In 2020, New Zealand held the world’s first national referendum on recreational cannabis legalisation. Fifty-one percent voted against it. The public considered the question directly and said no.

That same year, the New Zealand Medicinal’s Cannabis Scheme came into effect. Its architects described it as “a bill about cultivation, cannabidiol and compassion.” It would provide quality-controlled, CBD-based products to seriously ill patients: cancer sufferers, children with epilepsy, people in palliative care. In short, a narrow, humane reform. Nothing more.

Six years later, however, any registered doctor can prescribe any cannabis product to any patient for any condition. Private clinics operating on telehealth platforms now dominate the market, dispensing high-potency THC flower to front doors without a single face-to-face consultation. Some clinics offer subscription models. Furthermore, several pharmacies offer buy-now-pay-later. Researchers building the evidence base for this trajectory draw almost exclusively from surveys of self-identified drug users, a methodology so fundamentally compromised that it produced the published claim that most New Zealand drug users experience no harmful effects at all.

The public voted against recreational cannabis legalisation. Yet they got it anyway, with a prescription attached. As a result, the normalisation of a harmful substance has been advancing steadily ever since.

The Research Foundation Is a Farce

Before examining what the scheme has become, we need to look hard at the evidence base its defenders rely upon, because it does not hold.

Massey University researchers administer the New Zealand Drug Trends Survey annually, and it sits at the centre of nearly all published analysis of the scheme’s performance. The survey recruits between 9,000 and 25,000 respondents per year, roughly 80% of whom identify as past-six-months cannabis consumers. Participation is opt-in. In addition, researchers promote it through drug-user networks and communities. No random sampling takes place. No control group exists. Nothing screens respondents or blocks anyone from filling it in, for any reason.

This is not a population survey. Rather, researchers have constructed a self-portrait of a community that has pre-screened itself for tolerance of drug use and absence of serious personal harm. People who developed psychosis, lost employment, ended relationships, or withdrew from the online spaces where the survey circulates do not appear in the data. They cannot. The instrument simply cannot reach them.

The predictable consequence of building policy on such data arrived in January 2026, when researchers published findings under the headline: Majority of NZ drug and alcohol users experience no negative effects. The New Zealand Herald reported that almost two-thirds of survey respondents said they experienced no harmful consequences. Commentators described the finding as “groundbreaking.”

Groundbreaking it is not. Researchers asked a self-selected community of active drug users whether drugs had harmed them, then treated the answer as epidemiologically significant. In effect, the logic resembles surveying the prison population and publishing findings showing that most inmates consider themselves innocent. Researchers constituted the sample precisely from people whose circumstances reflect a particular relationship with the subject under investigation.

This is not a peripheral methodological concern. On the contrary, public health policy built on data that systematically excludes the harmed, the dependent, and the seriously ill will systematically fail them. Moreover, policy built this way produces a false public picture, one that presents drug use as broadly benign and frames anyone who raises harm concerns as out of step with the evidence.

Professor Wayne Hall, Emeritus Professor at the National Centre for Youth Substance Use Research at the University of Queensland, has identified the broader pattern. Researchers increasingly position survey data from satisfied consumers as evidence of efficacy, what Hall calls the subversion of the pharmaceutical regulatory system that modern medicine has relied upon for decades. “We’ve got Trumpian medicine,” Hall said at a 2025 NDARC webinar. “If patients think it works, that’s good enough evidence.”

Good enough it is not. Consequently, a scheme built on that standard is built on nothing.

What Was Promised, What Was Built

In March 2021, the Medicinal Cannabis Agency verified its first products under the scheme: CBD-dominant oral liquids, the category carrying the clearest therapeutic evidence base. Cannabidiol has established efficacy in reducing seizures in Dravet and Lennox-Gastaut syndromes. That early focus on oral liquid formulations reflected, however imperfectly, a genuine therapeutic intention.

Since then, however, researchers Marta Rychert and Chris Wilkins documented in a peer-reviewed analysis published in the New Zealand Medical Journal in October 2024 just how far the market has drifted. Of the 47 verified products under the scheme, 62% are now THC-dominant. Dried flower accounts for 27 of those 47 products. For example, some vaporiser products carry 25 to 26% THC by weight, potency levels comparable to legal recreational markets in the United States and Canada, and far stronger than anything circulating in New Zealand’s illegal market. By contrast, forensic analysis of seized illegal samples found THC potencies sitting between just 1% and 13.4%.

