Digital GLP 1 care study of 8127 women reports 15.5% weight loss in six months

A Nature Medicine paper on digital GLP 1 care followed 8,127 women in the UK who started weight loss injections through a remote, self pay programme. Almost all of them received tirzepatide. After six months the average weight loss was 15.5%, and 91.4% were still on treatment, the authors report. The study had no comparison group, and two private companies that ran the programme took part in the work, so the numbers describe what happened inside one service rather than what the drug or the app caused.

What the programme looked like

The team, led by John Deanfield at the Institute of Cardiovascular Science at University College London, recruited adult women online between 27 November 2024 and 31 January 2025. Of 16,121 people who passed a first screen, 4,025 never completed the baseline blood test and 308 were excluded after clinical review for abnormal results or other medical reasons. That left 8,127 women who were prescribed and sent medicine. Everything happened at a distance. Women filled in a structured questionnaire each month before a refill was approved, joined weekly video check ins with trained staff, confirmed their weight under supervision and scanned their medication. Clinicians reviewed each woman monthly and raised the dose only when they judged it useful, rather than on a fixed timetable.

Average starting weight was about 104 kg and average body mass index 38.2, so most participants had obesity, and ages ran from 18 to 79. Around 88% were white, while 37.6% lived in the two most deprived fifths of the Index of Multiple Deprivation, which the authors read as a sign that the price did not shut out poorer women. Fewer than 20 women started on semaglutide. The rest began on tirzepatide, a once weekly peptide that acts on both the GIP and GLP 1 receptors. Because doses were held back when weight was already falling or side effects were a problem, the average amount taken at six months stayed well below the highest licensed level. That matters for anyone comparing these results with the large trials.

Digital GLP 1 care results at six months

Among the 7,297 women weighed at six months, mean weight fell by 16.1 kg, or 15.5% of starting weight, and 99.3% lost some weight. The paper reports that 96.1% lost at least 5%, 83.1% at least 10%, 53.5% at least 15% and 22.4% at least 20%. About 58% of the six month loss came in the first three months. Blood tests moved the same way. In 6,245 women with paired HbA1c results, the average dropped by 3.3 mmol per mol. Of 155 women whose baseline value sat in the diabetes range, 131 returned to the normal range, and 546 of 567 with prediabetes did the same. Total cholesterol, LDL cholesterol and triglycerides all fell as well.

Of 155 women whose baseline value sat in the diabetes range, 131 returned to the normal range, and 546 of 567 with prediabetes did the same.

Treatment retention was high by the standards of routine care. At six months 7,429 women, or 91.4%, were still taking the drug and 698 had stopped. Side effects were common. In all, 5,092 women, or 62.7%, reported at least one, and among them the worst severity reported was mild for 36.8%, moderate for 55.4% and severe for 7.7%. The authors identified no serious adverse events caused by treatment. Women who used a recommended package of tips for handling side effects stopped less often, 5.0% against 7.4%, although that comparison was not randomised. Eleven women became pregnant and stopped at once. HDL cholesterol, often called good cholesterol, dipped slightly, which the authors link to the active weight loss phase.

How the cohort compares with the trials

The weight curve resembles the early months of the large randomised trials of semaglutide and tirzepatide, the authors write, even though most women here took less drug. That comparison is informal. The trials assigned people at random to drug or placebo and followed them for well over a year, while this cohort has six months of data and no control group at all. Women who pay for a structured service and pass a screening process are likely to be more motivated than average, and the authors say so themselves. Weight was also only known for women who were weighed at six months, and 830 of the 8,127 were not. A sensitivity analysis that reweighted for missing data gave similar answers.

Who ran and paid for the work also matters. The study was funded by a grant from The Chicago Community Trust through a donor advised fund, administered at University College London, and the paper says the funder had no role in the analysis. Two private UK companies jointly operated the programme, one employee of each is a co author, and the company behind the service issued the press release that announced the paper. Several academic authors declare consulting ties to drug makers, including Novo Nordisk and Eli Lilly. None of this makes the figures wrong. It does mean the study was built to show that this model of care can work, not to test it against anything else. More studies of this kind sit in our clinic section.

What digital GLP 1 care cannot answer yet

The limits section of the paper is blunt. Without a comparison group, it cannot say how much of the weight loss came from the medicine, from the monitoring and coaching around it, or from the kind of woman who signs up. Medical history and other medicines were self reported and not checked against health records. Six months covers the steepest part of the weight curve but says nothing about whether women keep the weight off, stay on treatment, or regain weight after stopping, which earlier studies show is common. Blood pressure, waist size and body composition were not measured, so it is unknown how much of the lost weight was fat rather than muscle.

For readers, the takeaway is narrower than the press release suggests. Real world evidence from women is welcome, because large outcome trials of these drugs enrolled mostly men, and this is one of the biggest women only cohorts reported so far. The results show that a remote service with strict monitoring can keep most people on treatment for half a year while recording large average weight loss. They do not show that such services are safer or better than in person care, and they apply to fairly healthy women who could pay, not to people with serious obesity related illness. Longer follow up and a comparison arm would be the next tests. Our figures page collects numbers we track across studies.

Conclusion

This digital GLP 1 care cohort adds a large, women only data set to a field led by trials. Mean weight loss of 15.5% and retention above 90% at six months are striking, but they come from a self selected, paying group, without a comparison arm, in a study the service operators helped to run. Read it as evidence that remote care can be organised at scale, not as proof that it beats other routes.

Source Deanfield J, Wong JLC, Katsoulis M and colleagues. Real-world, digital-first delivery of GLP-1 receptor agonist therapy in women. Nature Medicine, 16 October 2026. DOI 10.1038/s41591-026-04687-x

Amino Weekly reports research. It is not medical advice and never gives doses.