Successful implementation of stroke risk screening for sickle cell anemia in the DISPLACE study: results of a cluster randomized trial

Background: Stroke risk screening using transcranial Doppler (TCD) is a critical evidence-based tool for children with sickle cell anemia (SCA) that has been poorly implemented in the United States. The Dissemination and Implementation of Stroke Prevention Looking at the Care Environment (DISPLACE) study was designed to improve rates of stroke risk screening for SCA using interventions informed by an extensive multi-level barriers and facilitators assessment. This report describes the final outcomes of a large, randomized implementation trial comparing two intervention arms: 1) an application designed to track TCD implementation, ProviderMinder™, versus 2) ProviderMinder™ plus a single coordinator intervention. All sites additionally received a rebranding and educational intervention. The primary outcome was the difference in stroke risk screening rates between intervention arms. The intervention group was compared to four sites that did not implement either intervention and to their baseline rates as secondary outcomes. Methods: The initial part of DISPLACE included 28 sites from which 16 sites with poor stroke risk screening implementation were included in the trial and randomized to intervention arms. All sites entered patient data into a secure, customized electronic database and were required to use ProviderMinder™ for stroke risk screening data entry. Three sites were unable to adopt ProviderMinder™ and a fourth site from the original DISPLACE cohort was added to this group, resulting in thirteen intervention sites and four non-implementing sites (NIS). NIS collected data retrospectively for the same period as the implementation trial. A generalized quasi-likelihood Poisson mixed effects regression model compared screening rates between groups and timepoints while controlling for baseline screening rates and site size. Unadjusted stroke risk screening rates were also compared via two-proportion Z-tests for all outcomes. Results: The intervention-by-timepoint interaction indicated statistically significant improvement for the ProviderMinder™ arm relative to the combined intervention arm (difference of 10.0%) and for the intervention group (both arms) compared to NIS (difference of 15.9%). Screening rates increased by 28.0% from baseline to intervention, with an overall rate of 76.8%. Conclusions: Our intervention approach in DISPLACE significantly improved stroke risk screening for children with SCA, with procedure-patient tracking emerging as an important component for improving care. Trial registration: Clinical trial number: ClinicalTrials.gov; NCT04173026; 6/4/2020; https://clinicaltrials.gov/study/NCT04173026?cond=NCT04173026&rank=1