in LDL-C in both study groups compared with statin therapy alone.^8,9 However, it was unclear whether the LDL-C and TC reductions were clinically significant.

Limitations

The study design did not permit randomization prior to the treatments, restricting our ability to account for some confounding factors, such as diet, exercise, other antihyperlipidemic medication, and medication adherence, which may have affected LDL-C, HDL-C, TG, and TC levels. Differences in baseline characteristics—particularly major risk factors, such as hypertension, diabetes, and tobacco use—also could have confounding affect on lipid levels and ASCVD events. Additionally, patients prescribed alirocumab monotherapy may have switched from statin or ezetimibe therapy, and the washout period was not reviewed or recorded, which could have affected the lipid panel results.

The small sample size of this study also may have limited the ability to detect significant differences between groups. A direct comparison of alirocumab monotherapy vs ezetimibe plus statin therapy has not been performed, making it difficult to prospectively evaluate what sample size would be needed to power this study. A posthoc analysis was used to calculate power, which was found to be only 17%. Many patients were excluded due to a lack of laboratory results within the study period, contributing to the small sample size. 

Another limitation was the reliance on documentation in CPRS and JLV. For example, having documentation of the specific AEs for the 2 patients who discontinued alirocumab or ezetimibe could have helped determine the severity of the AEs. Several patients were followed by non-VA clinicians, which could have contributed to limited documentation in the CPRS and JLV. It is difficult to draw any conclusions regarding ASCVD events and AEs that led to treatment discontinuation between alirocumab monotherapy and ezetimibe plus statin therapy based on the results of this retrospective study due to the limited number of events within the 52-week period.

CONCLUSIONS

This study found that there was no statistically significant difference in LDL-C reduction to < 70 mg/dL between alirocumab monotherapy and ezetimibe plus statin therapy in a small population of veterans with ASCVD, with a higher percentage of participants in both groups achieving that goal in 25 to 52 weeks. There also was no significant difference in percentage change in HDL-C or TG or in incidence of ASCVD events and AEs leading to treatment discontinuation. However, there was a statistically significant difference in percentage reduction for LDL-C and TC during 52 weeks of alirocumab monotherapy vs ezetimibe plus statin therapy.

Although there was no significant difference in LDL-C reduction to < 70 mg/dL, targeting this goal in patients with ASCVD is still clinically warranted. This study does not support a change in current VA criteria for use of alirocumab or a change in current guidelines for secondary prevention of ASCVD. Still, this study does indicate that the efficacy of alirocumab monotherapy is similar to that of ezetimibe plus statin therapy in patients with a history of ASCVD and may be useful in clinical settings when an alternative to ezetimibe plus statin therapy is needed. Alirocumab also may be more effective in lowering LDL-C and TC than ezetimibe plus statin therapy in veterans with ASCVD and could be added to statin therapy or ezetimibe when additional LDL-C or TC reduction is needed.

Atherosclerotic cardiovascular disease (ASCVD) is a significant cause of morbidity and mortality in the United States. ASCVD involves the buildup of cholesterol plaque in arteries and includes acute coronary syndrome, peripheral arterial disease, and events such as myocardial infarction and stroke.¹ Cardiovascular disease (CVD) risk factors include high cholesterol levels, elevated blood pressure, insulin resistance, elevated blood glucose levels, smoking, poor dietary habits, and a sedentary lifestyle.²

According to the Centers for Disease Control and Prevention, about 86 million adults aged ≥ 20 years have total cholesterol levels > 200 mg/dL. More than half (54.5%) who could benefit are currently taking cholesterol-lowering medications.³ Controlling high cholesterol in American adults, especially veterans, is essential for reducing CVD morbidity and mortality.

The 2018 American College of Cardiology/American Heart Association (ACC/AHA) guideline recommends a low-density lipoprotein cholesterol (LDL-C) target goal of < 70 mg/dL for patients at high risk for ASCVD. Very high-risk ASCVD includes a history of multiple major ASCVD events or 1 major ASCVD event and multiple high-risk conditions (eg, age ≥ 65 years, smoking, or diabetes).⁴ Major ASCVD events include recent acute coronary syndrome (within the past 12 months), a history of myocardial infarction or ischemic stroke, and symptomatic peripheral artery disease. 

The ACC/AHA guideline suggests that if the LDL-C level remains ≥ 70 mg/dL, adding ezetimibe (a dietary cholesterol absorption inhibitor) to maximally tolerated statin therapy is reasonable. If LDL-C levels remain ≥ 70 mg/dL, adding a proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitor, such as alirocumab, is reasonable.⁴ The US Departments of Veterans Affairs/US Department of Defense guidelines recommend using maximally tolerated statins and ezetimibe before PCSK9 inhibitors due to established long-term safety and reduction in CVD events. 

