Confounders and co-interventions identified in non-randomized studies of interventions.

OBJECTIVES:To identify potential confounders and co-interventions systematically to optimise control of confounding for three non-randomized studies of interventions (NRSI) designed to quantify bleeding in populations exposed to different dual antiplatelet therapy (DAPT). STUDY DESIGN AND SETTING:Systematic review, interviews, and surveys with clinicians. We searched Ovid Medline, Ovid Embase, and the Cochrane Library to identify randomized-controlled trials and cohort studies of DAPT interventions. Two researchers independently screened citations, identified eligible studies and extracted data. We conducted individual semi-structured interviews with six cardiologists and six cardiac surgeons to elicit factors clinicians consider when they prescribe DAPT. We administered two online surveys for members of professional cardiology and cardiac surgery organisations. RESULTS:We screened 2,544 records, identified 322 eligible studies, and extracted data from 47. We identified 10 co-interventions and 70 potential confounders: review 31 (91%); interviews 19 (56%); surveys 31 (91%). 16/34 (47%) were identified by all three methods while, 3/34 (9%) were picked up by one method only. CONCLUSION:The review identified the majority of factors, but the interviews identified hard-to-measure factors such as perceived patient adherence and local prescribing culture. The methods could, in principle, be widely applied when designing or reviewing non-randomized studies of interventions (NRSI).