Two figures come out of the same survey question, and only one of them gets repeated. Coverage of these drugs for weight loss at the very largest employers rose sharply in a year. Coverage at everyone else did not, and the employers standing outside say they intend to stay there. Which of those two facts applies depends entirely on the size of the company issuing the health plan, which is not something a reader chooses. The alternative route, and why it is priced the way it is, sits in the coverage article.
How common it is, stated with its sample
The KFF 2025 Employer Health Benefits Survey interviewed 1,862 randomly selected non-federal public and private employers with ten or more workers. Davis Research LLC conducted the field work between January and July 2025, and the overall response rate was 13%.[1] That is the instrument behind every percentage in this section, and it is an estimate across firms rather than a census of plans.
Among firms offering health benefits with 200 or more workers, the survey found that 16% of firms with 200 to 999 workers, 30% of firms with 1,000 to 4,999 workers, and 43% of firms with 5,000 or more workers cover GLP-1 agonists when used primarily for weight loss.[1] The figure for the largest band was 28% in the 2024 survey, and KFF reports the increase to 43% as statistically higher than the prior year.
The half of the trend that is rarely quoted
The same survey put a forward-looking question to firms with 200 or more workers that do not currently cover these drugs for weight loss. Just 1% described themselves as very likely to begin within the next twelve months. Twenty-four percent said somewhat likely, 67% said not likely, and 8% did not know.[1]
Those two results are not in conflict. Adoption moved quickly inside a band of very large employers and has not moved much elsewhere, and the employers outside report no intention of joining. A reader at a thousand-person company reading that coverage is expanding is reading about a different band than the one paying their claims.
Covered does not settle what is dispensed
Among the firms that do cover these medications for weight loss, 34% require enrollees to meet with a dietitian, case manager or therapist, or to participate in a lifestyle program, in order to receive the coverage.[1] That condition sits before the prescription rather than after it, and a plan that satisfies the word “covered” can still require weeks of scheduled appointments first.
Employers also report that the benefit has behaved unexpectedly. Fifty-nine percent of covering firms with 5,000 or more workers, and 44% of covering firms with 1,000 to 4,999 workers, said use of these medications for weight loss was higher than expected. Sixty-six percent and 43% of those groups respectively said the coverage had a significant impact on the health plan’s prescription drug spending.[1] Those are the answers that precede a benefit redesign, and the arithmetic a reader should run against any route is in the annual-cost article.
What a carve-out is, and why it is not a formulary decision
A carve-out in benefits language is an explicit exclusion of a drug or a therapeutic class from the pharmacy benefit, written into the plan document rather than decided at the claims desk. It is not a judgment that a prescription is unnecessary, and it does not respond to clinical evidence about the patient, because there is no covered category for the claim to land in. A denial under a carve-out and a denial under medical policy read similarly and are entirely different objects.
Who writes that exclusion matters. Sixty-seven percent of covered workers, including 80% of covered workers at firms with 200 or more workers, are enrolled in plans that are self-funded, meaning the employer pays for health services directly from its own funds rather than purchasing insurance.[1] In a self-funded arrangement the insurer whose name is on the card is administering someone else’s money, and the coverage decision belongs to the employer.
Level-funded plans are a third category
Smaller employers have been moving into an arrangement that is neither fully insured nor conventionally self-funded. Thirty-seven percent of covered workers in firms with 10 to 199 workers are in a level-funded plan, which combines a relatively small self-funded component with stop-loss insurance that caps the employer’s liability. KFF notes that, unlike insured plans, these arrangements use health status in rating and underwriting and are not required to provide all of the essential health benefits mandatory for insured plans.[1]
That is worth knowing before assuming a small-employer plan is governed by the rules that apply to a plan bought on an exchange. The benefit design question and the regulatory question have the same answer here, and it is the plan document.
