Bayesian joint modeling of the health profile and demand of home care patients

The estimation of uncertain future patient demands is a key factor for the appropriate planning of human and material resources in health care facilities, where unplanned demand variations may impact the quality of schedules and, consequently, of the provided services. This issue is even more important for health services that are provided outside of hospitals, e.g. for home care (HC) services, where patients are assisted for longer periods and additional planning decisions related to the service delivery in the territory must be taken. With the goal of helping HC management to make robust decisions, we propose a Bayesian model for the estimation and prediction of both the demand for care and the history of health conditions for patients under the charge of HC services. In particular, in this study, we jointly model the temporal evolution of patient care profiles and the weekly number of visits required to nurses. The model is built so that the prediction can be easily computed by means of a Gibbs sampler. To shed light on the features and the applicative impact of our model, we have applied it to data collected from one of the largest Italian HC providers