As a result, flower products accounted for 40% of all prescriptions written in the year to April 2024. THC-dominant products now hold the single largest prescription category, overtaking both balanced and CBD-dominant products for the first time.

Rychert and Wilkins noted that these products compare directly to cannabis sold in legal recreational markets in jurisdictions New Zealand’s public voted, in 2020, not to emulate. THC at these concentrations is not a gentle therapeutic adjunct. Instead, clinicians deliver a potent psychoactive substance with a well-documented harm profile through a medical channel that asks almost nothing of those writing the scripts.

Meanwhile, the Medicinal Cannabis Agency verifies only that products are stable, consistent, and contaminant-free. Producers face no requirement to demonstrate clinical trial evidence of efficacy. Whether their products work, for which conditions, at what doses, and with what side effects, nobody has proven for the vast majority of indications attracting prescriptions. Hall put it plainly: “If you can prescribe these drugs without evidence of efficacy and safety, why would you bother funding clinical trials?” Nobody does.

The Clinic Model and Its Conflicts

General practitioners did not want to prescribe cannabis. Studies found approximately two in three declined patient requests, for clinically reasonable reasons: limited trial evidence, unfamiliarity with dosing, discomfort with dried herb formulations, and genuine concern about adverse effects.

In response, private cannabis clinics moved in to fill the gap, rapidly and profitably. By mid-2022, approximately 40 physicians operated across 11 clinics, starting in Auckland and Wellington before expanding nationally through telehealth. Patients can now receive a prescription, access a dispensary, and arrange home delivery without leaving the house. Furthermore, subscription models automate the process entirely. Some pharmacies now offer buy-now-pay-later financing for products that are, by any reasonable definition, intoxicating and habit-forming.

The conflicts of interest this model generates are not speculative. A clinic that prescribes only cannabis has a direct commercial incentive to prescribe cannabis. Moreover, at least one New Zealand operation runs as a subsidiary of a licensed cannabis producer, collapsing the distance between commercial interest and clinical judgement entirely. Regulators explicitly identified this form of vertical integration as problematic and prohibited it in the design of the recreational cannabis market voters rejected in 2020. Yet in the therapeutic market that replaced it, the same structure has operated without apparent challenge.

Survey data from 2025 found clinic patients were disproportionately younger, more likely to receive THC-dominant flower prescriptions, more likely to carry a longer history of non-medical use, and more likely to receive a recommendation for a specific brand. None of those patterns appeared among patients who accessed prescriptions through GPs. Similarly, Australian research examining the same clinic model found patients were more likely to meet diagnostic criteria for cannabis use disorder and reported lower satisfaction with information about harms, benefits, and consideration of alternative treatments.

Australia’s Health Practitioner Regulation Agency has moved against high-volume prescribers, describing single-medicine clinic models as operations that “may take advantage of consumer demand” and “may be putting profit ahead of patient welfare.” By contrast, New Zealand regulators have shown far more restraint. To date, no publicly reported disciplinary proceedings have targeted any cannabis prescriber in New Zealand, despite evidence of short consultation times, inadequate screening, and prescribing to patients with contraindicated histories.

Marketing an Intoxicant as Medicine

Cannabis clinic advertising in New Zealand has moved well beyond anything clinical. Research presented at the 2025 NDARC webinar identified three categories of promotional practice warranting serious scrutiny.

Therapeutic scope expansion is the first. At least one clinic has suggested in promotional material that cannabis can help patients manage cost-of-living pressures and economic anxiety, a claim with no clinical basis that positions an intoxicating substance as a general remedy for ordinary life stress. Simply put, that is not medicine. Clinics are normalising drug use and dressing it in clinical language.

Targeting vulnerable populations is the second. In particular, clinic advertising has specifically targeted women living with endometriosis, a condition that takes an average of ten years to diagnose in New Zealand. Clinics direct women failed by inadequate standard healthcare pathways toward a commercial prescribing service offering a product of unproven efficacy for their condition. Reaching out to the desperate and the unheard is a deliberate marketing strategy, not a compassionate one.