Employers do not think this is what workers worry about
The same survey asked large firms to assess how concerned they believe their employees are about several aspects of plan management. On the complexity of prior authorization requirements, 12% of large firms believed employee concern was high and 32% believed it was moderate, while 31% judged it low and 16% judged it none. On the number of denied claims, only 6% believed concern was high and 12% moderate, against 50% low, 14% none, and 17% who did not know.[1]
Those are employer perceptions rather than employee reports, and the survey is explicit about that. The distribution still matters, because the party that chooses whether to attach a utilization control is the party estimating how much the control will be felt, and on denied claims roughly one large firm in six has no estimate at all.
Step therapy, and the patient already on treatment
Step therapy requires a documented trial of one treatment before a plan will pay for another. The clearest published statement of how such a program must be run appears in the Medicare Advantage regulation, which binds plans applying step therapy to Part B drugs. It requires that a plan “apply step therapy only to new administrations of Part B drugs, using at least a 365 day lookback period”, that providers and enrollees be educated about the policy before it takes effect, and that the program be reviewed and approved by a pharmacy and therapeutics committee whose members include practicing physicians and pharmacists free of conflict.[2]
The lookback is the provision that protects a person mid-treatment: a patient with a year of therapy behind them is not a new administration. Nothing in that section reaches a self-funded employer plan, and no federal rule imports it there. Whether a given employer plan carries an equivalent continuity provision is a question for that plan’s own pharmacy policy document, and it is worth asking before a plan year turns over rather than after.
What federal law does fix for an employer plan is the clock. Under the Department of Labor claims-procedure regulation, a pre-service claim must be decided within a reasonable period appropriate to the medical circumstances but no later than 15 days after receipt, extendable one time by up to 15 further days where the plan determines the extension is necessary for reasons beyond its control and notifies the claimant first.[3] A new step-therapy requirement arriving in January therefore has a legally bounded, but not short, resolution window.
When coverage is taken away
Withdrawal has been measured once in a useful way. A cross-sectional study published in 2026 examined employees of a healthcare system that discontinued GLP-1 obesity coverage, analyzing 247 adults who had been prescribed the drugs for obesity and lost it (mean age 49.27, mean body-mass index 32.85, 90.28% female). Afterward, 49.39% felt devalued and 64.38% believed their employer did not value their health.[4]
The stratified result is the one that travels. Non-Hispanic Black and Hispanic employees were less likely than non-Hispanic White employees to report a worsened relationship with their employer, adjusted odds ratios of 0.29 and 0.34, while employees earning $150,000 or more were roughly 2.5 times as likely to report a negative impact (adjusted odds ratio 2.57).[4] The paper’s own reading is that the lower reported impact among those groups may reflect disparities in perceived job mobility. Reported dissatisfaction tracked the ability to leave, not the size of the loss.
That study is a single employer, self-reported, and cross-sectional, so it establishes a pattern rather than a rate. What stopping does to weight is a separate question, answered directly in the withdrawal trials.
The persistence numbers employers are reacting to
Benefit decisions are being made against real-world continuation rather than trial completion. A 2026 commentary in a managed-care journal summarizes the published picture: in obesity cohorts without diabetes, approximately one-third of members remain on therapy at one year, and among those with type 2 diabetes, nearly half discontinue by twelve months and approximately 70% by twenty-four.[5]
An employer looking at those figures sees a benefit where most of the spending buys a course that is abandoned. A patient looking at the same figures sees a treatment whose cost is the commonest reason people stop. Both readings are supported, and they lead to opposite decisions, which is why coverage in this class has been unstable in a way it is not for most drugs.
What a reader can actually do with this
Three documents settle more than any survey can: the plan’s own formulary for the current year, its medical policy for weight management, and its pharmacy prior-authorization criteria. The machinery behind that third document is set out in the prior-authorization article, and what to do when a request comes back denied is in the appeals article. Public coverage follows different rules entirely, described in the public-coverage article.
Where no benefit exists, the purchase moves to cash, and most of that market sells compounded preparations. Compounded drugs are not FDA-approved, and the agency does not review them for safety, effectiveness or quality before they are dispensed. What those sellers publish is recorded in the seller write-ups, ordered in the price board, and established the way the methodology describes.
Nothing above describes any particular employer. Two people at the same company, in different plans or different states, can get different answers in the same week, and a survey estimate across 1,862 firms cannot narrow that. It can only say which way the field is moving, and for whom.