Data feedback for quality improvement of stroke care - CAPTURE stroke experience

Background: Feedback on "quality indicators" collected from chart audit is a widely used quality-improvement strategy. Skeptics argue that temporal change in practice patterns or improvement of documentation as a result of implementation of chart audit can affect change in quality indicators rather than change in practice due to feedback. This study compared the effectiveness in quality improvement for stroke care of chart audit and feedback versus chart audit only. Methods: Data were examined on quality indicators constructed from the performance measures abstracted from chart audit for 1953 stroke patients admitted to 13 hospitals. Seven of the 13 hospitals were provided with feedback on quality indicators from baseline audit, and 6 hospitals had chart audit only. Results: Least-squares estimation was used to estimate the adjusted mean change in quality indicators between follow-up and baseline audits and differences in mean change between feedback and nonfeedback groups. Differences in adjusted mean change (with 90% confidence interval) were: deep vein thrombosis prophylaxis, 8.46 (-9.46 to 26.38); dysphagia screening, -3.78 (-21.37 to 13.81); antithrombotic therapy within 48 hours of hospitalization, 3.63 (-6.59 to 13.84); discharged on antithrombotic, -0.31 (-6.26 to 5.63); patients with atrial fibrillation discharged on warfarin/coumadin, 44.73 (-13.14 to 102.60); lipid screening, 19.93 (2.99 to 36.86); and smoking counseling, 17.47 (-12.13 to 47.08). Conclusions: Although not statistically significant, results suggest a potential for improvement with data feedback. There is a need for evaluation of the effectiveness of a multifaceted approach in a community setting.