Cause-brand marketing is the third: cannabis businesses aligning themselves with socially desirable causes, including drug reform advocacy, harm reduction messaging, diversity and inclusion initiatives, and charitable giving, to build consumer trust and brand loyalty. Consequently, research found cannabis-specific corporate social responsibility messaging ranked among the factors most likely to influence purchasing decisions, sitting closely behind price and convenience. In other words, companies selling an intoxicating substance leverage harm reduction language to sell more of it.

Formal prohibition applies to advertising medicinal cannabis products in New Zealand, as regulators class them as unconsented medicines. Even so, cannabis clinics appear to have successfully argued that they advertise services rather than products, a distinction the existing regulatory framework has proved unable or unwilling to contest.

The Harm Evidence Nobody Wants to Talk About

The Drug Trends Survey tells one story. Hospital records, however, tell another.

Research from Daniel Marrén’s group in Ontario, drawing on complete Canadian hospital system records, documented steep increases in patients presenting to emergency departments and requiring hospitalisation for cannabis-related psychosis. The trend is not ambiguous. Furthermore, Hall noted at the 2025 NDARC webinar that Australian surveys of patients presenting with first-onset psychosis show growing prevalence of medical cannabis prescriptions, including prescriptions for high-potency THC products given to patients for whom clinicians should have identified clear contraindications.

New Zealand, meanwhile, lacks the data infrastructure to identify equivalent domestic patterns in usable form. Coders record adverse outcome notifications inadequately. No mechanism links prescribed medicinal cannabis reliably to downstream health presentations. As a result, the Drug Trends Survey fills that vacuum, and regulators treat absence of evidence as evidence of absence.

High-frequency, high-dose THC use carries real risks: cannabis use disorder, psychosis, adverse impacts on cognition and daily functioning, impairment of the capacity to drive, work, and parent. Researchers established these associations in the context of recreational use, but the substance is the same substance. Importantly, a therapeutic framing does not alter the pharmacology. Cannabis at 25% THC does not become safer because a clinic wrote a prescription for it.

The Precedent Being Set

The New Zealand’s Medicinal Cannabis Scheme is not only a domestic policy matter. Beyond that, it shows precisely how a harmful substance moves through a therapeutic framework, step by deliberate step, until the original intention becomes unrecognisable.

Hall describes the pattern with precision. Compassion opens the door. A high-profile case, a sick child, a cancer patient, a palliative care story, creates the political conditions for a narrow, CBD-focused framework. Subsequently, lobbyists push to expand the range of conditions, broaden prescribing authority, and increase product diversity. Both New Zealand and Australia have now reached the endpoint: any doctor can prescribe any cannabis product to any patient for any reason, with no requirement to demonstrate it works.

The precedent reaches beyond cannabis. Notably, Hall observed that psychedelic advocates have followed the medical cannabis model using virtually identical framing and regulatory strategies. GLP-1 agonists like semaglutide, increasingly prescribed off-label at scale, generate comparable regulatory challenges. “We’ve seen that the medical cannabis has created a precedent,” Hall said, “that’s been followed by the psychedelic advocates of psychedelic medical use.”

Every time a harmful substance acquires a therapeutic identity, the public conversation about its risks contracts. As a consequence, prevention work becomes harder. People working to reduce community harm now operate in a landscape where a medical system that has stopped asking whether things work actively legitimises the substances they are trying to address.

Equity Promised, Inequity Delivered

The scheme’s architects framed it partly as an equity measure. Māori, lower-income users, and marginalised communities would gain access to quality-controlled products and proper medical oversight, escaping an illegal market that criminalised them disproportionately.

In practice, however, prescription data for the year to April 2024 tells a different story. Māori make up 17.4% of the population and use cannabis medicinally at higher rates than other ethnicities, yet they received only 12.9% of prescriptions. Similarly, lower-income earners, women, and people seeking cannabis for substance dependency treatment all appear in prescription data at rates below their prevalence among medicinal cannabis users.

Monthly therapy costs run from NZ$120 to NZ$400, with nothing subsidised. Disability Allowance support exists in theory but remains practically inaccessible. In addition, patients must exhaust a long list of prior treatments before regulators consider cannabis eligible.

As a result, people priced out stay in the illegal market, still criminally liable under the Misuse of Drugs Act, still exposed to products of unknown potency. Some access so-called green fairy networks: female-dominated, compassionately motivated informal supply chains operating entirely outside the regulated framework. Those networks exist as a direct verdict on the scheme’s equity performance. Ultimately, the people who most needed compassionate access built their own systems because the official one shut them out.