Generic statins and ezetimibe are administered orally and widely available. In contrast, PCSK9 inhibitors have unknown long-term safety profiles, require subcutaneous injection once or twice monthly, and are significantly more expensive. They also require patient education on proper use while providing comparable or lesser relative risk reductions.²

These 3 classes of medication vary in their mechanisms of action to reduce LDL.^5,6 Ezetimibe and several statin medications are included on the Veterans Affairs Sioux Falls Health Care System (VASFHCS) formulary and do not require review prior to prescribing. Alirocumab is available at VASFHCS but is restricted to patients with a history of ASCVD or a diagnosis of familial hypercholesterolemia, and who are receiving maximally tolerated statin and ezetimibe therapy but require further LDL-C lowering to reduce their ASCVD risk. 

Studies have found ezetimibe monotherapy reduces LDL-C in patients with dyslipidemia by 18% after 12 weeks.⁷ One found that the percentage reduction in LDL-C was significantly greater (P < .001) with all doses of ezetimibe plus simvastatin (46% to 59%) compared with either atorvastatin 10 mg (37%) or simvastatin 20 mg (38%) monotherapy after 6 weeks.⁸

Although alirocumab can be added to other lipid therapies, most VASFHCS patients are prescribed alirocumab monotherapy. In the ODYSSEY CHOICE II study, patients were randomly assigned to receive either a placebo or alirocumab 150 mg every 4 weeks or alirocumab 75 mg every 2 weeks. The primary efficacy endpoint was LDL-C percentage change from baseline to week 24. In the alirocumab 150 mg every 4 weeks and 75 mg every 2 weeks groups, the least-squares mean LDL-C changes from baseline to week 24 were 51.7% and 53.5%, respectively, compared to a 4.7% increase in the placebo group (both groups P < .001 vs placebo). The authors also reported that alirocumab 150 mg every 4 weeks as monotherapy demonstrated a 47.4% reduction in LDL-C levels from baseline in a phase 1 study.⁹Although alirocumab monotherapy and ezetimibe plus statin therapy have been shown to effectively decrease LDL-C independently, a direct comparison of alirocumab monotherapy vs ezetimibe plus statin therapy has not been assessed, to our knowledge. Understanding the differences in effectiveness and safety between these 2 regimens will be valuable for clinicians when selecting a medication regimen for veterans with a history of ASCVD.

METHODS

This retrospective, single-center chart review used VASFHCS Computerized Patient Record System (CPRS) and Joint Longitudinal Viewer (JLV) records to compare patients with a history of ASCVD events who were treated with alirocumab monotherapy or ezetimibe plus statin. The 2 groups were randomized in a 1:3 ratio. The primary endpoint was achieving LDL-C < 70 mg/dL after 4 to 12 weeks, 13 to 24 weeks, and 25 to 52 weeks. Secondary endpoints included the mean percentage change from baseline in total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), LDL-C, and triglycerides (TG) over 52 weeks. The incidence of ASCVD events during this period was also assessed. If LDL-C < 70 mg/dL was achieved > 1 time during each time frame, only 1 incident was counted for analysis. Safety was assessed based on the incidence of any adverse event (AE) that led to treatment discontinuation.

Patients were identified by screening the prescription fill history between October 1, 2019, and December 31, 2022. The 52-week data collection period was counted from the first available fill date. Additionally, the prior authorization drug request file from January 1, 2017, to December 31, 2022, was used to obtain a list of patients prescribed alirocumab. Patients were included if they were veterans aged ≥ 18 years and had a history of an ASCVD event, had a alirocumab monotherapy or ezetimibe plus statin prescription between October 1, 2019, and December 31, 2022, or had an approved prior authorization drug request for alirocumab between January 1, 2017, and December 31, 2022. Patients missing a baseline or follow-up lipid panel and those with concurrent use of alirocumab and ezetimibe and/or statin were excluded. 

Baseline characteristics collected for patients included age, sex, race, weight, body mass index, lipid parameters (LDL-C, TC, HDL-C, and TG), dosing of each type of statin before adding ezetimibe, and use of any other antihyperlipidemic medication. We also collected histories of hypertension, hyperlipidemia, diabetes, chronic kidney disease, congestive heart failure, and smoking or tobacco use status. The baseline lipid panel was the most recent lipid panel documented before starting alirocumab or ezetimibe plus statin therapy. Follow-up lipid panel values were gathered at 4 to 12 weeks, 13 to 24 weeks, and 25 to 52 weeks following initiation of either therapy.

