In this study, we briefly described a large urban Meals on Wheels program's adoption of the Patient Health Questionnaire-9 (PHQ-9) as its depression-screening tool. Then we reported the assessment outcomes with respect to the rates, severity, and correlates of depressive symptoms. The sample consisted of 736 MOW clients. Bivariate analysis, with chi(2) statistics, was performed to examine differences in the rates and severity of depressive symptoms by age group, gender, race/ethnicity, and cognitive status. Negative binomial regression analysis was used to determine the correlates of depression symptom severity. Of the sample, 17.5% had clinically significant depressive symptoms (PHQ-9>> or = 10), and 8.8% had probable major depressive disorder (MDD). A significantly higher proportion of those under age 60 years was found to have clinically significant depressive symptoms and probable MDD. The multivariate regression results show that age, gender, race/ethnicity, income, cognitive impairment, number of chronic medical conditions, and the nutritional risk score were significant predictors of the severity of depression symptoms. Implications of and recommendations for incorporating a valid depression-screening tool into social service agencies' existing assessment process are discussed.