Medical data analysis and predictive modelling for forecasting in public health and medicine / MEDPROG

The project and its objectives

MEDPROG is an interdisciplinary research group with a three-year timeframe, funded under an internal research grant scheme at the Westphalian University of Applied Sciences. The research group involves the Department of Computer Science and Communication (FB 3) and the Department of Economics and Information Technology (FB 5) at Westphalian University of Applied Sciences, as well as the research focus on Health Economics and Quality of Life at the Institute of Work and Technology at Westphalian University of Applied Sciences.
The research group focuses on health-related, small-scale data analyses. To this end, expertise from computer science, economics and health sciences is brought together. The aim is to develop mathematical forecasting models based on socio-demographic data and health data from heterogeneous data sources, and to derive concrete recommendations for action from these. The findings are intended to contribute both to prevention and to needs assessment in regional healthcare provision.

The approach

A specific case study illustrates the methodological approach: in collaboration with the City of Gelsenkirchen, a three-hour on-site workshop was held to develop, among other things, a key research question: Is there a link in Gelsenkirchen between emergency call-outs and heat – and can affected neighbourhoods or age groups be identified? The scientific literature supports the relevance of this question: studies demonstrate, amongst other things, an increase in hospital admissions via the ambulance service during heatwaves. Few studies have been identified for the German context to date – this is where MEDPROG specifically comes in. Data from various sources are brought together for the analysis: municipal social data, weather data from stations and the German Weather Service, as well as fire brigade emergency response data. Data analysis is carried out using cluster analyses, regression analyses, time series analyses and classification methods. A key challenge lies in the heterogeneity of the data in terms of format, level of aggregation and degree of digitisation. The forecasting models developed pursue two specific objectives: in the area of prevention, they enable the identification of relevant areas for action – for example, in the field of demand planning, the models help to identify future care needs at an early stage – such as to take into account the link between hot days and rescue operations in duty rosters.

The research group comprises

  1. Prof. Dr Laura Anderle, Department of Computer Science and Communication (FB 3), Westphalian University of Applied Sciences Gelsenkirchen, Bocholt and Recklinghausen
  2. Prof. Dr Marina Arendt, Department of Economics and Information Technology (FB 5), Westphalian University of Applied Sciences Gelsenkirchen, Bocholt and Recklinghausen
  3. Dr Peter Enste, Director of the Research Focus on Health Economics and Quality of Life, IAT at the Westphalian University of Applied Sciences Gelsenkirchen, Bocholt and Recklinghausen
  4. Prof. Dr Katja Zeume, Department of Computer Science and Communication (FB 3), Westphalian University of Applied Sciences Gelsenkirchen, Bocholt and Recklinghausen