As presented at the PHA Innovation & Adoption Summit · Washington, DC · Oct 21–22, 2026
By the time you learn a patient isn't responding to their GLP-1, you've paid for a year of medication. We can tell you in eight weeks.
Behavioral intelligence for obesity programs. EW2Health reads the behavior behind the drug, so health plans, employers and care teams know who is on track by week 8, not month 12.
The drug works. The behavior fails.
Employers and health plans have paid for GLP-1s for years. Since July 1, 2026, Medicare covers Wegovy and Zepbound for obesity through its GLP-1 Bridge demonstration. The spend is only growing.
40%
of American adults live with obesity: about 107 million people
CDC/NCHS, NHANES 2021–2023
11.4%
of total claims on US corporate plans now go to weight-loss GLP-1s
IFEBP Pulse Survey, 2026
2 in 3
people stop GLP-1 treatment for obesity within the first year
Gasoyan et al., JAMA Netw Open, 2025
~2/3
of the weight lost comes back within a year of stopping
Wilding et al., Diabetes Obes Metab, 2022
For a payer, that means paying full price for a very partial outcome.
What we found
The signal is there. Nobody is reading it.
110,000+
weight journeys of people on GLP-1 treatment, analyzed on Mayo Clinic Platform data.
A handful of typical journeys. Weight curves fall into a few typical trajectories.
They separate early. Between week 3 and week 8, you can see who is on track and who isn't.
Behavior is the strongest signal. People who keep weighing in, stay. People who stop weighing in drop out, weeks before they stop the drug.
42% vs 10%
reach ≥5% weight loss at 6 months, with vs. without an early monitoring habit
~10×
higher odds of dropping out without the habit
5.7 vs 1.9
active months on treatment
EW2Health real-world clinic cohort, intention-to-treat; n=400 (outcome), 942 (active months), 241 (dropout, 95% CI 5.2–19). Association, not causation.
What we built
The layer that reads it: capture, predict, act.
01
Capture
Our connected monitor, Sinque, shows direction, not a number. People stop fearing the scale, so they keep stepping on it, almost daily. The design is patented in the US (US 11,633,160).
02
Predict
Our models match each person to their likely trajectory within the first two months.
03
Act
The care team gets a weekly work queue: who needs attention, and whether it's a treatment issue or a behavior issue.
The AI says where to look. The clinician decides.
That's how one care manager can follow hundreds of members.
Explainable: every flag traces back to the member's own trend.
No diagnostic claims: Sinque supports the care team; it doesn't replace clinical judgment.
Privacy by design: reporting to plans and employers is aggregated and de-identified; data terms are agreed before any pilot.
The value
The signal doesn't save money on its own. The workflow you build around it does.
Sinque tells you who is likely not responding by month 2 instead of month 12. The savings come from what your program does next: a dose change, another therapy, or more intensive support. Nobody is removed from care.
Redirect earlier
Stop paying months of medication for members who aren't responding, and move them to what fits.
Recover members who are drifting
Reach people when they stop weighing in, weeks before they stop the drug.
Focus care-team time
A weekly work queue lets one care manager follow hundreds of members, and spend time where it changes the outcome.
How we work with you
Your costs, your workflow, our data.
Every program is different, so we don't forecast savings from generic assumptions. We start from your real numbers and help you design the workflow that turns the week-8 signal into savings.
01
Share your real costs.
Medication, clinical follow-up and program costs for one population. Aggregated and de-identified.
02
We model your savings together.
We build your forecast with you on your own numbers, so you can see what's realistic before you commit.
03
Implementation consultancy.
We help you design the best workflow: who gets flagged, who acts, how members are redirected, and how it fits your care-management model and vendors.
04
Pilot and measure.
Start with one population. First read-out at week 8, measured against your own cost data.
Implementation consultancy is included. Getting the workflow right is what generates the savings, so we do it with you.
Want to run a first estimate for your program? Leave your name and email to open our calculator.
You'll get instant access to our sample cost calculator for employer and health plan GLP-1 programs. Then we'll follow up to build a real forecast on your own costs.
Questions we hear
Sinque.
The drug works.
We let you see the behavior.