Health. Reconnected.
The cost of dropout isn't just time and money. It's the findings you can't defend.
Every dropout is a participant who was willing to contribute to the evidence base — and a trial that couldn't retain them.
Every existing tool depends on the participant doing something. That's the failure mode.
15-30%
dropout in late-phase trials
80%
of trials delayed by 1+ month
$600K-$8M
per day of delay
Helping trial teams stay connected to how participants are doing
Operational intelligence from changing behavioural trajectories
No questionnaires. No wearables. No participant effort.
What we learn from:
- Sleep regularity
- Social contact patterns
- Smartphone entropy
- Night-time activity
Operational
Helps trial teams notice when participants may need additional support — enabling proactive, caring outreach at the site level.
Scientific
Continuous objective endpoint data. Behavioural markers as pre-registered or exploratory outcomes in your trial design.
What changes with Rumii?
Participants don't drop out because of the disease. They drop out because of what the disease does to their behaviour — and nobody has the context to intervene.
Sophie, 28 — enrolled in a Phase III oral JAK inhibitor trial for Crohn's
Week 12 of 52-week study · responding well · CDAI improving · next visit in 4 weeks
Week 12
Study visit — responding well
Sophie’s CDAI has improved since baseline. She reports feeling better. Endoscopy shows mucosal improvement. The drug appears to be working. Next visit in 4 weeks.
Week 14
Life stress escalates
Sophie’s relationship breaks down. She moves back in with her parents. Sleep becomes fragmented. Activity contracts. She stops socialising. The behavioural preconditions for a stress-driven flare are building. Nobody in the trial team knows.
Week 15
Flare symptoms emerge
Abdominal pain returns. Urgency increases. Fatigue worsens. Sophie’s first thought: the drug has stopped working. She doesn’t connect the flare to the behavioural disruption of the past two weeks.
Week 15.5
Sophie withdraws
She emails the site coordinator: “I don’t think the medication is working anymore.” A participant who was genuinely responding to a novel therapeutic has been lost — not because of treatment failure, but because a stress-driven flare was indistinguishable from non-response.
Outcome: Participant lost
Responding to treatment, withdrew due to misattributed flare. One fewer completer in the efficacy analysis. Recruitment extension likely.
Intelligence products
Retention Risk Signal
Identifies when a participant may be struggling — before they disengage — so site teams can offer timely, proactive support.
Retention Analytics
Cohort-level wellbeing insights across sites and arms. Surfaces patterns that help teams anticipate and address retention challenges before they materialise.
Behavioural Health Endpoint Analytics
Publishable, continuous objective endpoint data. Sleep regularity, social contact patterns, and behavioural entropy as pre-registered or exploratory outcomes.
Deployment models
SDK Embedded (Decentralised)
Rumii SDK embedded in your existing trial app or eCOA platform. Runs quietly in the background with zero additional burden on participants.
Standalone App (Site-Based)
Rumii app deployed directly to participants at site. Continuous wellbeing insight with optional participant-facing support features.
SDK deployment in action
Your app knows when they stop opening it. It doesn't know why. Rumii's SDK provides the behavioural intelligence layer that turns reactive re-engagement into proactive support.
Nadia, 31 — enrolled in an RCT evaluating a CBT-based digital therapeutic
PHQ-9: 14 (moderate depression) · active arm · 12-week programme · 3 of 8 modules completed
Week 4
Engagement steady
Nadia has completed 3 modules. She logs in 4–5 times per week. PHQ-9 at week 4: 11 — early improvement. The product dashboard shows her as an engaged participant.
Week 5
Life disruption — invisible to the app
Nadia’s hours at work increase. She’s put on night shifts. Sleep becomes fragmented. Social contact drops. Her daily structure disintegrates. None of this is visible inside the app.
Week 6
In-app engagement drops
One login. No module completion. The dashboard flags declining engagement but has no context. A generic push notification fires: “You’re making great progress! Time for your next module.” Nadia ignores it. It feels tone-deaf.
Week 7
Withdrawal phenotype emerges
Nadia hasn’t opened the app in 9 days. Sleep is severely dysregulated. She’s socially withdrawn. The depression the DTx was treating has worsened — driven by behavioural collapse, not treatment failure.
Week 9
Lost to follow-up
Nadia doesn’t complete the week 8 assessment. She’s formally classified as a non-completer. Her data is missing from the primary endpoint analysis. The trial has lost a moderate-severity participant.
Outcome: Participant lost
Classified as non-completer. Data missing from efficacy analysis. Non-random dropout skews remaining completers toward milder cases, weakening the evidence.