Estimating Weekly National Opioid Overdose Deaths in Near Real Time Using Multiple Proxy Data Sources.

ABSTRACT:

Importance

Opioid overdose is a leading public health problem in the United States; however, national data on overdose deaths are delayed by several months or more.

Objectives

To build and validate a statistical model for estimating national opioid overdose deaths in near real time.

Design, setting, and participants

In this cross-sectional study, signals from 5 overdose-related, proxy data sources encompassing health, law enforcement, and online data from 2014 to 2019 in the US were combined using a LASSO (least absolute shrinkage and selection operator) regression model, and weekly predictions of opioid overdose deaths were made for 2018 and 2019 to validate model performance. Results were also compared with those from a baseline SARIMA (seasonal autoregressive integ

SUBMITTER: Sumner SA 

PROVIDER: S-EPMC9305381 | biostudies-literature | 2022 Jul

REPOSITORIES: biostudies-literature