To our knowledge, the present study is the largest study to date on self-reported use of CaM as substitution for prescription drugs in a European sample. Further, the study is the first to explore the type of cannabis used as a substitute across different types of prescription drugs, providing valuable insights into user perspectives on CaM as a substitute for prescription drugs in treating somatic and mental health conditions.
Our findings show that substitution of prescription drugs is a leading motive among users of CaM in a Danish convenience sample, and that this practice is more common among women, people with reduced working capacity due to illness, and people with chronic pain and other somatic conditions. We found that pain medication by far was the most common type of prescription drug substituted, followed by anti-depressants and arthritis medication, and that Tramadol was the most common class of pain medication that was substituted with CaM. Furthermore, the self-reported substitution effect was considerable, as the vast majority of substitution users reported either a substantial decrease in, or cessation of, prescription drug use as a consequence of their use of CaM. Across prescription drug categories, we found that CBD-oil was the most common form of cannabis used as substitution, except for people who substituted anti-psychotic medicines, where the use of ‘hash, pot or skunk’ was the most prevalent. A majority of the substitution users reported that CaM was much more effective in treating their conditions compared to the prescription drugs that they had used, and a large majority reported that the side effects from their use of prescription drugs were much worse compared to the side effects when using CaM.
Prevalence of substitution and characteristics of substitution users
Studies from Canada and the USA indicate that use of CaM as a substitute for prescription drugs is prevalent among users with legal access to cannabis [24, 26,27,28, 30, 64, 84, 85]. Our study adds to this literature by showing that substitution of prescription drugs is also a prevalent motive among users of CaM in a European sample, where the majority of substitution use takes place outside a legal medical setting. While a recent international cross-sectional survey found that substitution of prescription drugs was not significantly related to having legal access to medical cannabis , we found that users with legal access to CaM were significantly more likely to be substitution users compared to users without legal access. This difference may be partly explained by the fact that in Denmark, legal medical cannabis is only recommended for specific conditions when other treatment options have proven inadequate . Our finding that users with limited recreational experience with cannabis are just as likely to use cannabis as a substitute for prescription drugs as experienced recreational cannabis users is in line with previous research on substitution as a motive among prescription CaM users . Overall, the present study adds to the growing body of research indicating that use of cannabis as a substitute for prescription drugs is a prevalent phenomenon.
Our finding that women are more likely than men to substitute prescription drugs with CaM resembles previous studies [24, 28, 29] and indicates that there may be a more general pattern of gender differences related to the use of CaM. Of note, a cross-sectional survey of cannabis users from the USA found that female users report significantly lower frequency and quantity of cannabis use and significantly higher rate of medicinal use for anxiety, irritable bowel syndrome, nausea, anorexia, and migraines compared to men . Further, a qualitative study on legal medical cannabis use, also from the USA, found that gender may influence patterns and practices of use, as the narrative of female users included more collaboration with health care providers compared to the narrative of male users . While other studies have found that older CaM users are more likely to be substitution users , our study showed that a substitution motive of CaM was not significantly related to age. The fact that substitution users in our study were more likely to be on disability pension or in reduced employment may indicate that these people encounter treatment barriers in the healthcare system that ‘push’ them towards alternative types of treatment. One such barrier could be the cost of treatment, as the current instability in the prescription drug market with large fluctuations in price [88, 89] may further an economic substitution motive among some low-income patients. However, the use of unregulated cannabis may still be more expensive for most compared to the price of prescription drugs, as government reimbursement of medicine costs in Denmark are substantial . Furthermore, we found that CaM users who treat chronic pain were more likely to be substitution users, which may indicate insufficient treatment options for this patient group, as those suffering from non-malignant chronic pain are potentially undertreated [91, 92].
Type of prescription drug(s) substituted and the impact of substitution
Our finding that pain medications, particularly opioids, were the most common prescription drugs substituted with CaM, followed by anti-depressives and arthritis medication, corresponds with our findings on chronic pain, arthritis, and depression as the most frequent motives for use of CaM. These findings reflect findings from other survey studies; in a large international cross-sectional survey on cannabis users, Corroon et al.  found that opioids (13.6%), anxiolytics (12.7%), and antidepressants (12.7%) were the most common classes of prescription drugs substituted with cannabis, and a Canadian cross-sectional survey of medical cannabis patients found that opioid medications accounted for 35.3% of all prescription drug substitutions, followed by antidepressants (21.5%) .
