Evidence brief: regulation, safety and stakeholder support for Colombian cannabis policy

Colombia’s effort to build a workable regulatory framework for medicinal cannabis has received a rigorous scientific roadmap, thanks to a new study that systematically connects research evidence with the people who must act on it. Published in the journal Health Research Policy and Systems, the study documents a comprehensive knowledge translation initiative led by researchers at the Unit of Evidence and Deliberation for Decision Making (UNED) at the University of Antioquia in Medellín. Rather than simply reviewing the literature, the team designed and executed a four-part strategy to mobilize scientific evidence, public opinion and stakeholder perspectives directly into the hands of Colombia’s Ministry of Health, offering one of the most detailed real-world demonstrations to date of how evidence can be moved into policy in a complex and politically sensitive domain.

The research was motivated by a persistent problem in health policy: the gap between what science knows and what decision-makers use. Medicinal cannabis is a particularly acute case. Colombia has moved to regulate cannabis for medical and scientific purposes, but formulating rules about who can access products, what levels of psychoactive compounds they may contain and how health professionals should be trained remains contested. The study’s authors partnered with decision-makers from the Ministry of Health, who helped define three policy elements around which all evidence would be organized: the limits that regulations should place on cannabis products, the effectiveness and safety of medicinal cannabis for clinical uses, and the training and communication strategies needed to support implementation. This co-production of research questions with policymakers is a hallmark of evidence-informed policymaking and distinguishes the approach from conventional academic reviews.

The methodology unfolded in four sequential components. First, the team identified the best available scientific evidence on the three agreed policy elements and synthesized it into a formal evidence brief, a structured document designed to present research findings in a form that is accessible and useful for policy audiences. Second, the researchers convened citizens panels with twenty members of the public, who deliberated on the findings of the evidence brief and voiced their concerns and priorities. Third, a policy dialogue was organized with thirty-two key stakeholders drawn from organizations with a direct interest in medicinal cannabis regulation, ranging from health authorities to patient groups and industry representatives. Finally, the team conducted semi-structured interviews with four decision-makers to document how they perceived the process and whether they could trace any policy developments back to it. The study protocol was approved by the Ethics Committee of the Institute of Medical Research at the University of Antioquia and classified as minimal-risk research, with all participants providing informed consent.

The evidence brief itself yielded findings that were more nuanced than the polarized public debate around cannabis might suggest. On the question of regulatory limits, the synthesis showed that permissible thresholds of tetrahydrocannabinol, or THC, the primary psychoactive compound in cannabis, vary considerably across jurisdictions that have legalized medicinal cannabis, but commonly fall in a narrow band between 0.2 and 1 percent. This finding matters because THC thresholds determine which products can be sold outside strictly medical channels, how they are taxed and controlled, and where the boundary lies between medicinal and recreational markets. The variability among countries indicates that there is no single internationally agreed scientific standard for where such a line should be drawn, leaving regulators to weigh evidence alongside social, economic and enforcement considerations.

On effectiveness and safety, the evidence base supported the use of medicinal cannabis for a limited but meaningful set of indications. The synthesized research found support for its use in managing chronic pain, as an adjunctive treatment for refractory epilepsy, that is, epilepsy that does not respond to conventional anti-seizure medications, and for alleviating chemotherapy-induced nausea and vomiting in cancer patients. However, the researchers were careful to flag an important caveat: the underlying evidence was of low to moderate quality, and significant uncertainty persists. This grade of evidence means that while the direction of effect appears favorable for these conditions, the precise magnitude of benefit, the optimal dosing and the long-term safety profile remain incompletely characterized. The finding underscores a recurring tension in drug policy, in which regulatory decisions must often be made before the scientific evidence is as robust as clinicians and patients would like.

The third policy element, training and communication, produced findings with direct operational implications. The evidence showed that educational strategies can be effective at increasing clinical confidence among healthcare personnel who are asked to prescribe or advise on medicinal cannabis. Yet when researchers examined the current state of professional preparation, healthcare workers themselves reported that their training on the topic was insufficient. This mismatch, between what education can achieve and what professionals currently receive, identifies a concrete leverage point for regulators: any regulatory framework for medicinal cannabis will likely underperform if the clinicians implementing it do not feel adequately informed about its pharmacology, indications, contraindications and legal boundaries.

The deliberative components of the study added a dimension that systematic reviews alone cannot capture. When twenty citizens deliberated on the evidence brief and thirty-two stakeholders engaged in structured policy dialogue, a striking consensus emerged around a single barrier: the lack of clear and accessible information. Participants across both groups identified this information deficit as the primary obstacle to informed decision-making about medicinal cannabis, whether by patients choosing treatments, clinicians advising them, or regulators setting rules. From this shared diagnosis, the deliberations generated a set of priorities that included professional training programs, the development of evidence-based clinical guidelines, and the establishment of monitoring mechanisms to track outcomes once regulations take effect. These recommendations give Colombian policymakers a concrete, citizen- and stakeholder-endorsed agenda for the implementation phase.

The interviews with decision-makers provided an early measure of whether the entire exercise had any policy traction. While the study reports that the process was perceived positively and that participants associated it with policy developments, the researchers temper their conclusions with honest acknowledgment of the limits of what evidence and dialogue can resolve. Despite regulatory advancements in cannabis, they note, uncertainties remain about how regulatory limits should be set and what consequences different threshold choices may produce. The dearth of high-quality studies, they argue, continues to obscure the efficacy and safety profile of medicinal cannabis, meaning that regulators must act under conditions of partial knowledge. Their solution is not to wait for perfect evidence but to pair incremental regulation with monitoring systems that can generate better data over time, creating a feedback loop between policy and research.

The significance of the study extends well beyond Colombia’s borders. Knowledge translation, the deliberate effort to make research usable by decision-makers, has become a growing field within health policy research, and rigorous evaluations of translation strategies in real policy contexts are rare. Most studies in this area either document the content of evidence syntheses or the mechanics of stakeholder engagement, but few combine both with a clear chain of evidence linking scientific findings, citizen deliberation, stakeholder dialogue and decision-maker reflection. By structuring its entire inquiry around policy elements defined jointly with the Ministry of Health, the Colombian team has offered a template that other countries contemplating cannabis regulation, or indeed any contested health policy, could adapt. The work was funded by Minciencias, Colombia’s national science agency, under a project explicitly designed to support evidence-informed policymaking at the national level, signaling institutional commitment to this model of governance.

What the study ultimately illustrates is that the hardest questions in medicinal cannabis policy are not purely scientific. Science can tell us that THC thresholds in regulated markets cluster between 0.2 and 1 percent, that the evidence for chronic pain, refractory epilepsy and chemotherapy-induced nausea favors cautious use, and that training improves clinician confidence. But deciding where Colombia’s own threshold should fall, how uncertainty should be communicated to patients, and how quickly prescriber education should be scaled up requires deliberation among people with different values, interests and levels of expertise. By building a process in which citizens, stakeholders and officials all engaged with the same evidence brief, the University of Antioquia researchers have shown that these conversations can be structured, documented and fed into policy in a transparent way. Their findings suggest that the most important product of such a process may not be any single recommendation but the creation of shared, accessible information, precisely the resource that participants identified as the missing ingredient in Colombia’s medicinal cannabis debate. As more countries move to legalize or regulate cannabis for medical use, the Colombian experience offers both a method and a caution: evidence can inform policy, but only if deliberate machinery exists to carry it there, and only if that machinery honestly acknowledges where the evidence ends and judgment begins.