Telecare and Unplanned Hospitalization in Scotland: Evidence From Linked Survey and Administrative Data

In the face of increasing health care costs, policy makers in Scotland are interested in coming up with innovative ways of reducing unplanned hospital admissions. This paper investigates whether the use of telecare devices could be one such way. Unlike the previous studies looking into the same issue, we link the Scottish Homecare Census data to three other information sources-including the Scottish Morbidity Records-and estimate the treatment effect using time series analysis. We also employ an estimation strategy that controls for various methodological challenges associated with the analysis of observational data. A key highlight of our investigation is that the type of telecare device matters in as far as reducing unplanned hospitalization is concerned. For instance, when we conduct an analysis of the effect of using a community alarm, we find community alarm users to have a lower likelihood of being hospitalized as emergency cases than non-users, holding other factors constant. When we, however, consider the other telecare devices, the results show the opposite effect.