Medical marijuana laws were associated with modest reductions in opioid prescribing among veterans with chronic pain, according to research published September 28 in the Journal of General Internal Medicine. Led by Zachary L. Mannes, the study adds evidence to an important policy debate: whether expanding legal access to cannabis changes how patients and clinicians manage persistent pain. Its findings suggest that prescribing patterns can shift, although the implications for individual patients require closer examination.
For veterans seeking relief, the distinction between fewer prescriptions and better treatment is consequential. Reducing medication exposure can be valuable when pain remains controlled and daily functioning improves. A prescription decline, however, cannot independently establish either outcome. The new research is best understood as evidence about treatment patterns within a changing legal environment. Evaluating its clinical significance requires considering pain relief, medication safety, substance use disorders and the quality of care available when treatment changes.
How Researchers Examined Cannabis Laws and Prescribing
The researchers analyzed Veterans Health Administration records from 2013 through 2022 for patients aged 18–75 with chronic pain. They used a staggered-adoption difference-in-differences approach to compare prescribing changes around state cannabis-law enactment. This is a policy study rather than a trial assigning patients cannabis or a placebo. Its exposure was the legal environment in which patients lived, making it important to distinguish access to marijuana from documented use of a particular product.
In general, difference-in-differences methods ask whether an outcome changed differently in a population affected by a policy than in a comparison population. That is more informative than simply observing that prescriptions fell after a law passed: prescribing might also have declined without legalization. Nevertheless, interpreting such estimates causally depends on assumptions about how the groups would otherwise have changed. Other policies, shifts in clinical practice and differences between states can complicate that interpretation. A useful reading therefore separates the estimated association from possible explanations for it, including substitution, changing patient preferences or clinician decisions.
The Reductions Were Modest and Varied Across Measures
Medical laws were associated with a 0.79-percentage-point reduction in receipt of at least 30 days of opioids and a 0.34-point reduction in long-term therapy. Higher-dose prescribing and opioid–benzodiazepine co-prescribing declined less than in comparison states. After recreational legalization, relative to medical-only laws, reductions were 0.14 points for the 30-day measure and 0.33 points each for higher-dose prescribing and co-prescribing. Reductions were generally more consistent at ages 65–75.
Percentage points describe an absolute difference, which should not be confused with a relative percentage reduction. A 0.79-point difference corresponds mathematically to approximately eight people per 1,000 meeting a prescribing measure, rather than a 79% reduction or eight prescriptions eliminated. The mixed direction across measures also matters: an overall reduction in prescribing does not ensure that every higher-risk pattern improves at the same pace. For evaluating policy, those distinctions prevent a modest finding from becoming an exaggerated claim about widespread opioid replacement.
Why Fewer Prescriptions Do Not Automatically Mean Better Pain Care
Pain treatment is judged by what patients can do and how they feel, alongside medication-related risks. A person who reduces opioid use while maintaining mobility, sleep and independence may experience a meaningful improvement. Someone who loses access to medication while remaining in severe pain presents a very different clinical picture. Both situations could contribute to a declining prescription count. Assessing the consequences of a policy therefore requires outcomes that follow patients beyond the pharmacy record, including whether they receive effective alternatives and remain engaged in care.
This principle is reflected in the CDC’s 2022 opioid-prescribing guideline, which emphasizes individualized decisions and weighing benefits against risks. It advises against abruptly discontinuing opioids or rapidly reducing doses in most circumstances, and supports collaborative, gradual tapering when a change is appropriate. It also recognizes the importance of nonopioid medications and nonpharmacological treatments. Applied to the cannabis debate, that guidance means a population-level prescribing association should not become a blanket instruction to switch treatments. The relevant question is whether a particular change improves a patient’s overall condition safely.
Clinical Trials Support Some Pain Benefits, With Important Limits
There is a clinical basis for investigating cannabis as part of pain management. A 2021 systematic review led by Li Wang in The BMJ examined 32 randomized trials involving 5,174 adults. It found small to very small improvements in pain relief, physical functioning and sleep with non-inhaled medical cannabis or cannabinoids compared with placebo. Treatment also increased several adverse effects, including dizziness and drowsiness. These findings support the possibility of benefit for some patients while placing realistic boundaries around its average size. They do not establish that every formulation, route of administration or pain condition responds similarly.
A 2024 review by Haron M. Jeddi and colleagues in BMJ Open compared cannabis and opioids through a network meta-analysis of randomized trials. Low-certainty evidence suggested little difference in pain relief, while other findings favored cannabis for treatment discontinuations caused by adverse effects. However, most comparisons were indirect: only one trial directly compared the two approaches. The distinction matters because similar results across separate placebo-controlled trials do not establish that a patient taking opioids can switch successfully to cannabis. Such evidence helps frame treatment questions, but it cannot supply a universal substitution protocol.
