gms | German Medical Science

MAINZ//2011: 56. GMDS-Jahrestagung und 6. DGEpi-Jahrestagung

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V.
Deutsche Gesellschaft für Epidemiologie e. V.

26. - 29.09.2011 in Mainz

Risk assessment of time-varying factors on an acute event using the case-crossover method

Meeting Abstract

Suche in Medline nach

  • Peggy Sekula - Universitätsklinikum Freiburg, Freiburg
  • Martin Schumacher - Universitätsklinikum Freiburg, Freiburg

Mainz//2011. 56. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (gmds), 6. Jahrestagung der Deutschen Gesellschaft für Epidemiologie (DGEpi). Mainz, 26.-29.09.2011. Düsseldorf: German Medical Science GMS Publishing House; 2011. Doc11gmds072

doi: 10.3205/11gmds072, urn:nbn:de:0183-11gmds0728

Veröffentlicht: 20. September 2011

© 2011 Sekula et al.
Dieser Artikel ist ein Open Access-Artikel und steht unter den Creative Commons Lizenzbedingungen ( Er darf vervielf&aauml;ltigt, verbreitet und &oauml;ffentlich zug&aauml;nglich gemacht werden, vorausgesetzt dass Autor und Quelle genannt werden.



The analysis of a case series might be advantageous if both, the event and risk factors of interest, are rare and, thus, even case-control studies reach their limitations concerning feasibility and ability of risk estimation. Such situation arose when the risk of drugs to cause the severe cutaneous adverse reactions Stevens-Johnson syndrome and toxic epidermal necrolysis was assessed using the data obtained in the case-control study EuroSCAR [1]. Since some drugs were uncommonly used in the population, the numbers of exposed controls were partly so low that risk estimations were difficult. Furthermore, an important proportion of cases had to be excluded from the analysis due to lack of appropriate controls. Although risk assessment is still of interest, only cases are included in the subsequent study.

The case-crossover method was proposed for the analysis of time-varying factors with transient impact on an acute event utilizing a case series [2]. Different risk estimators are available [3], [4]. This method could be an alternative for risk assessment in the described situation. To investigate this, the EuroSCAR-data were re-analysed. For some drugs, the results were in agreement with those of the case-control analysis. However, there were differences that could not completely be explained by the data situation and the applied methodology. So, for example, the estimates of the case-crossover analysis were lower on average than those of the case-control analysis.

To reach a deeper understanding, a simulation model based on a Markov chain was developed. This model allows a realistic simulation of the described situation. Furthermore, conclusions about risk estimation based on the properties of Markov chains are possible. Drug exposure corresponds to a transient state of the Markov chain. By choosing several transition probabilities, different patterns of drug use can be considered. Dependent on the exposure status, an event, the skin reaction, can occur with pre-specified probability. Based on this model, different scenarios are considered in which, for example, the stationarity assumption requested by one of the estimators is violated or a confounder influences the process.

The aim of the simulation study is to compare the different estimators of the case-crossover analysis and the case-control analysis. Since the dynamic development of the whole cohort is simulated, cases and controls can be selected and analysed accordingly. Results of the simulation study for different scenarios will be presented and discussed.


Mockenhaupt M, et al. Stevens-Johnson syndrome and toxic epidermal necrolysis: Assessement of medication risks with emphasis on recently marketed drugs. The EuroSCAR-study. J Invest Dermatol. 2008;128(1):35-44.
Maclure M. The case-crossover design: a method for studying transient effects on the risk of acute events. Am J Epidemiol. 1991;133(2):144-153.
Greenland S. A unified approach to the analysis of case-distribution (case-only) studies. Stat Med. 1999;18(1):1-15.
Vines SK, Farrington CP. Within-subject exposure dependency in case-crossover studies. Stat Med. 2001;20(20):3039-3049.