Please activate JavaScript!
Please install Adobe Flash Player, click here for download


58 ISCB 2014 Vienna, Austria • Abstracts - Oral PresentationsWednesday, 27th August 2014 • 9:00-11:00 Monday25thAugustTuesday26thAugustThursday28thAugustAuthorIndexPostersWednesday27thAugustSunday24thAugust observations. The choices lead to a focus on direct or indirect, conditional or marginal effects of a particular type of exposure for specific (sub)popu- lations over a time horizon.We examine what is kept fixed and let loose for the practical question `what if exposure had been different’. The choices come with different meaning and distinct technical chal- lenges. We consider how practical questions and answers differ under the no unmeasured confounders assumption relying on outcome and/or pro- pensity score models and/or matching; or the instrumental variables as- sumption in a marginal or (double) conditional set-up including principle strata. We point to tutorials on the separate methods, line up overlap and differences in a principled fashion and by example. We zoom in on effec- tiveness research aiming to learn about drug effects (such as statins) from electronic health records and on the evaluation of quality of care in terms of hospital outcomes. Presented by Els Goetghebeur for STRATOSTG7 including Saskia Le Cessie, Erica Moodie, Ingeborg Waernbaum et al.  

Pages Overview