Depression is a prevalent and disabling disease (1,2). Effective treatments for depression exist, as do effective strategies for providing these treatments (3,4,5). These strategies—which involve patient screening, patient and provider education, and care management—are relatively cost-effective, although they do increase total health care costs (6,7,8,9). However, many depressed patients do not receive appropriate care (1,10,11).
Increasing the rates of effective care for depression would yield considerable benefits for patients and society (12,13,14). However, these increases depend on the actions of various stakeholders, such as patients, health care providers, insurance carriers, and employers and other purchasers of health insurance, who will act only if they are motivated to do so (15).
Researchers have identified several barriers that may keep key stakeholders from seeking or providing effective depression care (14,16,17,18). However, little empirical evidence exists about the relative importance of different barriers.
In our study, we investigated the market for increasing the rate of evidence-based treatment of depression in Pittsburgh. Earlier research, conducted as part of the Pittsburgh Regional Health Initiative, has documented shortcomings in depression care in the Pittsburgh area (19). We interviewed employers, medical and behavioral health insurance carriers, and health care providers in Pittsburgh about their interest in and strategies for improving care for depression. Our goal was to help inform future efforts to increase the rate of effective depression care, both in Pittsburgh and nationally.
Our study was guided by the economic concept of a "market." Here, the market is defined by the intersection of patient demand for treatment and provider willingness to supply it and by purchaser demand for depression-related health insurance benefits and carrier willingness to supply them.
Demand by patients for depression treatment may be affected by a number of factors. For instance, depressed patients may not realize that they have a treatable medical condition, they may not know where to find effective treatment, they may lack the energy or motivation to seek treatment because of their illness, they may dislike the treatments that are available to them, or they may face financial constraints in paying for care, particularly if they have limited or no health insurance coverage.
Providers will supply depression treatment if doing so is economically worthwhile. However, as noted above, research has shown that effective treatment strategies include patient screening and proactive care management in addition to the treatment per se (5). Providers incur additional costs if they provide these associated services, yet most health insurance plans do not cover them, and patients may be unwilling or unable to pay for them directly (12,17).
Health insurance defines the terms by which beneficiaries are able to gain access to care and clinicians are able to provide it. We assessed factors that affect the demand for and the supply of health insurance benefits that are effective for depression.
Several factors may limit whether stakeholders demand health insurance that provides parity for medical and mental illness and, in particular, health insurance that covers depression-related care management. For instance, purchasers may be reluctant to increase the amount that they currently pay for depression-related services (20,21,22), they may underestimate the costs of depression or the benefits of treatment (23), or they may underestimate consumers' health insurance preferences when designing health benefits (24).
Furthermore, employers and health plans with relatively generous depression-related benefits may disproportionately attract persons with depression, which would drive up the costs of the health plan. This potential for adverse selection could discourage employers from offering these benefits and undermine the financial viability of generous health plans.
We identified potential interviewees in three categories: employers, insurance carriers, and health care providers. We focused mainly on the largest employers and insurers in the Pittsburgh area, because of their influence on the market. Providers were selected on the basis of previous personal contact with the research team. For reasons of cost and access, we did not include consumers in this study.
Our stakeholder sample was selected purposively, but with the goal of including variation in each category. We interviewed directors of employee benefits from three manufacturers, a bank, an urban hospital system, and a rural hospital as well as a representative of the region's small business council (seven employers); the medical directors from two of the largest medical carriers in Pittsburgh (two medical health insurance carriers) and the medical directors from each carrier's affiliated behavioral health carrier (two behavioral health insurance carriers); the directors of a large community mental health center and a large behavioral health system and psychiatrists from two large private group clinical practices (four behavioral health care providers); and primary care physicians who practiced in urban, suburban, and rural settings (four primary health care providers).
The study team contacted 17 stakeholders directly. All these stakeholders agreed to participate and were included. In addition, the Pittsburgh Business Group on Health contacted nine of its employer members about our study; two such members expressed interest and were included in our sample of employers. No follow-up was conducted for the seven members who did not respond.
We sent each stakeholder written material that described the study and scheduled an interview. All interviews were conducted in person, except for one employer who preferred a telephone interview. We assured respondents that neither they nor their organizations would be identified in public reports, except with their permission. Institutional review boards at Rand and the University of Pittsburgh approved the study protocol.
