Epidemiologic studies suggest that African-American women may be less likely to obtain mental health services. Racial differences were explored in wanting and obtaining mental health services among women in an equal access primary care clinic setting after adjusting for demographics, mental disorder symptoms, and a history of sexual trauma. Participating in the study were women veterans at a primary care clinic at the Durham Veterans Affairs Medical Center. Consecutive women patients (n = 526) between the ages of 20 and 49 years were screened for a desire to obtain mental health services. Patients were given the Primary Care Evaluation of Mental Disorders questionnaire (PRIME-MD) and a sexual trauma questionnaire. Mental health service utilization was monitored for 12 months. The median age of the women was 35.8 years; 54.4% of them were African-American. African-American women expressed a greater desire for mental health services than whites, yet mental health resources at the clinic were similarly used by both racial groups. African-American women may want more mental health services; however, given an equal access system, there were no racial differences in mental health use. (J Natl Med Assoc. 2000;92:231-236.)
Key words: women * veterans * utilization mental health services
Epidemiologic studies suggest that there may be racial differences in the use of outpatient mental health services. While African-American women are thought to be less likely to obtain mental health services than white women despite reported higher rates of mental illness, reasons for the possible discrepancy are not well developed.1,2 This relatively lower use of mental health services is typically attributed to access to care barriers such as cost and insurance coverage, decreased desire for services, and stigma associated with obtaining mental health services or attending a predominantly white clinic.
The main objective of this study was to measure potential racial variation in patients' desire for and actual receipt of mental health services in a women's health clinic. For the duration of this study, care for women veterans was free in our clinic, allowing the effect of cost and insurance coverage to be controlled. The clinic had equal proportions of African-American women and Caucasian women, thus minimizing the perception that care was available for only a select group. Finally, mental health care was provided during the same days, times, and in the same clinic location as gynecological care and primary medical care, therefore reducing the stigma of attending a 'mental health clinic.' In this study environment, we examined whether patients who have current symptoms of mental disorders would obtain mental health services independent of race.
PATIENTS AND METHODS
Study Setting
The Women Veterans' Comprehensive Health Center (WVCHC) at the Durham Veterans Affairs Medical Center (DVAMC), Durham, NC opened in March, 1994 and is affiliated with Duke University Medical Center. Primary care, including general internal medicine, gynecology, and mental health services are available to women veterans who live within a 2-hour radius of the DVAMC. Additionally, specialty care in gynecology and breast cancer evaluation and treatment is offered to women veterans in the Southeastern states. The study was conducted from July 1, 1994 through December 31, 1996. During the first 3 years of clinic operation, which includes the study period, any woman veteran was eligible for care free of charge.
Patient Population
We administered a questionnaire that queries mental disorders and desire for mental health services to 528 consecutive women, ages of 20 to 49, on their initial visit to the WVCHC. Two women were excluded because they did not identify their race. Women were informed that the questionnaire was confidential, would not become part of their permanent medical record, and would be used to assess their medical care needs. Questionnaires were administered as part of an initial clinical assessment, and were then reviewed weekly by a team of mental health clinicians, including a clinical social worker, psychologist, and psychiatrist. Women who requested mental health services and women who reported symptoms of mental disorders were offered a mental health evaluation. Patients who had been referred to the center for specialty care from other VAMCs were excluded to ensure that patients in this study were presenting for primary care. Patients seeking primary care in the WVCHC were self-referred, having heard about the center from veterans' organizations and friends.
Data Instruments and Measures
On the initial visit questionnaire, the following demographic variables were measured: age, race, marital status, and education. We also administered the PRIME-MD questionnaire (PQ)3 and the Trauma Questionnaire (TQ).4 Additionally, to assess which patients wanted a mental health referral, the following question was included at the bottom of the survey: 'Would you like one of the Women's Mental Health Clinicians to review any concerns or active problems with you?'
