One number answers this, and it is 2.8 points. Coaching alongside a GLP-1 produced 6.0% weight loss in twelve weeks, against 3.3% on usual care [1]. That is the paper’s adjusted between-group difference, across 245 adults. Nobody was randomized into it, which is the limit on the whole result.
Coaching gets sold beside these drugs, sometimes inside the subscription. The question is what twelve weeks of one adds. The answer is a real number wrapped in two soft measurements. That is a familiar shape for anything claiming to protect muscle. Read the design before the headline.
The setup
245 adults with obesity who had started a GLP-1 within the previous eight weeks. [1] 151 received personalized, strength-focused activity coaching delivered virtually. 94 got usual care. Twelve weeks.
Everyone was on the drug. The question is only what the program added on top.
What it added
Weight: 6.0% lost against 3.3%. The adjusted between-group difference is 2.8 points, p < 0.001. That is the strongest result here.
Body fat percentage fell 3.3 points against 1.6. Muscle mass percentage rose 0.6 points against 0.2. Psychosocial and behavioral measures improved more too.
What measured it
Home bioelectrical impedance scales pass a small current through the body and estimate composition from the resistance.
Drink a glass of water and the reading moves. For total weight they are fine. For a four-tenths-of-a-point difference in muscle share, they are not the instrument you would choose.
Nobody drew lots
This was quasi-experimental, and people were not randomized into coaching.
Whoever ends up in a structured activity program is already more likely to exercise. They keep appointments, and they stay on the drug. That is the same selection that separates the people who persist from the majority who stop. An unrandomized 2.8 points is partly the program and partly the kind of person in it. This design cannot split them.
The randomized evidence is about support, not subscriptions
A network meta-analysis of 42 randomized adolescent trials asked a cleaner question [2]. Every drug tested did better with structured lifestyle treatment attached than given alone. That held for GLP-1 drugs, metformin, orlistat and phentermine-topiramate alike.
Semaglutide with counseling gave the largest BMI reduction, 8.31 points. Its interval runs from -12.33 to -4.28, so hold it loosely. Lifestyle treatment on its own moved BMI by 3.85 points, and the drug alone was never the winning arm.
Those participants were adolescents at a median age of 14.5. Follow-up ran six to twelve months. Neither figure converts to an adult buying a subscription.
What a digital program measured on itself
One Australian digital clinic added an AI support agent and analyzed 16,556 semaglutide patients afterwards [3]. Raw six-month adherence rose, 53.2% against 47.3%. The adjusted model in the same paper found the opposite direction.
Post-launch was independently associated with higher odds of attrition, OR 1.178 (95% CI 1.052–1.318). One study, two answers, and the crude figure is the one a press release carries. Higher program cost predicted attrition at roughly two and a half times the odds.
What a buyer should take
That resistance training during weight loss is worth doing is not in dispute. It does not require a subscription.
What these studies do not establish is that a paid companion program beats doing the same thing yourself. Thirty interviewed patients said support was the part that varied most [4]. So two sellers charging alike can deliver very different care.
None of these papers reports a price. So there is no cost-per-point to compute — and what a GLP-1 actually costs already has enough lines on it. Ask who assembled the evaluation before treating it as a reason to add another. That is how any favorable comparison should be read.