AbstractDifferent approaches to assessment of outcome of psychiatric illness generally are critically appraised. The article examines research design, outcome criteria and statistical analysis, including newer statistical approaches. Measurement of outcome is complex and a more detailed and sophisticated approach needs to be adopted than in the past. Some of the special problems associated with outcome studies of depressive illness in the elderly are discussed. Specific recommendations are suggested for future outcome studies of depression in the elderly. Such studies should be prospective, recruit a wide range of patient severity and make greater use of newer statistical analyses, eg survival analysis and proportional hazards model. Outcome criteria should be precisely defined, measured using standardized instruments and be comprehensive (clinical, physical health, psychosocial, etc.). Attention should be given to inter‐episode symptomatolog
展开▼