Estimating the prevalence of depression in family practice using variant methods.

: Prevalence estimates for depression in primary care vary depending on diagnostic methods and classification criteria. The present study assessed the prevalence of depression in new, female, family practice patients using self-report and office visit data. Psychological and somatic symptoms and physician interventions were used to create classification criteria. Prevalence was higher by self-report than by physician assessment. The single checklist item "depression" appeared to yield a valid prevalence estimate. Agreement between self-report and physician recognition was low. Prevalence estimates were enhanced when single-visit patients were excluded. The findings suggest that patients who report depression by questionnaire may differ from those admitting depression to physicians; therefore, patient and physician characteristics are likely to contribute to the underrecognition of depression in primary care.