Objective: To determine whether depression in older medical inpatients predicts mortality.
Method: Medline, PsycINFO, Embase, and the Cochrane Database of Systematic Reviews were searched for potentially relevant articles; the bibliographies of relevant articles were searched for additional references. Retrieved studies were screened to meet five inclusion criteria. Validity of studies was assessed according to four criteria adapted from the Evidence-Based Medicine Working Group. Data were abstracted from each study and tabulated. Data synthesis involved a qualitative meta-analysis.
Results: Many of the studies had methodological limitations. Six reported that depression predicted increased mortality, five reported that depression did not predict mortality, and one reported that depression predicted decreased mortality when there was a history of prior depression. Unadjusted risk ratios for death ranged from 0.60 to 12.6; adjusted risk ratios ranged from 0.42 to 7.4. The disparate findings may be explained in part by differences in the proportions of young older patients and men enrolled in the different studies.
Conclusion: The evidence that depression in older medical inpatients predicts mortality is inconclusive. There is a need for further studies that pay attention to design, populations enrolled, and analysis.