One of the largest real-world analyses of medical cannabis use among people with cancer has found substantial numbers of patients reporting meaningful improvements in pain, nausea, vomiting, disturbed sleep, appetite, anxiety and depression after beginning treatment. The Minnesota Office of Cannabis Management analyzed 6,621 cancer patients who purchased cannabis through the state’s medical program between July 2015 and December 2023, drawing on purchase records and symptom assessments completed before dispensary visits. Among patients who entered the program with moderate-to-severe symptoms, nearly 40 percent improved in nausea and disturbed sleep, almost half improved in vomiting, and roughly three in ten experienced clinically meaningful pain reduction within four months.
Published September 10, 2026, the report represents a major expansion of an earlier Minnesota cancer-cannabis study and provides an unusually large window into how patients use cannabis outside a tightly controlled clinical trial. The findings are encouraging, but they require an important qualification: this was an observational analysis without a placebo or non-cannabis control group. It therefore cannot prove that marijuana itself caused the improvements. Cancer treatment, changes in conventional medication, natural symptom fluctuations and selection effects could contribute. Even so, the scale of the dataset—and the consistency of improvement across several common cancer symptoms—makes the report an important addition to an evidence base that remains surprisingly limited despite widespread medical cannabis use.
Minnesota Followed Thousands of Cancer Patients Through Its Medical Cannabis Program
Minnesota is unusually well positioned to study real-world medical cannabis because its program systematically collects information from patients, health professionals and cannabis purchases. Between July 1, 2015, and December 31, 2023, 7,762 people enrolled with a cancer-related qualifying condition, and 6,621—or 85.3 percent—went on to purchase medical cannabis. The average patient was 58 years old, with ages ranging from one to 100. Men and women were represented almost equally. About 69 percent qualified because of severe or chronic cancer-related pain, 49 percent qualified for nausea or severe vomiting, and nearly 30 percent qualified for cancer-associated cachexia or severe wasting; patients could qualify under more than one category.
Before cannabis purchases, patients completed a standardized self-evaluation rating eight symptoms from zero to ten: anxiety, appetite loss, depression, disturbed sleep, fatigue, nausea, pain and vomiting. Researchers defined a 30 percent or greater decrease from baseline as clinically significant improvement, rather than counting small changes as treatment success. That threshold makes the results more meaningful than a simple survey asking whether patients “felt better,” although the measurements were still self-reported. Symptom analyses included 6,605 patients with sufficient baseline information.
Nearly One-Third of Patients With Significant Pain Reported Meaningful Relief
Pain was one of the most prevalent problems in the cohort. Approximately 82.7 percent of patients reported moderate-to-severe pain when entering the program. Within four months of their first cannabis purchase, 29.9 percent—1,631 patients—experienced at least a 30 percent reduction in pain. Among patients who improved and had enough continued follow-up data, 53.1 percent maintained that level of improvement for another four months. When all patients with moderate-to-severe baseline pain were used as the denominator, 12.1 percent both reached the 30 percent improvement threshold and sustained it for four additional months.
Those numbers are meaningful but do not settle the long-running debate over cannabis for cancer pain. Randomized trials have generally produced less impressive results than observational programs. The American Society of Clinical Oncology concluded in its evidence-based guideline that available controlled evidence remains insufficient to recommend cannabis or cannabinoids routinely for cancer pain. A small Minnesota randomized trial involving patients with stage IV cancer nevertheless found that adding medical cannabis to standard care was feasible and well tolerated, with signals suggesting improved pain control and reduced opioid requirements in the group receiving earlier access. The contrast illustrates why the new 6,621-patient report is useful: it shows what happens in routine practice, while randomized trials remain necessary to determine how much improvement is actually attributable to cannabis.