Māori are more likely to use cannabis medicinally, less likely to hold a prescription, and more likely to face arrest for cannabis-related offences. The scheme has not reduced that disparity. On the contrary, on the available evidence, it has deepened it.

The Commercial Experiment Is Collapsing

In March 2026, Helius Therapeutics entered voluntary administration. One of New Zealand’s founding medicinal cannabis manufacturers, established in 2018, the company cited a “challenging commercial and regulatory environment” and immediately closed its East Tamaki manufacturing facility. Asset sales followed.

Helius is not an isolated case. Similarly, reports from early 2026 document stock shortages affecting thousands of patients and commercial pressure throughout the sector. Dispensings under the scheme grew from 4,875 in 2020 to 265,731 in 2025, a more than fiftyfold increase, yet volume growth produced no commercially sustainable domestic industry.

That outcome is not surprising. A market built on products requiring no efficacy evidence, run through clinics carrying structural conflicts of interest, serving a consumer base substantially motivated by non-medical factors, and overseen by a framework too under-resourced to enforce its own guidelines was never a pharmaceutical market. Instead, it was a commercial venture dressed in clinical clothing. Helius Therapeutics did not collapse because of bad luck. Rather, it collapsed because the foundations were missing from the start.

What Needs to Be Said

The New Zealand’s Medicinal Cannabis Scheme set out to provide therapeutic relief to seriously ill patients. Instead, it has become a lightly regulated, commercially driven market in high-potency THC products, resting on research that cannot withstand scrutiny, and failing the communities it claimed to serve.

Voters rejected recreational cannabis legalisation in 2020. Nevertheless, the scheme delivered it by another route, without taxation frameworks, age restrictions, or the public health infrastructure a transparent recreational market would require, and without the honest public debate a decision of this magnitude demands.

First, policymakers and researchers need to assess the Drug Trends Survey for what it actually is: an opt-in survey of drug users that cannot capture those most harmed, and is wholly unsuitable as the primary evidence base for national drug policy. Accordingly, harm evidence from Canadian hospital data, Australian adverse prescribing reports, and the established pharmacology of high-potency THC deserves the weight it has not received.

Second, direct regulatory intervention, not more guidelines but actual enforcement, must address the commercialisation of prescribing through cannabis clinics, vertical integration between producers and prescribers, telehealth prescribing without mandatory face-to-face assessment, and buy-now-pay-later financing for a habit-forming intoxicant.

Above all, cannabis is a harmful substance. The harm evidence is substantial. The evidence base for most therapeutic applications remains weak. Commercial interest, far more than public health, has shaped the architecture surrounding cannabis prescription in New Zealand. The communities bearing the costs of that arrangement, in dependence, in psychosis, in criminalisation, deserve a policy framework built on honesty rather than compassion as a marketing strategy.

Six years ago, a Minister of Health told parliament this was a bill about compassion. The record of six years says otherwise.

Sources

Rychert M, Wilkins C. “Implementation of the Medicinal Cannabis Scheme in New Zealand: six emerging trends.” New Zealand Medical Journal, October 2024.

NDARC Webinar: Six Years of the Medicinal Cannabis Scheme in New Zealand, Challenges Ahead. Associate Professor Marta Rychert; Professor Wayne Hall, University of Queensland, May 2026.https://www.unsw.edu.au/research/ndarc/resources/six-years-of-the-medicinal-cannabis-scheme-in-new-zealand

NZ Herald: “Mostly harmless: groundbreaking research shows most drug users experience no negative consequences.” Derek Cheng, January 2026.https://www.nzherald.co.nz/nz/politics/mostly-harmless-groundbreaking-research-shows-most-drug-users-experience-no-negative-consequences/premium/4326YXF5NNDOJIDFW3FBU34M6Y/

RNZ: “Medicinal cannabis company Helius Therapeutics shuts down.” 11 March 2026.https://www.rnz.co.nz/news/national/589275/medicinal-cannabis-company-helius-therapeutics-shuts-down

NZ Herald: “Medicinal cannabis industry under strain, facing shortages and closures.” Lachlan Rennie, 2026.https://www.nzherald.co.nz/nz/medicinal-cannabis-industry-under-strain-facing-shortages-and-closures/THMZ3FGD2VG7HOL6B7IYQ6RXHY/