Depressive disorders in primary care

John W. Williams, Caroline A. Kerber, Cynthia D. Mulrow, Anna Medina, Christine Aguilar

To assess the relative prevalence of subsyndromal depression (SubD) and major depression (MDD) in primary care patients and describe their associated functional impairments, and to define the operating characteristics of a short depression screen (SDS). Three primary care clinics: a university-affiliated Veterans Affairs clinic, a county general internal medicine clinic, and a community health center. Randomly selected adult patients (n = 221), aged > or = 30 years, with no history of psychiatric comorbidity, current substance abuse, major depressive disorder, chronic pain disorder, or dementia. The SDS and the Medical Outcomes Study Short Form 36 (SF-36) were interviewer-administered by an experienced bilingual research assistant to all subjects in the language of their choice. A physician administered independently the Structured Clinical Interview for DSM-III diagnosis (SCID) and the Hamilton Depression Rating Scale (HDRS) to all patients who exceeded a minimum threshold on the SDS and to a randomly selected sample of patients who had subthreshold scores. MDD was defined by DSM-III criteria and SubD was defined as two to four DSM-III criteria, of which one had to be depressed mood or anhedonia. Demographic and clinical characteristics of the patients were: Mexican American 53%, non-Hispanic white 38%, and African American 9%; men 68%; mean age 60 +/- 12.7 years; mean level of education 9.5 +/- 4.4 years; and hypertension 57%, diabetes mellitus 51%, and arthritis 45%. The prevalences of MDD and SubD (adjusted for sampling strategy) were 4% and 16%, respectively. For the patients who had MDD, the median HDRS score was 17 (interquartile range, 10-18), and for those who had SubD, the median HDRS score was 9 (interquartile range, 8-14). Compared with the patients who did not have depressive symptoms, those who had either MDD or SubD were significantly impaired in multiple domains of self-reported function. The sensitivity and specificity of the SDS for MDD were 100% (95% CI 57-100) and 72% (95% CI 63-81), respectively. For depressive disorders (MDD or SubD), the sensitivity was 66% (95% CI 49-83) and the specificity was 79% (95% CI 69-89). SubD was more prevalent than MDD in these primary care settings. Both MDD and SubD were associated with significant functional impairment. The sensitivity of the SDS was lower for identifying depressive disorders (MDD or SubD) than it was for identifying MD.