Bergen four-day OCD treatment beat an unguided online course, 42% to 14% in remission

Therapy for obsessive-compulsive disorder, or OCD, usually means a weekly appointment, kept up for months. In Germany a standard outpatient course may comprise up to 24 sessions, and finding a therapist trained in the right method can take a long time. A treatment devised in Bergen, Norway, squeezes the same core work into four consecutive days, and a randomized trial in Hamburg has now tested it against an active treatment for the first time.

The trial, published on October 19 in JAMA Network Open, randomly assigned 119 adults with OCD to the four-day treatment or to an eight-week online self-help course. The four days brought greater reductions in OCD symptoms, and the four-day group was still well ahead at the three-month check, even as the gap narrowed: 42% of those assessed were in remission, against 14% of those assessed after the course. The team at the University Medical Center Hamburg-Eppendorf, led by Lena Jelinek, ran the trial independently of the Norwegians who developed the method, and its results are noticeably less dazzling than theirs, which may be the most useful thing about them.

What happens in the four days of the Bergen treatment

OCD traps people in a loop of unwanted, distressing thoughts and the rituals they perform to quiet them: washing, checking, counting, repeating. It has a lifetime prevalence of 1% to 3% and is frequently undertreated. Exposure and response prevention, usually shortened to ERP, is the recommended first-line psychological treatment: a person deliberately faces the situations their obsessions are about and then holds back from the ritual. The Hamburg authors note that shortages of trained therapists, long waiting times, refusal and dropout all keep it from many of the people it could help.

The Bergen 4-day treatment keeps ERP and changes the schedule, compressing treatment traditionally delivered over several months into 4 days after a short preparation. In Hamburg, that preparation was two individual 50-minute sessions. Patients then met in small groups, typically 4 patients, each with a therapist of their own. The first day covered how OCD works and a personal plan of exposures. On days two and three, each person worked through exposure exercises in the real places and situations that set off their symptoms, with their therapist beside them all day, and kept practicing alone in the evening and the next morning. Relatives came to a session of their own on the third afternoon. The fourth day went on consolidating what had changed and on planning the next month of solo practice, and a booster session followed about three months later.

Throughout, the method asks people to lean in. Under what its developers call the Leaning into Anxiety protocol, patients are guided to intentionally engage with feared stimuli, choosing to approach the thing they dread.

Norway’s striking numbers for the Bergen treatment, and why they needed a harder test

The format made its name in Norway. By the autumn of 2016, Oslo University Hospital had a waiting list of 101 patients with OCD who were entitled to specialist care. The Bergen team offered all of them the four-day format, and 90 received it within eight days. At three months the remission rate was 67.7%, and Oslo University Hospital adopted the four days as its standard OCD care.

Almost all of that record came from studies with no comparison group. The one earlier randomized trial, also by the Norwegian developers, split 48 patients diagnosed with OCD between the four-day format, a self-help book and a waiting list. Its remission rate was 62.5%, against 6.3% and 0% in the other two arms. The Hamburg authors point out that waiting list comparisons may inflate estimated treatment effects, and that psychotherapy trials can be colored by the loyalties of the people running them.

That second worry has numbers behind it. A 2021 meta-analysis of ERP trials by Jemma Reid and colleagues at the University of Hertfordshire found that 28 of the 36 studies demonstrated suspected researcher allegiance, and the benefit tracked it closely: efficacy was strongly linked to researcher allegiance, and the eight trials without that suspicion showed no clear benefit. It does not follow that ERP fails, only that results reported by a method’s champions deserve a second look.

How the Hamburg trial set up a fairer test of the four-day treatment

Jelinek’s team changed three things. The comparison was an active treatment: an unguided OCD course of eight online modules over eight weeks, covering psychoeducation, exposure and response prevention, mindfulness, self-esteem and social skills. Nobody coached participants through it, though they could contact the study team about technical problems. The people who scored symptoms did not know who had received which treatment. And the starting score was taken before the preparatory sessions, where some earlier studies had taken it after, so improvement during preparation counted toward the treatment instead of quietly lowering the baseline.

The main measure was the Yale-Brown Obsessive Compulsive Scale, a clinician’s interview whose total score ranges from 0 to 40. A fall of at least 35% counted as a response, and a score of 12 or less as remission.

Between May 2024 and August 2025 the team screened 253 people. Fifty-three turned out not to be eligible and 81 declined to participate, which left 119: sixty went to the four-day treatment and 59 to the online course. Two thirds were women and the average age was 33. The earlier Norwegian trial had excluded anyone who had already had cognitive behavioral therapy, the broader family of talking treatments that ERP belongs to. This one took people as they come to clinics, and most participants had previous ERP experience. Both groups could keep any medication or therapy they were already receiving.

