Conference article

Ageing population and traditional consequences of ageing is expected to cause an increased number of recipients of community care services and socio-economic burden. To accommodate these changes, The Danish Health Authority has recommended several tools to help community personnel detect early signs of disease in the home care setting. These tools are used solely for detecting current deviations from the habitual health status, and no data analysis is performed in order to predict upcoming deviations. This paper describes a study protocol to investigate the potential of developing a data driven decision support model to predict unplanned, preventable hospitalizations. Machine learning techniques, such as logistic regression, will be applied on data from three various sources in order to predict which recipients of home care services are at risk of an unplanned, preventable hospitalization. If successful, the proposed model may facilitate earlier prediction and actions towards deviations from the individual citizen’s habitual health status and thereby increase the chance of prevention of hospitalization and functional decline.

Health Services for the Aged, Decision Support Techniques, Forecasting, Hospitalization.