The claim that medical cannabis laws reduce opioid deaths reversed when extended through 2017 | Research

The takeaway

Replicating a landmark 2014 study through 2017, the previously reported association between medical cannabis laws and reduced opioid deaths reversed from -21% to +23%, suggesting the original finding was likely spurious.

Policymakers, opioid crisis researchers, cannabis industry stakeholders, and public health scientists.

Association reversed from -21% to +23% with extended data

What the researchers found

Extending the Bachhuber et al. (2014) analysis from 2010 to 2017 using identical methods, the association between medical cannabis laws and opioid overdose mortality reversed from -21% to +23%. Neither recreational cannabis laws nor low-THC cannabis laws were associated with opioid mortality changes.

Why it matters

The original 2014 study was widely cited by the cannabis industry and advocates as evidence that medical cannabis reduces opioid deaths. This reversal demonstrates the danger of drawing causal conclusions from ecological correlations.

The numbers in context

Original finding (1999-2010): -21% opioid mortality. Extended finding (1999-2017): +23% opioid mortality. Medical cannabis used by about 2.5% of US population. Neither recreational nor low-THC laws showed association.

How the study worked

Replication and extension of Bachhuber et al. (2014) using the same state-level panel regression methods through 2017. Analyzed associations between medical cannabis law adoption and opioid analgesic overdose mortality rates.

What this study cannot tell us

Ecological (state-level) analysis cannot capture individual-level substitution. The reversal itself does not prove medical cannabis increases opioid deaths. The authors interpret the association in both directions as likely spurious.

How to read the evidence

Strong: direct replication using identical methods with extended data, published in PNAS.

When this study was published

Published in 2019.

The bigger picture

This is a cautionary tale about cherry-picking time periods in policy research. The opioid crisis has its own trajectory driven by factors far more powerful than cannabis access, making it implausible that cannabis policy explains large swings in overdose mortality.

Questions still open

Why did the association reverse? Is it because more states adopted cannabis laws during the most severe phase of the opioid crisis? Can individual-level studies better capture cannabis-opioid substitution effects?

Common questions

Does this mean medical cannabis increases opioid deaths?

No. The authors argue that neither the original negative association nor the reversed positive association is likely causal. With only 2.5% of the population using medical cannabis, it is implausible to have large effects on population-level opioid mortality.

Should we stop researching cannabis for pain?

No. Individual-level research on cannabis as an opioid alternative should continue. What should stop is the claim that state-level cannabis laws have been proven to reduce opioid overdose deaths.

Read the original research

Association between medical cannabis laws and opioid overdose mortality has reversed over time.

Proceedings of the National Academy of Sciences of the United States of America, 116(26), 12624-12626

View on PubMed ↗Publisher’s paper ↗

Citation

Shover, Chelsea L; Davis, Corey S; Gordon, Sanford C; Humphreys, Keith. (2019). Association between medical cannabis laws and opioid overdose mortality has reversed over time.. Proceedings of the National Academy of Sciences of the United States of America, 116(26), 12624-12626. https://doi.org/10.1073/pnas.1903434116