Industries / Health / Healthcare / Weight-loss clinics
Follow-up
Follow-up is what happens once treatment has started: whether people stay on a drug or program, what happens to their weight when they stop, and referral to weight-loss surgery and care after it.
7
Milestones achieved, from 1967 to 2026.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1967
Stomach bypass replaces bowel bypass
The jejunoileal bypass, common in the 1960s and 1970s, caused weight loss by shortening the working length of the gut; many patients later died of its complications, had it reversed or needed liver or kidney transplants. In 1967 E. E. Mason and C. Ito described gastric bypass, an operation on the stomach instead. Its later form, the Roux-en-Y gastric bypass, and the sleeve gastrectomy now make up about 90% of weight-loss operations worldwide, by the count of the American Society for Metabolic and Bariatric Surgery (ASMBS).
LessonHarm from an operation can appear years later, which is why a 1991 US National Institutes of Health panel called "lifelong medical surveillance after surgical therapy" a necessity.
Credit
E. E. Mason and C. Ito
2007
Surgery linked to fewer deaths
The Swedish Obese Subjects study followed 2,010 people who had weight-loss surgery and 2,037 matched people given usual care for an average of 10.9 years. Ten years on, the surgery groups kept off 14% to 25% of their weight, depending on the operation, while the usual-care group's weight changed by less than 2%. There were 101 deaths after surgery against 129, an adjusted hazard ratio of 0.71, meaning about 29% lower risk of death once differences between the groups were allowed for. The groups were matched, not randomized.
Credit
L. Sjöström and colleagues, Swedish Obese Subjects study
2013
Accreditation rule dropped
From 2006 Medicare paid for weight-loss surgery only at centers certified by the American College of Surgeons or the American Society for Bariatric Surgery. Using 2004–2009 hospital data from 12 states, a University of Michigan team found no improvement in complications or repeat operations among Medicare patients beyond trends already under way, and no difference between certified and other hospitals. On 24 September 2013 CMS removed the requirement, concluding that keeping it "would not improve health outcomes".
LessonA certificate earns its place only when outcomes show a difference.
Credit
J. B. Dimick, J. D. Birkmeyer and colleagues, who asked for the change; CMS
2022
Surgery criteria widened
A 1991 US National Institutes of Health consensus statement limited surgery to people with a BMI of 40 or more, or 35 or more with a condition such as high blood pressure, and most insurers still relied on it 30 years later. In October 2022 the ASMBS and the International Federation for the Surgery of Obesity and Metabolic Disorders (IFSO) replaced it: surgery is recommended from a BMI of 35 whatever other conditions a person has, should be considered from 30 with metabolic disease, from 27.5 for Asian people, and in selected children and adolescents.
Credit
ASMBS and IFSO (D. Eisenberg, S. A. Shikora and colleagues)
2022
Weight returns after stopping
In the off-treatment extension of Novo Nordisk's STEP 1 trial, 327 people who had lost a mean 17.3% of their weight on semaglutide regained about two-thirds of that loss within a year of stopping, and most gains in blood pressure, blood sugar and blood fats faded. In Eli Lilly's SURMOUNT-4 trial, people switched from tirzepatide to a placebo regained 14.0% of their weight over a year, while those who stayed on it lost a further 5.5%. An Oxford review of 37 studies with 9,341 people, published in January 2026, confirmed regain of about 0.4 kg a month after these medicines stop, faster than after behavioral programs.
LessonThe drugs control weight while they are taken; in the STEP 1 authors' words, the findings "confirm the chronicity of obesity".
Credit
J. P. H. Wilding and colleagues (STEP 1); L. J. Aronne and colleagues (SURMOUNT-4); S. West, S. A. Jebb, D. A. Koutoukidis and colleagues, University of Oxford
2025
Most stop within a year
In health records from a group of US health systems, 125,474 adults with overweight or obesity started a GLP-1 drug between 2018 and 2023. Within one year 64.8% of those without type 2 diabetes and 46.5% of those with it had stopped; greater weight loss and higher income went with staying on, and moderate or severe stomach side effects with stopping. In Israel, records of 34,309 adults who started Wegovy in early 2025 at Clalit Health Services showed that 67.6% had stopped within 365 days.
Credit
P. J. Rodriguez, E. J. Emanuel and colleagues, Truveta; E. Winter, G. Lavie and colleagues, Clalit Health Services
2026
Surgery falls as drugs rise
Using Epic's Cosmos database of US health records, an NYU team found that sleeve gastrectomy and gastric bypass operations peaked in the last quarter of 2022 and fell 39% by the last quarter of 2025. Over the same years the share of surgery patients who had been given semaglutide or tirzepatide before their operation rose from 0.2% to 35.3%.
