A circadian feedback app reduced mood episode recurrence against a sham

Design

Multicentre, double-blind, sham-controlled randomised trial, 12 months

Population

93 adults with major depressive disorder or bipolar disorder (80 analysed)

Primary outcome

Number of recurrent mood episodes per participant

Effect

Recurrence rate higher with sham, IRR 3.39 (95% CI 1.86–6.17); time to recurrence HR 3.03 (1.58–5.81)

A Korean multicentre trial randomised 93 adults with major depressive disorder or bipolar disorder to a smartphone app that used passive phone and wearable data to give personalised sleep and activity-timing feedback and short mood forecasts, or to a sham app that looked identical but gave feedback designed not to change behaviour. Both participants and assessors were blinded, and follow-up lasted 12 months.

In the 80 people analysed, recurrences were about three times more frequent with the sham app (incidence rate ratio 3.39, 95% CI 1.86–6.17). Days spent in an episode and time to first recurrence also favoured the active app, and no significant adverse effects were reported.

The design is stronger than most digital psychiatry trials because the control was a credible sham. But the sample is small, 13 randomised participants were not in the analysis, depression and bipolar disorder were pooled, and the app is not available outside the study. It supports what clinicians already advise, keeping sleep and daily routines regular, rather than a new tool to prescribe. It was published in June 2026.

• Ask about sleep and wake times, not just sleep quantity, at every mood disorder review.
• Encourage regular daily routines as part of relapse prevention in bipolar disorder and recurrent depression.
• Treat this app as research-stage: it is not available to prescribe.
• When patients use mood-tracking apps, ask what they do with the feedback and whether it helps.

Why it matters

A sham-controlled trial suggests rhythm-focused feedback, not just app use, may prevent relapse.

Don't overread it

Eighty analysed patients in one country, with two disorders pooled, is not enough to generalise the effect size.

The statistics, in plain English

An incidence rate ratio of 3.39 means the sham group had about three times as many recurrences per person as the active group. The confidence interval (1.86–6.17) is wide, which is typical of a small trial: the true effect could be much smaller. Analysing 80 of 93 randomised people (modified intention-to-treat) can bias results if those left out differ between arms.

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.