The PRIME-MD questionnaire was developed by Pfizer, Incorporated, and was designed to help primary care physicians recognize common mental illnesses in their primary care practices.3 The PQ is a 1-page self-administered instrument consisting of 26 YES/NO questions about symptoms and/or signs present during the previous month. It is divided into five areas: mood, anxiety, somatoform, alcohol, and eating disorders. The questionnaire serves as an initial symptom screen for mental disorders and can be used with a 12-page clinician evaluation guide to aid in interpretation. The PQ is very specific (>80%) for four diagnoses commonly seen in primary care: depression, panic disorder, eating disorder, and alcohol abuse.3 When compared with other case-finding instruments for depression in primary care, the PQ is comparable to the Beck Depression Inventory, Center for Epidemiologic Studies Depression Screen, and Zung Self-Assessment Depression Scale.3,5 The PRIME-MD uses the Diagnostic and Statistical Manual of Mental Disorders in either the III revised or IV edition as a criterion standard.6 The following variables were measured on the PQ survey instrument: 15 common physical symptoms, control over eating, depression, anxiety, panic, alcohol dependence, and overall health. Physical symptoms were counted, and divided into two groups. Women endorsing three or more symptoms were given a score of 1, and those endorsing fewer than three physical symptoms were given a score of 0. Each mental health symptom was scored as 1, indicating endorsement of the symptom, or as 0, indicating no symptoms (see Ref. 3 for more details).
The TQ was developed as part of a nationwide VA study to assess the history of lifetime trauma.6 The TQ has been used extensively at the VA National Center for Posttraumatic Stress Disorder and at VA Womens' Health Centers and validated compared to a clinician interview.4,7 Three questions were used from the TQ to measure early life sexual trauma, rape, and battering:
1. 'Were you ever sexually assaulted or touched in a sexual way by a person five or more years older than you when you were younger than 13?'
2. 'Have you ever had an experience where someone used force or the threat of force to have sexual relations with you against your will?'
3. 'At any time, has a partner, spouse or significant other ever actually hit you, kicked you, or physically hurt you in any way?' These questions were considered jointly as they are highly correlated.
Two orthogonal indicator variables were created. The first, Trauma 1, was scored as 1 if the individual had experienced exactly 1 type of traumatic event (rape, early rape, or battering) and as 0 otherwise. The second, Trauma 2, was scored as 1 if an individual experienced two or three of these events, and as 0 if none or one of these events occurred (see Ref. 4 for more details).
Utilization of mental health services in the VCHC was examined at the end of the 1996 calendar year, for the previous 12 months. Using the VA Decentralized Computerized Health Program (DHCP), we identified all women who used the mental health services. If a woman had at least one visit to any of the mental health providers, she was considered to have obtained mental health services. Wanting mental health help was scored as 1 if the individual indicated she wanted to review concerns or active problems raised on the questionnaire with a mental health clinician and as 0 if she did not want this review.
Statistical Analysis
Because the objective of this study was to measure potential racial variation in desire for and receipt of mental health services, we developed a clinical model and used logistic regression techniques to assess the association between trauma, physical symptoms, symptoms of depression, symptoms of an eating disorder, race, marital status, education, desire for mental health help, and utilization of mental health services.8 First, bivariate relationships were assessed, and then multivariate modeling was completed on the full clinical model using logistic regression.8 The full model did not include panic and anxiety because these variables were highly correlated with depression (p < 0.0001) and there were concerns of multicollinearity.
RESULTS
Overall patient demographic characteristics can be summarized as follows: the mean age was 35.8 years, 54.4% were African-American, 38.2% were currently married, and 25.1% were college graduates.
Initial analysis of the racial variation found that African-American and white women were not different in age distribution (Table 1). African-American women were less likely to be married (34% vs. 43%, p = 0.03) and to have completed college (21% vs. 30%, p = 0.02). There was no difference in distribution of mental disorder symptoms between the racial cohorts except African-American women were less likely to report alcohol problems, as assessed by a modified CAGE questionnaire, which is incorporated into the PQ (11% vs. 18%, p = 0.03). Although there were no significant differences between African Americans and whites, both groups reported high rates of somatic problems and anxiety. African-American women reported less trauma than white women (18% vs. 26%, p = 0.02). However, we found that African-American women wanted a mental health referral more frequently than white women (34% vs. 24%, p = 0.01). There were no differences in the proportions of African American and white women who actually obtained mental health services (31% vs. 27%, p = 0.31) (see Table 1).
