A geographical analysis of ambulatory care sensitive conditions among racial/ethnic groups within California counties

Preventable Hospitalizations

The timely and effective use of primary care can prevent the need for hospitalization. For example, properly managed diabetes can reduce the risk of amputations, seizures, and shock. Hospital admissions that may be preventable are termed “ambulatory care sensitive conditions” (ACSCs). There is evidence that people with greater access to outpatient care are less likely to be hospitalized for ACSCs.

Ecological analyses of ACSCs at the county or city-level often demonstrate disparities between racial/ethnic groups. To facilitate such analyses, the Agency for Healthcare Research and Quality (AHRQ) provides free software to generate Prevention Quality Indicators (PQIs)—a set of 14 indicators created from hospital inpatient discharge data to evaluate the quality of ambulatory care in counties and metropolitan areas.

Objectives

Our objective was to geographically report racial/ethnic ACSC rates at the sub-county level because large California counties often encompass heterogeneous communities that differ with respect to racial/ethnic composition, socio-economic status, and access to quality healthcare. We achieved this goal by modifying AHRQ's software to produce the PQIs at the Medical Service Study Area (MSSA)—areas based on US Census tracts and created through legislative authority to evaluate sub-county areas lacking sufficient medical services. Our ultimate goal was to improve the geographic presentation of ACSC rates to government stakeholders.

Methods

We constructed PQI denominators by aggregating race-specific 2007 census tract data for each MSSA, and then modified AHRQ's PQI SAS statistical software code to facilitate the use of MSSA population denominators. Using California inpatient discharge data, we calculated PQI rates stratified by race and MSSA. We mapped the rates using ArcView software, and performed two sets of analyses. First, we performed regression and discriminant analyses to identify sociological correlates with MSSA-level ACSC rates. We used the results to identify underlying community characteristics associated with higher or lower PQI rates and to characterize each of the MSSAs. Second, we explored methods of spatially transforming zip code level patient data to MSSA aggregations to ensure statistical reliability, and to provide a closer mapping to real community-level forces.

Results

Racial/ethnic-specific rates for most of the PQIs vary across MSSAs. Higher rates indicating poor access to primary care are associated with areas of lower socio-economic status for one or more particular racial/ethnic groups. PQIs modified to use the Census-based MSSA geography are more relevant to illustrating the socio-economic disparities in healthcare because of finer spatial granularity and the linkage to Census demographics.