How many people with OCD reached remission after four days

The four-day group improved far more. Just after treatment, its average symptom score sat about 8 points below the online group’s on the 40-point scale, and three months later the gap was still between 5 and 6 points, a large effect at both times by the usual yardsticks. Counted person by person, 35 of 57 in the four-day group had responded at three months, against 11 of 44 who took the course, and 24 of 57 were in remission, against 6 of 44. That is about four people in ten in remission, against about one in seven.

Self-rated OCD symptoms, depression, quality of life and day-to-day functioning all moved the same way. Participants rated the four days far more highly as well: treatment satisfaction was markedly greater than with the course.

The online course still did real good. Its users also demonstrated clinically meaningful within-group improvements, and their scores kept falling after the eight weeks were over, which is why the distance between the two groups shrank between the end of treatment and the three-month check. That fits the wider evidence. A 2022 review of 11 trials led by Hissei Imai at Kyoto University found unguided computer-based self-help for OCD more effective than a waiting list or psychological placebo, although people abandoned it about twice as often and the quality of evidence was very low.

Set beside Norway, though, the Hamburg numbers are clearly smaller. The authors compare their 61% response and 42% remission with earlier Scandinavian studies, uncontrolled ones reporting response rates of approximately 80% and remission near 70%. They expected the gap. Studies without a control group, scored by raters who may guess what treatment someone had, tend to inflate results, and the Hamburg volunteers were a more mixed group, many of them previously treated. Even the team’s own German pilot of the format, which also lacked a comparison group, found 54% were in remission at 3-month follow-up.

Why packing exposure into four days might help

The study was not built to explain why the format works, but Jelinek and colleagues offer four ideas. The concentrated delivery of exposure over consecutive days may reduce opportunities for avoidance, leaving little time between sessions for old habits to return. Each person has a clinician at their side for the hardest moments. The group and the relatives’ session may add motivation and help people carry what they learned into ordinary life. And choosing to lean into anxiety may change how people relate to it. The authors suspect these work together, and none was tested on its own.

The result also fits a larger pattern. Yingying Wang and colleagues at Vrije Universiteit Amsterdam ran a network meta-analysis, a method that pools trials and lets treatments be compared even when no single study tested them side by side, across 61 randomized trials involving 3,710 patients with OCD. Most ways of delivering cognitive behavioral therapy were more effective than unguided self-help, and in their ranking time-intensive formats came out top for effectiveness while unguided self-help came last. The Hamburg study put those two ends of the ranking against each other directly, and the outcome matched the prediction.

What the Hamburg OCD trial left untested

The comparison most people would care about is missing. The four days were measured against a self-guided course, and the two arms differed enormously in therapist time, so part of the advantage may simply be a great deal of one-to-one support set against almost none. The authors say plainly that direct comparisons with standard face-to-face ERP require future head-to-head trials.

They list further limits themselves. The study was conducted at a single center, and follow-up so far runs only three months, with a 12-month analysis still to come. Neither patients nor therapists could be kept unaware of their assignment, which can shape expectations. How faithfully people worked through the online course may have varied across participants, and other treatment carried on in both arms, so its contribution cannot be separated out. Fewer participants from the online course came back for the three-month assessment, 44 of 59 against 57 of 60 in the four-day group, so its figures rest on a smaller base. Possible harms were recorded but left out of this report. And little is known about the 81 people who declined beyond their gender, which leaves open who would sign up for four intensive days if a clinic offered them.

Where a four-day OCD treatment would sit in care

Delivery is its own hurdle. The format requires concentrated therapist resources over 4 consecutive days, with one therapist for every patient, which many services would struggle to arrange. The authors counter that formats should be compared on the total therapist time each completed course requires, and standard outpatient therapy in Germany runs to as many as 24 sessions, with extensions possible.

They do not cast the two treatments as rivals. “Rather than representing competing approaches, these findings suggest the complementary roles of the 2 interventions within stepped-care frameworks,” they write, with internet therapy as an accessible first step and the four-day format as a more intensive option for individuals requiring more intensive care.

That is close to how existing guidance is built. The UK’s National Institute for Health and Care Excellence recommends stepped care for adults with OCD: brief therapy that includes ERP, up to 10 therapist hours per patient, for mild cases, and for moderate ones a choice between a course of an SSRI, a common type of antidepressant, and more intensive therapy with ERP, more than 10 therapist hours per patient. The guideline predates the Bergen format and does not mention it. Whether intensive treatment suits any one person is decided with their clinician.

At three months, four supported days outperformed a self-guided course by a wide margin; whether they also outperform weekly therapy is the trial still to be run.