Title: Methods and applications for the secondary use of claims data from the Austrian health insurance system
Language: English
Authors: Endel, Florian 
Qualification level: Diploma
Advisor: Filzmoser, Peter 
Issue Date: 2021
Number of Pages: 165
Qualification level: Diploma
Background Previous studies suggest that a significantly increased risk to suffer frommyocardial infarction can be observed for parents in comparison with adults withoutchildren. However, definitive evidence is lacking and insufficient to adapt clinical practiceguidelines on that basis. Furthermore, actual cases seem to be rare and backing data isnot available.Objectives The main objective of this work is to gather evidence on the relative risk ofmyocardial infarction in parents compared with couples without children in a retrospective,observational cohort study.Methods Reimbursement data from the health care system are routinely collectedby Austrian social insurance institutions for administrative and accounting purposes.GAP-DRG, a linked research database holding data from the Austrian health and socialinsurance system covering several years is utilized to determine the number of potentialcases.Genealogical information is crucial but lacking from the data source. As a result, parentship is not encoded in the available administrative data. Therefore, a method forindirect deduction of this personal information is developed and implemented. Furthermore, information about individual comorbidities and the social-economic status arededuced.Identified cohorts are documented in detail and various statistical procedures are applied.Such methods include univariate statistics and cross-tabulations, decision trees, andmultivariate regression models. Resampling, balancing, and propensity score matchingare used to achieve more accurate estimates and robustness.Results In summary, no additional evidence was found to support the initial claim.In addition, unprecedented developments such as the disclosure of previously unknowndata quality problems in the administrative data source, coinsurance networks andgenealogical information, the handling of socioeconomic information, and morbidityscores are reported in detail.
Keywords: Health risk; Propensity scores; Regression; Robustness
URI: https://doi.org/10.34726/hss.2021.35400
DOI: 10.34726/hss.2021.35400
Library ID: AC16194153
Organisation: E180 - Fakultät für Informatik 
Publication Type: Thesis
Appears in Collections:Thesis

Files in this item:

Page view(s)

checked on Oct 17, 2021


checked on Oct 17, 2021

Google ScholarTM


Items in reposiTUm are protected by copyright, with all rights reserved, unless otherwise indicated.