Mental Disorders and Access to Medical Care in the United States

Recent years have seen a slowing in the growth of health care costs, a trend attributable in large part to the expansion of managed care (1,2). However, questions have arisen as to whether the rush toward cost containment may in part have been accomplished at the expense of access to health care for vulnerable populations (3, 4).

People with mental disorders may face particular difficulties in obtaining needed medical care. Mental disorders may represent preexisting conditions that make purchasing new health insurance difficult or make people fearful of losing existing benefits. Restricted provider panels, with or without accompanying utilization review, may prevent people from obtaining needed specialty care even when they have health insurance and a primary care provider.

This study used a large national survey to examine the experience of individuals with mental disorders in obtaining access to medical care. By examining a number of components of access, we hoped to obtain a fuller understanding of the potential barriers to health care for people with mental disorders.

METHOD

The sample was drawn from the 1994 National Health Interview Survey (5), which interviewed 45,705 households in 198 separate U.S. regions. The total of 116,179 individuals surveyed represented a 94.1% response rate to the survey. The 1994 survey included a detailed section on access to care, as well as a disability supplement (the National Health Interview Survey-Disability) that included questions about mental symptoms and mental illness. The sampling frame for this study included all respondents over the age of 18 who responded to both the core survey and the access supplement (N=77,183).

Independent variables consisted of mental illness, demographic variables, health insurance status, and physical health status. We defined respondents with mental disorders as those who reported a psychiatric or substance abuse disorder, significant anxiety or depressive symptoms, or use of psychiatric medications within the past year (N=7,409, or 9.6% of the total sample). The demographic variables of age, race, sex, income, and geographic region were available through the National Health Interview Survey and were included in all multivariate models. We adjusted for type of insurance coverage (fee-for-service, health maintenance organization [HMO], Medicare, Medicaid, uninsured) through use of the survey’s health insurance supplement. Physical health status was controlled for by using self-reported health (a 5-point scale ranging from excellent to poor) (6,7).

Dependent variables were insurance and primary care and actual receipt of medical services. We assessed insurance status (any insurance/no insurance) as the first measure of access (N=77,183). Among individuals with insurance (N=65,800), we included two additional measures of difficulty with obtaining or maintaining coverage: past denial of insurance due to a preexisting condition and the need to stay in the current job for 2 years or more for fear of losing health insurance. We also assessed whether the individual reported having a primary care provider.

Among individuals with insurance and a primary care provider (N=9,138), we used two self-report variables as markers of actual receipt of medical care: delay in seeking medical care because of cost and inability to obtain needed medical care. We conducted these analyses among individuals who had fair or poor health status, which we considered a proxy for need for medical services. (These analyses continued to control for fair versus poor health status.)

Because all outcomes were dichotomous, logistic regression was used to model each access measure as a function of the presence or absence of a mental disorder. All models controlled for demographic variables, type of insurance, and self-reported general health status. The SUDAAN statistical package, with appropriate weighting and nesting variables, was used for all statistical comparisons because of the complex stratified survey design.

RESULTS

A total of 7,409 individuals (9.6% of the sample) reported evidence of a mental disorder or use of psychiatric medications within the past year. Within this group, 6,406 (86.5%) reported a mood or anxiety disorder or significant mood or anxiety symptoms, 309 (4.2%) reported a psychotic disorder, 463 (6.2%) reported a substance abuse disorder, and 237 (3.2%) reported taking a medication for a mental health problem.

For the sample as a whole (i.e., before analyses of subgroups), lack of insurance was the most common barrier to receiving care (17.1%, N=13,198); the second most common barrier was lack of a usual source of care (14.8%, N=11,423). People who reported denial of insurance due to a preexisting condition represented 1.5% of the sample (N=1,158), and 2.3% (N=1,775) reported staying in their current job at least 2 years because they were afraid of losing their health benefits. Almost one-tenth of the people in the sample (9.5%, N=7,332) reported that they had delayed seeking needed care in the past year because of cost. A smaller but still substantial number (2.8%, N=2,161) reported being unable to obtain necessary medical care.

In multivariate models, people who reported mental disorders were no more or less likely to lack insurance than those without mental disorders (odds ratio=0.98, 95% confidence interval=0.94–1.03, p=0.20). However, they were significantly more likely to report problems in obtaining or maintaining their coverage. Insured people with mental disorders were approximately twice as likely to report having been denied insurance because of a preexisting condition (odds ratio=2.18, 95% confidence interval=1.77–2.68, p<0.0001) or having had to stay in their present job because they were afraid of losing health benefits (odds ratio=1.90, 95% confidence interval=1.62–2.23, p<0.0001). People with mental disorders were as likely as people without mental disorders to have a primary care provider (odds ratio=0.97, 95% confidence interval=0.94–1.01, p=0.27).