High-, moderate-, and low-intensity dosing of statin therapy and alirocumab dosing (75 mg every 2 weeks, 150 mg every 2 weeks, or 300 mg every 4 weeks) were recorded at the specified intervals. However, no patients in this study received the latter dosing regimen. ASCVD events and safety endpoints were recorded based on a review of clinical notes over the 52 weeks following the first available start date.

Statistical Analysis

The primary endpoint of achieving the LDL-C < 70 mg/dL goal from baseline to 4 to 12 weeks, 13 to 24 weeks, and 25 to 52 weeks after initiation was compared between alirocumab monotherapy and ezetimibe plus statin therapy using the χ² test. Mean percentage change from baseline in TC, HDL-C, LDL-C, and TG were compared using the independent t test. P < .05 was considered statistically significant. Incidence of ASCVD events and the safety endpoint (incidence of AEs leading to treatment discontinuation) were also compared using the χ² test. Continuous baseline characteristics were reported mean (SD) and nominal baseline characteristics were reported as a percentage.

RESULTS

There were 80 participants in this study: 20 in the alirocumab monotherapy group and 60 in the ezetimibe plus statin therapy group. More than 100 patients did not meet the prespecified inclusion criteria and were excluded. Mean (SD) age was 75 (8) years in the alirocumab group and 74 (8) years in the ezetimibe plus statin group. There was no significant differences in mean (SD) weight or mean (SD) body mass index. All study participants identified as White and male except for 2 patients in the ezetimibe plus statin therapy group whose race was not documented. Differences in lipid parameters were observed between groups, with mean baseline LDL-C, HDL-C, and TC higher in the alirocumab monotherapy group than in the ezetimibe plus statin therapy group, with significant differences in LDL-C and TC (Table 1).

Fourteen patients (70%) in the alirocumab monotherapy group had hypertension, compared with 31 (52%) in the ezetimibe plus statin therapy group. In both groups, most patients had previously been diagnosed with hyperlipidemia. More patients (60%) in the alirocumab group had diabetes than in the ezetimibe plus statin therapy group (37%). The alirocumab monotherapy group also had a higher percentage of patients with diagnoses of congestive heart failure and used other antihyperlipidemic medications than in the ezetimibe plus statin therapy group. Five patients (25%) in the alirocumab monotherapy group and 12 patients (20%) in the ezetimibe plus statin therapy group took fish oil. In the ezetimibe plus statin therapy group, 2 patients (3%) took gemfibrozil, and 2 patients (3%) took fenofibrate. Six (30%) patients in the alirocumab monotherapy group and 12 (20%) patients in the ezetimibe plus statin therapy group had chronic kidney disease. Although the majority of patients in each group did not use tobacco products, there were more tobacco users in the ezetimibe plus statin therapy group.

In the alirocumab monotherapy group, 15 patients (75%) were prescribed 75 mg every 2 weeks and 5 patients (25%) were prescribed 150 mg every 2 weeks. In the ezetimibe plus statin therapy group, 59 patients (98%) were prescribed ezetimibe 10 mg/d (Table 2). Forty-three patients (72%) were prescribed a high-intensity statin 10 received moderate-intensity (17%) and 7 received low-intensity statin (12%). Most patients were prescribed rosuvastatin (45%), followed by atorvastatin (42%), pravastatin (10%), and simvastatin (3%).

Primary Endpoint

During the 52-week study, more patients met the LDL-C goal of < 70 mg/dL in the alirocumab monotherapy group (70%) than in the ezetimibe plus statin therapy group (57%); however, the difference was not significant (P = .29). Of the patients prescribed alirocumab monotherapy who achieved LDL-C < 70 mg/dL, 15% achieved this goal in 4 to 12 weeks, 40% in 13 to 24 weeks, and 45% in 25 to 52 weeks. In the ezetimibe plus statin therapy group, 28% of patients achieved LDL-C < 70 mg/dL in 4 to 12 weeks, 31% in 13 to 24 weeks, and 41% in 25 to 52 weeks (Table 3).

Secondary Endpoints

During weeks 4 to 52 of treatment, the mean percentage change decreased in LDL-C (37.7% vs 21.4%; P = .01), TC (24.7% vs 12.5%; P = .01), and TG (0.9% vs 7.0%; P = .28) in the alirocumab monotherapy group and the ezetimibe plus statin therapy group, respectively (Table 4). The mean percentage change increased in HDL-C by 3.6% in the alirocumab monotherapy group and 1.8% in the ezetimibe plus statin therapy group (P = .36). During the study, ASCVD events occurred in 1 patient (5%) in the alirocumab monotherapy group and 3 patients (5%) in the ezetimibe plus statin therapy group (P = .99). The patient in the alirocumab monotherapy group had unstable angina 1 month after taking alirocumab. One patient in the ezetimibe plus statin therapy group had coronary artery disease and 2 patients had coronary heart disease that required stents during the 52-week period. There was 1 patient in each group who reported an AE that led to treatment discontinuation (P = .41). One patient stopped alirocumab after a trial of 2 months due to intolerance, but no specific AE was reported in the CPRS. In the ezetimibe plus statin therapy group, 1 patient requested to discontinue ezetimibe after a trial of 3 months without a specific reason noted in the medical record.

