Rozeta Simonovska (p), Egle Tafenau

Jean-Claude Thill

see extended abstract

Patrick Starkloff (p), Egle Tafenau

Jean-Claude Thill

see document

Anna-theresa Renner (p), Dieter Pennerstorfer

Jean-Claude Thill

The determinants of healthcare utilization are much debated, and while theoretical models exist, empirical analyses, especially for outpatient utilization, are still scarce. Theoretical models usually distinguish between (1) need determinants (such as age, gender and socio-economic status), (2) demand factors (e.g. health literacy and beliefs), and (3) supply-side factors (e.g. accessibility and availability of services). Most empirical socio- and health economic research in this field focuses on the socio-economic gradient (need), supplier-induced-demand (demand and supply) and competition among providers (supply). The latter two usually investigate the general availability of services, but not necessarily the geographic distribution of public and private services. We aim to fill this gap by linking research methods and approaches from health economics, regional science and geography. More specifically, we aim to investigate availability and spatial accessibility of public outpatient healthcare services and their differential effects on the utilization patterns of socio-economically heterogeneous patient populations. Furthermore, we are interested in the role of public versus private outpatient supply in determining patient flows between two regions. We are exploiting a panel data set on the patient flows between 115 political districts in the Austrian public outpatient sector from 2009 – 2016 and estimate a gravity-like model using different count data models (e.g. negative binomial model or Poisson pseudo maximum likelihood). Furthermore, we are using weighted Floating Catchment Areas (FCA) as a refined measure of accessibility of the regions. The FCA are computed in a 2-step approach using web-scraping data on the exact location of the all outpatient physicians in Austria, their opening hours and their specialisation as well as 250 × 250 grid data on population size. As first results point towards gender and age differentials in utilization behaviour for a given level of accessibility, we are going to analyse subsamples of the data (women and population over 75).