Objective: We aimed to identify (1) differences in opioid poisoning mortality among population groups, (2) geographic clusters of opioid-related deaths over time, and (3) health conditions co-occurring with opioid-related death in Ohio by computational analysis. Materials and Methods: We used a large-scale Ohio vital statistic dataset from the Ohio Department of Health (ODH) and U.S. Census data from 2010-2016. We surveyed population differences with demographic profiling and use of relative proportions, conducted spatiotemporal pattern analysis with spatial autocorrelation via Moran statistics at the census tract level, and performed comorbidity analysis using frequent itemset mining and association rule mining. Results: Our analyses found higher rates of opioid-related death in people aged 25-54, whites, and males. We also found that opioid-related deaths in Ohio became more spatially concentrated during 2010-2016, and tended to be most clustered around Cleveland, Columbus and Cincinnati. Drug abuse, anxiety and cardiovascular disease were found to predict opioid-related death. Discussion: Comprehensive data-driven spatiotemporal analysis of opioid-related deaths provides essential identification of demographic, geographic and health factors related to opioid abuse. Future research should access personal health information for more detailed comorbidity analysis, as well as expand spatiotemporal models for real-time use. Conclusion: Computational analyses revealed demographic differences in opioid poisoning, changing regional patterns of opioid-related deaths, and health conditions co-occurring with opioid overdose for Ohio from 2010-2016, providing essential knowledge for both government officials and caregivers to establish policies and strategies to best combat the opioid epidemic.
ASJC Scopus subject areas