An economic analysis of more than a decade of federal survey data found that adults with diabetes who started GLP-1 medications showed no measurable change in mental health, self-rated health, employment, or marital status compared with their own earlier circumstances.
The finding comes from a National Bureau of Economic Research working paper by University of Chicago Harris School of Public Policy economist Robert Kaestner and coauthor Cuiping Schiman. The paper was issued in May, was publicized by the University of Chicago in late July, and has circulated more widely in recent weeks.
Nothing in it disputes what these drugs do to the body. Improved blood sugar control, weight loss, and reduced cardiovascular risk are established in clinical trials and are not what this analysis examined. The question it asked is narrower and, for households weighing cost against expectation, arguably more personal: does the physical change carry over into how life actually feels and functions?
For patients paying out of pocket, or arguing with an insurer, the answer shapes what it is reasonable to expect from a monthly injection.
Inside the Analysis and Its Limits
The researchers used the Medical Expenditure Panel Survey, a nationally representative federal survey, and focused on adults with diabetes between 2012 and 2023. Rather than comparing GLP-1 users with non-users, they followed the same individuals over time and looked at whether a person's circumstances shifted after starting treatment. The analysis sample covered 4,284 observations for women and 3,672 for men.
That design choice is the core of the result. At first, GLP-1 users did look different from non-users on employment and marriage. Those differences largely disappeared once the analysis looked within individuals, suggesting the original gaps reflected who tends to be prescribed these drugs rather than what the drugs do. Results held whether the researchers looked at roughly one year of use or a two-year window.
The limitations are substantial, and the authors state them. The outcomes measured were blunt: whether someone became employed or unemployed, whether they married or divorced, and standard screening indicators of depression and psychological distress. Kaestner said in remarks released with the findings that those are important measures, but "they don't necessarily capture changes in self-esteem, relationship quality, or other day-to-day experiences."
This is observational research, not a randomized trial. A null result on these particular measures is not evidence that the drugs have no effect on well-being. The study also focused on adults with diabetes over a period when GLP-1 use for weight management alone was far less common, so it does not describe the larger population now taking these medications primarily to lose weight.
The paper is an NBER working paper and has not been peer reviewed. Kaestner reported having nothing to disclose.
The finding also sits against other evidence pointing in different directions. A separate 2026 working paper by Diamond, examining GLP-1 use for weight loss among women in the Understanding America Study, reported large increases in employment among women who were not working at baseline and in marriage or cohabitation among women who were single at baseline. That study has not been peer reviewed either, and no single analysis settles a question this broad.
Coverage Decisions This Finding Touches
The timing matters because of what is being debated. Insurers, employers and government programs are actively weighing how far to extend coverage for a drug class whose list prices run into the hundreds of dollars a month.
Medical benefits are well documented and carry the coverage argument on their own. Broader social and economic gains, if they existed and could be measured, would strengthen the case for wider access by suggesting returns beyond the clinic. This analysis found no such gains in the outcomes it could observe, which removes one line of argument without touching the clinical one.
That distinction is easy to lose in headlines. A finding that GLP-1 drugs did not measurably change employment status is not a finding that they do not work. It is a finding about what a large federal survey could detect in the lives of people with diabetes over one to two years.
Decisions at national scale are where analyses like this one get read most closely. The Centers for Medicare and Medicaid Services launched the Medicare GLP-1 Bridge on July 1, a time-limited demonstration that gives eligible Part D beneficiaries access to certain weight-loss GLP-1 medications for a $50 monthly copay through Dec. 31, 2027.
Guidance for Patients Already Taking These Medications
Nobody should stop or change a prescription based on a working paper. GLP-1 medications are prescribed for glycemic control, weight management, and cardiovascular risk reduction, and stopping abruptly can reverse metabolic gains. Questions about whether a medication is still the right choice belong with the prescribing clinician.
What the research does offer is a more realistic frame for expectations. Patients starting these drugs hoping that weight loss will resolve depression, repair a strained relationship, or change a work situation may be asking the medication to do something this analysis could not detect it doing. That is worth knowing before the first injection rather than after a disappointing year.
Anyone who experiences new or worsening mood symptoms while taking a GLP-1 medication should raise it with a clinician rather than assume it is expected. Mental health treatment and metabolic treatment address different problems, and improvement in one does not substitute for care in the other.
For households facing cost pressure, patient assistance programs, formulary appeals, and discussion of alternatives with a prescriber remain the practical routes. Prior authorization denials can often be appealed, pharmacists can explain what a specific plan covers, and the CMS guidance for providers sets out who qualifies under the Bridge demonstration.
The researchers expect the picture to develop. As use expands beyond diabetes and longer follow-up becomes available, broader effects may emerge in populations and timeframes this analysis could not reach. Whether the working paper is revised or published in a peer-reviewed journal will be the next milestone to watch.
What remains unknown is whether these drugs produce subtler benefits that survey instruments miss, and whether results differ for people taking them primarily for weight loss rather than diabetes.
Key Questions Answered
What did the analysis find? Among adults with diabetes, GLP-1 use was not meaningfully associated with changes in mental health, self-rated health, employment, or marital status.
Does this mean GLP-1 drugs do not work? No. The physical benefits, including blood sugar control, weight loss, and cardiovascular risk reduction, are established in clinical trials and were not the subject of this analysis.
What data was used? More than a decade of the Medical Expenditure Panel Survey, a nationally representative federal survey, covering adults with diabetes from 2012 through 2023.
Has the research been peer reviewed? No. It is a National Bureau of Economic Research working paper issued in May 2026 and publicized by the University of Chicago in late July.
What are the main limitations? The outcome measures were broad and may miss changes in self-esteem or relationship quality. The population studied was adults with diabetes, not people using GLP-1 drugs mainly for weight loss.
Should patients change their medication? No one should start or stop a prescription based on this analysis. Questions belong with a prescribing clinician.
Why does this matter for coverage? Insurers and government programs are weighing broader access. Evidence of social or economic gains would have added to the case, and this analysis did not find any it could measure.
