Modeling Multidimensional Transitions in Health Care

Conventional multistate life table accounting procedures are based on theoretical assumptions that are appropriate primarily for demographic events. Applying these approaches to the area of health care, however, may lead to serious biases given the frequent turnovers of events such as hospitalization and institutionalization. In addition, traditional approaches have been criticized for failing to capture population heterogeneity. This research introduces a new algorithm to estimate multistate life table indicators regarding health care use, taking advantage of the availability of information on average lengths of stay in hospitals and nursing homes. The survival analysis approach is used to estimate age-specific transition probabilities in order to address the issue of population heterogeneity.