In the multivariate model, race was not significantly associated with receipt of mental health services. Results from the adjusted logistic regression showed that desire for a mental health referral (p = 0.001), depressive symptoms (p = 0.0001), and endorsing two or more traumas (p = 0.0003) were significantly associated with obtaining mental health services (Table 2).
DISCUSSION
This study suggests that racial differences among women veterans exist in the desire for mental health services in bivariate analyses. Yet, after adjusting for mental disorder symptoms, a history of trauma, and a desire for mental health services, we found a trend toward African Americans seeking more services in the WVCHC. There were 44 African American (15.4% of all African American) women who requested mental health services and received it as compared to 25 whites (10.4% of all whites) women who requested mental health services and received it.
Although some studies have shown that AfricanAmerican women obtain less mental health services than other women,9-11 other studies have shown increased levels of use with office-based mental health practitioners and psychologists.12 One important question is whether financial barriers to care account for this. Although we did not find any significant racial differences in receipt of mental health services, this may be attributed to conducting the study within a VA medical center where financial barriers are less of an issue. Relatively little research has been conducted examining mental health utilization among women in VA medical centers, the largest fully integrated health care system in the U.S.13 Women veterans receive free health care at the WVCHC, including mental health services, in this equal access system. Our finding that there lacked a racial difference in seeking mental health care is similar to past studies that have noted that interethnic differences in mental health services use are the result of socioeconomic differences and are minimized when these factors are controlled.14,15 By showing that racial differences in use of outpatient psychiatric services do not differ in an equal access system, this study undermines explanations based on attitudinal factors intrinsic to specific ethnic groups.
There are additional factors that may potentially explain the lack of racial differences in use of mental health services. First, subjects were enrolled from the Durham VA WVCHC, which had equal proportions of African-American and white women. Second, mental health care was provided in the same clinic as other health care. Thus, the clinical study setting may have contributed to eliminating racial differences in mental health use by decreasing the stigma associated with seeking mental health services because mental health services were integrated into the primary care setting.
We found no significant racial differences in the endorsement of mental disorder symptoms such as depression and anxiety. In general, African-American women reported less trauma than white women. However, despite reporting less history of trauma and similar symptoms of mental illness, a greater proportion of African-American women indicated a desire for mental health services, and it appeared that African-American women were more likely to obtain mental health services (31% vs. 27%, p < 0.31) than whites. Additional research is needed to determine why, despite less endorsement of mental disorder symptoms, African-American women sought more mental health services.
In this study and other studies of mental health service utilization, service delivery itself, including its main constituents (practitioners, administrators, and policy makers), was omitted. Little is known about how factors such as 'clinic environment' and service availability affect the likelihood of wanting and obtaining outpatient mental health treatment among African-American and white women. Further studies are needed to examine this and to consider larger, representative samples of women veterans.
Prior studies that have examined racial differences in receiving mental health services have been problematic in that studies rarely examined gender issues when examining racial utilization of health care. That is, few studies have considered the interaction between race and gender. Women tend to seek more health care than men, but it is possible that African-American women are not much different than white women in terms of seeking mental health care when equal or destigmatized access is afforded.
There are several potential limitations to this study that need to be recognized. First, there are difficulties in screening for a history of trauma, because women may not have the same definition for what is traumatic. Nevertheless, the TQ is widely used throughout veterans hospitals, and validation studies have been performed.47 Additionally, the age range of the study, 20 to 49 years, is limited. This range encompassed the majority of women who are seen in the women's clinic and was predefined to investigate those women with a higher likelihood of having experienced trauma. In a crosssectional study, it is impossible to speculate about timing of events or draw causative inferences. Finally, given the large catchment area, it is unknown whether the women who wanted mental health help but did not seek help within this clinic, were receiving services in another setting. It is known that weekly mental health visits are impossible for some women in our catchment area due to long driving distances.
