Open Dialogue did not delay relapse after a crisis, though admissions and experience favoured it

Design

Multisite cluster-randomised superiority trial, masked outcome assessment

Population

494 adults presenting in crisis to community mental health services in five English NHS trusts

Primary outcome

Time to first relapse after recovery from the index crisis, over 2 years

Effect

Marginal HR 0.95 (95% CI 0.67–1.32)

ODDESSI was a cluster-randomised trial in five NHS mental health trusts in England. Thirty clusters of general practices were assigned to Open Dialogue — network meetings with the person, their family and social network, and usually two continuing practitioners — or to usual care from functional community teams. It enrolled 494 adults presenting in crisis between 2019 and 2021.

The primary outcome, time to first relapse after recovery from the index crisis over two years, did not differ (marginal HR 0.95, 95% CI 0.67–1.32). Several secondary outcomes favoured Open Dialogue: lower probability of psychiatric admission and of re-referral to crisis or secondary services, and better self-rated recovery, quality of life and satisfaction. Social network size and quality did not change. Most serious adverse events (97%) were judged unrelated to either model.

For services, the honest reading is that the model's stated mechanism — changing the social network — did not show, and neither did its primary clinical benefit. The admission and experience findings are worth a further trial, not a service redesign on their own. The principle of seeing a person's family early in a crisis costs little and is compatible with any model.

• Invite family or chosen supporters to the first crisis assessment where the person agrees
• Keep the same clinicians through a crisis episode where rota allows
• Do not cite this trial as evidence that Open Dialogue prevents relapse
• Watch for further analyses on bed use and cost before commissioning decisions

Why it matters

A widely promoted model of crisis care failed its main test, and the secondary signals need confirming before they drive service change.

Don't overread it

The admission and satisfaction benefits were secondary outcomes, not what the trial was designed to show.

The statistics, in plain English

A hazard ratio of 0.95 with an interval from 0.67 to 1.32 means the trial is compatible with anything from a third fewer relapses to a third more, so it neither proves nor rules out a modest effect. The positive secondary outcomes were not the trial's main question, and testing many outcomes raises the chance that some look significant by chance.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.