Having a mental disorder conferred a 76% greater risk of having delayed seeking care because of cost (odds ratio=1.76, 95% confidence interval=1.48–2.09, p<0.0001). Individuals who reported mental disorders were more than twice as likely as people without mental disorders to have been unable to obtain needed medical care (odds ratio=2.30, 95% confidence interval=1.78–2.97, p<0.0001). Type of coverage—Medi­caid, Medicare, fee-for-service, or HMO—was not a significant predictor of access to medical care (χ²=1.00, df=3, p=0.39).

DISCUSSION

The findings of this study suggest that people with mental disorders may face substantial barriers to obtaining and maintaining both health insurance and necessary health care. While they were no more likely to be uninsured than those without mental disorders, they were more likely to have had difficulty in procuring their insurance (as represented by denial for a preexisting condition) and had more concerns about losing their insurance (reflected in staying in their jobs because of fear of losing health benefits). While they were equally likely to have a primary care provider, they were about twice as likely to have delayed seeking care or to have been unable to obtain needed medical care.

Before discussing the implications of these findings, we should mention two limitations to the use of the National Health Interview Survey for the study’s purposes. The use of self-report measures to identify mental disorders lacks the sensitivity of comprehensive diagnostic interviews. This method of case identification is the most likely explanation for the relatively low prevalence of mental disorders and may also result in a sample with greater levels of symptoms than the general population of people with mental disorders in the United States. Second, access to care is a complex construct that is difficult to capture through use of summary measures (8, 9). Nonetheless, the access survey does represent a state-of-the-art instrument that captures a number of domains not available in previous surveys (3).

Inasmuch as these findings were adjusted for physical health status, they suggest that many respondents had been denied insurance because of a preexisting mental disorder. In the current health care system, insurers’ fear of adverse selection—the risk that people with expensive, chronic conditions will differentially choose plans with more generous mental health benefits—provides a strong incentive to discourage enrollment from people with mental disorders (10). Despite the fact that 45 states now prohibit restrictions for preexisting conditions, as of 1995, 56% of enrollees in fee-for-service plans, and 71% of people enrolled in preferred provider organizations, still faced limitations in their benefits related to preexisting health conditions (11).

Because switching jobs often requires switching health insurance, people in treatment for mental disorders may be effectively locked into their jobs for fear of losing their coverage. The Health Insurance Portability and Accountability Act of 1996 (i.e., the Kassebaum-Kennedy Act) requires that insurance be portable if an employee changes jobs or an employer changes health plans and also places restrictions on denial of benefits due to preexisting conditions (12). However, authors have expressed concern that this law may add regulatory burden while failing to address the problems that underlie lack of access to insurance in the current system (13).

The study’s findings suggest that while improving rates of insurance coverage (14) or access to generalist providers (15) may be worthwhile goals, they may be insufficient to address the access gap for people with mental disorders. Even individuals with insurance may have substantial out-of-pocket expenses related to copayments and services not covered by their insurance. Furthermore, the problems faced by people with mental disorders in obtaining needed medical care mean that laws working toward parity of mental health and medical insurance benefits (16) may fall short of providing full access to either medical or mental health care.

Finally, the study’s results have implications for systems measuring access as an indicator of quality of care (17). Had we examined only measures typically used by those “report card” systems for monitoring care—for instance, presence of an outpatient medical visit—we might have concluded that people with mental disorders had access to care that was as good as or better than that of people without such disorders. The findings of this study suggest that it may be necessary to use measures of access that are more complete than those currently available in quality monitoring systems.

Equity of access to care has been described as one of the fundamental principles underlying a just health care system (18). The results of this study suggest that there may be a serious gap in access to medical care for people with mental disorders. Assessing and successfully addressing the gap will require maintaining a focus on people with mental disorders as a group at risk in the current health care system.

Received Oct. 22, 1997; revisions received Feb. 12 and May 12, 1998; accepted June 12, 1998. From the Northeast Program Evaluation Center, VA Connecticut Healthcare System, and the Departments of Psychiatry and Public Health, Yale University. Address reprint requests to Dr. Druss, Department of Psychiatry/116A, Yale University, 950 Campbell Ave., West Haven, CT 06516. Supported in part by grants from the Donaghue Medical Research Foundation and the National Alliance for Research on Schizophrenia and Depression.