Abstract

OBJECTIVES:

There are a paucity of long-term studies from the United States concerning predictors of outcome among depressed older community adults. This article examines predictors of depression in a biracial sample of older persons in Brooklyn, NY.

METHODS:

The authors conducted a naturalistic study of 110 persons aged 55 years and older living in randomly selected block groups who had a Center for Epidemiologic Studies-Depression (CES-D) score of > or = 8 at baseline. Persons were reassessed on an average of 3 years later. Their mean age was 69 years, 52% were women, and 35% were whites, and 65% were blacks, among whom 71% were African Caribbeans. Using George's Social Antecedent Model of Depression, the authors examined the impact of 13 predictor variables on two outcome measures: presence of either subsyndromal or syndromal depression (CES-D score > or = 8) and presence of syndromal depression (CES-D score > or = 16). To control for design effects, the authors used SUDAAN for the data analysis.

RESULTS:

On follow-up, 82% and 88% of subsyndromally and syndromally depressed persons at baseline, respectively, were depressed (CES-D > or = 8). In logistic regression, baseline depressive symptoms, baseline anxiety symptoms, greater increase in anxiety symptoms during the follow-up period, and higher locus of control were predictors of any level of depression. These four variables along with greater paranoid ideation and/or psychoses and more reliable social contacts were significant predictors of syndromal depression on follow-up. There were no inter- or intraracial differences in outcome.

CONCLUSION:

Depressed community elders in Brooklyn have highly unfavorable outcomes. Preventive strategies that target at-risk persons-i.e., especially those with baseline subsyndromal depression, greater anxiety symptoms, and more paranoid ideation and/or psychoses-may reduce the development of severe or persistent depression.