Medical Cannabis for Children in Israel: What the National Data Shows

Israel is one of the few countries with a national medical cannabis programme old enough, and centralised enough, to say what actually happens when children are treated. Paediatric licences run through the same Ministry of Health machinery as adult ones, every dispensing event is recorded in a single national database, and Israeli hospitals ran some of the world's first controlled trials of cannabinoids in children. That combination has produced an unusually honest picture — and it is more complicated than either the enthusiasm or the alarm around the subject suggests.

This guide is general information, not medical advice. Cannabinoid treatment in children carries real risks and should only be considered with a paediatric specialist. Nothing here is a recommendation to start, stop or change any treatment.

Which children are actually treated

A nationwide cohort study published in Frontiers in Pharmacology in November 2025 analysed every child under 12 who received medical cannabis in Israel between 2018 and 2022, drawn from the Medical Cannabis Unit's dispensing database via the Health Ministry's Timna research platform. It found 1,341 children across four indications (Frontiers in Pharmacology, 2025):

• Autism spectrum disorder — 751 children, by far the largest group, mean age 7.9 years
• Refractory epilepsy — 330 children, mean age 5.7 years
• Tourette syndrome — 165 children, mean age 8.3 years
• Paediatric cancer — 95 children, mean age 5.3 years

Ages ranged from 1 to 11. In autism, epilepsy and Tourette syndrome most children received high-CBD products; only in paediatric cancer did balanced or high-THC preparations predominate, reflecting their use for nausea and appetite alongside chemotherapy. That pattern is worth holding on to, because the public conversation about "cannabis for children" tends to imagine THC, while Israeli paediatric practice is overwhelmingly cannabidiol-led.

How a child gets a licence

The route is the adult route, applied more restrictively. Paediatric use sits under the Ministry of Health's Procedure 106, administered by the Medical Cannabis Unit (Yakar), and requires an authorised specialist physician to submit the request against defined clinical criteria. Licences are issued for a defined period, typically three to twelve months, after which they must be renewed (Frontiers in Pharmacology, 2025).

This is not a route a parent can initiate alone, and it is not a first-line therapy: paediatric indications are framed around refractory cases, where conventional treatment has failed. The general mechanics of applying, and of the April 2024 reform that moved most adult indications to the health funds, are covered in how to get a medical cannabis licence in Israel and medical cannabis licence renewal; for the indication list, see qualifying conditions.

Epilepsy: the strongest evidence

Refractory paediatric epilepsy is where cannabinoid treatment has the firmest footing, and Israel contributed early to that evidence base. A 2016 study in Seizure reported the Israeli clinical experience with CBD-enriched cannabis in children with intractable epilepsy, at a point when interest was running well ahead of the data (Seizure, 2016).

The regulatory position has since caught up with a purified product. Israel's Ministry of Health approved purified cannabidiol in 2021 as an add-on antiseizure medicine for patients over two years old with Dravet syndrome and Lennox-Gastaut syndrome — the same indications on which international regulators approved it. A multicentre retrospective study across five Israeli medical centres reviewed 139 children and young adults treated with purified CBD for drug-resistant epilepsy between 2018 and 2022; the commonest diagnoses were Lennox-Gastaut syndrome (37.4%) and Dravet syndrome (16.5%) (Pediatric Neurology, 2023).

Consistently with that, epilepsy showed the highest treatment persistence of the four paediatric indications in the national cohort — children stayed on it longer than children treated for anything else, which is the pattern one expects when a treatment is working.

Autism: the honest reading

Autism is the largest paediatric indication in Israel and the one where the evidence is least settled — a mismatch worth sitting with.

The landmark study is a placebo-controlled double-blind trial run at Shaare Zedek Medical Center in Jerusalem, published in Molecular Autism in 2021. It randomised 150 participants aged 5 to 21 to whole-plant cannabis extract at a 20:1 CBD:THC ratio, purified cannabinoids at the same ratio, or placebo, for 12 weeks with crossover (Aran et al., Molecular Autism, 2021).

