Only 25% of cannabis-coded ER visits were actually attributable to cannabis use | Research

The takeaway

A chart review of cannabis-coded ED visits found only 25.74% were actually at least partially attributable to cannabis, with these patients being more often young, Caucasian males.

Emergency medicine researchers, cannabis policy analysts, and health services researchers.

Only 25.74% of cannabis-coded ED visits were actually cannabis-related

What the researchers found

Of all ED visits with cannabis ICD-CM codes, detailed chart review determined only 25.74% were at least partially attributable to cannabis. These patients were more often young, Caucasian, and male compared to the overall ED population (all p values significant). The remaining ~75% had cannabis codes incidental to the visit reason.

Why it matters

Administrative data studies often use cannabis ICD codes to estimate cannabis-related ED burden. This study shows that 75% of such visits are not actually caused by cannabis, meaning published estimates of cannabis-related ED visits may be grossly inflated.

The numbers in context

25.74% actually attributable to cannabis; patients more often young, Caucasian, male (all p<0.05); ~75% had cannabis codes incidental to visit.

How the study worked

Retrospective chart review of ED visits identified by cannabis ICD-9 and ICD-10-CM codes between 2012 and 2016, with pre-specified attribution criteria and inter-rater reliability assessment.

What this study cannot tell us

Single hospital system; retrospective chart review with inherent subjectivity; 2012-2016 data may not reflect current patterns; attribution criteria were pre-specified but novel; cannot determine what fraction of the 25% were primarily vs secondarily caused by cannabis.

How to read the evidence

Moderate: systematic chart review with pre-specified criteria and inter-rater reliability, but single center.

When this study was published

Published 2020.

The bigger picture

This has major implications for cannabis policy debates. Claims about "dramatic increases in cannabis-related ER visits" based on ICD codes may be measuring coding practices rather than actual cannabis harms. Better attribution methods are needed.

Questions still open

How should cannabis-attributable ED visits be defined and measured? Are policy claims about cannabis ED burden based on inflated administrative data?

Common questions

How many ER visits are truly caused by cannabis?

This study found only about 26% of ED visits coded for cannabis were actually attributable to cannabis use. The rest had cannabis codes incidentally, meaning published statistics based on coding data may overestimate cannabis-related ER visits by about 4-fold.

Who goes to the ER for cannabis-related problems?

Among the visits truly attributable to cannabis, patients were more often young, Caucasian, and male compared to the overall ED population.

Read the original research

Why do patients come to the emergency department after using cannabis?

Clinical toxicology (Philadelphia, Pa.), 58(6), 453-459

View on PubMed ↗Publisher’s paper ↗

Citation

Shelton, Shelby K; Mills, Eleanor; Saben, Jessica L; Devivo, Michael; Williamson, Kayla; Abbott, Diana; Hall, Katelyn E; Monte, Andrew A. (2020). Why do patients come to the emergency department after using cannabis?. Clinical toxicology (Philadelphia, Pa.), 58(6), 453-459. https://doi.org/10.1080/15563650.2019.1657582