Home › Research › Dutch Protocol

Criticism of the Dutch Protocol (VUmc/Cohen-Kettenis)

The Dutch Protocol — developed at the VUmc Amsterdam under the leadership of Peggy Cohen-Kettenis and Annelou de Vries — became the global model for pediatric gender medicalization. N=70, no control group, one participant died perioperatively, and the study was continued. Cass (2024): "research not robust". An ideological belief elevated to an international standard without any validation.

What the protocol entailed

From around 1996, Cohen-Kettenis and colleagues developed a three-step protocol: psychological screening, puberty blockers from Tanner 2–3, cross-sex hormones from 16, and surgery from 18. At the time, it was intended for a specific group: early-onset, dysphoric, mentally stable adolescents — predominantly boys. A global treatment industry was built on that narrow basis.

The core studies

The evidence for the protocol came from two publications: De Vries et al. (2011) — 70 adolescents, follow-up until after hormones — and De Vries et al. (2014) — 55 of them, follow-up until after surgery. Two publications, one center, one cohort, no replication. On this meager basis , WPATH implemented SOC7 and SOC8 . See also the pediatric rollout from 2000 .

Problem 1: no control group

No group without medical treatment was compared. Consequently, no claim can be made that the treatment works better than psychotherapy alone. Biggs (2022) reanalyzed the original data and shows that a significant portion of the "improvement" was attributable to natural development, family context, and social affirmation — not to the medical intervention. The lack of a control group is not a practical inconvenience but a fatal methodological flaw. The Cass Review (2024) explicitly mentions this.

Problem 2: postoperative death, continued

One participant died perioperatively from a complication of vaginoplasty (necrotizing fasciitis after neo-vagina construction). The study was continued, the death downplayed. No other pediatric study would continue unchanged after a death. The fact that this did happen here shows how deeply ingrained the ideological belief in "true identity" was: a dead patient was acceptable collateral damage.

Problem 3: dropout and selection

Of the 70 who started with blockers, 15 dropped out (21%). The remaining group was strongly selected based on favorable characteristics: early onset, mentally stable, and a solid family context. The outcomes are therefore not generalizable — certainly not to the current population, which has a radically different composition (see ROGD and the shift after 2010 ). Yet this happens in every gender clinic worldwide.

Problem 4: Selective outcome measures

Gender dysphoria was measured lower post-hoc — an effect that is entirely expected (given large-scale social and medical affirmation). Mental well-being improved only marginally. Functional outcomes regarding work, relationships, fertility, and sexual function were not reported or were weakly reported. Biggs (2022) concluded after reanalysis: the improvements are smaller than the narrative suggests, and the damage items (bone health, brain development, sterilization) were systematically downplayed.

Problem 5: not replicable

The Karolinska Clinic attempted to replicate the Dutch Protocol — and failed. Its own Swedish outcome data were worse. Consequently, Karolinska discontinued routine application (2021). The same happened at the GIDS cohort in England; the Cass Review ultimately led to the closure of Tavistock GIDS (2024). A protocol that proved irreplicable in two countries remains fully in effect in the Netherlands.

What this means

The Dutch Protocol — the basis of international policy — is built on 55 carefully selected adolescents, without a control group, with a concealed death, selective outcome measures, and proven non-replicability. It has been scaled up internationally to tens of thousands of children with fundamentally different characteristics. That is not a scientific error — that is institutional capture. Dutch institutions continue to defend the protocol and actively deny Cass. Criticism is silenced; colleagues who express doubts are bullied out. See also the WPATH Files, which show exactly the same dynamic.