A Method for Selecting Communities for Efficient Equity-Informed Radiology Quality Improvement Interventions Applied...

Determine if recommendation for additional imaging (RAI) adherence is heterogeneous within similar socioeconomic strata and develop a more resource- and impact-efficient method for targeting community-based radiology quality interventions than those based on socioeconomic strata alone.

Secondary analysis of RAI from a closed-loop communication system at a large academic medical center from a study conducted 10/21/2019-6/30/2021. A potential adherence improvement index (PAII: non-adherence rate multiplied by total RAI without adherence) was developed to reflect potential intervention impact on a given community. PAII, RAI adherence, area deprivation index (ADI), and ADI variance were compared between the highest PAII and ADI quintiles. Geospatial mapping and multifactorial plots were generated to visualize community distributions including the intersections between various PAII and ADI thresholds.

545 communities with 2 or more RAI were included. The highest PAII quintile communities were spread across geographic regions and ADI quintiles with only 20% (22 of 109) of these communities in the highest ADI quintile. Maps and plots showed optimal target communities for a variety of resource thresholds. Median ADI (higher=more socioeconomic disadvantage) was 16 (IQR 9-21) in highest PAII quintile vs. 29.5 (26-33.8) in highest ADI quintile (p<.001). RAI adherence was 41%±13% vs. 74%±25% (p<.001), respectively. Mean PAII (higher=more potential impact) was 1.53±0.61 vs. 0.49±0.69 (p<0.001), respectively. ADI variance was higher in the highest PAII quintile vs. highest ADI quintile (F=2.12, p<0.001).

Equity-informed radiology quality interventions can likely be more resource- and impact-efficient if composite socioeconomic criteria are combined with additional community-level exposure information to focus interventions on high-risk communities.