We used semistructured interview guides for each stakeholder category. We asked all stakeholders whether they viewed depression as a significant problem, whether depression care was currently adequate, whether they preferred that depression be treated in primary care or specialty settings, and whether and how they faced incentives to improve depression care. Interview guides are available on request from the authors.
Only one of the seven employers whom we interviewed considered depression to be a major problem for their workforce. Employers identified cardiovascular and pulmonary conditions, musculoskeletal conditions, gastrointestinal conditions, and diabetes as the most costly in terms of direct and indirect costs. No employer identified depression as one of the organization's three most costly medical conditions, although two employers suggested that depression might underlie some gastrointestinal disorders. Also, several employers suggested that the stigma of depression might limit their ability to measure related costs accurately; for instance, physicians might be reluctant to record a diagnosis of depression.
All of the employers preferred, but did not require, that health maintenance organization plans be accredited by the National Committee for Quality Assurance. Beyond this accreditation, all the employers said that they did not base purchasing decisions on indicators of health care quality, for depression or for any other condition. Six of the seven employers stated that they offered relatively comprehensive mental health benefits, but none reported using mental health benefits as a basis for selecting insurance carriers or plans. None of the employers said that their employees demanded improved mental health benefits, including benefits related to depression.
None of the employers whom we interviewed screened for depression in the workplace or purchased disease management programs for depression, and they were not interested in doing so. All the employers said that they are receptive to quality improvement initiatives proposed by their health vendors, particularly cost-neutral ones, but that they generally did not ask vendors to develop such initiatives. When we asked employers how they would evaluate a hypothetical program that improved depression outcomes but increased direct health care costs, all employers said that they would want to see data on the return on investment of such interventions. All the employers also said that when assessing the return on investment of this hypothetical program, they would credit savings in other parts of the workplace—for example, higher productivity and lower turnover—against the increased direct heath care costs. However, they also said that measuring such gains would be difficult.
Five employers believed that depressed patients could be treated more effectively by mental health specialists than by primary care providers. However, they recognized that many depressed patients are seen by primary care providers and felt that primary care for depression could be improved. All employers believed that their health plans reimbursed primary care providers for claims with mental health diagnoses.
Finally, most employers reported that their companies' mental health costs were rising, particularly for antidepressant medications; in part, employers used data on rising antidepressant costs and use to support their views that untreated depression was not a major problem. To the extent that employers expressed an active interest in improving care for depression, it was with regard to physicians' prescribing behavior.
Medical insurance carriers
Both of the medical health insurance carriers reported that employers often requested information on plan quality but that employers rarely based purchasing decisions on quality indicators. Moreover, both of the medical health insurance carriers reported that quality improvement activities, including chronic disease management, did not appear to be a high priority for the employers that purchased their insurance plans.
Both of the medical health insurance carriers felt that collaborating with employers had a great deal of potential for improving care. However, both carriers said that employers rarely asked them to develop any particular quality improvement initiatives and that employers showed little interest in initiatives that the carriers proposed. Both carriers described initiating some depression-related projects on their own, for example, screening for postpartum depression.
Both of the medical health insurance carriers said that behavioral health benefits were not an important issue for most purchasers. Plan contracts were generally structured with behavioral health benefits carved out to a separate vendor, and health plans did not track the quality of behavioral health care in the same way that they track other quality indicators. Both carriers said that their plans reimbursed primary care providers for claims with mental health diagnoses.
Behavioral health insurance carriers
Both of the behavioral health insurance carriers we interviewed reported having little leeway in structuring insurance contracts and little direct contact with the employers who purchased their insurance plans. In practice, both of these carriers provided carved-out benefits in conjunction with one general health insurance carrier, instead of contracting directly with employers. These respondents believed that depression was prevalent and that care was currently inadequate for many patients. At the same time, they felt that employers were relatively uninterested in benefits for mental health care, including depression.
Both behavioral health insurance carriers preferred coordinating treatment with primary care providers but said that their contracts did not provide for such coordination. Both carriers described initiating some depression-related projects with their networks of specialty providers.