DISCUSSION

This study found no statistically significant difference in the incidence of reaching an LDL-C goal of < 70 mg/dL after alirocumab monotherapy initiation compared with ezetimibe plus statin therapy. This occurred despite baseline LDL-C being lower in the ezetimibe plus statin therapy group, which required a smaller reduction in LDL-C to reach the primary goal. Most patients on alirocumab monotherapy were prescribed a lower initial dose of 75 mg every 2 weeks. Of those patients, 30% did not achieve the LDL-C goal < 70 mg/dL. Thus, a higher dose may have led to more patients achieving the LDL-C goal.

Secondary endpoints, including mean percentage change in HDL-C and TG and incidence of ASCVD events during 52 weeks of treatment, were not statistically significant. The mean percentage increase in HDL-C was negligible in both groups, while the mean percentage reduction in TG favored the ezetimibe plus statin therapy group. In the ezetimibe plus statin therapy group, patients who also took fenofibrate experienced a significant reduction in TG while none of the patients in the alirocumab group were prescribed fenofibrate. Although the alirocumab monotherapy group had a statistically significant greater reduction in LDL-C and TC compared with those prescribed ezetimibe plus statin, the mean baseline LDL-C and TC were significantly greater in the alirocumab monotherapy group, which could contribute to higher reductions in LDL-C and TC after alirocumab monotherapy.Based on the available literature, we expected greater reductions in LDL-C in both study groups compared with statin therapy alone.^8,9 However, it was unclear whether the LDL-C and TC reductions were clinically significant.

Limitations

The study design did not permit randomization prior to the treatments, restricting our ability to account for some confounding factors, such as diet, exercise, other antihyperlipidemic medication, and medication adherence, which may have affected LDL-C, HDL-C, TG, and TC levels. Differences in baseline characteristics—particularly major risk factors, such as hypertension, diabetes, and tobacco use—also could have confounding affect on lipid levels and ASCVD events. Additionally, patients prescribed alirocumab monotherapy may have switched from statin or ezetimibe therapy, and the washout period was not reviewed or recorded, which could have affected the lipid panel results.

The small sample size of this study also may have limited the ability to detect significant differences between groups. A direct comparison of alirocumab monotherapy vs ezetimibe plus statin therapy has not been performed, making it difficult to prospectively evaluate what sample size would be needed to power this study. A posthoc analysis was used to calculate power, which was found to be only 17%. Many patients were excluded due to a lack of laboratory results within the study period, contributing to the small sample size. 

Another limitation was the reliance on documentation in CPRS and JLV. For example, having documentation of the specific AEs for the 2 patients who discontinued alirocumab or ezetimibe could have helped determine the severity of the AEs. Several patients were followed by non-VA clinicians, which could have contributed to limited documentation in the CPRS and JLV. It is difficult to draw any conclusions regarding ASCVD events and AEs that led to treatment discontinuation between alirocumab monotherapy and ezetimibe plus statin therapy based on the results of this retrospective study due to the limited number of events within the 52-week period.

CONCLUSIONS

This study found that there was no statistically significant difference in LDL-C reduction to < 70 mg/dL between alirocumab monotherapy and ezetimibe plus statin therapy in a small population of veterans with ASCVD, with a higher percentage of participants in both groups achieving that goal in 25 to 52 weeks. There also was no significant difference in percentage change in HDL-C or TG or in incidence of ASCVD events and AEs leading to treatment discontinuation. However, there was a statistically significant difference in percentage reduction for LDL-C and TC during 52 weeks of alirocumab monotherapy vs ezetimibe plus statin therapy.

Although there was no significant difference in LDL-C reduction to < 70 mg/dL, targeting this goal in patients with ASCVD is still clinically warranted. This study does not support a change in current VA criteria for use of alirocumab or a change in current guidelines for secondary prevention of ASCVD. Still, this study does indicate that the efficacy of alirocumab monotherapy is similar to that of ezetimibe plus statin therapy in patients with a history of ASCVD and may be useful in clinical settings when an alternative to ezetimibe plus statin therapy is needed. Alirocumab also may be more effective in lowering LDL-C and TC than ezetimibe plus statin therapy in veterans with ASCVD and could be added to statin therapy or ezetimibe when additional LDL-C or TC reduction is needed.