Cleveland Clinic Journal of Medicine

Coronary artery bypass graft (CABG) surgery is one of the most common and costly medical procedures performed in the United States.¹ However, up to one-half of post-CABG patients report significant increases in mood symptoms following surgery,² and these individuals are more likely to report poorer health-related quality of life (HRQoL) and worse functional status,³ and to experience higher risk of rehospitalizations⁴ and death⁵ despite a satisfactory surgical result.

Strategies to detect and then manage depression in CABG patients and in cardiac populations are of great interest given the potential for depression treatment to reduce cardiovascular morbidity.

In recognition of the prevalence and excess burdens associated with this condition, a recent American Heart Association (AHA) Science Advisory has advocated regular screening and treatment of cardiac patients for depression.⁶ Yet, the Advisory has been controversial,^7,8 as most depression treatment trials conducted in patients with cardiac disease have had less-than- anticipated impact on mood symptoms,^7,9–14 cardiovascular morbidity,^7,9,10,14 or mortality.^{7,9–11,13–15} Possible explanations include: (1) dependence solely on single antidepressant agents^9,14 that, in general, are often ineffective,¹⁶ untolerated, or otherwise discontinued by patients¹⁷; (2) reliance on psychologic counseling in elderly, medically ill populations who may be either unwilling or unable to adhere to successive face-to-face encounters with a therapist^10,13; (3) inadequate consideration of patients’ preferences for type and location of treatment^18,19; (4) insufficient treatment adherence^20,21; (5) perceived stigma of depression²²; (6) brief duration of treatment and followup^9,13,14; and (7) higher-than-expected spontaneous remission rates for depression.^10,14

Reprinted, with permission, from Effective Clinical Practice (Wagner EH. Eff Clin Pract 1998; 1:2-4). Copyright 1998 American College of Physicians. All rights reserved.

Over the past 15 years, numerous trials have supported use of the flexible real-world collaborative care approach to improve outcomes for depression^27,28 as well as a variety of other chronic medical conditions^29–32 and at a lower total cost of care.^33,34 This strategy is supported even outside the framework of a trial.^35,36 Moreover, collaborative care was the clinical framework³⁷ for a Robert Wood Johnson Foundation program to realign clinical and financial incentives for providing sustainable high-quality depression treatment in primary care.^38–41 It is also embraced by depression improvement initiatives supported by the MacArthur (http://www.depression-primarycare.org/)⁴² and Hartford (http://impact-uw.org/)⁴³ Foundations. In recognition, a National Heart Lung and Blood Institute–sponsored working group on the assessment and treatment of depression in patients with cardiovascular disease endorsed testing of collaborative care strategies for treating depression in combination with “usual cardiologic care” as a method to improve clinical outcomes.²³ Collaborative care has also emerged as an integral part of the “patient-centered medical home” model presently advocated by leading professional organizations to organize and reimburse PCPs for providing high-quality chronic illness care.⁴⁴

Despite this interest in collaborative care, to date, only the “Bypassing the Blues” (BtB) trial has reported the impact of this depression treatment strategy on the clinical outcomes of a population with cardiac disease.⁴⁵ In an effort to help disseminate collaborative care more broadly into routine practice as envisioned by the AHA Science Advisory, we pre sent the key design elements and main outcome findings from BtB, along with our efforts to improve upon and expand the model for testing in other cardiac conditions.

STUDY OVERVIEW

IDENTIFICATION OF DEPRESSION

Applying the two-step Patient Health Questionnaire (PHQ) depression screening strategy recently endorsed by the AHA Science Advisory,⁶ BtB recruited medically stable post-CABG patients prior to hospital discharge from seven Pittsburgh-area hospitals between 2004 and 2007. To support our recruitment efforts, we developed press releases, wall posters, newsletter articles, and brochures to inform physicians, hospital staff, patients and their families about the impact of depression on cardiovascular disease and our study (available for download at: www.bypassingtheblues.pitt.edu).

Study nurse-recruiters obtained patients’ signed informed consent to undergo screening with the two-item PHQ-2²² (“Over the past 2 weeks have you had: little interest or pleasure in doing things or “felt down, depressed, or hopeless?”).⁴⁷ We defined a positive PHQ-2 depression screen as patient endorsement of one or both of its items (90% sensitive and 69% specific for major depression among patients with cardiac disease when measured against the “goldstandard” Diagnostic Interview Schedule⁴⁸).

The psychologic and physical symptoms of depression often overlap with the post-CABG state (eg, fatigue, sleeplessness) and these elevations in depressive symptoms frequently remit spontaneously. Therefore, we administered the nine-item PHQ-9⁴⁹ over the telephone 2 weeks following hospital discharge to confirm the PHQ-2 screen. We required that patients score at least 10 to remain protocol-eligible, a threshold that signified at least a moderate level of depressive symptoms⁴⁹ and has been described as “virtually diagnostic” for depression among patients with cardiac disease (90% specific).⁴⁸

ASSESSMENT AND OUTCOME MEASURES

Upon confirmation of all protocol-eligibility criteria prior to randomization, we conducted a detailed baseline telephone assessment that included the SF-36⁵⁰ to determine mental (MCS) and physical (PCS) HRQoL, the 12-item Duke Activity Status Index (DASI)²⁶ to determine disease-specific physical functioning, and the 17-item Hamilton Rating Scale for Depression (HRS-D)²⁷ to track mood symptoms. Telephone assessors blinded as to randomization status readministered these measures at 2, 4, and 8 months’ followup and routinely inquired about any hospitalizations and mental health visits patients may have experienced since their last telephone assessment. Whenever they detected a potential “key event,” we requested a copy of relevant medical records from the hospital where the event occurred. These were then forwarded to a physician adjudication committee that was blinded as to the patient’s depression and intervention status to classify the nature of the event (cardiovascular, psychiatric, or “other”).

COLLABORATIVE CARE INTERVENTION

Following randomization, a nurse care manager telephoned each intervention patient to: (1) review his or her psychiatric history, including use of any prescription medications, herbal supplements, or alcohol to self-medicate depressive symptoms; and (2) provide education about depression, its impact on cardiac disease, and basic advice for managing the condition (eg, exercise, sleep, social contact, alcohol avoidance); and (3) assess the patient’s treatment preferences for depression.

Using a shared decision-making approach, patients then selected one or more of the following treatment options: (1) a workbook designed to impart self-management skills for managing depression⁵¹; (2) antidepressant pharmacotherapy, primarily a selective serotonin-reuptake inhibitor (SSRI) chosen according to patient preference, prior usage, and insurance coverage, but prescribed by the patient’s PCP⁴⁶; (3) referral to a local mental health specialist in keeping with the patient’s insurance coverage; and (4) “watchful waiting” if the patient’s mood symptoms were only mildly elevated and he or she had no prior history of depression.

