Sam Doerken1, Maja Mockenhaupt2, Luigi Naldi3, Martin Schumacher4, Peggy Sekula4. 1. Institute for Medical Biometry and Statistics, Faculty of Medicine and Medical Center - University of Freiburg, Freiburg, Germany. doerken@imbi.uni-freiburg.de. 2. Dokumentationszentrum schwerer Hautreaktionen (dZh), Medical Center, University of Freiburg, Freiburg, Germany. 3. USC di Dermatologia, Azienda Ospedaliero Papa Giovanni XXIII, Bergamo, Italy. 4. Institute for Medical Biometry and Statistics, Faculty of Medicine and Medical Center - University of Freiburg, Freiburg, Germany.
Abstract
BACKGROUND: The case-crossover design is an attractive alternative to the classical case-control design which can be used to study the onset of acute events if the risk factors of interest vary in time. By comparing exposures within cases at different time periods, the case-crossover design does not rely on control subjects which can be difficult to acquire. However, using the standard method of maximum likelihood, resulting risk estimates can be heavily biased when the prevalence to risk factors is very low (or very high). METHODS: To overcome the problem of low risk factor prevalences, penalized conditional logistic regression via the lasso (least absolute shrinkage and selection operator) has been proposed in the literature as well as related methods such as the Firth correction. We apply and compare several penalized regression approaches in the context of a case-crossover analysis of the European Study of Severe Cutaneous Adverse Reactions (EuroSCAR; 1997-2001). RESULTS: Out of 30 drugs, standard methods only correctly classified 17 drugs (including some highly implausible risk estimates), while penalized methods correctly classified 22 drugs. CONCLUSION: Penalized methods generally yield better risk classifications and much more plausible risk estimates for the EuroSCAR study than standard methods. As these novel techniques can be easily implemented using available R packages, we encourage routine use of penalized conditional logistic regression for case-crossover data.
BACKGROUND: The case-crossover design is an attractive alternative to the classical case-control design which can be used to study the onset of acute events if the risk factors of interest vary in time. By comparing exposures within cases at different time periods, the case-crossover design does not rely on control subjects which can be difficult to acquire. However, using the standard method of maximum likelihood, resulting risk estimates can be heavily biased when the prevalence to risk factors is very low (or very high). METHODS: To overcome the problem of low risk factor prevalences, penalized conditional logistic regression via the lasso (least absolute shrinkage and selection operator) has been proposed in the literature as well as related methods such as the Firth correction. We apply and compare several penalized regression approaches in the context of a case-crossover analysis of the European Study of Severe Cutaneous Adverse Reactions (EuroSCAR; 1997-2001). RESULTS: Out of 30 drugs, standard methods only correctly classified 17 drugs (including some highly implausible risk estimates), while penalized methods correctly classified 22 drugs. CONCLUSION: Penalized methods generally yield better risk classifications and much more plausible risk estimates for the EuroSCAR study than standard methods. As these novel techniques can be easily implemented using available R packages, we encourage routine use of penalized conditional logistic regression for case-crossover data.
Authors: Julie Papay; Nancy Yuen; Greg Powell; Maja Mockenhaupt; Thomas Bogenrieder Journal: Pharmacoepidemiol Drug Saf Date: 2011-12-05 Impact factor: 2.890
Authors: Marta Avalos; Yves Grandvalet; Nuria Duran Adroher; Ludivine Orriols; Emmanuel Lagarde Journal: Stat Med Date: 2012-03-15 Impact factor: 2.373
Authors: Chan Zeng; Sophia R Newcomer; Jason M Glanz; Jo Ann Shoup; Matthew F Daley; Simon J Hambidge; Stanley Xu Journal: Am J Epidemiol Date: 2013-09-27 Impact factor: 4.897
Authors: Elizabeth Mostofsky; Joel Schwartz; Brent A Coull; Petros Koutrakis; Gregory A Wellenius; Helen H Suh; Diane R Gold; Murray A Mittleman Journal: Am J Epidemiol Date: 2012-07-31 Impact factor: 4.897
Authors: Andrew D A C Smith; Jon Heron; Gita Mishra; Mark S Gilthorpe; Yoav Ben-Shlomo; Kate Tilling Journal: Epidemiology Date: 2015-09 Impact factor: 4.822
Authors: Dara Rouholiman; Jamison G Gamble; Sylvie D Dobrota; Ellen M Encisco; Ashish G Shah; Francisco J Grajales Iii; Larry F Chu Journal: JMIR Res Protoc Date: 2018-03-26