Study-ready. Partner-ready.
No patients enrolled yet.
12-week randomized protocol on whether real-time pattern interrupt during GLP-1 maintenance reduces weight regain after discontinuation. Protocol drafted, IRB pathway mapped, N=80 target. We’re looking for telehealth prescribers, obesity-medicine clinics, and research labs to run it with us.
Status
Protocol
Drafted v0.9
IRB pathway
Designed for minimal-risk expedited review (Cat. 7)
Sample size
N = 80 (1:1 randomized)
Duration
12 weeks active + 90d follow-up
Status
Open for partner enrollment
GLP-1s suppress appetite while you take them.
The autopilot is still there when you stop.
Published discontinuation literature shows ~two-thirds of weight loss returns within a year of stopping a GLP-1 (Wilding et al., 2022 — STEP 1 extension, NEJM / Diabetes, Obesity and Metabolism), with the largest regain occurring in the first 90 days. The drug suppressed the hunger signal, but it never touched the late-night kitchen loop, the stress-eat reflex, or the “I deserve this” script. When the suppression lifts, the script is right where the user left it.
COYL is built to interrupt that script in real time. The hypothesis is that training the interrupt during the medicated window builds behavioral muscle memory that survives discontinuation. That’s the question this study answers.
Three arms. Twelve weeks. One real question.
Recruitment
N = 80
Randomized 1:1 to intervention or control. Stratified by baseline BMI and GLP-1 type.
Intervention arm
Rx + COYL Premium
GLP-1 prescription per usual care, plus 12 weeks of COYL Premium access (rescue flows, recovery engine, precision interrupts, pattern detection).
Control arm
Rx alone
GLP-1 prescription per usual care. Waitlist offered Premium after the 12-week active phase. Reduces ethical concern around withholding a promising consumer tool.
Week 0
Baseline weight, consent, randomization
Weeks 1–12
Active intervention, in-app data + weekly weight
Week 13–24
Discontinuation window, +30/+60/+90 weight
Week 25
Dataset lock, analysis, readout
Pre-specified. Reproducible from the events table.
Primary
Weight regain at 90 days post GLP-1 discontinuation. Effect-size estimation, not powered for confirmatory inference. H₁: COYL + Rx arm regains less weight than Rx-alone arm.
Secondary
Program-adherence rate (% of weeks with ≥3 logged check-ins), late-night-eating frequency from self-report, time-from-slip-to-recovery in hours, study retention at week 12.
Exploratory
Correlation between Patterns Defeated count (in-app metric) and weight outcome. Excuse-category distribution shifts pre/post intervention. Self-trust score trajectory.
Sample size justification: feasibility + effect-size estimation, not confirmatory inference. N = 80 powered to detect a ~3 kg differential weight regain at 90 days (effect size d ≈ 0.55, α = 0.05, two-sided, 80% power) assuming 20% attrition. A confirmatory replication would target N ≥ 200.
Inclusion
- +Adults 18–65, BMI 27–45
- +Prescribed a GLP-1 (semaglutide, tirzepatide, liraglutide) within the last 90 days
- +iOS or Android smartphone, English fluency
- +Willing to log weight at baseline, week 12, and at +30 / +60 / +90 days post-discontinuation
Exclusion
- −Active or historical eating disorder (within last 12 months)
- −Current psychiatric crisis or active suicidal ideation
- −Concurrent enrollment in another behavior-change study
- −Pregnant, lactating, or planning pregnancy during study window
Minimal risk. Expedited pathway. Partner-ready DUA.
Risk profile
Designed as a minimal-risk behavioral study. No medical intervention, no medication change, no PHI collected without explicit consent. Structured for expedited IRB review category 7 (research on individual or group characteristics or behavior), pending board determination.
Privacy & data
De-identified outcome data only, exchanged under a Data Use Agreement (DUA). No raw chat content shared with partner. HIPAA-aligned handling end-to-end. BAA available for prescriber partners covered as Business Associates. 90-day retention post-readout, then deletion or anonymized aggregate only.
Withdrawal
Participants may withdraw at any time. In-app data deletion within 30 days of withdrawal request. Withdrawal does not affect their COYL Premium access — they keep the product.
Adverse events
Behavioral study; no expected adverse events from the intervention itself. Standard safety reporting for any participant-reported concern, escalated to PI within 24h.
COYL provides
- •Premium product access for all enrolled participants (12 weeks active + 90-day follow-up window)
- •Engineering integration (deep links from partner platform, optional SSO)
- •De-identified outcome dataset, partner-shared under DUA
- •Co-authored manuscript, primary authorship negotiable
- •Statistical analysis support (pre-registered SAP)
- •Branded onboarding + cohort identification
Partner provides
- •IRB submission (commercial IRB acceptable — typical $3–5k fee, COYL covers if needed)
- •Recruitment from your existing GLP-1 patient cohort
- •Weight measurements at baseline, week 12, and three follow-up windows (telehealth scales acceptable)
- •Optional: short clinician interview qualitative arm
Eight months from kickoff to readout.
- 01
Month 1
Protocol finalization, partner DUA, IRB submission
- 02
Month 2
IRB approval (expedited timeline ~3–6 weeks), recruitment ramp
- 03
Months 3–5
Active 12-week intervention, in-app + scale data collection
- 04
Months 6–7
90-day post-discontinuation follow-up, dataset lock, analysis
- 05
Month 8
Readout, manuscript draft, conference abstract submission
Bring the cohort.
We’ll bring the protocol.
One email opens it. We respond within two business days with the full protocol pack, draft DUA, and a kickoff call link. Co-authored publication welcome. De-identified data sharing under a DUA welcome.
COYL is a behavioral support tool. It is not a medical device, treatment, or therapy. The study described here is a behavioral feasibility study, minimal-risk category, conducted under IRB oversight. Not a substitute for professional medical care.