Nausea and Vomiting Produced Some of the Strongest Results
The most striking findings involved gastrointestinal symptoms. Among patients entering the program with moderate-to-severe nausea, 39.9 percent experienced at least a 30 percent improvement within four months. Of those with follow-up after improving, 65.4 percent maintained that response for another four months. Vomiting produced an even stronger signal: 47.2 percent of patients with moderate-to-severe vomiting improved within four months, and 74.4 percent of responders with follow-up maintained the improvement. Overall, about one quarter of patients who began with significant vomiting both improved and sustained that improvement for at least four months.
This finding is consistent with the area where cannabinoid medicine already has its strongest oncology evidence. Controlled trials have shown that cannabinoids can reduce refractory chemotherapy-induced nausea and vomiting in some patients. In a randomized, double-blind crossover trial published in Annals of Oncology, an oral preparation containing equal amounts of THC and CBD improved control of chemotherapy-induced nausea and vomiting when added to standard antiemetic treatment, although cannabinoid-related adverse effects were more common. ASCO now states that cannabis or cannabinoids may improve refractory chemotherapy-induced nausea and vomiting when added to guideline-recommended antiemetic therapy, while emphasizing that they should not replace established first-line antiemetics.
Sleep, Anxiety and Depression Also Improved for Many Patients
Disturbed sleep was almost as prevalent as pain. 82.8 percent of cancer patients reported moderate-to-severe sleep disturbance at baseline, and 38.5 percent reached the 30 percent improvement threshold within four months. Among responders who remained under observation, 60.5 percent maintained the improvement for four additional months. Anxiety improved in 40.1 percent of patients who began with moderate-to-severe symptoms, while depression improved in 43.6 percent. About 28 percent of patients with substantial fatigue also improved, although sustained improvement in fatigue was less common than for the other symptoms.
The sleep findings are particularly relevant because insomnia can amplify pain, fatigue, anxiety and overall distress during cancer treatment. Yet randomized evidence suggests cannabinoid effects on sleep may be modest when studied under controlled conditions. A systematic review of randomized trials found only a very small average improvement in sleep disturbance among patients with chronic cancer pain. ASCO consequently regards the evidence for most supportive-care outcomes other than refractory chemotherapy-related nausea and vomiting as uncertain. The Minnesota data therefore provide a strong reason for additional trials, rather than proof that cannabis should become a standard oncology sleep treatment.
Appetite Improved, but Cannabis Has Not Been Proven to Reverse Cancer Cachexia
Appetite loss was another major problem. About 71 percent of patients had moderate-to-severe lack of appetite when they entered the program, and 38.2 percent improved by at least 30 percent within four months. Among those who improved and remained in follow-up, 62.8 percent sustained that response. The effect is clinically interesting because loss of appetite and cachexia can seriously undermine strength, treatment tolerance and quality of life in advanced cancer.
However, improved appetite should not be confused with evidence that cannabis reverses cancer cachexia. The Minnesota report itself notes that previous systematic reviews have not established meaningful improvements in caloric intake, body weight or cachexia from cannabinoids. The program also did not record patients’ body-weight changes or detailed nutritional outcomes, so the new findings concern self-reported appetite symptoms, not reversal of wasting syndrome.
Cancer Patients Used a Wide Range of Products, Often With High THC Levels
The size of Minnesota’s purchase database provides another unusual feature of the study. Researchers analyzed 35,859 sales transactions involving 121,865 individual products purchased during patients’ first year in the program. Inhaled products—including vaporized cannabis and, after its introduction into the program, cannabis flower—accounted for 48 percent of purchases. Edible or other enteral products such as capsules, oral solutions and gummies accounted for another 42.5 percent, with smaller shares coming from oromucosal and topical formulations.
THC-predominant products were common. Among inhaled products that could be classified by THC-to-CBD ratio, 58.9 percent were high or very high in THC relative to CBD. Among swallowed products, that figure reached 82.4 percent. This diversity is both a strength and a weakness of the research. It reflects how patients actually consume cannabis, but it prevents the study from identifying one cannabinoid ratio, dose or delivery method responsible for the observed improvements. A patient using balanced THC:CBD capsules may have a very different pharmacological experience from someone using high-THC flower, yet both contribute to the same broad outcome analysis.
