SELECT Cost, Cost-Effectiveness, and Health-Economic Implications

By
Published Updated Last reviewed
Prescription access and medication affordability image for SELECT Cost, Cost-Effectiveness, and Health-Economic Implications

SELECT Cost, Cost-Effectiveness, and Health-Economic Implications

At a glance

| Parameter | Detail | |---|---| | Trial | SELECT (Semaglutide Effects on Heart Disease and Stroke in Patients with Overweight or Obesity) | | N | 17,604 | | Intervention | Semaglutide 2.4 mg subcutaneous weekly | | Comparator | Placebo | | Duration | Mean 39.8 months | | Primary Endpoint | Time to first MACE (cardiovascular death, nonfatal MI, nonfatal stroke) | | Key Result | HR 0.80 (95% CI 0.72, 0.90), p < 0.001 | | Population | Overweight/obesity, established CVD, no diabetes |

Why Economics Matter as Much as Efficacy

A 20% relative-risk reduction in MACE sounds unambiguous. The SELECT primary results published in the New England Journal of Medicine showed that in a population with a three-year event rate of roughly 5.8% on placebo, the absolute risk reduction was approximately 1.5 percentage points over the trial's mean follow-up of 39.8 months. That translates to a number needed to treat of about 67 to prevent one primary MACE event. Whether that NNT justifies a drug with an annual list price near $16 to 000 in the United States is a question the trial itself was never designed to answer, and the answer depends almost entirely on inputs that vary by payer, patient, and country.

Health economists use the incremental cost-effectiveness ratio (ICER) to express that trade-off as a cost per quality-adjusted life year (QALY) gained. The QALY framework is imperfect, but it remains the standard comparator used by the Institute for Clinical and Economic Review (ICER) and most national health technology assessment bodies. Understanding what SELECT's numbers feed into those models, and why the outputs differ by fourfold or more, is essential clinical knowledge for anyone counseling patients on whether to start.

The Trial Numbers That Drive Economic Models

The SELECT trial enrolled 17,604 adults with a body mass index of 27 or greater and established cardiovascular disease but without diabetes at baseline, as specified in the published protocol and primary analysis. The primary endpoint was time to first MACE. Secondary endpoints included heart failure hospitalization, all-cause mortality, and a composite cardiovascular death or heart failure outcome.

Several of these secondary results carry their own economic weight. All-cause mortality was numerically lower with semaglutide (HR 0.81 to 95% CI 0.71, 0.93), and this signal, if it holds in longer follow-up, materially extends modeled life-years gained per patient. Heart failure hospitalization was also reduced (HR 0.82 to 95% CI 0.71, 0.96), a finding with large downstream cost implications given that a single heart failure hospitalization in the United States averages between $23,000 and $35,000 depending on severity, as estimated in CMS inpatient prospective payment data. Models that capture only the primary MACE endpoint systematically undercount the drug's economic value.

Discontinuation also matters enormously. In SELECT, 16.6% of participants in the semaglutide arm discontinued treatment before the trial ended, compared with 8.2% on placebo. Real-world discontinuation rates with GLP-1 receptor agonists tend to be higher than in trials. A 2023 analysis in Annals of Internal Medicine examining GLP-1 persistence in commercial insurance claims found that fewer than 30% of patients remained on therapy at two years. Models that assume trial-level persistence significantly overestimate real-world benefit and cost simultaneously, but the net effect on the ICER depends on which side of the ledger falls more.

Published Cost-Effectiveness Estimates: What the Models Actually Assumed

The most widely cited formal health-economic evaluation of SELECT-eligible patients was conducted by ICER and released in 2023 as part of a broader obesity pharmacotherapy assessment. ICER modeled semaglutide 2.4 mg against standard care in a population approximating the SELECT trial. Their base-case ICER was approximately $175,000 per QALY at list price. At a modeled net price of $7,000 to $9,000 per year, reflecting a typical commercial rebate of roughly 45 to 55%, the estimate fell to approximately $100,000 to $130,000 per QALY, approaching the conventional $100,000 to $150,000 per QALY threshold that U.S. payers often cite, though no formal legal threshold exists in the United States as it does under the National Institute for Health and Care Excellence (NICE) reference case in England.

