Dissertations, Theses, and Capstone Projects

Date of Degree


Document Type


Degree Name



Public Health


Nancy Sohler

Committee Members

Katarzyna Wyka

Glen Johnson

Alexander Walley

Subject Categories



heroin; mortality; naloxone; overdose; poisoning; prevention


Background: Drug overdose mortality is the leading cause of injury death in both the United States (US) and New York City (NYC). Heroin-related overdoses make up the majority of overdoses in NYC. Since 2006, when a law was passed that allowed for layperson administration of naloxone, an opioid antagonist, heroin-related overdose deaths have decreased in NYC. No studies to date have investigated a possible association between the implementation of this intervention and heroin-related overdose mortality.

Objectives: To investigate the possible association between overdose prevention programs (OPPs) and heroin-related overdose mortality in NYC, using interrupted time series and geospatial analytic techniques.

Methods: Using surveillance of NYC accidental drug poisoning deaths (2000-2012), a demographic profile of heroin-related overdose deaths was described prior to implementation of OPP (January 2000 - June 2006) and after implementation (July 2006 - December 2012). Interrupted time series (ITS) analyses tested for a difference in level and trend of heroin-related mortality, comparing the post-OPP period with the pre-OPP period, for NYC as a whole. Geospatial patterns of heroin-related overdose mortality were described before and after implementation of OPP. After mapping OPP sites, NYC neighborhoods were stratified by naloxone penetration level, and using multivariable regression, we tested the hypothesis that neighborhoods with greater naloxone penetration experienced steeper declines in heroin-related overdose mortality, after controlling for neighborhood characteristics. We calculated street walking distance from the OPP to the location of each overdose fatality in one neighborhood, the Lower East Side of Manhattan, to test the hypothesis that risk increases with increasing distance. We mapped overdose rate by census tract and conducted Poisson regression.

Results: 2,142 heroin-related overdose deaths occurred in the 6.5 years prior to implementation of NYC's OPP, and 1,764 occurred in the 6.5 years after implementation, representing a 22.4% reduction in the age-adjusted mortality rate. We found, using ITS, that the level of heroin-related overdose mortality decreased by 16% (not statistically significantly different from no decrease) following implementation. When analyses were limited to only those parts of NYC with OPP, we found that neighborhoods with greatest OPP penetration saw greater decreases in overdose mortality rates, compared to neighborhoods without OPPs (-3.1 compared with -0.8). In the Lower East Side, we found that census tracts located furthest from the OPP had statistically significantly higher overdose mortality rates compared with census tracts closer to the OPP. The census tract where the OPP was located experienced the greatest decrease in heroin-related overdose death from pre-OPP to post (from 7.8 to 1.31 per 100,000 population). An individual is 1.22 times more likely to die from a heroin overdose for every 1,000 feet away from the OPP (p=0.0002).

Conclusions: While time series analysis of NYC as a whole did not find a statistically significant change in the level of heroin-related overdose mortality after implementation of OPP, when OPP locations were geocoded, only one-third of NYC neighborhoods had any OPP in the six years following implementation. In analyses limited to those neighborhoods of NYC with OPPs, we found statistically significant associations between OPP and heroin-related overdose mortality risk. This suggests that OPP may be contributing to decreased heroin-related overdose mortality in NYC.

Included in

Epidemiology Commons