[Predictive factors for institutionalization in the context of geriatric planning in the Gironde department (France)].

The objective of this study was to investigate the predictive factors over five years follow-up of institutionalization, relevant for starting interventions in the area of Gironde (South Western France). This study deals with a longitudinal study of community residents aged 65 years and older. Of the 2792 subjects initially included in the cohort in year 1988, 2493 were selected for this study, leading to the subjects whose type of housing was known at least once over the five year follow-up (1989-1993). The statistical analysis was made with a discrete Cox proportional hazard model. Of the 2493 subjects, 150 entered an institution during the five years of follow-up, leading to a cumulative incidence of 7.1%. In the multivariate analysis, the predictive factors found in this study were age, absence of phone in the house, living in home for the aged, cognitive decline evaluated with the Mini-Mental-Status-Examination (MMSE), dependency for one item of the Lawton's scale (Instrumental Activities of Daily Living or IADLs), namely the inability to do shopping, and limited social contact with the family. These different indicators are a basis for decision making to target home care services fitted to the needs of the area.