Afterward, the nurse care manager telephoned the patient approximately every other week during the acute phase of treatment to practice skills imparted through workbook assignments, monitor pharmacotherapy, promote adherence with recommended care, and suggest adjustments in treatment as applicable. Depending upon the patient’s motivation to complete workbook assignments and whether he or she accepted antidepressant pharmacotherapy, these followup contacts typically lasted 15 to 45 minutes and continued for 2 to 6 months. The patient subsequently transitioned to the “continuation phase” of treatment, during which the care manager contacted him or her less frequently until the end of our 8-month intervention.

WEEKLY CASE REVIEW

Following discussion, the clinical team typically formulated one to three treatment recommendations that the nurse conveyed to the patient via telephone. As PCPs were responsible for prescribing all medications and dosage adjustments, we conveyed pharmacologic recommendations to them via telephone or fax. PCPs could accept or reject these recommendations at their discretion. If the patient demonstrated little response, had complex psychosocial issues (eg, impending divorce), or had an uncertain diagnosis (eg, bipolar disorder), we typically recommended referral to a mental health specialist. At quarterly intervals and at the end of the 8-month intervention, we mailed the PCP a summary of the patient’s progress that included antidepressant dosages, PHQ-9 scores, and other pertinent information.⁴⁶

PROMOTING MEDICATION ADHERENCE

To promote adherence with our treatment recommendations, our nurse care managers offered to call in antidepressant prescriptions to patients’ pharmacies under their PCP’s verbal orders, and then forwarded an order sheet for the PCP to sign and return to document it.

Some patients agreed to a trial of antidepressant pharmacotherapy but then declined or quickly discontinued it because of cost, side effects, or concerns about dependence, safety, or stigma. In these instances, particularly if the patient remained symptomatic, care managers attempted to overcome the patient’s reluctance using various motivational interviewing approaches. Care managers also provided educational materials, including the workbook,⁵¹ to mitigate any concerns, and emphasized they would monitor the patient’s clinical status closely and report back to the clinical team and the patient’s PCP for ongoing guidance. The care manager also informed the PCP of the patient’s reason(s) for nonadherence, raising the possibility that the clinician could help overcome the patient’s resistance.

OUTCOMES

Self-reported measures

PROCESSES OF CARE

Of the 150 patients randomized to our collaborative care intervention, 146 (97%) had one or more telephone care manager contacts and 83% had three or more contacts by the 4-month followup. At the 8-month conclusion of our intervention, the median number of care manager contacts per patient was 10 (range: 1–28). The proportion of intervention patients using antidepressants also increased from 15% at baseline to 44% by 8 months, and 4% reported a visit to a mental health specialist. In comparison, 31% (P = .05) and 6% (NS) of usual-care patients, respectively, were using an antidepressant or saw a mental health specialist during this period.⁴⁵

HEALTH SERVICES UTILIZATION

Depressed patients reported a similar 8-month incidence of all-cause (33% intervention vs 32% usual care) and cardiovascular-cause (15% vs 18%) rehospitalizations by randomization status. However, male intervention subjects tended to have a lower incidence of cardiovascular-cause rehospitalizations than men randomized to usual care (13% vs 23%; P = .07) and one that was similar to that of nondepressed BtB male post-CABG patients (13%). Notably, we did not observe a similar pattern among female patients enrolled in BtB. To better examine the “business case” for treating post-CABG depression, we are presently analyzing claims data from Medicare and from two large western Pennsylvania insurance providers and hope to report these analyses shortly.

DISCUSSION

BtB was the first trial to examine the impact of a real-world collaborative care strategy for treating depression in post-CABG patients or in any other cardiac population. The generalizability of our treatment strategy is enhanced by multiple design features including: (1) use of a brief, validated, two-stage PHQ depression screening procedure that was endorsed by the AHA and can be routinely implemented by nonresearch clinical personnel; (2) a centralized telephone-delivered intervention; (3) reliance on a variety of safe, effective, simple-to-dose and increasingly generic pharmacotherapy options, a commercially available workbook, and community mental health specialists to deliver step-up care; (4) consideration of patients’ prior treatment experiences, current care preferences, and insurance coverage when recommending care; (5) use of trained nurses as care coordinators across treatment delivery settings and providers across state lines; and (6) an informatics infrastructure designed to document and promote delivery of evidence-based depression treatment, care coordination, and efficient internal operations.

The ES improvement in HRS-D we observed in the BtB trial was at the upper end of a meta-analysis of 37 collaborative care trials for depression involving 12,355 primary care patients (ES: 0.25; 0.18–0.32).²⁷ It compared favorably with the improvements reported by the ENRICHD (Enhancing Recovery in Coronary Heart Disease Patients) randomized trial (ES: 0.22; 0.11–0.33),¹⁰ the SADHART (Sertraline Antidepressant Heart Attack Randomized Trial) (ES: 0.14; −0.06–0.35),⁹ and the citalopram arm of the CREATE (Canadian Cardiac Randomized Evaluation of Antidepressant and Psychotherapy Efficacy) trial (ES: 0.29; 0.05–0.52).¹³ However, our ES improvement was smaller than those generated by the more laborintensive and face-to-face interventions provided by Freedland et al’s trial of cognitive behavioral therapy (CBT) for post-CABG depression (ES: 0.73; 0.29–1.20; N = 123),¹⁵ the COPES (Coronary Psychosocial Evaluation Studies) trial of problem-solving therapy (ES: 0.59; 0.18–1.00) that was the first to report a significant reduction in major adverse cardiac events from treating depression,^15,56 or a recent meta-analysis of psychologic treatments in patients with medical disorders (ES: 1.00; 0.57–1.44).⁵⁷

Although the BtB intervention focused on depressed post-CABG patients, it is also generalizable to patients with other cardiovascular conditions. Moreover, the model can be readily adapted into practices at a variety of integrated health care delivery systems.⁵⁸ Therefore, we believe collaborative care interventions such as ours will become more widespread as elements of the 2010 Affordable Care Act are phased in.

FUTURE DIRECTIONS

Despite positive outcomes on HRQoL and mood symptoms generated by BtB and other recent trials,^15,56 it remains unclear whether effective depression treatment can reduce cardiovascular morbidity and mortality. Given the trend toward a reduced incidence of rehospitalization for cardiovascular causes among depressed male patients in BtB and findings from COPES⁵⁶ and other trials,⁷ we believe a comparative effectiveness trial of reasonable size (N < 2,000 study subjects) and cost will require an intervention capable of producing an ES reduction in mood symptoms of at least 0.50. Furthermore, because of declines in morbidity and mortality over the past decade following CABG surgery and myocardial infarction,¹ we also believe heart failure remains the only prevalent cardiovascular disorder for which to conduct this future comparative effectiveness trial.