All four of the primary care providers and all four of the mental health specialty providers believed that depression care needed to be improved. All the primary care providers said that neither health plans nor their colleagues supported quality improvement efforts for depression. All the primary care providers also stated that health plans evaluated them in certain clinical areas, such as mammography and immunizations, but not regarding depression. Mental health specialists participated in quality improvement programs in their organizations but viewed the current efforts as being more bureaucratic than substantive.
Both types of providers said that patients rarely asked them about the quality of mental health care.
Primary care providers generally preferred to refer complex or severe depression to specialists, and all primary care and specialty providers viewed collaboration between primary and specialty care as the preferred approach. Yet primary care and behavioral health providers each reported having problems gaining access to the other group and being able to coordinate care between the two groups. All eight of the primary care and behavioral health providers thought these barriers arose partly from the structure of carve-out contracts and the low reimbursement rates. Four primary care providers reported having difficulty at times getting reimbursed for claims with mental health diagnoses. Our questions about reimbursement practices asked about plans in general, not the practices of the specific insurance carriers interviewed.
Our findings suggest that there is currently little demand among purchasers for improving depression care and little interest among insurers and providers in improving care in the absence of purchaser demand. Even stakeholders who identified depression as an important problem did not agree about who should be responsible for addressing it.
Employers reported that they primarily look to their vendors to initiate quality improvement efforts, whereas insurance carriers reported that such improvement efforts were more likely to occur if initiated by their employer customers. Health care providers, in turn, reported feeling powerless to initiate change. Although we were unable to interview consumers directly, none of the stakeholders reported any significant demands from consumers for improved depression care.
In terms of our conceptual framework, further research is needed to determine whether the lack of purchaser demand that we found in Pittsburgh reflects information deficits, such as underestimating the costs of depression, the benefits of treatment, or the preferences of employees; concern about adverse selection; or other factors. These different issues are likely to require different solutions. For instance, it may be possible to address information deficits through additional research and education. But those same strategies may actually reinforce adverse selection, which may require risk adjustment or other policy interventions.
Although we selected our stakeholder sample to be diverse, our findings may reflect the characteristics of the particular stakeholders we surveyed or of the Pittsburgh area. For instance, several of the employers we surveyed had health plans that served more retirees than active employees; presumably employers' interests in addressing depression differ for various types of beneficiary. For these reasons and because of the sample size and exploratory nature of our study, we think it is important to replicate this study elsewhere.
During this project, one study investigator participated in a policy workshop with multiple national employers, behavioral health plans, and pharmacy benefit firms about patterns of use of antidepressant medication. Participating employers expressed concern that current patterns of use were not cost-effective and stated that they had an interest in understanding prescribing patterns for antidepressants and in promoting efficient use. However, even these stakeholders could not come to a consensus on who would be responsible for improving care. Also, both employers and vendors were uncertain about cost-effective strategies that employers could implement to increase the rate of effective depression care in their covered population. Furthermore, like the firms we interviewed in Pittsburgh, these employers generally did not express strong concern about untreated depression among their beneficiaries.
Overall, the absence of a clear locus of responsibility for improving depression care lends considerable inertia to the status quo. Our findings suggest that it will be difficult to increase rates of effective care at the population level without increasing demand. However, improving depression care seems unlikely unless employers and consumers believe that depression is a significant problem, that it is necessary and feasible to improve the quality of care for depression, and that the benefits of improvement exceed the costs.
This work was supported by a grant from the Jewish Healthcare Foundation of Pittsburgh, with additional support from grant MH-30915 from the National Institute of Mental Health through the Mental Health Intervention Research Center at the University of Pittsburgh, the Robert Wood Johnson Foundation's National Program on Depression in Primary Care, and the MacArthur Foundation's Child Mental Health Initiative. The authors thank Michael T. Flaherty for his support and insights and for his helpful facilitation of local stakeholder interviews and to Chris Whipple for her help in facilitating interviews.
Dr. Schoenbaum is affiliated with RAND, 1200 South Hayes Street, Arlington, Virginia 22202 (e-mail, email@example.com). Dr. Kelleher is with the Columbus Children's Research Institute at Ohio State University in Columbus. Dr. Lave is with the Graduate School of Public Health and Dr. Pincus is with the department of psychiatry of the University of Pittsburgh. Dr. Keyser and Dr. Pincus are with Rand in Pittsburgh.