Improving quality and outcomes of stroke care in hospitals: Protocol and statistical analysis plan for the Stroke123 implementation study.

Rationale The effectiveness of clinician-focused interventions to improve stroke care is uncertain. Aims To determine whether an organizational intervention can improve the quality of stroke care over usual care. Sample size estimates To detect an absolute 10% difference in overall performance (composite outcome), a minimum of 21 hospitals and 843 patients per group was determined. Methods and design Before and after controlled design in hospitals in Queensland, Australia. Intervention Externally facilitated program (StrokeLink) using outreach workshops incorporating clinical performance feedback, patient outcomes (survival, quality of life at 90-180 days), local barrier assessments to best practice care, action planning, and ongoing support. Descriptive and multivariable analyses adjusted for patient correlations by hospital (intention-to-treat method). Context Concurrent implementation of financial incentives to increase stroke unit access and use of the Australian Stroke Clinical Registry for performance monitoring. Study outcome(s) Primary outcome: net change in composite score (i.e. total number of process indicators achieved divided by the sum of eligible indicators for each cohort). SECONDARY OUTCOMES: change in individual indicators, change in composite score comparing hospitals that did or did not develop action plans (per-protocol analysis), impact on 90-180-day health outcomes. Sensitivity analyses: hospital self-rated status, alternate cross-sectional audit data (Stroke Foundation). To account for temporal effects, comparison of Queensland hospital performance relative to other Australian hospitals will also be undertaken. Discussion Twenty-one hospitals were recruited; however, one was unable to participate within the study time frame. Workshops were held between 11 March 2014 and 7 November 2014. Data are ready for analysis.