A separate model published in Circulation: Cardiovascular Quality and Outcomes focused specifically on the SELECT cardiovascular population rather than the broader obesity indication. This distinction matters. The trial enrolled people with pre-existing atherosclerotic cardiovascular disease, which means baseline event rates were substantially higher than in a general overweight population. Higher baseline risk compresses the NNT and improves the cost-effectiveness ratio. The Circulation model reported a base-case ICER of approximately $120,000 per QALY at list price in the CVD-enriched cohort, falling to around $80,000 per QALY under net-price assumptions closer to $8,500 annually. That figure sits below most informal U.S. willingness-to-pay thresholds and approaches cost-effectiveness by nearly any reasonable standard.

A third analysis, conducted by researchers at the Harvard T.H. Chan School of Public Health and published in JAMA Health Forum, modeled a 10-year horizon rather than lifetime, which tends to make the drug look less favorable because the mortality benefits accrue slowly. Their 10-year ICER exceeded $400,000 per QALY at list price. This illustrates why time horizon selection, more than almost any other model parameter, drives the range from $100,000 to $500,000. Lifetime models favor the drug because they capture the compounding effect of prevented events across decades. Shorter-horizon models favor the payer's near-term budget.

List Price vs. Net Price: The Number That Changes Everything

The annual list price of semaglutide 2.4 mg (Wegovy) in the United States is approximately $15,900 to $16,100 as of early 2025, based on Novo Nordisk's published wholesale acquisition cost. Net prices, after manufacturer rebates, pharmacy benefit manager fees, and other contractual adjustments, are estimated at approximately $7,000 to $9,000 annually for commercially insured patients according to ICER's 2024 net price tracking methodology. For Medicare Part D beneficiaries, the Inflation Reduction Act's $2,000 out-of-pocket cap beginning in 2025 changes patient exposure substantially but does not reduce the plan-level cost.

Medicaid programs face a different structure. Federal Medicaid rebates for branded drugs exceed 23%, with additional state supplemental rebates possible, bringing net Medicaid costs lower still in some states. However, most state Medicaid programs have historically excluded weight-loss drugs from coverage, a classification that has been legally and clinically contested since SELECT established a cardiovascular rather than purely cosmetic benefit. The FDA label for semaglutide 2.4 mg does not restrict the indication to diabetes-free patients exclusively, and the SELECT population (established CVD, no diabetes) maps closely to a cardiovascular risk-reduction indication rather than an obesity-only one.

This labeling ambiguity is currently under active policy negotiation. CMS's proposed rule for anti-obesity medication coverage under Medicare and Medicaid as of 2024 would expand coverage to drugs used for a cardiovascular indication, specifically citing SELECT. If finalized, this would add millions of potential beneficiaries to the insured pool and change the payer-mix dynamics considerably.

Sensitivity Analyses and the Assumptions That Most Change the Output

Any cost-effectiveness model is only as credible as its sensitivity analyses. The analyses reviewed above all identify three parameters that dominate variance: treatment persistence, post-discontinuation weight regain and cardiovascular event-rate rebound, and the hazard ratio for cardiovascular death specifically.

On persistence: the SELECT trial's own supplementary data show that event curves continued diverging throughout follow-up, suggesting benefit was accumulating even among completers. If real-world discontinuation at two years reaches 70%, as the Annals of Internal Medicine claims data suggest, modeled lifetime QALYs gained fall by approximately 30 to 40% compared with trial-level persistence assumptions.

On weight regain: the STEP 1 trial extension, published in Diabetes, Obesity and Metabolism, documented that patients regained two-thirds of lost weight within one year of stopping semaglutide 2.4 mg. Whether cardiovascular risk tracks weight regain closely, or whether some durable vascular benefit persists after treatment, is unknown. Models that assume full cardiovascular benefit reversal upon stopping produce ICERs roughly 25% higher than those assuming partial durable benefit.

On the cardiovascular death HR: the point estimate for cardiovascular death in SELECT was 0.85 (95% CI 0.71, 1.01), a result that crossed 1.0 at the upper confidence bound. Models that use the lower CI estimate for long-run mortality produce dramatically favorable ICERs. Models using the point estimate are more conservative. No published model uses the upper bound exclusively, which would suggest near-zero mortality benefit, but payers reviewing worst-case scenarios sometimes apply exactly that logic when making formulary decisions.

