Research watch · Research · 23 October 2026
From offer to care: what the HARD trial in healthcare workers teaches us
A French randomised trial, published on 20 October 2026, found no benefit at six months from an offer of EMDR to healthcare workers in distress. This result tells us very little about how effective EMDR is. It tells us a great deal, however, about the gap between a treatment that is offered and a treatment that is actually received.
By Marie-Agnès Thulliez · Author · Publishing Director
Reading time: 6 min
During the pandemic, many schemes offered psychological support to healthcare workers. The idea seemed simple: identify those who are struggling, offer them an effective psychotherapy, and wait for improvement. The HARD trial, conducted in France under the leadership of Professor Wissam El-Hage, is a reminder that real access to care matters just as much as the quality of the treatment on offer.
The study in brief
This is a pragmatic randomised trial embedded in a cohort, a design known as a trial within cohort. Of 849 healthcare professionals exposed to COVID-19 and followed up by online questionnaires every three months for a year, 237 exceeded the thresholds set for post-traumatic stress, depression or burnout. They were randomly allocated to one of two arms: an offer of access to EMDR sessions, or usual care alone.
| Element | What the trial shows |
|---|---|
| Population | Healthcare professionals exposed to COVID-19 (cohort of 849, of whom 237 were randomised) |
| Intervention evaluated | An offer of access to EMDR, compared with usual care |
| Primary outcomes | Post-traumatic stress, depression and burnout at six months |
| Main result | No statistically significant difference between the groups |
| Secondary results | Favourable signals at twelve months for depressive symptoms and post-traumatic avoidance, exploratory in nature |
| Level of evidence on the efficacy of EMDR | Low |
| Lessons on acceptability and implementation | Moderate |
This methodological point is decisive. The trial compares people who were offered EMDR with people who were not offered it, rather than people treated with EMDR with untreated people. The original protocol planned twelve sessions with a certified therapist and a far larger sample than was eventually recruited (around 900 people planned for randomisation, compared with 237).
From offer to treatment: a very narrow funnel
The most instructive figure in the study is the proportion of people who actually received what they were offered.
In other words, around two in three people in the "EMDR" group received no sessions at all. And among those who did, the median number was seven sessions (between four and twelve for the middle half of participants), well short of the twelve set out in the protocol, with wide variation from one person to another. Comparing this group with usual care therefore amounts to comparing two groups that, for the most part, were not treated with EMDR. Seen in this light, the absence of a difference is hardly surprising.
An offered treatment and a received treatment are two different things
A trial that evaluates an offer measures the effectiveness of a care pathway: it depends on the treatment itself, but also on whether it is accepted and started, on the number of sessions actually attended, and on fidelity to the protocol. A null result at this level does not allow us to conclude that the treatment is ineffective when it is genuinely carried through. Nor does it allow us to conclude the opposite.
Four concepts to keep distinct
- Acceptability: does the person say yes to what is offered? Here, 62% of those approached.
- Engagement: does the person actually start treatment? In the end, fewer than one in three.
- Dose received: how many sessions actually took place, compared with what the protocol planned?
- Fidelity to the protocol: does the treatment delivered match the treatment described? It was not systematically monitored in this trial.
These distinctions apply well beyond EMDR. They hold for any psychotherapy evaluated in real-world conditions, whether cognitive behavioural therapy, trauma-focused therapy or brief intervention. Pragmatic trials all ask the same question: does a treatment that works when it is followed still work when it is offered to people who are exhausted, overstretched and often reluctant to acknowledge that they are struggling?
Limitations to bear in mind
- Recruitment far below what was planned, and therefore insufficient statistical power to detect a modest effect.
- A low proportion of participants actually treated in the EMDR arm.
- A number of sessions that varied widely from one person to another.
- No systematic monitoring of fidelity to the protocol.
- A very specific population and context (healthcare workers, a public health crisis), which limit generalisation.
- The impossibility of clearly separating the effect of EMDR from barriers to access and engagement.
As for the signals observed at twelve months on depression and avoidance, they are encouraging but secondary and exploratory. They should be read as leads for future research rather than as a demonstration.
What this changes for practice
For practitioners, this study shifts attention from the technique to the path that leads to it. Among exhausted healthcare workers, the barriers are well known: lack of time, incompatible working hours, guilt about "looking after oneself" while colleagues keep going, fear of how peers will see them, and a fatigue that makes any additional commitment costly.
Several levers lie squarely within clinical work:
- taking care over the first contact and the explanation of the treatment, giving the person time to make it their own;
- identifying ambivalence early, rather than discovering it when an appointment is missed;
- adapting practical arrangements (time slots, remote sessions) while keeping the structure of the protocol;
- tracking the number of sessions actually delivered, as well as changes in scores.
Frequently asked questions
Does this study show that EMDR does not work for depression or burnout?
No. It allows no conclusion either way about the efficacy of EMDR, since most of the group concerned did not receive the treatment. What it mainly evaluates is an access scheme.
Can this study be cited as evidence in favour of EMDR for depression?
No, not that either. The signals observed at twelve months are secondary results from an underpowered trial. They justify further studies, and call for caution in any claim.
Why did so many people not start treatment?
The trial does not allow the causes to be pinpointed precisely. The context (healthcare workers in the midst of a public health crisis, workload, exhaustion) does, however, make practical and psychological barriers to engagement plausible.
What is a trial "within a cohort"?
It is a trial in which participants are drawn from a cohort that is already being followed up. Only those randomised to the experimental arm are offered the intervention; the others continue to be followed up as before. This design is close to real-world conditions, but it measures the effect of an offer rather than that of a treatment received.
References
El-Hage, W., Mansy, L., Gissot, V., et al. (2026). Acceptability and delivery of an offer of access to EMDR for depression, posttraumatic stress, and burnout in healthcare workers exposed to COVID-19: a randomized controlled trial within a cohort. Scientific Reports. Published 20 October 2026. https://doi.org/10.1038/s41598-026-73880-2
Caille, A., Allemang-Trivalle, A., Blanchin, M., … Hingray, C., & El-Hage, W. (2023). EMDR for symptoms of depression, stress and burnout in health care workers exposed to COVID-19 (HARD): A study protocol for a trial within a cohort study. European Journal of Psychotraumatology, 14(1), 2179569. https://doi.org/10.1080/20008066.2023.2179569 · Full text (HAL)
This article summarises the scientific evidence available at the date of publication. It is no substitute for clinical training, supervision, or the advice of a health professional.
