Impact of racial residential segregation on “hotspots” of preventable hospitalizations in Maryland

Hospitalizations for ambulatory care-sensitive conditions (ACSCs) are considered preventable.  They serve as a proxy for a lack of access to primary care.  Preventable hospitalizations (PHs) are more prevalent among African American patients, in urban cores and rural areas.  However, few studies have empirically assessed the presence of geographic clusters of PHs, or “hotspots”.  Furthermore, few studies examine contextual effects on PH hotspots, particularly segregation which has been positively associated with a number of poorer health outcomes.  Segregation is thought to shape the availability of healthcare resources, a significant predictor of PHs.  It is important to identify areas with high rates of PHs so that public health practitioners can efficiently target resources.  Knowledge of PH hotspots allows researchers to characterize possible contextual and compositional effects on PHs.  This study seeks to assess the presence of PH hotspots in the State of Maryland by race.  Data for all hospitalizations in 2010 in Maryland were collected, including patient race/ethnicity, zip code of residence and primary diagnosis.  Hospitalizations for ACSCs were considered preventable.  The rate of PHs (per 100,000 population) for every zip code in Maryland was calculated.  Segregation was measured by the Dissimilarity Index and was calculated for each zip code.  The spatial scan statistic was used to identify contiguous groups of zip codes with higher than unadjusted PH rates and PH rates after adjusting for segregation.  The rate of total PHs in Maryland was 1,735.7 per 100,000.  There were a total of ten hotspots.  The largest included 27 zip codes with a radius of 30.5 kilometers.  The PH rate for hotspots ranged from 5,022.4 per 100,000 (RR: 2.90, p<0.001) to 2,170.5 per 100,000 (RR: 1.25, p<0.001). After adjusting for segregation, 15 distinct clusters were detected, but the size of the clusters decreased such that the majority of hotspots only included one zip code.  These results suggest that a social-contextual factor affects the clustering of high PH rates in Maryland.  Public health practitioners can utilize this work to target resources and possibly develop culturally relevant interventions.