Because an improvement of at least 0.50 ES in mood symptoms is higher than the ES improvements presently generated by collaborative care treatment approaches, it is critical to develop new interventions that blend the scalability and patient acceptability of telephone-delivered collaborative care with the greater efficacy of more intensive face-to-face counseling strategies. To address this need, we are investigating how best to incorporate Internet-delivered computerized cognitive behavioral therapy (CCBT) and other online strategies for treating depression into the BtB model. CCBT is a new and evolving technology that can improve patients’ access to personalized, convenient, and effective treatment for depression.⁵⁹ Used primarily in the United Kingdom, Australia, and the Netherlands, CCBT has attracted growing interest by US investigators.⁶⁰ Importantly, some CCBT programs are able to produce the ES improvements in mood symptoms needed to potentially demonstrate a reduction of cardiovascular morbidity⁶¹ and do so reliably, at scale, and at low cost compared with more labor-intensive methods of care.^62–64 Still, pilot testing of this innovative treatment approach is necessary to evaluate: (1) whether CCBT will be as effective among depressed patients with cardiovascular disease as among those recruited from primary care settings; (2) how best to integrate CCBT within a collaborative care program linked to cardiovascular patients’ usual sources of cardiac and primary care; and (3) whether incorporating Internet-delivered CCBT into a “traditional” collaborative care program that provides active follow-up, pharmacotherapy monitoring, and mental health specialty referral as options provides either no additional benefit (ES ∼0.30), benefit approaching that of CCBT alone (ES: ∼0.60),⁶¹ or an additive or synergistic benefit approaching face-to-face CBT (ES: ≥ 0.80).^15,65 Findings from these studies could also have profound implications for changing the way both cardiovascular and mental health conditions are treated⁶⁶ and direct further attention to the emerging field of e-mental health by other US investigators.⁶⁰

Coronary artery bypass graft (CABG) surgery is one of the most common and costly medical procedures performed in the United States.¹ However, up to one-half of post-CABG patients report significant increases in mood symptoms following surgery,² and these individuals are more likely to report poorer health-related quality of life (HRQoL) and worse functional status,³ and to experience higher risk of rehospitalizations⁴ and death⁵ despite a satisfactory surgical result.

Strategies to detect and then manage depression in CABG patients and in cardiac populations are of great interest given the potential for depression treatment to reduce cardiovascular morbidity.

In recognition of the prevalence and excess burdens associated with this condition, a recent American Heart Association (AHA) Science Advisory has advocated regular screening and treatment of cardiac patients for depression.⁶ Yet, the Advisory has been controversial,^7,8 as most depression treatment trials conducted in patients with cardiac disease have had less-than- anticipated impact on mood symptoms,^7,9–14 cardiovascular morbidity,^7,9,10,14 or mortality.^{7,9–11,13–15} Possible explanations include: (1) dependence solely on single antidepressant agents^9,14 that, in general, are often ineffective,¹⁶ untolerated, or otherwise discontinued by patients¹⁷; (2) reliance on psychologic counseling in elderly, medically ill populations who may be either unwilling or unable to adhere to successive face-to-face encounters with a therapist^10,13; (3) inadequate consideration of patients’ preferences for type and location of treatment^18,19; (4) insufficient treatment adherence^20,21; (5) perceived stigma of depression²²; (6) brief duration of treatment and followup^9,13,14; and (7) higher-than-expected spontaneous remission rates for depression.^10,14

Reprinted, with permission, from Effective Clinical Practice (Wagner EH. Eff Clin Pract 1998; 1:2-4). Copyright 1998 American College of Physicians. All rights reserved.

Over the past 15 years, numerous trials have supported use of the flexible real-world collaborative care approach to improve outcomes for depression^27,28 as well as a variety of other chronic medical conditions^29–32 and at a lower total cost of care.^33,34 This strategy is supported even outside the framework of a trial.^35,36 Moreover, collaborative care was the clinical framework³⁷ for a Robert Wood Johnson Foundation program to realign clinical and financial incentives for providing sustainable high-quality depression treatment in primary care.^38–41 It is also embraced by depression improvement initiatives supported by the MacArthur (http://www.depression-primarycare.org/)⁴² and Hartford (http://impact-uw.org/)⁴³ Foundations. In recognition, a National Heart Lung and Blood Institute–sponsored working group on the assessment and treatment of depression in patients with cardiovascular disease endorsed testing of collaborative care strategies for treating depression in combination with “usual cardiologic care” as a method to improve clinical outcomes.²³ Collaborative care has also emerged as an integral part of the “patient-centered medical home” model presently advocated by leading professional organizations to organize and reimburse PCPs for providing high-quality chronic illness care.⁴⁴

Despite this interest in collaborative care, to date, only the “Bypassing the Blues” (BtB) trial has reported the impact of this depression treatment strategy on the clinical outcomes of a population with cardiac disease.⁴⁵ In an effort to help disseminate collaborative care more broadly into routine practice as envisioned by the AHA Science Advisory, we pre sent the key design elements and main outcome findings from BtB, along with our efforts to improve upon and expand the model for testing in other cardiac conditions.

STUDY OVERVIEW

IDENTIFICATION OF DEPRESSION

Applying the two-step Patient Health Questionnaire (PHQ) depression screening strategy recently endorsed by the AHA Science Advisory,⁶ BtB recruited medically stable post-CABG patients prior to hospital discharge from seven Pittsburgh-area hospitals between 2004 and 2007. To support our recruitment efforts, we developed press releases, wall posters, newsletter articles, and brochures to inform physicians, hospital staff, patients and their families about the impact of depression on cardiovascular disease and our study (available for download at: www.bypassingtheblues.pitt.edu).

Study nurse-recruiters obtained patients’ signed informed consent to undergo screening with the two-item PHQ-2²² (“Over the past 2 weeks have you had: little interest or pleasure in doing things or “felt down, depressed, or hopeless?”).⁴⁷ We defined a positive PHQ-2 depression screen as patient endorsement of one or both of its items (90% sensitive and 69% specific for major depression among patients with cardiac disease when measured against the “goldstandard” Diagnostic Interview Schedule⁴⁸).

The psychologic and physical symptoms of depression often overlap with the post-CABG state (eg, fatigue, sleeplessness) and these elevations in depressive symptoms frequently remit spontaneously. Therefore, we administered the nine-item PHQ-9⁴⁹ over the telephone 2 weeks following hospital discharge to confirm the PHQ-2 screen. We required that patients score at least 10 to remain protocol-eligible, a threshold that signified at least a moderate level of depressive symptoms⁴⁹ and has been described as “virtually diagnostic” for depression among patients with cardiac disease (90% specific).⁴⁸

ASSESSMENT AND OUTCOME MEASURES

Upon confirmation of all protocol-eligibility criteria prior to randomization, we conducted a detailed baseline telephone assessment that included the SF-36⁵⁰ to determine mental (MCS) and physical (PCS) HRQoL, the 12-item Duke Activity Status Index (DASI)²⁶ to determine disease-specific physical functioning, and the 17-item Hamilton Rating Scale for Depression (HRS-D)²⁷ to track mood symptoms. Telephone assessors blinded as to randomization status readministered these measures at 2, 4, and 8 months’ followup and routinely inquired about any hospitalizations and mental health visits patients may have experienced since their last telephone assessment. Whenever they detected a potential “key event,” we requested a copy of relevant medical records from the hospital where the event occurred. These were then forwarded to a physician adjudication committee that was blinded as to the patient’s depression and intervention status to classify the nature of the event (cardiovascular, psychiatric, or “other”).

