1 INTRODUCTION
Approximately 1 in 10 cannabis users report using cannabis for medicinal purposes (3.1 million Americans and 330 000 Britons use cannabis for medical purposes).1 As of 2020, 19 European countries, Canada, and 33 states in the United States, allow the use of cannabis for medical conditions.2 The expansion of medical cannabis indications has increased quickly over the last 5 years for a variety of indications including chronic pain, multiple sclerosis, and seizure disorders.3 This expansion of medicinal use contradicts evidence from systematic reviews and meta‐analyses of randomized controlled clinical trials (RCTs) which identified no strong evidence of benefit of cannabis for patient outcomes in many of these conditions.3, 4 To provide more evidence on the true effects of medicinal cannabis use, there has been a recent focus on evaluating the safety and efficacy of medical cannabis through RCTs. Although RCTs demonstrate the efficacy of a treatment, they are often limited in their generalizability to the real world due to their focus on a narrow group of individuals with similar characteristics consuming a standardized dose or product.5 Thus, RCTs are often unable to explain precisely for whom the drug is effective or safe once it is used by a broader population.
Real‐world evidence (RWE) studies provide the opportunity to fill gaps in our understanding where RCTs may fall short. Using electronic health records or linked large databases of routinely collected health information like insurance claims and physician fee codes, these studies provide new opportunities to develop broader evidence to support clinical decisions in a timely fashion. In this commentary, we will argue that RWE for medical cannabis is not being utilized and should be. It can be used as an opportunity to help fill the gaps in our understanding of medical cannabis effectiveness and safety. In addition, we offer solutions of how to improve commonly used healthcare databases to increase the capacity for the real‐world study of medical cannabis.
2 OPPORTUNITY
RWE plays a central role in health care decision‐making, particularly in monitoring post‐market safety of medications after they have received regulatory approval. These post‐market studies are typically required by regulators.6 RWE studies are less expensive and faster to complete than trials and, if appropriately designed, can provide valuable and complementary information to RCTs. For example, observational studies are commonly used to signal safety issues with drugs during their post‐marketing phase and examining clinical factors such as, drug–drug and drug co‐morbidity interactions.7, 8
RWE would help to fill current evidence gaps in medical cannabis to account for limitations of RCTs. First, populations may differ in the real world from those included in RCTs. For example, patients who use cannabis for medical purposes often have psychiatric and medical complexities which would exclude their enrollment in most RCT studies. Further, the patient populations receiving medical cannabis have changed greatly over time (e.g., changes in age group and new medications), and RWE can account for these factors, as repeating RCTs may not be feasible. For instance, more older adults are using cannabis for medicinal reasons in place of prescription medications.9 However, the vast majority of cannabis research is completed in younger populations (mean age of 45 to 50 years across most studies) and either does not include the elderly at all or does not report on them separately.10 The lack of data in this population and others with multiple comorbidities is critical, as these groups may be more likely to have rarer or more serious adverse effects. Second, RWE studies can allow for a long observation window to provide insights into toxicity, long‐term safety, and rare adverse events in large, representative patient populations. This is important given the often short observation periods used in cannabis clinical trials, with most RCT study durations ranging from 3 to 14 weeks.3
The combination of evidence from RCTs (rigorous efficacy data) and RWE (real‐world utilization and safety) would allow healthcare providers to make more evidence‐based decisions in assessing for the appropriateness of initiating or continuing medicinal cannabis for therapeutic purposes.
3 BARRIERS AND POTENTIAL SOLUTIONS
3.1 Data issues
The current capabilities do not exist in any jurisdiction, to our knowledge, to conduct RWE studies to examine medical cannabis in a similar fashion to other treatments. As such, work is needed to develop the necessary infrastructure to make these studies feasible. RWE makes use of multiple real‐world data sources, such as clinical registries, claims data, and electronic health records. To roll out these data infrastructures on a population level, we first need to ascertain what data sources will be utilized. Medical cannabis is challenging as, depending on the jurisdiction, patients may be able to acquire cannabis for medical purposes, from differing legal source. This means that both traditional “medical dispensers” and “recreational dispensers” may be providing cannabis for medical purposes. Further to this, the definition of “medical use” may vary by jurisdiction. Additionally, there needs to be a link between medical cannabis dispensing records and other healthcare databases (such as medication dispensing data, electronic medical records, physician services, and hospital inpatient and outpatient data). To date, cannabis is not an approved medication in most jurisdictions and does not have a product identifier as a drug (e.g., National Drug Code [NDC]), limiting the ability to link the dispensing of medical cannabis with healthcare databases to generate evidence through large population‐based studies.
