Background: The present study examines transitions in the functional status and discharge destination of new nursing home admissions who remain at least 100 days, and ascertains baseline covariates associated with transition patterns.
Methods: Using a fully observed, continuous-time Markov chain model for maximum likelihood estimation of probability intensities, transition processes are characterized. The long-stay cohort (n = 9,541) was derived from a sample of elders newly admitted to 48 National Health Corporation (NHC) nursing homes between 1983 and 1987. Assessment of functional status, using a modified Katz ADL scale, occurred during the first, second, and third months, and the second, third, and fourth quarters after admission. Four types of residential changes (e.g., mortality) were examined for the latter three quarters.
Results: While stability was the predominant pattern during the first 90 days in the nursing home, 51.5% of residents experienced a change in function. The probability of change was higher for modest (one level) rather than substantial change, and for such change to represent improvement rather than decline. Over 25% of this long-stay sample exited in the second quarter, 37% of them returning home. The rates for returns to home and for mortality were strongly related to functional level. Several sociodemography variables (e.g., age, source of payment), diagnostic indicators (e.g., cancer), and orientation status were consistently associated with transition rates within functionally homogeneous groups.
Conclusions: Overall, findings underscore the heterogeneity and complexity of transition patterns for a subgroup historically treated as likely to decline or remain stable, and viewed as "permanent" residents. The model was useful in describing the volatile transition process among older long-stayers.