Among cancer patients in this study, we observed an improvement in symptoms of anxiety and depression for individuals who remained in the MMCP for more than 30 days and purchased enteral cannabis products. Overall, the THC dose was not strongly associated with anxiety or depression, while greater CBD doses were associated with greater improvements in anxiety but not depression scores.

This study suggests that the dose of CBD, the ratio of THC to CBD, and the route of administration of cannabis may all influence self‐reported anxiety in cancer survivors. Specifically, our study identified a positive association between CBD and reduced anxiety, consistent with existing literature [18]. The anxiolytic effects of cannabis products were more pronounced with enteral products, which had higher CBD levels than any other route of administration in this study. Previous studies with inverse U‐shaped dose–response curves have concluded that 300 mg may be the most efficacious CBD dose for anxiety [33]. Additional studies are needed to confirm whether providers may want to encourage cancer survivors to try cannabis products with relatively high CBD relative to THC doses and the use of enteral products for anxiety relief. Additionally, on the ESAS scales, a 1‐point improvement is considered clinically significant for anxiety [30]. So, while this study found a significant relationship between CBD and anxiety reduction, clinical significance may be modest.

Interestingly, THC dose was not significantly associated with a change in anxiety for the patients in our study. This contradicts clinical studies that have suggested that THC may have anxiogenic effects, especially at high doses [34]. Several factors may have contributed to these findings. For instance, cancer‐associated differences in the endocannabinoid system may have reduced THC‐related changes in anxiety [35]. Furthermore, prolonged use of THC may lead to the development of tolerance, potentially resulting in diminished effects on anxiety in regular users [36].

Our findings suggest that enteral products play an important role in improving feelings of depression. A key feature of enteral products is their ability to produce sustained effects over a longer period of time relative to other routes of administration; the duration of effects of enteral products is around 4–12 h [37]. The maximum effect of enteral products may be less predictable due to delayed onset [38]. Enteral products also can be more intense than other routes due to the conversion of THC to its more potent active metabolite, 11‐hydroxy‐tetrahydrocannabinol (11‐hydroxy‐THC), in the liver [39]. The use of topical products in addition to enteral products may have resulted in greater depressive symptom reduction due to the possible soothing effects associated with massaging a topical product into the skin [40, 41].

Existing literature supports the potential for CBD to relieve depressive symptoms, an effect seen in this study [42]. Conversely, THC is thought to reduce CB1 receptor activity over time, leading to diminished therapeutic impacts and potentially contributing to depression [43]. For this study, we only evaluated symptom scores after day 30 of MMCP enrollment. Any initial changes in depression related to THC may have dissipated before the 30‐day time point.

The implications of these findings for clinical practice are to move cannabinoid prescribing for cancer‐related mood symptoms toward more structured protocols. This study supports a CBD‐dominant, enteral‐first approach for patients with persistent psychological distress. Because there are no existing guidelines for cannabis use for anxiety in cancer patients, a helpful starting point for cannabinoid‐based anxiety relief may be the existing guidelines for cannabinoid titration in pain management, which recommend initiating CBD at low doses (e.g., 5 mg twice daily) and gradually titrating upward while monitoring tolerability and symptom response [44]. However, this must be tested in a clinical trial. While titrating, it may be helpful to implement the use of validated assessment scales, such as the ESAS used in this study, at baseline and at regular intervals. Safety monitoring should include assessment of unwanted cognitive effects and interactions with concurrent medications such as warfarin, buprenorphine, and tacrolimus [45]. Discontinuation should be considered if a clinically significant improvement is not achieved within the trial period.

A limitation of our dataset is that the cannabis transaction data were used as a proxy for product use and adherence, despite possible variability in use [46]. However, transaction data may be more accurate than patient‐reported dose estimates of cannabis, which are often inaccurate [47]. Objective transaction data from dispensaries allow for more precise tracking of the timing and contents of cannabis products for each patient. This, in turn, enhances confidence in inferences drawn from patient‐reported symptom responses, as standardized patient surveys were collected with each purchase. A second limitation is that patient cannabis use at baseline, prior use, and use outside of the MMCP purchases was unknown. Patients with prior cannabis use may have been able to tolerate greater doses and may have even required greater doses to perceive any effect compared to patients without previous exposure. Moreover, this dataset does not include a systematically collected objective assessment of what anxiety or depression medications were used by patients, nor adverse effects, which could have impacted symptom scores over time. Lastly, self‐reported depression and anxiety levels were subjective and not evaluated clinically [48].

Patients were recruited exclusively from Minnesota, which may reduce the generalizability of these findings to other patient populations. Even within a region, the differences between patients' cancer type, stage of cancer, and treatment regimens may affect neuropsychiatric symptom severity and sensitivity to CBD and THC. In this study, most patients did not report whether their cancer was metastatic, which may have impacted their symptom scores [49]. Also, the requirement of two post‐30‐day visits may have introduced selection bias for those who had greater longevity.

Future studies must account for the heterogeneity of cancer patients in general. Age, comorbidities, prior cannabis experience, and genetic variability in cannabinoid receptor expression in this population may impact the efficacy and tolerability of cannabinoids [50, 51, 52]. Even the preference for specific routes of administration may be confounded by patient characteristics; for example, fast‐acting inhalation or oromucosal products may be preferred by patients with more severe acute‐onset symptoms [53]. The integration of these factors into research and clinical decision‐making can help optimize therapeutic benefits, minimize risks, and advance more patient‐centered care in oncology.