[Factors predicting return home of hospitalized aged patients]

The identification of predictors of return home and nursing home placement in elderly patients, at the moment of admission to a Geriatric Convalescence Unit, may allow to make a discharge planning for these patients, and also may improve efficiency in the unit.

A number of 445 patients over 65 years were consecutively admitted to the unit. Analyzed variables at admission were; age, sex, diagnostic, comorbidity, functional status prior to admission, degree of independence in activities of daily living (Barthel index), cognitive function (Folstein Mini-Mental) and social support (modified version of social-familiar evaluation scale of Gijón). All these variables were used in two models of logistic regression analysis: 1 [symbol: see text] (dependent variable: return home after discharge versus other destinations), and 2 [symbol: see text] (dependent variable: nursing home placement on discharge versus other destinations).

The variables significatively associated with greater probability of at home discharge were: moderate dependency for activities of daily living (Barthel index = 31-59), odds ratio (OR) = 2.46 (95% confidence interval: 1.43-4.22); mild dependency (Barthel index > or = 60), OR = 3.67 (1.96-6.87); normal cognitive function (Mini-Mental > or = 21), OR = 3.14 (1.76-5.60), and a favourable social support (social-familiar scale < 12), OR = 7.33 (3.18-16.91). The only variable significatively associated with the probability of nursing home placement after discharge was an impaired social function (social-familiar scale > or = 12), OR = 4.53 (2.25-9.12).

Probability of return home after discharge is associated with functional status, cognitive function and social support. Definitively nursing home placement was only related with de degree of familiar support.