Point process modeling of drug overdoses with heterogeneous and missing data

DOI10.1214/20-AOAS1384MaRDI QIDQ2233140zbMATH OpenOpenAlexFDO

Authors Xueying Liu, Jeremy Carter, Brad Ray, George O. Mohler

Publication date 14 October 2021

Published in The Annals of Applied Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2010.06080

zbMATH Keywords

expectation maximization algorithmnonnegative matrix factorizationopioid overdosepoint processsemisupervised learning

Mathematics Subject Classification ID

Point processes (e.g., Poisson, Cox, Hawkes processes) (60G55) Missing data (62D10) Applications of statistics to biology and medical sciences; meta analysis (62P10)

Abstract: Opioid overdose rates have increased in the United States over the past decade and reflect a major public health crisis. Modeling and prediction of drug and opioid hotspots, where a high percentage of events fall in a small percentage of space-time, could help better focus limited social and health services. In this work we present a spatial-temporal point process model for drug overdose clustering. The data input into the model comes from two heterogeneous sources: 1) high volume emergency medical calls for service (EMS) records containing location and time, but no information on the type of non-fatal overdose and 2) fatal overdose toxicology reports from the coroner containing location and high-dimensional information from the toxicology screen on the drugs present at the time of death. We first use non-negative matrix factorization to cluster toxicology reports into drug overdose categories and we then develop an EM algorithm for integrating the two heterogeneous data sets, where the mark corresponding to overdose category is inferred for the EMS data and the high volume EMS data is used to more accurately predict drug overdose death hotspots. We apply the algorithm to drug overdose data from Indianapolis, showing that the point process defined on the integrated data outperforms point processes that use only homogeneous EMS (AUC improvement .72 to .8) or coroner data (AUC improvement .81 to .85).We also investigate the extent to which overdoses are contagious, as a function of the type of overdose, while controlling for exogenous fluctuations in the background rate that might also contribute to clustering. We find that drug and opioid overdose deaths exhibit significant excitation, with branching ratio ranging from .72 to .98.

Recommendations

• A multivariate spatiotemporal change-point model of opioid overdose deaths in Ohio
• Estimating the burden of the opioid epidemic for adults and adolescents in Ohio counties
• Modeling and estimation of multi-source clustering in crime and security data
• Continuous Time-Interaction Processes for Population Size Estimation, with an Application to Drug Dealing in Italy
• Line and point cluster models for spatial health data

Cites work

• Estimation of Space–Time Branching Process Models in Seismology Using an EM–Type Algorithm
• Modeling and estimation of multi-source clustering in crime and security data
• Self-exciting point process modeling of crime
• Stochastic Declustering of Space-Time Earthquake Occurrences

Cited in

(5)

• A multivariate spatiotemporal change-point model of opioid overdose deaths in Ohio
• Estimating the burden of the opioid epidemic for adults and adolescents in Ohio counties
• Spatio-temporal Hawkes point processes: a review
• A modified two-process Knox test for investigating the relationship between law enforcement opioid seizures and overdoses
• Spatiotemporal-network point processes for modeling crime events with landmarks

Describes a project that uses

Uses Software

• SEKS-GUI

This page was built for publication: Point process modeling of drug overdoses with heterogeneous and missing data

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2233140)