We often attribute variations in medical care and outcomes to system, hospital, or patient factors. However, in their March 2020 BMJ Quality and Safety paper titled "Physician-level variation in clinical outcomes and resource use in inpatient general internal medicine: an observational study”, Verma et al. found that perhaps some differences in outcomes could be attributed to physician-level variations.
The authors conducted an observational study of ~100k GIM admissions to 7 hospitals in Canada over a 5 year period, evaluating outcomes such as inpatient mortality, LOS, 30d readmission, and diagnostic imaging. They divided the 135 attending physicians into 4 quartiles based on performance, and found that after propensity score matching the patients (ie, controlling for patient-level differences), there were statistically significant physician-level differences in LOS (1.2 days between highest and lowest quartile) and imaging use (0.32 tests per admission). They concluded that these differences may represent practice variation that highlight QI opportunities.
The Dashboard group is currently conducting a similar study, and taking it one step further to identify characteristics that may be associated with being in a high- or low-performing quartile (such as years as an attending or relative time spent on direct-care service), which could help guide help guide individualized feedback and coaching.
-Saj Patel
Verma AA, Guo Y, Jung HY, et al. BMJ Qual Saf 2021;30:123–132.