Most Reported Side Effects Were Mild, but Adverse Effects May Be Underestimated
Side effects were reported by 962 patients, or 14.5 percent of the cohort, generating 2,582 adverse-effect reports. Approximately 59.5 percent were rated mild, 34.4 percent moderate and 6.2 percent severe. Dry mouth was the most frequently reported specific side effect, followed by sedation or drowsiness and mental clouding or “foggy brain.” Among the 159 severe side-effect reports, fatigue was most common, followed by dry mouth, nausea and headache.
The researchers caution that the 14.5 percent figure likely underestimates adverse reactions. Patients completed the evaluation before making another purchase. Someone who experienced an unpleasant reaction and stopped using cannabis altogether might never return to complete another survey, meaning the event could disappear from this dataset. This is a particularly important limitation when interpreting the apparently low severe-side-effect rate. THC-containing products can cause dizziness, sedation, impaired concentration, anxiety and other effects that may be especially relevant for older or medically fragile cancer patients.
The Study’s Size Is Impressive, but It Cannot Prove Cannabis Caused the Improvements
The most important limitation is the absence of a control group. Every participant included in the principal analysis purchased cannabis. Researchers therefore could not compare symptom trajectories with otherwise similar cancer patients who did not use marijuana. The program also did not collect cancer type, cancer stage or treatment information, meaning researchers could not determine whether patients were receiving chemotherapy, immunotherapy, radiation, surgery, hospice care or other interventions while their symptoms changed.
Loss to follow-up was substantial as well. Roughly one quarter of the patients purchased cannabis only once, while cancer patients may stop participating because chemotherapy ended, their symptoms improved, cannabis did not work, the products were too expensive, access was difficult or their disease progressed. Those very different reasons can distort long-term response estimates. The report therefore provides strong real-world association data, but not randomized evidence of efficacy.
The Findings Strengthen the Case for Better Cancer-Cannabis Trials
The Minnesota findings arrive as the National Cancer Institute is trying to close exactly this evidence gap. In 2026, NCI researchers described a coordinated effort supporting five large prospective observational studies examining cannabis among cancer patients receiving active systemic treatment. Researchers plan to examine pain, nausea, sleep, anxiety, quality of life, treatment outcomes and potential harms across more diverse cancer populations. The initiative reflects an unusual situation in oncology: patient use has become common faster than high-quality research has been able to determine which cannabis preparations work, for whom and at what doses.
ASCO’s current position remains appropriately cautious. Cannabis should not be used as a treatment for cancer itself, and patients should not substitute it for chemotherapy, immunotherapy, surgery, radiation or other established cancer therapies. Evidence is strongest for cannabinoids as an adjunct in refractory chemotherapy-induced nausea and vomiting; for pain, sleep, appetite, anxiety and other symptoms, evidence remains incomplete. Clinicians are nevertheless encouraged to ask patients about cannabis use openly and without judgment because unreported use can affect symptom management, medication decisions and safety.
A Major Real-World Signal, Not the Final Word
The new Minnesota report is important because of its sheer scale. Across 6,621 medical cannabis patients with cancer, meaningful symptom improvements repeatedly appeared after treatment began: 29.9 percent improved in pain, 39.9 percent in nausea, 47.2 percent in vomiting, 38.5 percent in disturbed sleep, 38.2 percent in appetite loss, 40.1 percent in anxiety and 43.6 percent in depression. Many patients who improved and remained in the program maintained those gains for months.
Those results should neither be dismissed as anecdote nor promoted as proof that marijuana treats every cancer-related symptom. They sit in the middle: a large government dataset showing that thousands of real patients reported clinically meaningful changes, alongside randomized evidence that remains considerably more cautious. For cancer care, that may be the most useful interpretation. Medical cannabis appears capable of helping some patients manage difficult symptoms, but researchers still need controlled studies that identify which cannabinoids, doses and delivery methods actually provide the benefits—and which patients are most likely to experience them safely.
