Comparison of statewide substance use disorder prevalence estimates from survey data vs electronic health record data

Prevalence estimates for substance use disorders (SUD) in the United States are primarily based on the National Survey on Drug Use and Health (NSDUH) which often lags, making real-time interventions difficult. Electronic health record (EHR)-derived estimates can provide near real-time estimates. We aimed to compare EHR-based state-level estimates with those derived from the NSDUH survey.

2023 EHR SUD prevalence data from the Minnesota Electronic Health Record Consortium (MNEHRC), a network of health systems across Minnesota, were compared to 2023 NSDUH prevalence data for Minnesota.

MNEHRC estimated alcohol use disorder (AUD) prevalence to be 3.80% while NSDUH reported 11.40%. MNEHRC prevalence for opioid use disorder (OUD) was 1.30% while NSDUH reported it as 1.76%. NSDUH reported a prevalence of 1.56% for cocaine use and 0.85% for methamphetamine use, while MNEHRC data found a prevalence of cocaine use of 0.32% and psychostimulant use of 0.83%. Age trends varied among the two methods. NSDUH estimated the highest substance use prevalence in the 18-25 age group with a decline in the > 26 group while MNEHRC showed an increasing prevalence with age for all substances.

Estimates of OUD and methamphetamine use prevalence were similar across the two approaches, whereas those for AUD and cocaine use differed. These findings suggest that EHR data may capture some substance use conditions accurately enough to proxy population level prevalence.