Evidence for Cannabis Helping Patients Reduce Opioids Remains Uncertain
The more specific question—whether adding cannabis allows patients to reduce opioids—has proved difficult to answer. A 2021 BMJ Open review led by Atefeh Noori included five randomized trials and 12 observational studies. The randomized trials enrolled patients with cancer pain and instructed them to maintain their opioid doses, limiting their usefulness for testing opioid reduction. Observational studies suggested a possible reduction, but the certainty of that evidence was very low. The review concluded that opioid-sparing effects remained uncertain. This illustrates why a study’s design must match the clinical claim being made.
Individual-level research has also produced less encouraging findings. Gabrielle Campbell and colleagues followed 1,514 Australians with chronic non-cancer pain who had been prescribed opioids, reporting their results in The Lancet Public Health in 2018. Over four years, they found no evidence that cannabis use reduced prescribed opioid use or increased opioid discontinuation. Because this was observational research, it did not settle the effectiveness of a standardized, supervised cannabis intervention. It nevertheless challenges the assumption that choosing cannabis for pain reliably leads to reduced opioid treatment, and demonstrates why policy associations and individual treatment outcomes must be evaluated separately.
Earlier Legalization Studies Have Reached Different Conclusions
The new veterans findings join a literature in which results depend partly on the population, period and outcome examined. In a 2018 JAMA Internal Medicine study, Hefei Wen and Jason Hockenberry analyzed Medicaid prescribing data from 2011 through 2016. Medical marijuana laws were associated with approximately 5.9% lower opioid prescribing rates, while adult-use laws were associated with a further reduction of approximately 6.4%. Those results strengthened interest in cannabis access as a potential influence on prescribing. They concerned prescription rates among Medicaid enrollees, however, rather than the same patient-level measures used in the veterans analysis.
A broader 2024 study by Hai V. Nguyen and colleagues in JAMA Health Forum reached a different conclusion. Using state-level data from 2006 through 2020, it found no statistically significant overall association between medical or recreational cannabis law implementation and opioid prescribing or total opioid-overdose mortality. It did identify a possible reduction in synthetic-opioid deaths associated with recreational laws. These differing findings discourage treating legalization as a uniform intervention with a fixed effect everywhere. A law’s implementation, the population studied and the statistical handling of overlapping policy changes can all influence what an analysis detects.
Veterans Research Also Raises Questions About Substance Use Disorders
Especially relevant is a separate 2025 study by Mannes and colleagues in JAMA Health Forum. Examining VHA records from 2005 through 2022, that analysis associated medical and recreational cannabis laws with greater diagnosed opioid use disorder prevalence, with more pronounced increases among certain older groups and patients with chronic pain. It did not support legalization as a means of reducing the burden of opioid use disorder. As with other observational findings, the association should not be converted into proof that cannabis caused individual patients to develop the condition.
A 2023 study led by Deborah Hasin in The Lancet Psychiatry also found that cannabis-law enactment was associated with larger increases in diagnosed cannabis use disorder among VHA patients with chronic pain than among those without it. Together, these studies show why prescription volume cannot stand in for the entire safety picture. It is possible for a population to receive fewer opioid prescriptions while experiencing other substance-related problems. Evaluating expanded cannabis access therefore requires tracking benefits and harms simultaneously, including whether patients develop impaired control over cannabis use or continue using despite adverse consequences.
What the Findings Mean for Veterans and Future Research
The practical implication is to make treatment discussions more specific. For a veteran considering cannabis, useful questions concern the type of pain, current medicines, treatment goals and how improvement will be measured. Goals might include walking farther, sleeping more consistently or participating in rehabilitation. Tracking those outcomes alongside medication use would help distinguish meaningful relief from a change in prescriptions alone. This is an implication of the evidence as a whole, rather than a treatment regimen tested by the new study. It also avoids making opioid reduction the only definition of success.
Future research can move the debate forward by connecting legal access, actual cannabis exposure and patient outcomes. Studies should document product composition, frequency of use, opioid dose changes, pain interference and adverse events, while following patients long enough to assess whether benefits persist. Trials designed to permit carefully supervised medication adjustments would address a question that fixed-dose studies cannot. The new findings make that work more relevant: prescribing may change when cannabis becomes legally accessible, but the clinical value of those changes depends on whether veterans ultimately experience safer, more effective and more sustainable pain care.