Credit
A. Kozato, K. R. Chhabra and colleagues, NYU Grossman School of Medicine
Watch for
Switching to a pill
In the ATTAIN-MAINTAIN trial, people who had lost weight on injected tirzepatide or semaglutide were switched for 52 weeks to orforglipron, a daily GLP-1 pill, or to a placebo. Those switched from tirzepatide kept a mean 74.7% of their loss, against 49.2% on placebo; those switched from semaglutide kept 79.3%, against 37.6%. The results appeared in Nature Medicine on 12 May 2026. Eli Lilly, which makes orforglipron and tirzepatide, sponsored the trial, and the lead author is a paid Lilly consultant.
Achieved whena trial not sponsored by the maker, or a regulator's published review, confirms that switching to the pill keeps most of the loss for at least a year.
Claimed by
Eli Lilly and investigators at Weill Cornell Medicine, May 2026
Open
01
A tracker for maintenance trials
A public table, refreshed every month, of registered trials that test a way to keep weight off after a weight-loss drug: staying on it, tapering it, switching to a pill, a diet or exercise program, a device or an endoscopic procedure. Each row gives the strategy, sponsor, size, status, length of follow-up and, once posted or published, the weight result, with links to the registry entry and the paper.
The briefCredit · unclaimed02
Count everyone who started
An open calculator and a short reporting standard that show how a service's average weight loss changes with who is counted: only people still enrolled and reporting at one year; everyone who started, with dropouts' last weights carried forward; or everyone, with dropouts assumed to regain at published rates. It runs on synthetic member records tuned to published stopping and regain rates, and comes with a coded list of published real-world studies showing which counting rule each one used.
The briefCredit · unclaimed
Open 01 · the brief
A tracker for maintenance trials
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public table, refreshed every month, of registered trials that test a way to keep weight off after a weight-loss drug: staying on it, tapering it, switching to a pill, a diet or exercise program, a device or an endoscopic procedure. Each row gives the strategy, sponsor, size, status, length of follow-up and, once posted or published, the weight result, with links to the registry entry and the paper.
Why it is still open
A systematic review published in October 2026 found 18 reports from 16 families of studies and described them in words, without a pooled estimate, because their measures, follow-up and added treatments differed. On 20 November 2026 a ClinicalTrials.gov search for obesity trials of these drugs with maintenance, withdrawal, cessation or tapering in the title returned 26 studies, some of them off topic; 10 were recruiting or not yet recruiting, and 2 had posted results. Lilly's completed ATTAIN-MAINTAIN trial was published in May 2026 but showed no posted results on the registry.
Built with what exists
The ClinicalTrials.gov API (version 2) and the WHO International Clinical Trials Registry Platform for registrations; Europe PMC for published results; an AI model such as Claude to draft the strategy coding from each registration, checked by a person; a static page on GitHub Pages.
First step
Pull the 26 ClinicalTrials.gov records found here, drop those that are not about keeping weight off after a drug, and code the rest by strategy.
Done looks like
Every matching registration in both registries coded by two people, with their agreement reported; for each strategy, the number of trials, the number of people enrolled and the number with results; monthly updates for a year, each change logged.
Open 02 · the brief
Count everyone who started
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open calculator and a short reporting standard that show how a service's average weight loss changes with who is counted: only people still enrolled and reporting at one year; everyone who started, with dropouts' last weights carried forward; or everyone, with dropouts assumed to regain at published rates. It runs on synthetic member records tuned to published stopping and regain rates, and comes with a coded list of published real-world studies showing which counting rule each one used.
Why it is still open
Most people stop these drugs within a year (64.8% of US adults without diabetes in one large records study), so a result counted only among people who stay describes those who stayed. WeightWatchers' 29% figure, for example, comes from 3,325 members who reported a week-52 weight. The international guideline on estimands, ICH E9(R1), where an estimand is the exact effect a study sets out to measure, defines such choices for drug trials, but a search of Europe PMC in November 2026 found no tool or reporting standard that applies them to weight-loss services.
Built with what exists
Python or R; the published one-year stopping curves (Rodriguez and colleagues, 2025; Winter and colleagues, 2026) and regain rates (West and colleagues, 2026) to generate synthetic records; the estimand terms of ICH E9(R1) to name each counting rule; a small web page for the calculator.
First step
Build 10,000 synthetic records that match the US one-year stopping rates and compute the average loss among completers and among everyone who started.
Done looks like
The calculator reproduces the published one-year stopping rates within 2 percentage points and reports, for a synthetic cohort, the average loss under at least four counting rules; at least 30 published real-world studies coded by two people for the rule they used, with agreement reported; code and data public.