Tirzepatide and healthcare costs after 55: what the new claims study found

What the source reports

A retrospective matched-cohort study published in Diabetes, Obesity and Metabolism examined healthcare spending among US adults older than 55 who started tirzepatide for obesity or overweight without diabetes. People who remained on treatment had lower non-drug healthcare costs over time than matched adults who did not start a GLP-1 or dual GIP/GLP-1 medicine.

Depending on the statistical method, the estimated difference during months 6 to 12 was $145 or $181 per person per month. During months 12 to 18, it was $319 or $607 per person per month. Lower inpatient and emergency-department use contributed to the difference.

Those figures do not mean tirzepatide saved every person that amount, and they are not a direct calculation of total treatment value. The researchers analysed group-level insurance claims, and the reported healthcare-cost totals deliberately excluded the cost of tirzepatide itself.

Who was included in the study

Researchers used de-identified US insurance claims from the Komodo Healthcare Map. They identified 15,843 adults older than 55 who initiated tirzepatide between November 2023 and September 2025 and matched each person with one untreated control, producing 31,686 people in the comparison.

The average age was 64.5 years. Participants had obesity, or overweight with at least one related complication, and did not have diabetes. Everyone had at least 12 months of continuous insurance enrolment before entering the analysis.

The matching process considered measured baseline demographics, health conditions, obesity-related complications and previous healthcare use. Matching can make observational groups more comparable, but it cannot balance characteristics that were absent, poorly coded or unknown in the claims data.

How the cost comparisons were made

The study used two approaches because people had different lengths of follow-up. Its primary analysis applied inverse-probability-of-censoring weighting, a method intended to reduce bias when participants leave an analysis at different rates. A secondary pairwise analysis compared matched pairs while both remained observable.

In the weighted analysis, the estimated monthly difference was $145 during months 6 to 12 and $319 during months 12 to 18. The pairwise method produced larger estimates of $181 and $607, respectively. Agreement in direction across the two methods is useful, while the wide distance between their later estimates shows that the size of the association is less settled.

The primary analysis also recorded lower hospital-admission and emergency-department rates among tirzepatide users across follow-up periods. In the secondary analysis those rates were numerically lower but not statistically significant, so the utilisation finding was not equally robust under both methods.

The medicine cost was left out

The most important qualification is in the study’s definition of cost. Claims data did not provide the net price paid for tirzepatide, so the researchers excluded the medicine from total healthcare spending. The results therefore estimate possible offsets elsewhere in care, not whether treatment reduced total spending after the drug bill was included.

Lilly’s announcement compared the estimates with a proposed monthly price under a US Medicare access programme. That policy comparison depends on a particular negotiated price and American coverage rules; it cannot be transferred directly to the NHS, another country or an individual insurance plan.

A full cost-effectiveness assessment would need the actual medicine price, administration and monitoring costs, quality-of-life effects, clinical outcomes and a longer time horizon. This claims study answers a narrower question about spending patterns among people observed in routine US care.

Staying on treatment creates a difficult comparison

The headline finding concerned sustained tirzepatide use. People who remain on a medicine may differ from those who stop or never begin: they may tolerate it better, have more stable coverage, engage with healthcare differently or have social and financial advantages that claims matching cannot fully measure.

The statistical weighting attempted to address unequal follow-up, but no observational method can guarantee that all these differences have been removed. Lower spending could reflect tirzepatide, characteristics of people able to remain on treatment, changes in care around treatment initiation, or a combination of factors.

Routine claims also record billing events rather than detailed clinical measurements. They cannot show whether an avoided admission was caused by weight change, improved symptoms, unrelated events or differences in coding and care access.

This was not a trial of clinical benefit

The study did not randomly assign treatment, and its main outcome was healthcare cost rather than weight, physical function, cardiovascular events or survival. It therefore cannot establish that tirzepatide caused fewer medical emergencies or improved health in this particular cohort.

Tirzepatide has evidence from randomized trials for defined clinical outcomes and is approved for specific uses, but those separate findings do not turn this cost association into a randomized result. Each claim needs to be judged using the design and population that produced it.

The cohort also excluded people with diabetes and was limited to insured US adults older than 55. Results may differ in people with diabetes, younger adults, those without continuous insurance, or health systems with different prices and pathways.

Funding and interpretation

The paper was written by researchers affiliated with Eli Lilly, the manufacturer of tirzepatide, and its public announcement uses the findings to support an access and reimbursement argument. Industry involvement does not invalidate a study, but it makes transparent attention to methods, endpoints and framing especially important.

The large matched cohort, two analytic approaches and explicit treatment of censoring are strengths. The excluded medicine price, potential residual confounding, selection of sustained users, reliance on billing records and relatively short follow-up limit how far the result can be taken.

The most defensible reading is that the study found an association worth testing and incorporating into broader economic research—not proof that tirzepatide pays for itself or guarantees fewer hospital visits.

What remains uncertain

It remains unclear how total costs change when the actual net medicine price is included, whether the association persists over several years, and how results differ between people who continue, pause or discontinue treatment.

Randomized evidence or carefully designed independent studies would help clarify whether lower acute-care use is caused by treatment. Longer analyses should also examine clinical outcomes, quality of life, equity, treatment persistence and the different costs borne by patients, insurers and public health systems.

The current result cannot determine whether coverage expansion would save money at population level, because eligibility, uptake, negotiated prices, monitoring and real-world continuation would all affect that calculation.

Bottom line

Among 15,843 matched pairs of US adults older than 55 without diabetes, sustained tirzepatide use was associated with lower healthcare spending outside the cost of the medicine. The estimated difference grew during longer observed follow-up and appeared to be driven partly by less hospital and emergency care.

The finding is relevant to the debate about the wider value of obesity treatment, but it is not evidence that tirzepatide reduced total spending or caused better outcomes. Its price was excluded, the design was observational, and the analysis focused on people who stayed on treatment.