Enough that one drug class now takes a share of the budget several times its share of the prescriptions. A nationwide analysis of Croatia found GLP-1 drugs at 16.1% of non-insulin antidiabetic prescriptions by 2024 and 42.7% of the money spent on that category, which came to €42.07 million [1]. That gap between share of use and share of spend is what a premium drug looks like on a national ledger. It is the same arithmetic asked person-by-person in whether a GLP-1 is worth its price.
One country, fourteen years
Between 2010 and 2024, total non-insulin diabetes drug consumption in Croatia more than doubled. Within that category, GLP-1 drugs went from nothing at market entry to 15.99 defined daily doses per 1,000 inhabitants per day. Semaglutide led on both volume and spend, followed by dulaglutide and liraglutide.
That reimbursement boundary is the practical finding for a buyer, because it is not a Croatian quirk. It is the structure of most coverage everywhere, including the US plans tracked in what insurance actually covers. Two limits apply before the euro figure travels anywhere: it is one country of about four million people with its own negotiated prices, and the prescription-to-spending ratio reflects those negotiated prices too.
How large the eligible population is
The reason budget holders are nervous is the size of the denominator. A cross-sectional analysis applied Australia’s approved indications to its 2022 National Health Survey. It found 39.7% of adults meeting the chronic weight management criteria, 95% CI 38.4% to 41.0%, or about 7.8 million people [2]. Of those, 2.9 million had no weight-related comorbidity, 3.3 million had one, and 1.7 million had at least two.
The second indication is far narrower. Up to 338.9 thousand of that group, 95% CI 271.3 to 406.5 thousand, also met the criteria for secondary prevention of cardiovascular disease. That is roughly 4% of the larger pool, by division of the two published figures. Which of the two indications a person falls into changes everything about what a system will fund, without changing anything about the drug.
What the spending buys back, and when
A nationally representative US household survey covering 7,144 person-years, of which 275 were sustained GLP-1 users, found non-drug medical spending lower by $2,586 per person-year among users and acute-care spending lower by $2,019 [3]. Total medical spending was higher, with a coefficient of 0.3182, and the published abstract truncates before the corresponding dollar figure.
The authors describe this as no evidence of same-year cost savings alongside a distinct shift in spending composition. The money moved out of hospitals and acute care and into the pharmacy, and rather more arrived at the pharmacy than left the hospital. Everything in that analysis is same-year, and the models that find savings put them years out, which is set out in what a GLP-1 saves the system.
Why a national figure matters to a cash buyer
Because the reductions land on somebody else’s ledger and the increase lands on yours. A system that watches one class take two fifths of a category budget reaches for coverage rules first, and every tightening pushes more people onto the self-pay prices this desk publishes. How far employer coverage actually reaches in the US is counted in the nineteen percent of large firms.
Eligibility and access are also different questions, and the second decides everything. Most people who meet the criteria never start, and when a study actually offered the drug to eligible patients, most of them declined on price, as counted in who turns a prescription down.