CONCLUSION
The goal of any health care practice should be to provide the best care possible to the patient. This means providing qualified clinicians and necessary equipment for provision of care. But delivering quality medical care also requires knowing the needs of the patients who will utilize the practice. The women veterans in this study are of particular interest as they are users of VA primary health care services and both whites and African Americans are equally represented. It is important in planning for mental health resources to ask whether or not they want the service. Additionally, once financial barriers and stigmas are removed, racial differences in utilization dissipate.
[Reference]
REFERENCES
[Reference]
1. Regier DA, Narrow WE, Rae DS, Manderscheid RW,, Locke BZ, Goodwin FK. The de Facto US mental and addictive disorders service system: Epidemiological Catchment Area prospective I-year prevalence rates of disorders and services. Arch Gen Psychiatry. 1993;50:85-94.
2. Padgett DK, Patrick C, Burns BJ, Schlesinger HJ. Ethnic differences in use of inpatient mental health services by Blacks, Non-African-Americans, and Hispanics in a national insured population. Health Service Res. 1994;29:135-153.
3. Spitzer RL, Williams JBW, Kroenke K, et al. Utility of a new procedure for diagnosing mental disorders in primary care: the Prime-MD 1000 study. JA4.1994;272:1749-1756.
4. McIntyre LM, Butterfield MI, Parsey K, et al. Validation of a trauma questionnaire in women veterans. J Gen Internal iMed. 1999;14:186-189.
[Reference]
5. Mulrow CD, Williams JW, Gerety MB, Ramirez G, Montiel OM, Kerber C. Case-finding instruments for depression in primary care settings. Ann Intern Med. 1995;122:913-921.
6. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition. Washington, DC: American Psychiatric Association; 1987.
7. Wolfe J. Trauma, traumatic memory, and research: where do we go from here? J Traumatic Stress. 1995;8:717-726. 8. Hosmer D, Lemeshow S. Applied Logistic Regression. New York: John Wiley & Sons; 1989.
9. HuT, Snowden LR, Jerrell JM, et al. Ethnic population in public mental health: services choice and level of use. Amj Public Health. 1991;81:1429-1434.
10. Temkin-Greener H, Clark KT. Ethnicity, gender, and utilization of mental health services in a Medicaid population. Soc Sci Med. 1988;26:989-996.
11. Taube CA, Kessler LG, Burns BJ. Estimating the probability and level of ambulatory mental health service use. Health Service Res. 1986;Part 11:321-340.
12. Wood WD, Sherrets SD. Requests for outpatient mental health services: a comparison of Whites and Blacks. Comer Psychiatry. 1984;25:329-334.
13. Kizer KW. The 'New VA': a national laboratory for health care quality management. Am,JMed Qual. 1999;14:13-20. 14. Wells KB, Golding JM, Hough RL. Factors affecting
probability of use of general and medical health and social/ community services for Mexican Americans and non Hispanic Whites. Med Care. 1988;26:441-452.
15. Scheffler RM, Miller AG. Demand analysis of mental health service among ethnic subpopulations. Inquiry. 1989;26: 202-215.
[Author Affiliation]
Hayden B. Bosworth, PhD, Kelly S. Parsey MD, MHS, Marian 1. Butterfield, MD, MPH, Lauren M. McIntyre, MStat, PhD, Eugene Z. Oddone, MD, MHS, Karen M. Stechuchak, MS, and Lori A. Bastian, MD, MPH
Durham, North Carolina
[Author Affiliation]
2000. From the Women Veterans' Comprehensive Health Center (HB, MB, LB, LMM, KP); Center for Health Services Research in Primary Care, Durham Veterans Affairs Medical Center (HB, MB. LB, LMM, KS, KP, EO); Departments of Medicine (HB, LB, KP, EO), Psychiatry, Family Medicine (LMM), and Division of Biometry, Department of Community and Family Medicine (LMM); Duke University Medical Center. Requests for reprints should be addressed to Hayden B. Bosworth, PhD, Health Services Research and Development, Building 16, Room 70, Veterans Affairs Medical Center (152, 508 Fulton St., Durham, NC 27705.
