Because roughly half of the people who try do not get one. In a US national cohort of 4,555 adults, 49.9% of those who attempted treatment were unsuccessful [1]. Failing to obtain the drug is the median experience rather than a personal failure. The three measured reasons are money, coverage and supply, and clinical doubt is not among them.
Four in five American adults have heard of these drugs. Fewer than one in ten has used one. That gap is the entire reason a self-pay price desk exists. It is the distance between a published price and a filled prescription, and it runs through the sellers publishing two prices.
Who is using one, and who qualifies
Eligibility here means a BMI of 30 or more, or 27 with at least one comorbidity. Among the 1,883 respondents who met it, 13.4% had used a GLP-1 in the previous three years. Among those who did not, 4.0% had. Awareness across the whole cohort was 81.6%.
One number deserves reporting without commentary. Use among people who did not meet the FDA criteria was 4.0%. That is what the respondents said. How any of them obtained a prescription is not something this study measured, and not something this desk will guess at.
The income pattern repeats in diabetes
The same association turns up where the clinical case is strongest. Among US adults with type 2 diabetes, income above the median and private insurance each raised the odds of using semaglutide [2]. Those data are from 2022, and the direction has not reversed since.
A diabetes diagnosis is the strongest indication this drug class has. If income still sorts who gets it there, it sorts everywhere, which is what the income analysis measured directly.
The prescriber is not the obstacle
Asked whether these drugs work, 98% of responding Swedish primary care physicians agreed that they do [3]. The barriers they named were cost and shortages.
That rules one explanation out. A refusal is rarely a clinical verdict on the drug, and the survey says so in the physicians’ own answers. The Swedish system is not the American one. The mechanism transfers even where the payer does not.
What the data cannot say
Two limits set the boundary. The cohort is self-reported rather than drawn from prescription records, so it captures what people say they did. It is a single snapshot from December 2023.
Supply and self-pay pricing have both moved since. That is why this desk re-walks its own price rows rather than citing old ones. The reasoning is in why falling prices and annual plans do not mix.
If you are comparing self-pay prices at all, you are already past the step where most people stopped. The next figure to check is what the seller bills against what it advertises, in the headline-versus-billed comparison.