COLLABORATIVE CARE INTERVENTION

Following randomization, a nurse care manager telephoned each intervention patient to: (1) review his or her psychiatric history, including use of any prescription medications, herbal supplements, or alcohol to self-medicate depressive symptoms; and (2) provide education about depression, its impact on cardiac disease, and basic advice for managing the condition (eg, exercise, sleep, social contact, alcohol avoidance); and (3) assess the patient’s treatment preferences for depression.

Using a shared decision-making approach, patients then selected one or more of the following treatment options: (1) a workbook designed to impart self-management skills for managing depression⁵¹; (2) antidepressant pharmacotherapy, primarily a selective serotonin-reuptake inhibitor (SSRI) chosen according to patient preference, prior usage, and insurance coverage, but prescribed by the patient’s PCP⁴⁶; (3) referral to a local mental health specialist in keeping with the patient’s insurance coverage; and (4) “watchful waiting” if the patient’s mood symptoms were only mildly elevated and he or she had no prior history of depression.

Afterward, the nurse care manager telephoned the patient approximately every other week during the acute phase of treatment to practice skills imparted through workbook assignments, monitor pharmacotherapy, promote adherence with recommended care, and suggest adjustments in treatment as applicable. Depending upon the patient’s motivation to complete workbook assignments and whether he or she accepted antidepressant pharmacotherapy, these followup contacts typically lasted 15 to 45 minutes and continued for 2 to 6 months. The patient subsequently transitioned to the “continuation phase” of treatment, during which the care manager contacted him or her less frequently until the end of our 8-month intervention.

WEEKLY CASE REVIEW

Following discussion, the clinical team typically formulated one to three treatment recommendations that the nurse conveyed to the patient via telephone. As PCPs were responsible for prescribing all medications and dosage adjustments, we conveyed pharmacologic recommendations to them via telephone or fax. PCPs could accept or reject these recommendations at their discretion. If the patient demonstrated little response, had complex psychosocial issues (eg, impending divorce), or had an uncertain diagnosis (eg, bipolar disorder), we typically recommended referral to a mental health specialist. At quarterly intervals and at the end of the 8-month intervention, we mailed the PCP a summary of the patient’s progress that included antidepressant dosages, PHQ-9 scores, and other pertinent information.⁴⁶

PROMOTING MEDICATION ADHERENCE

To promote adherence with our treatment recommendations, our nurse care managers offered to call in antidepressant prescriptions to patients’ pharmacies under their PCP’s verbal orders, and then forwarded an order sheet for the PCP to sign and return to document it.

Some patients agreed to a trial of antidepressant pharmacotherapy but then declined or quickly discontinued it because of cost, side effects, or concerns about dependence, safety, or stigma. In these instances, particularly if the patient remained symptomatic, care managers attempted to overcome the patient’s reluctance using various motivational interviewing approaches. Care managers also provided educational materials, including the workbook,⁵¹ to mitigate any concerns, and emphasized they would monitor the patient’s clinical status closely and report back to the clinical team and the patient’s PCP for ongoing guidance. The care manager also informed the PCP of the patient’s reason(s) for nonadherence, raising the possibility that the clinician could help overcome the patient’s resistance.

OUTCOMES

Self-reported measures

PROCESSES OF CARE

Of the 150 patients randomized to our collaborative care intervention, 146 (97%) had one or more telephone care manager contacts and 83% had three or more contacts by the 4-month followup. At the 8-month conclusion of our intervention, the median number of care manager contacts per patient was 10 (range: 1–28). The proportion of intervention patients using antidepressants also increased from 15% at baseline to 44% by 8 months, and 4% reported a visit to a mental health specialist. In comparison, 31% (P = .05) and 6% (NS) of usual-care patients, respectively, were using an antidepressant or saw a mental health specialist during this period.⁴⁵

HEALTH SERVICES UTILIZATION

Depressed patients reported a similar 8-month incidence of all-cause (33% intervention vs 32% usual care) and cardiovascular-cause (15% vs 18%) rehospitalizations by randomization status. However, male intervention subjects tended to have a lower incidence of cardiovascular-cause rehospitalizations than men randomized to usual care (13% vs 23%; P = .07) and one that was similar to that of nondepressed BtB male post-CABG patients (13%). Notably, we did not observe a similar pattern among female patients enrolled in BtB. To better examine the “business case” for treating post-CABG depression, we are presently analyzing claims data from Medicare and from two large western Pennsylvania insurance providers and hope to report these analyses shortly.

DISCUSSION

BtB was the first trial to examine the impact of a real-world collaborative care strategy for treating depression in post-CABG patients or in any other cardiac population. The generalizability of our treatment strategy is enhanced by multiple design features including: (1) use of a brief, validated, two-stage PHQ depression screening procedure that was endorsed by the AHA and can be routinely implemented by nonresearch clinical personnel; (2) a centralized telephone-delivered intervention; (3) reliance on a variety of safe, effective, simple-to-dose and increasingly generic pharmacotherapy options, a commercially available workbook, and community mental health specialists to deliver step-up care; (4) consideration of patients’ prior treatment experiences, current care preferences, and insurance coverage when recommending care; (5) use of trained nurses as care coordinators across treatment delivery settings and providers across state lines; and (6) an informatics infrastructure designed to document and promote delivery of evidence-based depression treatment, care coordination, and efficient internal operations.

The ES improvement in HRS-D we observed in the BtB trial was at the upper end of a meta-analysis of 37 collaborative care trials for depression involving 12,355 primary care patients (ES: 0.25; 0.18–0.32).²⁷ It compared favorably with the improvements reported by the ENRICHD (Enhancing Recovery in Coronary Heart Disease Patients) randomized trial (ES: 0.22; 0.11–0.33),¹⁰ the SADHART (Sertraline Antidepressant Heart Attack Randomized Trial) (ES: 0.14; −0.06–0.35),⁹ and the citalopram arm of the CREATE (Canadian Cardiac Randomized Evaluation of Antidepressant and Psychotherapy Efficacy) trial (ES: 0.29; 0.05–0.52).¹³ However, our ES improvement was smaller than those generated by the more laborintensive and face-to-face interventions provided by Freedland et al’s trial of cognitive behavioral therapy (CBT) for post-CABG depression (ES: 0.73; 0.29–1.20; N = 123),¹⁵ the COPES (Coronary Psychosocial Evaluation Studies) trial of problem-solving therapy (ES: 0.59; 0.18–1.00) that was the first to report a significant reduction in major adverse cardiac events from treating depression,^15,56 or a recent meta-analysis of psychologic treatments in patients with medical disorders (ES: 1.00; 0.57–1.44).⁵⁷

Although the BtB intervention focused on depressed post-CABG patients, it is also generalizable to patients with other cardiovascular conditions. Moreover, the model can be readily adapted into practices at a variety of integrated health care delivery systems.⁵⁸ Therefore, we believe collaborative care interventions such as ours will become more widespread as elements of the 2010 Affordable Care Act are phased in.