Further, medical cannabis dispensing needs to be standardized to provide an opportunity to collect meaningful data. Prior to a cannabis claim submission, the infrastructure should be developed such that we are able to collect information about the composition of key biologically active ingredient composition (i.e., tetrahydrocannabinol and cannabidiol), directions for use (e.g., one inhalation twice daily), and route of administration (e.g., oral tablet, capsule, oil or edibles product, or dried products for vaporization or smoking, or topical products). This provides an opportunity to collect richer data by understanding patients’ reasons for use, along with important product and administration information.
3.2 Policy and regulatory issues
For proper medical cannabis reporting to occur, there needs to be a shift in policy and support from regulators. The policy changes would mandate the reporting, what is reported, and support for the development of a cannabis monitoring system (or leverage currently existing systems). To date, there are no regulations that state a claim must be submitted upon dispensation of medical cannabis and how to report the claim. For good infrastructure to exist for RWE, there is a need for policy changes from a regulator that mandates that any time medical cannabis is dispensed, a claim is entered to a monitoring system. A possible solution is to leverage monitoring systems (e.g., Ontario’s Narcotic Monitoring System [NMS]) available in most jurisdictions that have been built for narcotics to also collect medical cannabis dispensing information. From a regulatory framework, we can learn from the opioid experience by using the opioid monitoring database, and we can similarly mandate cannabis reporting leveraging already established infrastructure and data standards.11 Similar solutions should be considered for countries that have established robust drug claims infrastructure and/or pharmaceutical monitoring systems. However, for countries without robust drug claims systems and data sources, potential solutions could include developing linkable registries or monitoring wholesale supplies and distribution.
4 CAUTIONS
While RWE can reflect the real world, there are still limitations that need to be addressed. To gain meaningful conclusions from RWE studies, we must rely on accurate classification of medical cannabis exposure (as listed in administrative claims). However, there are a variety of circumstances in which these prescription claims may not reflect the actual medication exposure of patients, in turn resulting in spurious conclusions. Potential sources of drug exposure misclassification include non‐users misclassified as users and users misclassified as non‐users.12 Lessons can be learned from the field of pharmacoepidemiology. One source of misclassification is that there will be patients with medical cannabis authorizations or prescriptions that are filled but never taken, those initially taken and then discontinued, and those taken (as needed) or intermittently, leading to non‐users of medical cannabis being misclassified as users. An approach to minimize the impact of these misclassified individuals is to require evidence of a second prescription fill within a fixed period of time to increase the likelihood that patients are actually using the medication.13, 14 Another source of error is when patients pay for prescription medications out of pocket, over the counter, or are hospitalized, these medical cannabis users will be misclassified as non‐users of medical cannabis in the database. The same limitation also applies to those using cannabis illegally or recreationally. In administrative databases that include medications on a formulary (e.g., in Canada or U.S. Medicare), we believe individuals who are prescribed medical cannabis to treat a medical indication may limit their use to a therapeutic purpose and not recreational, largely due to the added cost.12
5 CONCLUSION
RWE for medical cannabis is under‐utilized, creating a missed opportunity to enrich healthcare decision‐making that is urgently needed. RWE is a robust approach to study the use of medical cannabis in the real world and can help inform future decisions in policy and practice. To expand the use of RWE to examine medicinal cannabis, we need to invest in capacity and resource building and also require regulatory support that this data should be collected and shared. Currently, cannabis is being permitted for medical purposes, and population‐level use is at highest‐reported levels in many jurisdictions, but requirements for ongoing study of its effects are not built into current policy. This gap in real‐world safety and effectiveness of medicinal cannabis may have widespread clinical impact long‐term, leading to unnecessary use and potential harm to patients.
COMPETING INTERESTS
There are no competing interests to declare.