The fact that Tramadol was the most common type of prescription drug substituted in our study may in part be explained by the prevalent use of Tramadol in treatment of non-malignant pain in Denmark . However, it may also be related to the decision by the Danish Medicines Agency to surveil prescription patterns of Tramadol among physicians in September 2017, following a rise in Tramadol use  and growing concern among clinicians’ regarding the abuse potential of Tramadol . It is likely that this decision reduced access to prescribed Tramadol for some patients, who subsequently turned to CaM in order to treat their pain condition. In fact, the number of Tramadol users in Denmark decreased for the first time since 2008 in 2017 , and continued to decrease from 2017 to 2018 (23% reduction in Tramadol users) . Indeed, problematizing and reducing opioid prescriptions leaves a vacuum that may motivate some patients to seek other therapies such as CaM . Interestingly, the increased use of Tramadol in Denmark earlier occurred as a consequence of another vacuum in the management of non-malignant pain caused by the problematization of nonsteroidal anti-inflammatory drugs (NSAIDs) and the discovery of serious long-term side effects of these drugs . Thus, there seems to be a “cycle of vacuums” in the treatment of non-malignant pain, underscoring the need to rethink the management of non-malignant pain in the Danish health care system , as this patient group is potentially undertreated .
Our findings on the cessation of and substantial decrease in prescription drug use show a considerable reported substitution effect related to the use of CaM, which is echoed in other studies on the reported substitution effect of CaM among patients with access to medical cannabis. In a survey of dispensary members in New England, the majority reported a decreased use of opioids (76.7%), anxiety medication (71.8%), migraine medication (66.7%) and sleep medication (65.2%) . Similarly, in a survey of American medical cannabis users with chronic pain, the majority of users reported complete cessation of opioids (72%) benzodiazepines (68%), NSAIDs (44%), gabapentanoids (74%), disease-modifying antirheumatic drugs (80%), Serotonin–Norepinephrine Reuptake Inhibitors (78%), and Selective Serotonin Reuptake Inhibitors (80%) . The link to a decrease in opioid use has also been shown over time in a small cohort study in New Mexico, comparing 37 chronic pain patients enrolled in a medical cannabis program to 29 non-enrolled chronic pain patients over the course of 21 months . Findings from this study showed clinically and statistically significant associations between medical cannabis enrollment and opioid prescription cessation and reduction, as well as improved quality of life. Thus, findings from our study add to the growing body of research indicating that from a user perspective, CaM has a substantial substitution effect for a variety of prescription drugs, most notably opioids.
Type of cannabis used as a substitute for prescription drugs
The findings that, in our sample, CBD-oil is the most prevalent type of cannabis used as a substitute for prescription drugs, and that one third of the substitution users used CBD-oil only, are in accordance with findings from a recent survey of 1.483 medicinal CBD users in the USA where CBD was used as a specific therapy for medical conditions, particularly pain and inflammatory disorders, as well as anxiety, depression, and sleep disorders . In the same study, the majority (65.3%) reported that CBD treated their condition(s) moderately or very well without the use of conventional medicine, and 30.4% reported that CBD was effective in combination with conventional medicine. Our findings are also in accordance with a recent Italian study, where the unintended legalization of CBD-based cannabis products with less than 0.6% THC was associated with a significant decrease in the sale of prescription drugs, particularly of anxiolytics, sedatives, and anti-psychotics . This is of particular public health interests, as CBD has been shown to have a better safety profile in terms of side effects and abuse potential, relative to THC [42, 98, 99]. In line with this, a recent review and meta-analysis of clinical trials found that CBD was well tolerated and had few serious side effects across medical conditions . The safety profile of CBD-based cannabis products may also be superior in terms of toxicology and abuse potential compared to some of the prescription medication that is substituted, such as opioids and benzodiazepines [49, 52, 53].
From a public health perspective, the problematic aspects of medicinal use of low THC/high CBD cannabis products are also, and maybe more, related to the fact that these products are unregulated and used without medical supervision . Indeed, use of unregulated cannabis products increases the risk of consuming hazardous contaminants, such as fungi, bacteria, pesticides or heavy metals [1, 102] or consuming a product with undesired psychoactive effects. Interestingly, a recent examination of CBD-oils available in Denmark by Department of Forensic Medicine in Odense, revealed that 38% of CBD-oils tested contained between 0.2% THC and 1.2% THC, despite being advertised as below 0.2% THCFootnote 1 . Thus, users of CBD-oil may unknowingly use products that are illegal to consume.
Our finding that users of CaM, who substituted anti-depressants or anti-psychotics, are significantly more inclined to use “hash, pot or skunk” compared to other substitution users is of particular interest to public health, as the THC content in skunk is high, and the THC content in cannabis resin (hash) and herbal cannabis has increased markedly in the last decades in Europe  and the USA [105, 106]. In Denmark, we found a threefold increase in THC concentration in seized hash from 2000 (mean: 8.3%) to 2017 (mean: 25.3%), while CBD levels remained stable (mean around 6%) . This trend is concerning, as increasing evidence suggests that exposure to high-THC and low-CBD cannabis products is associated with higher risk of cannabis-related harms, such as cannabis dependence [108,109,110], psychosis [111, 112], and cognitive impairment [99, 113], compared to low-THC and high-CBD products. The magnitude of the problem is further underlined by findings from naturalistic studies on users of smoked cannabis products, which indicate that users do not fully adjust their use to differences in THC concentration, suggesting that users of more potent products are exposed to higher levels of THC [114, 115]. Thus, it is possible that use of high THC cannabis products as a substitute for prescription drugs may exacerbate the condition that is the target of the treatment, particularly in relation to treatment of psychotic disorders.