FUTURE DIRECTIONS

Despite positive outcomes on HRQoL and mood symptoms generated by BtB and other recent trials,^15,56 it remains unclear whether effective depression treatment can reduce cardiovascular morbidity and mortality. Given the trend toward a reduced incidence of rehospitalization for cardiovascular causes among depressed male patients in BtB and findings from COPES⁵⁶ and other trials,⁷ we believe a comparative effectiveness trial of reasonable size (N < 2,000 study subjects) and cost will require an intervention capable of producing an ES reduction in mood symptoms of at least 0.50. Furthermore, because of declines in morbidity and mortality over the past decade following CABG surgery and myocardial infarction,¹ we also believe heart failure remains the only prevalent cardiovascular disorder for which to conduct this future comparative effectiveness trial.

Because an improvement of at least 0.50 ES in mood symptoms is higher than the ES improvements presently generated by collaborative care treatment approaches, it is critical to develop new interventions that blend the scalability and patient acceptability of telephone-delivered collaborative care with the greater efficacy of more intensive face-to-face counseling strategies. To address this need, we are investigating how best to incorporate Internet-delivered computerized cognitive behavioral therapy (CCBT) and other online strategies for treating depression into the BtB model. CCBT is a new and evolving technology that can improve patients’ access to personalized, convenient, and effective treatment for depression.⁵⁹ Used primarily in the United Kingdom, Australia, and the Netherlands, CCBT has attracted growing interest by US investigators.⁶⁰ Importantly, some CCBT programs are able to produce the ES improvements in mood symptoms needed to potentially demonstrate a reduction of cardiovascular morbidity⁶¹ and do so reliably, at scale, and at low cost compared with more labor-intensive methods of care.^62–64 Still, pilot testing of this innovative treatment approach is necessary to evaluate: (1) whether CCBT will be as effective among depressed patients with cardiovascular disease as among those recruited from primary care settings; (2) how best to integrate CCBT within a collaborative care program linked to cardiovascular patients’ usual sources of cardiac and primary care; and (3) whether incorporating Internet-delivered CCBT into a “traditional” collaborative care program that provides active follow-up, pharmacotherapy monitoring, and mental health specialty referral as options provides either no additional benefit (ES ∼0.30), benefit approaching that of CCBT alone (ES: ∼0.60),⁶¹ or an additive or synergistic benefit approaching face-to-face CBT (ES: ≥ 0.80).^15,65 Findings from these studies could also have profound implications for changing the way both cardiovascular and mental health conditions are treated⁶⁶ and direct further attention to the emerging field of e-mental health by other US investigators.⁶⁰

Coronary artery bypass graft (CABG) surgery is one of the most common and costly medical procedures performed in the United States.¹ However, up to one-half of post-CABG patients report significant increases in mood symptoms following surgery,² and these individuals are more likely to report poorer health-related quality of life (HRQoL) and worse functional status,³ and to experience higher risk of rehospitalizations⁴ and death⁵ despite a satisfactory surgical result.

Strategies to detect and then manage depression in CABG patients and in cardiac populations are of great interest given the potential for depression treatment to reduce cardiovascular morbidity.

In recognition of the prevalence and excess burdens associated with this condition, a recent American Heart Association (AHA) Science Advisory has advocated regular screening and treatment of cardiac patients for depression.⁶ Yet, the Advisory has been controversial,^7,8 as most depression treatment trials conducted in patients with cardiac disease have had less-than- anticipated impact on mood symptoms,^7,9–14 cardiovascular morbidity,^7,9,10,14 or mortality.^{7,9–11,13–15} Possible explanations include: (1) dependence solely on single antidepressant agents^9,14 that, in general, are often ineffective,¹⁶ untolerated, or otherwise discontinued by patients¹⁷; (2) reliance on psychologic counseling in elderly, medically ill populations who may be either unwilling or unable to adhere to successive face-to-face encounters with a therapist^10,13; (3) inadequate consideration of patients’ preferences for type and location of treatment^18,19; (4) insufficient treatment adherence^20,21; (5) perceived stigma of depression²²; (6) brief duration of treatment and followup^9,13,14; and (7) higher-than-expected spontaneous remission rates for depression.^10,14

Reprinted, with permission, from Effective Clinical Practice (Wagner EH. Eff Clin Pract 1998; 1:2-4). Copyright 1998 American College of Physicians. All rights reserved.

Over the past 15 years, numerous trials have supported use of the flexible real-world collaborative care approach to improve outcomes for depression^27,28 as well as a variety of other chronic medical conditions^29–32 and at a lower total cost of care.^33,34 This strategy is supported even outside the framework of a trial.^35,36 Moreover, collaborative care was the clinical framework³⁷ for a Robert Wood Johnson Foundation program to realign clinical and financial incentives for providing sustainable high-quality depression treatment in primary care.^38–41 It is also embraced by depression improvement initiatives supported by the MacArthur (http://www.depression-primarycare.org/)⁴² and Hartford (http://impact-uw.org/)⁴³ Foundations. In recognition, a National Heart Lung and Blood Institute–sponsored working group on the assessment and treatment of depression in patients with cardiovascular disease endorsed testing of collaborative care strategies for treating depression in combination with “usual cardiologic care” as a method to improve clinical outcomes.²³ Collaborative care has also emerged as an integral part of the “patient-centered medical home” model presently advocated by leading professional organizations to organize and reimburse PCPs for providing high-quality chronic illness care.⁴⁴

Despite this interest in collaborative care, to date, only the “Bypassing the Blues” (BtB) trial has reported the impact of this depression treatment strategy on the clinical outcomes of a population with cardiac disease.⁴⁵ In an effort to help disseminate collaborative care more broadly into routine practice as envisioned by the AHA Science Advisory, we pre sent the key design elements and main outcome findings from BtB, along with our efforts to improve upon and expand the model for testing in other cardiac conditions.

STUDY OVERVIEW

IDENTIFICATION OF DEPRESSION

Applying the two-step Patient Health Questionnaire (PHQ) depression screening strategy recently endorsed by the AHA Science Advisory,⁶ BtB recruited medically stable post-CABG patients prior to hospital discharge from seven Pittsburgh-area hospitals between 2004 and 2007. To support our recruitment efforts, we developed press releases, wall posters, newsletter articles, and brochures to inform physicians, hospital staff, patients and their families about the impact of depression on cardiovascular disease and our study (available for download at: www.bypassingtheblues.pitt.edu).

Study nurse-recruiters obtained patients’ signed informed consent to undergo screening with the two-item PHQ-2²² (“Over the past 2 weeks have you had: little interest or pleasure in doing things or “felt down, depressed, or hopeless?”).⁴⁷ We defined a positive PHQ-2 depression screen as patient endorsement of one or both of its items (90% sensitive and 69% specific for major depression among patients with cardiac disease when measured against the “goldstandard” Diagnostic Interview Schedule⁴⁸).

