Characterizing the geographic variation and risk factors of fatal prescription opioid poisoning in New Hampshire, 2003–2007

Abstract

The state of New Hampshire has one of the highest national rates of fatal prescription opioid poisoning in the United States. The study objective was to identify spatial patterns and potential predictors of New Hampshire's fatal prescription opioid poisonings using a straightforward spatial methodology that could be adopted by public health practitioners. Death certificate data were collected from fatal prescription opioid poisoning cases in New Hampshire from 2003 to 2007. Deaths were aggregated to corresponding five-digit ZIP Code Tabulation Areas (ZCTAs), and age–sex standardized mortality rates were calculated per ZCTA. The demographic and socioeconomic characteristics of cases and ZCTAs were evaluated. A geographic information system-based weighted-average method smoothed rates and accounted for the small-number problem. The mean center and standard distance of annual rates illustrated spatiotemporal trends. An ordinary least squares regression evaluated the association between Gaussian-transformed rates and spatial exposure covariates. Due to remaining residual spatial autocorrelation, a simultaneous autoregressive (SAR) spatial error model was used to investigate the association. From 2003 to 2007, 451 unintentional fatal prescription opioid poisonings occurred. Cases varied by sex, age, marital status, and employment status. The smoothed rate map indicates that high fatal poisoning rates concentrated in the state's eastern-central region. The mean centers and standard distance circles of the annual rates indicate stable smoothed rates over the analysis period. SAR model results suggest that the mean transformed rate varied significantly with a ZCTA's median household income and percent with employment-related disability. These results can be used to develop a policy addressing the soaring rate of prescription opioid poisonings in the United States.