Open Dialogue did not shorten relapse, but cut admissions

Design

Multisite, two-arm, cluster-randomised controlled superiority trial, England

Population

494 adults presenting in crisis to community mental health services across 30 clusters

Primary outcome

Time to first relapse after initial recovery over 2 years

Effect

Marginal hazard ratio 0.95 (95% CI 0.67 to 1.32); lower admissions and better experience as secondary outcomes

The ODDESSI trial randomised 30 clusters across five NHS trusts in England, enrolling 494 adults presenting in crisis, to Open Dialogue or treatment as usual. Open Dialogue is a service-wide model built on network meetings with the person, their social network and usually two practitioners who stay with them through care.

On the primary outcome — time to first relapse after initial recovery over two years — there was no difference (marginal hazard ratio 0.95, 95% CI 0.67 to 1.32). But the secondary picture favoured Open Dialogue: lower rates of psychiatric admission and of re-referral to crisis or secondary services, and better self-rated recovery, quality of life and satisfaction. Serious adverse events were fewer in the Open Dialogue arm, and almost all were judged unrelated to the intervention.

So the honest reading is mixed. Open Dialogue did not change the relapse course, but it changed how care was experienced and how acute beds were used. Whether that justifies the service redesign it requires is the open question the authors leave.

• Cluster-randomised across 30 clusters and five NHS trusts; 494 adults in crisis.
• No difference in time to first relapse: hazard ratio 0.95 (95% CI 0.67 to 1.32).
• Open Dialogue lowered psychiatric admissions and re-referral to crisis or secondary care.
• Self-rated recovery, quality of life and service satisfaction improved.
• Serious adverse events were fewer in the Open Dialogue arm and almost all judged unrelated.

Why it matters

It tests a model promoted for better outcomes and finds its value is in bed use and experience, not in the relapse course it is often sold on.

Don't overread it

The admission and experience gains were secondary outcomes in a trial whose primary endpoint was negative, so they are hypothesis-supporting rather than definitive.

The statistics, in plain English

The relapse hazard ratio of 0.95 with an interval spanning 1.0 means no detectable difference; the secondary gains are real but secondary, so they support the model more weakly than a positive primary outcome would.

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.