The psychologic and physical symptoms of depression often overlap with the post-CABG state (eg, fatigue, sleeplessness) and these elevations in depressive symptoms frequently remit spontaneously. Therefore, we administered the nine-item PHQ-9⁴⁹ over the telephone 2 weeks following hospital discharge to confirm the PHQ-2 screen. We required that patients score at least 10 to remain protocol-eligible, a threshold that signified at least a moderate level of depressive symptoms⁴⁹ and has been described as “virtually diagnostic” for depression among patients with cardiac disease (90% specific).⁴⁸

ASSESSMENT AND OUTCOME MEASURES

Upon confirmation of all protocol-eligibility criteria prior to randomization, we conducted a detailed baseline telephone assessment that included the SF-36⁵⁰ to determine mental (MCS) and physical (PCS) HRQoL, the 12-item Duke Activity Status Index (DASI)²⁶ to determine disease-specific physical functioning, and the 17-item Hamilton Rating Scale for Depression (HRS-D)²⁷ to track mood symptoms. Telephone assessors blinded as to randomization status readministered these measures at 2, 4, and 8 months’ followup and routinely inquired about any hospitalizations and mental health visits patients may have experienced since their last telephone assessment. Whenever they detected a potential “key event,” we requested a copy of relevant medical records from the hospital where the event occurred. These were then forwarded to a physician adjudication committee that was blinded as to the patient’s depression and intervention status to classify the nature of the event (cardiovascular, psychiatric, or “other”).

COLLABORATIVE CARE INTERVENTION

Following randomization, a nurse care manager telephoned each intervention patient to: (1) review his or her psychiatric history, including use of any prescription medications, herbal supplements, or alcohol to self-medicate depressive symptoms; and (2) provide education about depression, its impact on cardiac disease, and basic advice for managing the condition (eg, exercise, sleep, social contact, alcohol avoidance); and (3) assess the patient’s treatment preferences for depression.

Using a shared decision-making approach, patients then selected one or more of the following treatment options: (1) a workbook designed to impart self-management skills for managing depression⁵¹; (2) antidepressant pharmacotherapy, primarily a selective serotonin-reuptake inhibitor (SSRI) chosen according to patient preference, prior usage, and insurance coverage, but prescribed by the patient’s PCP⁴⁶; (3) referral to a local mental health specialist in keeping with the patient’s insurance coverage; and (4) “watchful waiting” if the patient’s mood symptoms were only mildly elevated and he or she had no prior history of depression.

Afterward, the nurse care manager telephoned the patient approximately every other week during the acute phase of treatment to practice skills imparted through workbook assignments, monitor pharmacotherapy, promote adherence with recommended care, and suggest adjustments in treatment as applicable. Depending upon the patient’s motivation to complete workbook assignments and whether he or she accepted antidepressant pharmacotherapy, these followup contacts typically lasted 15 to 45 minutes and continued for 2 to 6 months. The patient subsequently transitioned to the “continuation phase” of treatment, during which the care manager contacted him or her less frequently until the end of our 8-month intervention.

WEEKLY CASE REVIEW

Following discussion, the clinical team typically formulated one to three treatment recommendations that the nurse conveyed to the patient via telephone. As PCPs were responsible for prescribing all medications and dosage adjustments, we conveyed pharmacologic recommendations to them via telephone or fax. PCPs could accept or reject these recommendations at their discretion. If the patient demonstrated little response, had complex psychosocial issues (eg, impending divorce), or had an uncertain diagnosis (eg, bipolar disorder), we typically recommended referral to a mental health specialist. At quarterly intervals and at the end of the 8-month intervention, we mailed the PCP a summary of the patient’s progress that included antidepressant dosages, PHQ-9 scores, and other pertinent information.⁴⁶

PROMOTING MEDICATION ADHERENCE

To promote adherence with our treatment recommendations, our nurse care managers offered to call in antidepressant prescriptions to patients’ pharmacies under their PCP’s verbal orders, and then forwarded an order sheet for the PCP to sign and return to document it.

Some patients agreed to a trial of antidepressant pharmacotherapy but then declined or quickly discontinued it because of cost, side effects, or concerns about dependence, safety, or stigma. In these instances, particularly if the patient remained symptomatic, care managers attempted to overcome the patient’s reluctance using various motivational interviewing approaches. Care managers also provided educational materials, including the workbook,⁵¹ to mitigate any concerns, and emphasized they would monitor the patient’s clinical status closely and report back to the clinical team and the patient’s PCP for ongoing guidance. The care manager also informed the PCP of the patient’s reason(s) for nonadherence, raising the possibility that the clinician could help overcome the patient’s resistance.

OUTCOMES

Self-reported measures

PROCESSES OF CARE

Of the 150 patients randomized to our collaborative care intervention, 146 (97%) had one or more telephone care manager contacts and 83% had three or more contacts by the 4-month followup. At the 8-month conclusion of our intervention, the median number of care manager contacts per patient was 10 (range: 1–28). The proportion of intervention patients using antidepressants also increased from 15% at baseline to 44% by 8 months, and 4% reported a visit to a mental health specialist. In comparison, 31% (P = .05) and 6% (NS) of usual-care patients, respectively, were using an antidepressant or saw a mental health specialist during this period.⁴⁵

HEALTH SERVICES UTILIZATION

Depressed patients reported a similar 8-month incidence of all-cause (33% intervention vs 32% usual care) and cardiovascular-cause (15% vs 18%) rehospitalizations by randomization status. However, male intervention subjects tended to have a lower incidence of cardiovascular-cause rehospitalizations than men randomized to usual care (13% vs 23%; P = .07) and one that was similar to that of nondepressed BtB male post-CABG patients (13%). Notably, we did not observe a similar pattern among female patients enrolled in BtB. To better examine the “business case” for treating post-CABG depression, we are presently analyzing claims data from Medicare and from two large western Pennsylvania insurance providers and hope to report these analyses shortly.

DISCUSSION

BtB was the first trial to examine the impact of a real-world collaborative care strategy for treating depression in post-CABG patients or in any other cardiac population. The generalizability of our treatment strategy is enhanced by multiple design features including: (1) use of a brief, validated, two-stage PHQ depression screening procedure that was endorsed by the AHA and can be routinely implemented by nonresearch clinical personnel; (2) a centralized telephone-delivered intervention; (3) reliance on a variety of safe, effective, simple-to-dose and increasingly generic pharmacotherapy options, a commercially available workbook, and community mental health specialists to deliver step-up care; (4) consideration of patients’ prior treatment experiences, current care preferences, and insurance coverage when recommending care; (5) use of trained nurses as care coordinators across treatment delivery settings and providers across state lines; and (6) an informatics infrastructure designed to document and promote delivery of evidence-based depression treatment, care coordination, and efficient internal operations.

