Coming off a weight-loss jab: people left to usual care regained about 11% of their weight in four months, those given ready meals or a coaching app about 4%

GLP-1 drugs such as semaglutide and tirzepatide help many people lose a fifth of their body weight, but plenty stop taking them, because of cost, lost insurance cover, side effects or simply having reached their goal. Between a fifth and a half of users stop within a year, and the weight tends to come back fast, most steeply in the first three or four months. What should happen in that window has barely been tested. A pilot trial from UT Southwestern in Dallas and Tufts University, published in Obesity Science & Practice on 7 October 2026, is one of the first to try.

Between March 2024 and August 2025 the researchers enrolled 39 adults who had lost more than 10% of their weight on a GLP-1 drug (18% on average) and had stopped within the previous 30 days. Most were women, average age about 50. Fourteen got usual care, whatever their own clinic offered. Thirteen could choose 10 dietitian-designed 'weight wellness' ready meals a week from a menu of more than 70, delivered free to their door for four months; lunches and dinners ran from 160 to 460 kcal. Twelve got a free subscription to Noom, a commercial app that teaches habits for managing hunger, cravings and eating cues and asks for about 15 minutes a day of lessons, food logging and weigh-ins. Everyone was sent a Bluetooth bathroom scale, so weights came straight from home rather than from memory.

Over four months, the model estimated that the usual-care group regained 10.8% of their body weight, close to 3% a month. The meals group regained 3.6% and the app group 3.6%, both about 7 percentage points less than usual care, a difference the trial could be fairly confident about. Nothing else separated the groups: diet-quality scores, how healthy people felt, how well they said they stuck to the plan and how satisfied they were all came out much the same. People on the meals plan ate about eight of their ten meals a week; app users opened it on fewer than three days a week, well short of the daily use asked of them.

This was a feasibility study, not proof. It fell short of its 60-person target, five people had no final weigh-in, and nobody could be blinded to which help they got. The usual-care group was also not like the others: all 14 were women, they were lighter at the start and had lost more on the drug (21% against about 16 to 18%), and people who lose more tend to regain faster. Adjusting for weight lost on the drug and starting weight shrank the gap only a little, but with so few people that cannot settle it. Four months is short, too, and the trial cannot say whether the help keeps working once the meals stop arriving. After this study was finished, the lead author received a grant from Noom for a larger, longer one; the authors say the company had no role in this trial or its report. The take-away for now is modest but practical: the weeks after stopping a GLP-1 drug look like the time to have a plan for food, not to go it alone.