Health Disparities in Care for Depression Possibly Obscured by the Clinical Significance Criterion

The DSM-IV clinical significance criterion for psychiatric disorder requires either distress or disability. This criterion has undergone a subtle but potentially crucial modification in two key community mental health surveys—the National Institute of Mental Health Epidemiologic Catchment Area Program and the National Comorbidity Survey—that have served as the authoritative sources for estimates of treatment need in the United States. For symptoms to meet diagnostic criteria, the Diagnostic Interview Schedule (used in the Epidemiologic Catchment Area study) and the revised Composite International Diagnostic Interview (used in the National Comorbidity Survey) each required that endorsed symptoms be associated with either reports of consulting a doctor or other professional or interference in the respondent’s life or activities. This revised criterion confounds diagnosis with receipt of care and impairment. It might therefore yield underestimates of mental health care need for populations facing significant barriers to accessing this care by failing to diagnose persons who did not receive care or who had a different threshold at which they felt they were impaired. We investigated this possibility by comparing rates of major depression between African American and white subjects using data from a large, nationally representative community sample that included a diagnostic inventory for depression.

Method

The National Health Interview Survey, conducted by the National Center of Health Statistics and the Centers for Disease Prevention and Control, is the major data collection instrument used to assess the general health of the U.S. resident civilian noninstitutionalized population. It is an annual health survey of a nationally representative sample of households that in 1999 consisted of 30,801 adults interviewed in their homes by trained interviewers from the U.S. Bureau of the Census. The 1999 survey instrument included the depression module of the Composite International Diagnostic Interview–Short Form that we scored using the recommended algorithms (http://www3.who.int/cidi/CIDISFScoringMemo12-03-02.pdf). Depression was present in subjects 18 years of age or older who felt sad, blue, or depressed for 2 weeks or more in a row in the past 12 months and whose feelings had lasted half the day or longer everyday or almost everyday, and who had three or more of six symptoms during the 2-week period: lost interest in most things, more tired than usual, unintentional weight changes, trouble falling asleep, trouble concentrating, felt worthless or no good, or thought a lot about death. Subjects were asked “How much did these problems interfere with your life or activities: a lot, some, a little or not at all?” Impairment was present in those who responded “a lot”. Finally, subjects received care if they indicated telling a doctor or other professional (psychologist, social worker, counselor, nurse, clergy or other helping professional) about the problems.

Rates of major depression were calculated by race and gender defining major depression as: 1) symptoms only, 2) symptoms and impairment, 3) symptoms and medical care, and 4) symptoms and either impairment or medical care. Chi-square statistics tested for differences in rates. Risk ratios and their 95% confidence intervals (CI) are used to present effect sizes. All statistical analyses were conducted using SUDAAN (version 9) (Research Triangle Institute, Research Triangle Park, N.C.) to accommodate the complex sample design of the NHIS and the provided sampling weights.

Results

The sample consisted of 26,515 white/other individuals and 4,286 African American respondents. As seen in Table 1 , when diagnosis of depression was based on symptom counts alone, there were no differences in prevalence between white/other subjects compared with African American individuals (risk ratio=1.13, 95% CI=0.95–1.35). When operationalized as symptoms plus either receipt of care or impairment ( Figure 1 ), the clinical significance criterion differentially resulted in an increased prevalence of depression for white relative to African American subjects overall (risk ratio=1.43, 95% CI=1.14–1.78), for women (risk ratio=1.44, 95% CI=1.13–1.82), and for men (risk ratio=1.55, 95% CI=1.02–2.35). The effect was replicated for receipt of care overall (risk ratio=1.64, 95% CI=1.26–2.13), for women (risk ratio=1.73, 95% CI=1.32–2.26), and for men (risk ratio=1.60, 95% CI=0.98–2.61). For endorsement of impairment, the effect held overall (risk ratio=1.29, 95% CI=1.01–1.65) and for women (risk ratio=1.36, 95% CI=1.04–1.79), but not men (risk ratio=1.25, 95% CI=0.80–1.95). These bivariate gender and race findings were essentially unaffected when reexamined in multivariate logistic regression analyses that included covariates for gender, race, age and family income (data not shown).

Enlarge table

Discussion

Findings suggest caution in interpreting assessments of depression that require either receipt of care or self-reported impairment as a criterion for a diagnosis. Both clinical significance criterion substantially lowered estimates of major depression in African American respondents relative to white/other individuals. This effect might be explained either by differences in access to mental health care or preferences related to seeking treatment when it is available. The impact of endorsement of impairment might be attributable to differences in the self-evaluation of the effects of psychiatric symptoms or in the identification of role limitations as being due to psychiatric symptoms versus the effects of negative contextual factors associated with race and ethnic identity. Overall, when diagnosis is not made because of failure to receive care, then estimates of racial/ethnic differences in unmet need will be artificially reduced. Specifically, by making the impairment criterion a requirement for diagnosis, we may be effectively concluding that at similar levels of symptoms, African American individuals are less in need of care for depression than are white/other subjects. The present data do not allow determination whether racial differences in the prevalence of depression with the impairment criterion should be interpreted as unmet treatment needs for African American subjects or overutilization of care by white individuals, but there is ample documentation that depression in the community is in general undertreated (1) . Effects were generally found for both men and women, but the greater prevalence of depression among women held regardless of whether the impairment criterion was applied. Regardless, these findings call into the question the use of the criteria for clinical significance when ethnic/racial comparisons are being made. More generally, these results add to the range of conceptual and empirically based objections (2 – 5) to the manner in which clinical significance and therefore diagnosis are established in the scoring of interview-based measures of depression used in community surveys. Further research is needed to determine the extent to which depressive symptoms represent need for treatment regardless of reported impairment or treatment, and an obvious first step would be to determine if the clinical significance criterion improves the prediction of benefit from established treatments.