The ES improvement in HRS-D we observed in the BtB trial was at the upper end of a meta-analysis of 37 collaborative care trials for depression involving 12,355 primary care patients (ES: 0.25; 0.18–0.32).²⁷ It compared favorably with the improvements reported by the ENRICHD (Enhancing Recovery in Coronary Heart Disease Patients) randomized trial (ES: 0.22; 0.11–0.33),¹⁰ the SADHART (Sertraline Antidepressant Heart Attack Randomized Trial) (ES: 0.14; −0.06–0.35),⁹ and the citalopram arm of the CREATE (Canadian Cardiac Randomized Evaluation of Antidepressant and Psychotherapy Efficacy) trial (ES: 0.29; 0.05–0.52).¹³ However, our ES improvement was smaller than those generated by the more laborintensive and face-to-face interventions provided by Freedland et al’s trial of cognitive behavioral therapy (CBT) for post-CABG depression (ES: 0.73; 0.29–1.20; N = 123),¹⁵ the COPES (Coronary Psychosocial Evaluation Studies) trial of problem-solving therapy (ES: 0.59; 0.18–1.00) that was the first to report a significant reduction in major adverse cardiac events from treating depression,^15,56 or a recent meta-analysis of psychologic treatments in patients with medical disorders (ES: 1.00; 0.57–1.44).⁵⁷

Although the BtB intervention focused on depressed post-CABG patients, it is also generalizable to patients with other cardiovascular conditions. Moreover, the model can be readily adapted into practices at a variety of integrated health care delivery systems.⁵⁸ Therefore, we believe collaborative care interventions such as ours will become more widespread as elements of the 2010 Affordable Care Act are phased in.

FUTURE DIRECTIONS

Despite positive outcomes on HRQoL and mood symptoms generated by BtB and other recent trials,^15,56 it remains unclear whether effective depression treatment can reduce cardiovascular morbidity and mortality. Given the trend toward a reduced incidence of rehospitalization for cardiovascular causes among depressed male patients in BtB and findings from COPES⁵⁶ and other trials,⁷ we believe a comparative effectiveness trial of reasonable size (N < 2,000 study subjects) and cost will require an intervention capable of producing an ES reduction in mood symptoms of at least 0.50. Furthermore, because of declines in morbidity and mortality over the past decade following CABG surgery and myocardial infarction,¹ we also believe heart failure remains the only prevalent cardiovascular disorder for which to conduct this future comparative effectiveness trial.

Because an improvement of at least 0.50 ES in mood symptoms is higher than the ES improvements presently generated by collaborative care treatment approaches, it is critical to develop new interventions that blend the scalability and patient acceptability of telephone-delivered collaborative care with the greater efficacy of more intensive face-to-face counseling strategies. To address this need, we are investigating how best to incorporate Internet-delivered computerized cognitive behavioral therapy (CCBT) and other online strategies for treating depression into the BtB model. CCBT is a new and evolving technology that can improve patients’ access to personalized, convenient, and effective treatment for depression.⁵⁹ Used primarily in the United Kingdom, Australia, and the Netherlands, CCBT has attracted growing interest by US investigators.⁶⁰ Importantly, some CCBT programs are able to produce the ES improvements in mood symptoms needed to potentially demonstrate a reduction of cardiovascular morbidity⁶¹ and do so reliably, at scale, and at low cost compared with more labor-intensive methods of care.^62–64 Still, pilot testing of this innovative treatment approach is necessary to evaluate: (1) whether CCBT will be as effective among depressed patients with cardiovascular disease as among those recruited from primary care settings; (2) how best to integrate CCBT within a collaborative care program linked to cardiovascular patients’ usual sources of cardiac and primary care; and (3) whether incorporating Internet-delivered CCBT into a “traditional” collaborative care program that provides active follow-up, pharmacotherapy monitoring, and mental health specialty referral as options provides either no additional benefit (ES ∼0.30), benefit approaching that of CCBT alone (ES: ∼0.60),⁶¹ or an additive or synergistic benefit approaching face-to-face CBT (ES: ≥ 0.80).^15,65 Findings from these studies could also have profound implications for changing the way both cardiovascular and mental health conditions are treated⁶⁶ and direct further attention to the emerging field of e-mental health by other US investigators.⁶⁰

Depression has been studied extensively in relation to cardiovascular disease.^1–3 In addition to depression, anger⁴ and anxiety⁵ also may promote coronary artery disease (CAD), suggesting that emotional distress in general may be related to increased cardiovascular risk. Evidence indicates that the general distress shared across depression, anger, and anxiety predicts CAD, even after controlling for each of these specific negative emotions.⁶

THE CONCEPT OF TYPE D PERSONALITY

Lately, there is a renewed interest in broad individual differences in general distress and heart disease.⁷ Since psychologic factors often cluster together in individual patients, biobehavioral research may benefit from the identification of discrete personality subtypes.⁸ This focus on the identification of psychologically vulnerable patients who are at increased risk for adverse outcomes has led to the introduction of the distressed⁹ or type D¹⁰ personality profile in cardiovascular research. This personality construct is defined as follows:

“The type D (distressed) personality profile refers to a general propensity to psychological distress that is characterized by the combination of negative affectivity and social inhibition.”¹⁰

Negative affectivity, or the tendency to experience negative emotions across time and situations, is a major determinant of emotional distress in cardiac patients.^9,10 Patients who score high on this trait frequently report feelings of dysphoria, worry, and tension. Social inhibition, or the tendency to inhibit the expression of emotions or behavior, is a major determinant of social distress.^9,10 Patients who score high on this trait tend to avoid negative reactions from others.

Both traits define psychologically vulnerable patients and can be assessed with the type D scale (DS14).¹⁰ This brief measure consists of a seven-item negative affectivity subscale (eg, I often feel unhappy) and a seven-item inhibition subscale (eg, I am inhibited in social interactions), and has a clear two-factor structure and good reliability (Cronbach’s α = .88 and .86). Patients are classified as type D if they score 10 or higher on both DS14 subscales.¹⁰ The prevalence of type D personality ranges between 20% and 40% across different types of cardiovascular conditions.

The type D construct was designed for the early identification of chronically distressed patients. This article reviews (1) the risk of adverse events associated with type D, (2) the extent to which type D is distinct from depression, (3) the biologic pathways of type D, and (4) the implications of the type D personality profile.

RISK ASSOCIATED WITH TYPE D

The relationship between type D personality and adverse events has also been investigated in other cardiovascular conditions. Type D has been associated with poor prognosis in patients with peripheral arterial disease,¹⁷ but evidence for the prognostic role of type D in patients with chronic heart failure is mixed. In a study of patients with heart failure following myocardial infarction, type D predicted cardiac death independent of disease severity¹⁸; in a study of heart failure patients who underwent cardiac transplantation, type D was associated with early allograft rejection and increased mortality.¹⁹ However, type D was not associated with cardiac death in a recent, larger heart failure study.²⁰ The link between psychologic factors and heart failure is complex³ and may be less obvious than the type D-CAD link.²⁰ Type D has also been associated with the occurrence of life-threatening arrhythmias following implantable cardioverter defibrillator (ICD) treatment,²¹ and it has been shown to predict an increased risk for mortality in ICD patients, independent from shocks and disease severity.²²

The wide range in odds ratios and confidence intervals indicates disparity in data across these type D studies (Table 1). We recently performed a metaanalysis of prospective studies between 1996 and 2009 to provide a more reliable estimate of the risk associated with type D. In this analysis, type D was associated with a threefold increased risk of adverse events²³; the confidence interval of this pooled odds ratio ranged from 2.7 to 5.1. In addition, type D personality was associated with a threefold increased risk (range, 2.6 to 4.3) of emotional distress over time.²³ From the recent studies that were not included in this meta-analysis, one reported negative findings²⁰ and three others positive findings^16,21,22 on the risk associated with type D.

