The timing of change: patterns in transitions in functional status among elderly persons.

Data from the Longitudinal Study on Aging (LSOA) were analyzed to describe the heterogeneity of functional status transitions over 2-years (single-state model), and to explore whether changes in status in the previous two to four year period enhance the prediction of a subsequent transition (two-state model). Multivariate logistic regression with a robust estimate of variance was used to analyze a 7-category nominal response of: functional status (4 levels), institutionalization, death, and missing. Weighted percentages for functional status transitions and stability (unchanged status) showed that unchanged or improved functioning were at least as common as death or worsened functional status. Initial disability level, morbidities, and self-rated health were the strongest predictors of disability status after 2-years. The two-state model revealed that a previous transition (positive or negative) increases the risk for a subsequent transition, independent of initial disability level. The predictive and explanatory quality of the two-state model is substantially improved over single-state models, particularly from its ability to identify subgroups of individuals with marked functional status instability.