The results were genuinely mixed:

• The registered primary outcome — behavioural problems on the Home Situation Questionnaire — showed no difference between groups.
• On the co-primary Clinical Global Impression scale, disruptive behaviour was much or very much improved in 49% on whole-plant extract versus 21% on placebo (p = 0.005).
• Social Responsiveness Scale scores improved by a median of 14.9 points on whole-plant extract versus 3.6 on placebo (p = 0.009).
• There were no treatment-related serious adverse events, but somnolence was reported by 28% and decreased appetite by 25% on whole-plant extract, against 8% and 15% on placebo.

The authors' own conclusion is the fair summary: the preparations were well tolerated, but "evidence for efficacy of these interventions are mixed and insufficient," and further testing is recommended. Anyone citing this trial as proof that cannabis treats autism is overreading it; anyone citing the null primary outcome as proof it does nothing is underreading it.

The finding parents should know: most families stop

The most practically important result in the Israeli data is not about efficacy at all. It is about how long treatment lasts.

Across all four indications, fewer than 60% of children were still on treatment at three months, and in paediatric cancer and Tourette syndrome persistence fell below 50% at that point. Extending the window, roughly 60–70% of children discontinued within six months. In the autism sub-analysis, only 26% were still being treated at twelve months (Frontiers in Pharmacology, 2025).

Children discontinued more readily than adolescents or young adults on the same treatments — the adjusted hazard ratio for discontinuation among young adults was 0.83 relative to children (95% CI 0.71–0.96).

An open-label sub-study that followed 59 children with biweekly monitoring recorded why. The leading reasons were adverse events and perceived lack of efficacy, with bureaucratic and logistical difficulty — supply gaps, licence renewal, cost — accounting for a further share over the longer term. Where doses were reduced, 58% of those reductions were linked to adverse events, predominantly psychiatric, neurological and general reactions such as fatigue, restlessness, agitation and crying. The authors estimate conservatively that 20–30% of children treated for autism experienced adverse effects that changed the course of their therapy.

Two further details cut against the standard "start low, go slow" picture. Many trajectories involved reducing the THC:CBD ratio and the dispensed amount over time rather than titrating up — a signature of adverse effects rather than optimisation. And persistence was markedly better in the closely monitored open-label group than in routine care at the six-month mark, before converging once follow-up intensity equalised, which suggests that the quality of medical follow-up is doing real work independent of the medicine.

What this means in practice

For a family weighing this, the Israeli data supports a small number of concrete points:

• Expectations are the main failure mode. Discontinuation clusters early and is driven substantially by hoped-for improvement not materialising. Agreeing explicit, observable treatment goals with the specialist before starting is the single highest-value step.
• Follow-up intensity matters. Close monitoring measurably improved persistence. A treatment plan without scheduled review is a weaker plan.
• Adverse effects are common and often behavioural. Somnolence, appetite change, restlessness and agitation are the things to watch for, and they frequently prompt dose reduction rather than escalation.
• The indications are not equivalent. Refractory epilepsy has approved purified-CBD treatment and the best persistence; autism has a large treated population and unsettled efficacy evidence.
• Plan for the paperwork. Licence renewal, supply and cost were themselves recorded causes of discontinuation, not footnotes.

The bottom line

Israel licenses cannabis for children in tightly defined refractory conditions, overwhelmingly as high-CBD preparations, through a specialist-led process under Procedure 106. The evidence is strongest in refractory epilepsy, where a purified cannabidiol product has been approved since 2021 for Dravet and Lennox-Gastaut syndromes. In autism — the largest paediatric group by some distance — the best Israeli trial found real signals on secondary measures and nothing on its primary one.

And in routine practice, most children stop within six months, usually because of side effects or because the improvement families were hoping for did not arrive. That is not an argument against trying it in appropriate cases. It is an argument for going in with realistic expectations, a specialist who will follow up properly, and an agreed point at which you will stop.

For the wider patient picture, see our Patient Access hub; for the Israeli institutions behind this evidence base, see Israel's cannabis research institutions.