COMPARING DEPRESSION AND TYPE D

Clinical evidence shows that, after adjustment for depression, type D remained a predictor of adverse cardiac events in CAD.^16,24,25 Following ICD implantation, anxious type D patients were at risk of ventricular arrhythmias, whereas depression did not predict arrhythmias.²¹ Type D also exerts an adverse effect on patients’ health status following coronary bypass surgery,²⁶ heart failure,²⁷ or myocardial infarction,²⁸ adjusting for depressive symptoms. Type D is related to biomarkers of increased stress levels independent of depression^29–31 and, unlike depression, type D is not confounded by the severity of cardiac disorder.³²

Following myocardial infarction, only one of four distressed patients met criteria for both type D and depression; most had one form of distress but not the other.³² Research in healthy³³ and in cardiac³⁴ populations confirmed that items from depression and type D scales reflect different distress factors. After adjustment for depression at baseline, type D also predicted the incidence,³⁵ persistence,³⁶ and severity^37,38 of depression and anxiety. However, these findings do not imply that depression and type D are antonymous perspectives or that one perspective is better than the other in predicting outcomes; rather, we would like to argue that both constructs represent complementary perspectives that have added value.²³

BIOLOGIC PATHWAYS OF TYPE D

Other studies found that type D was associated with inflammatory dysregulation. In healthy adults, type D has been related to higher concentrations of C-reactive protein.⁴¹ In heart failure patients, type D is associated with increased plasma levels of the proinflammatory cytokine tumor necrosis factor (TNF)-α and its soluble receptors 1 and 2.^46,47 Increased TNF-α levels may cause suppression of bone-marrow–derived endothelial progenitor cells (EPCs) that play an important role in maintaining vascular integrity. The negative affectivity component of type D has been shown to predict decreased circulating EPC counts in healthy individuals⁴⁸; another study found that these EPC numbers were reduced by more than 50% in heart failure patients with a type D personality.⁴⁹ Type D personality is also associated with an increased oxidative stress burden in patients with chronic heart failure.²⁹ Studies on genetic linkage⁵⁰ and heritability⁵¹ further support biologic underpinnings of the type D construct.

Regarding pathways that may explain the effect of type D, some issues are of special interest. First, genetic factors contribute to stability in type D personality, but environmental factors may induce changes in type D characteristics over time.⁵¹ Hence, given this role of environmental influences over time, behavioral intervention would be feasible and useful in type D patients. Second, type D can promote heart disease indirectly through behavioral pathways. Type D has been associated with a sedentary lifestyle,^41,52 an unhealthy diet,⁵³ and a passive coping style.^54,55 Poor adherence to medical treatment^56,57 and reluctance to consult clinical staff⁵⁸ may jeopardize the working relationship with type D patients in clinical care. Intervention may focus on the management of these behavioral risk factors in type D patients. Third, many of these biologic^{40–43,48,50,51} and behavioral^41,52–54 pathways have also been documented in healthy type D individuals, which suggests that these associations cannot be explained away by the confounding effect of underlying cardiovascular disease.

CLINICAL IMPLICATIONS OF TYPE D

The findings from type D research have a number of clinical implications. Type D is associated with an increased risk of adverse events,²³ chronic distress,^35–38 and suicidal ideation.⁵⁹ Type D may also have an adverse effect on the outcome of invasive treatment.^{14,19,21,22,24,26,60}

Type D was associated with mortality and morbidity at 9 months¹⁴ and 2 years²⁴ following coronary artery stenting, and with impaired health status 1 year following bypass surgery.²⁶ Type D also predicted mortality and allograft rejection following heart transplantation,¹⁹ and an increased risk of ventricular arrhythmia²¹ and mortality²² in ICD patients. Researchers from the Cleveland Clinic have shown that type D is a risk factor for anxiety in ICD patients.⁶⁰

Regarding the DSM-IV classification by the American Psychiatric Association,⁶¹ type D qualifies for the diagnosis “psychological factors affecting medical condition” (Section 316). In keeping with this classification, the diagnostic category type D affects (1) the course of cardiovascular conditions,²³ (2) the treatment of these conditions,^56,57 and (3) the working relationship with medical staff.⁵⁸ At present, no clinical trial has examined whether intervention for distress among type D patients alters their risk for adverse events. Nevertheless, some have argued that it is plausible for type D patients to learn new strategies to reduce their level of general distress.⁶² Previous research with patients experiencing symptoms like those of type D patients suggests that psychotherapy, social skills training, stress management, and relaxation training may reduce stress in these patients and improve their ability to express their emotions to others.⁶² Others have suggested that stress management training, including communication skills and problem-solving, may further improve the risk profile and health in cardiac patients.⁶³

It is possible that type D patients may benefit from close monitoring of their clinical condition and from aggressive management of their risk factor profile to prevent adverse clinical events. Cardiac rehabilitation is an effective approach to treating risk factors and enhancing well-being in CAD.^63,64 A few studies have examined the effect of cardiac rehabilitation in type D patients. One study found a significant decrease in the social inhibition component of type D following cardiac rehabilitation, but there was no change in the prevalence of type D at 1-year follow-up.⁶⁵ Although the type D profile tends to remain stable during rehabilitation,^65,66 evidence shows that type D patients who participate in cardiac rehabilitation improve in physical and mental health status.⁶⁶ Cardiac rehabilitation may also ward off further deterioration in negative affect,⁶⁷ which, in turn, has been associated with better survival in patients who participated in rehabilitation.⁶⁸ Future studies need to examine the effect of cardiac rehabilitation and other personalized approaches to treatment in type D patients.

CONCLUSIONS

General distress shared across negative emotions^6,23 may partly account for the role of depression, anxiety, and anger in cardiovascular disorders.^1–5 Some cardiac patients are more likely to experience distress than others. Type D may identify these psychologically vulnerable patients who tend to experience general distress.²³ This propensity to general distress differs from depression, predicts adverse outcomes, is linked to plausible biologic pathways, and highlights the chronic nature of psychologic distress in some cardiac patients.

After adjustment for depression, type D remains significantly associated with an increased risk of adverse events in patients with CAD.^16,24,25 However, this association is less obvious in patients with heart failure, and type D did not predict survival in one heart failure study.²⁰ Although initial findings suggest a number of plausible biologic and behavioral pathways, more research is needed to explain the adverse effect of type D on cardiovascular outcomes. Future research also needs to investigate whether type D patients may benefit from close monitoring of their risk factors and a more personalized approach to behavioral and cardiac treatment.

Overall, the current understanding of type D indicates that general distress should not be ignored in the link between mind and heart, and that cardiovascular patients who have a type D personality profile are particularly vulnerable to the adverse clinical effects of general distress. The DS14¹⁰ is a brief, well-validated measure of type D that could be incorporated into clinical research and practice to identify patients who are at risk of chronic distress and poor prognosis.