Area Based Socioeconomic Analysis for 2001 and 2010 Pneumonia Hospitalizations in New Mexico (CSTE Health Disparities Pilot Project)

BACKGROUND:  The Harvard Gecoding Data Analytic Method developed by Nancy Krieger et al. is a technique of applying area based socioeconomic measures to case level data.  The New Mexico Department of Health (NMDOH) is one of the eleven sites selected by the Council for State and Territorial Epidemiologists (CSTE) Health Disparities Project to implement a pilot project using Harvard’s methods.  The NMDOH will be using Hospital Inpatient Discharge Data (HIDD) to assess socioeconomic difference in pneumonia hospitalization rates during 2001 and 2010. 

METHODS:  Addresses of patients hospitalized for pneumonia in 2001 and 2010 will be assessed for completeness and quality using the program Semaphore ZP4.  Addresses will be geocoded to the US Census tract using a protocol developed by the NMDOH Geocoding/Small Area Analysis Workgroup.  Methods include evaluation for quality, standardization of addresses, manual review and correction.  Analysis techniques defined by the Harvard Analytic Method will be applied to this geocoded data set.  This method includes the aggregation of case data into numerators by Census tract, assigning population denominator data from the US Census into age cells within Census tracts, assigning percent below area-level poverty indicator categories to numerators and denominators by Census tract, creating percent below poverty level categories and case and population counts for each strata and age category, and finally creating age-standardized pneumonia rates for each percent below poverty level. These rates will then be compared to those calculated by race and ethnicity.

RESULTS:  Rates by socioeconomic measures and racial and ethnic categories will be compared to determine differences in hospitalization rates for pneumonia. 

CONCLUSIONS:  This pilot project is a stepping stone towards more extensive area based socioeconomic analysis.  Next steps include applying these methods to additional years in the HIDD, other hospitalizations and other databases.