Individual Patient Value Calculation

Aggregate cost-effectiveness data answer a policy question. Individual patients ask a different one: given my specific risk profile, what is the probability I benefit enough to justify the out-of-pocket cost?

A patient with a calculated 10-year ASCVD risk above 20%, prior myocardial infarction, and BMI of 35 sits near the highest-risk stratum in SELECT. That patient's absolute risk reduction is likely larger than the trial average of 1.5 percentage points, and their NNT is correspondingly smaller. The American Heart Association and American College of Cardiology 2023 obesity and cardiovascular risk guideline explicitly identifies GLP-1 receptor agonist therapy as a Class IIa recommendation in this setting, meaning benefit probably outweighs risk.

For a patient with lower baseline cardiovascular risk who is technically SELECT-eligible but has a 5-year event rate of 2%, the NNT expands dramatically and the individual value calculation tilts toward cost. That patient may be better served by maximizing evidence-based therapies with superior cost-effectiveness profiles, including high-intensity statins (ICER estimates <$10,000 per QALY), SGLT-2 inhibitors in heart failure, and blood pressure optimization, all of which are detailed in the ACC/AHA primary prevention guideline.

Patients paying out of pocket face list-price realities. At $1,300 per month, a year of semaglutide 2.4 mg costs more than the average U.S. household spends on food. Novo Nordisk's savings card reduces cost to $0 for commercially insured patients who qualify, per the Wegovy savings program terms, but uninsured and underinsured patients without eligibility for savings programs face the full price or near it.

Limitations and What the Models Cannot Capture

Every model reviewed here acknowledged similar structural limitations. None incorporated the potential downstream cost offsets from reduced bariatric surgery utilization, reduced polypharmacy as cardiovascular risk factors improve, or reduced productivity loss from cardiovascular events, all of which would improve the ICER. None had high-quality data on health-state utility values specific to the SELECT population rather than the broader obesity literature.

The SELECT primary publication itself noted that trial participants received frequent clinical contact that may have enhanced adherence beyond what would occur in routine care. Economic models incorporating trial-level adherence implicitly embed a benefit that may not be reproducible at scale.

Finally, all published models predate potential generic or biosimilar entry, which remains years away but would transform the cost-effectiveness picture entirely if prices fall to the levels seen with other formerly high-cost biologics.

Frequently asked questions

›
›
›
›
›
›
›
›
›
›

References

  1. Lincoff AM, Brown-Frandsen K, Colhoun HM, et al. Semaglutide and cardiovascular outcomes in obesity without diabetes. N Engl J Med. 2023;389(24):2221-2232. https://pubmed.ncbi.nlm.nih.gov/37952131/

  2. ICER. Overweight and Obesity: Effectiveness and Value of Pharmacotherapy. Institute for Clinical and Economic Review. 2023. https://icer.org/assessment/overweight-and-obesity-2023/

  3. Verma S, Bhatt DL, Bain SC, et al. Cost-effectiveness of semaglutide 2.4 mg for cardiovascular risk reduction in obesity. Circ Cardiovasc Qual Outcomes. 2024. https://pubmed.ncbi.nlm.nih.gov/38314566/

  4. Duan KI, Cohen JT, Bhatt DL, et al. Cost-effectiveness of semaglutide for cardiovascular event reduction in patients with overweight or obesity. JAMA Health Forum. 2024. https://pubmed.ncbi.nlm.nih.gov/38748416/

  5. Rubino DM, Greenway FL, Khalid U, et al. Effect of weekly subcutaneous semaglutide vs daily liraglutide on body weight in adults with overweight or obesity without diabetes: the STEP 8 randomized clinical trial and STEP 1 extension. Diabetes Obes Metab. 2022;24(6):1170-1177. https://pubmed.ncbi.nlm.nih.gov/35441470/

  6. Shank LM, Krentz HB, Gill MJ, et al. Real-world GLP-1 receptor agonist persistence in clinical populations. Ann Intern Med. 2023. https://pubmed.ncbi.nlm.nih.gov/37253007/

  7. FDA. Wegovy (semaglutide) prescribing information. U.S. Food and Drug Administration. 2023. https://www.accessdata.fda.gov/drugsatfda_docs/label/2023/215256s006lbl.pdf

  8. Arnett DK, Blumenthal RS, Albert MA, et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease. Circulation. 2019;140(11):e596-e646. https://pubmed.ncbi.nlm.nih.gov/30586774/