Premature retirement due to back disorders. Using propensity score matching to combine cross-sectional and longitudinal datasets

Maria Weske, Deutschen Rentenversicherung Hessen and University of Marburg
Markus Thiede, Federal Institute of Occupational Safety and Health
Ulrich O. Mueller, University of Marburg

To Analyze the Career of premature retired (PMR) on the cause of back disorders it is necessary to combine information out of different sources. The German Research Datacenter of the Federal Insurance (FDZ-RV) holds Data with different information. The Longitudinal Dataset include retired as well as a sample of the working population. The Data allows to analyze episodes of unemployment, disability and employment. The Influence of these parts of the career are on the main focus of the analysis. The longitudinal Data didn’t had any information of the reason for the PMR. To get the diagnoses which lead to PMR the FDZ-RV holds a cross-sectional Dataset in which the PMR with diagnosis are contained. To analyze the career of PMR with back disorders this two sources are combined through statistical matching. Therefore the Propensity Score Matching (PSM) technique is used. This allows controlling for covariates and reducing the selection bias. To get equal distribution as in the original datasets a frequency matching will be used. This will lead to a valid result. At the end with this new dataset the career of PMR can be analyzed and contributed with the working population and the old-age retired. There can be differences expected in the described episodes of the career and also in other variables.

  See extended abstract

Presented in Poster Session 2