rethinkTHC Search
Menu
Study breakdown

Denver researchers created a 95.7% accurate system for tracking marijuana-related ER visits in real time

Cross SectionalModerate evidence
The takeaway

A syndromic surveillance definition using 6 keywords and 5 diagnosis codes identified marijuana-related ER visits with 95.7% accuracy in Denver hospitals, providing a scalable method for real-time public health monitoring.

Read this if you're interested in how public health systems track the impact of marijuana legalization.

95.7% accuracy for identifying marijuana-related ER visits in Denver hospitals

What the researchers found

Researchers developed and validated a case definition for identifying marijuana-related ER visits using structured and free-text data from 15 Denver hospitals. The initial definition achieved 92.7% positive predictive value (PPV) in a validation period of 126,646 ER visits.

After refinement, the final case definition achieved 95.7% PPV in a second validation period of 140,932 visits. The final definition contained just 6 keywords for marijuana or derivatives and 5 diagnosis codes for cannabis use, abuse, dependence, poisoning, and lung disease.

This tool enables near-real-time monitoring of marijuana-related ER visits, which is critical for tracking the public health impact of legalization.

Why it matters

As marijuana policies change rapidly, public health systems need reliable, efficient ways to monitor health impacts. This tool demonstrates that existing ER data systems can be leveraged for near-real-time marijuana surveillance with high accuracy, without requiring new data collection infrastructure.

The numbers in context

Round 1: 524 matches from 126,646 visits, PPV 92.7%. Round 2: 698 matches from 140,932 visits, PPV 95.7%. Final definition: 6 keywords + 5 diagnosis codes. 15 Denver hospitals.

How the study worked

Applied a syndromic case definition to BioSense 2.0 data from 15 Denver hospitals. Two rounds of validation with manual record review to determine true and false positives. Iterative refinement between rounds improved the PPV from 92.7% to 95.7%.

What this study cannot tell us

Validated in one metropolitan area (Denver). The PPV measures accuracy of identified cases but does not capture sensitivity (cases missed). Free-text data quality varies across hospitals. The system identifies "marijuana-related" visits, not necessarily visits caused by marijuana.

How to read the evidence

Validated surveillance methodology with two rounds of testing. Demonstrates high accuracy but validated in a single metro area.

When this study was published

Published in 2017. Syndromic surveillance of cannabis-related health events has been adopted by additional jurisdictions since.

The bigger picture

The ability to monitor marijuana-related health events in near-real time is essential for evidence-based policy. Without reliable surveillance, debates about legalization's health impact rely on anecdote and outdated data. This methodology can be adopted by any jurisdiction with syndromic surveillance infrastructure.

Questions still open

  • What is the sensitivity of this definition (how many marijuana-related visits does it miss)? Would the same keywords and codes work in other cities? Can this approach distinguish between recreational and medical cannabis-related visits?

Common questions

Why do we need to track marijuana ER visits?
As marijuana policies change, public health systems need data on whether legalization increases health-related emergencies. Real-time tracking allows officials to identify trends, emerging problems (like edible overconsumption), and vulnerable populations.
How does the system work?
It searches ER records for specific marijuana-related keywords and diagnosis codes. When a match is found, it flags that visit as potentially marijuana-related. Manual review confirmed the system correctly identifies relevant visits about 96% of the time.

Read the original research

Validation of a Syndromic Case Definition for Detecting Emergency Department Visits Potentially Related to Marijuana.

Public health reports (Washington, D.C. : 1974), 132(4), 471-479

Citation

DeYoung, Kathryn; Chen, Yushiuan; Beum, Robert; Askenazi, Michele; Zimmerman, Cali; Davidson, Arthur J. (2017). Validation of a Syndromic Case Definition for Detecting Emergency Department Visits Potentially Related to Marijuana.. Public health reports (Washington, D.C. : 1974), 132(4), 471-479. https://doi.org/10.1177/0033354917708987

Explore the wider topic