In sum, there is considerable complexity related to the use of cannabis as a substitute for prescription drugs, as an evaluation of whether cannabis is in fact “the lesser of two evils” depends on various factors, such as the cannabinoid composition in the type of cannabis used, the dosage, and the type of prescription drug that is substituted. Recently, another “lesser of two evils”-dynamic is emerging from cannabinoid research, as it is increasingly plausible that low-THC cannabis products are less harmful compared to high-THC cannabis products, and that increased availability of low-THC cannabis products may hold a potential for harm reduction among users of high-THC cannabis products . In line with this reasoning, the unintended legalization of low THC-cannabis products in Italy, was not only associated with a significant decrease in prescription drug use, but also with a decrease in confiscations of illegal cannabis and drug-related arrests , indicating a substitution effect of introducing low THC-cannabis products on the consumption of conventional illegal cannabis products.
Experienced effects and side effects
Findings from our sample show that most substitution users find CaM more effective in managing their condition(s) compared to prescription drugs, and that an overwhelming majority found CaM to have a better side effect profile compared to the prescription drugs that they had been prescribed for their condition(s). This is in line with recent findings from Canada and the USA. In a cross-sectional survey, Canadian medical cannabis users listed, ‘relative safety of cannabis to prescription drug’, ‘fewer adverse side effects’ and ‘better symptom management’ as their top three reasons for using CaM as a substitute for prescription drugs . In a survey of medical cannabis patients in California who used cannabis as a substitute/in conjunction with opioid-based pain medication, 80% found that cannabis was more effective than opioids for pain, and 92% that the side effects of cannabis were more tolerable than opioids . Moreover, in a survey of American medical cannabis users with chronic pain, respondents listed ‘fewer side effects’ and ‘better symptom management’ as their top reasons for using medical cannabis as a substitute for prescription drugs . Lastly, a qualitative study on cannabis users in San Francisco showed similar perceptions among substitution users who found cannabis to be a safer and more effective alternative compared to prescription drugs . It is interesting that substitution users in our and several other studies rate the side effect profile of CaM higher than its effects, when comparing cannabis to prescription drugs. This suggests that substitution users may have the same “ lesser of two evils”-perspective on the medical utility of cannabis that was documented among some physicians in Israel , where cannabis becomes medicine not only on the basis of what it is in terms of effects, but on what it is not in terms of side effects when compared to prescription drugs. Considering the growth in use of CaM, it is likely that this perspective will result in an increasing number of people seeking information and advice about the effectiveness of CaM and use CaM as a substitute for prescription drugs, even in the absence of rigorous clinical trials and despite lack of legal access to medical cannabis. Future research is needed to assess effects and side effects of long-term use of CaM from longitudinal studies. Furthermore, placebo-controlled clinical efficacy trials are needed to explore the effects of cannabis beyond placebo, and current barriers to whole plant cannabis research need to be addressed .
Our study has several limitations. First, this self-selected convenience sample is limited by selection bias, as it likely weighs towards successful users of CaM, users with internet access, a familiarity with online surveys, users engaged with the topic on social media, and with the resources necessary to answer such surveys. Therefore, the survey may not be representative of the population of CaM users or substitution users . Second, the data used in the study may be subject to self-reporting biases, such as recall bias, confirmation bias, placebo effects or social desirability bias . For example, users are likely to have optimistic expectations regarding the efficacy of CaM  and may exaggerate positive effects of CaM or under-report adverse effects. Also, recall bias may be more salient for the small group of respondents who were not current users. Third, although duplicates were excluded from the analyses, we cannot rule out multiple responses from the same person, as IP-addresses were not accessible to researchers. Fourth, the cross-sectional study design lacks a temporal dimension, and we do not know if the reported cessation or reduction in prescription drug use is sustained over time. Fifth, findings on prescription drugs substituted may be skewed by the fact that “sleep medication”, “ADHD medication” and “anxiety medication” were not presented to respondents as independent categories. Sixth, the experienced effects and side effects of CaM and prescription drugs could potentially be affected by pharmacokinetic interactions between cannabis products and prescription drugs when taken simultaneously [120, 121]. Also, the reported effects and side effects were a mapping of substitution users experiences with various types of prescription drugs and different subtypes of cannabis, which likely vary in terms of effect and side effect profile.