CoChat
Sign Up
Cited by 75Diabetes Treatment and ManagementPharmaceutical Economics and PolicyHealth Systems, Economic Evaluations, Quality of Life

Evaluation of Out-of-Pocket Costs and Treatment Intensification With an SGLT2 Inhibitor or GLP-1 RA in Patients With Type 2 Diabetes and Cardiovascular Disease

Open access
Luo, Jing, Feldman, Robert, Callaway Kim, Katherine, Rothenberger, Scott, Korytkowski, Mary, Hernandez, Inmaculada, Gellad, Walid F.·2023·JAMA Network Open
DOI 10.1001/jamanetworkopen.2023.17886PMC PMC10261995
Open this paper in CoChat — free accountOpen PDF

Abstract

Importance: The latest guidelines continue to recommend sodium-glucose cotransporter 2 (SGLT2) inhibitors and glucagon-like peptide-1 receptor agonists (GLP-1 RAs) for patients with type 2 diabetes (T2D) and established cardiovascular disease (CVD). Despite this, overall use of these 2 drug classes has been suboptimal. Objective: To assess the association of high out-of-pocket (OOP) costs and the initiation of an SGLT2 inhibitor or GLP-1 RA among adults with T2D and established CVD who are treated with metformin-treated. Design, Setting, and Participants: This retrospective cohort study used 2017 to 2021 data from the Optum deidentified Clinformatics Data Mart Database. Each individual in the cohort was categorized into quartiles of OOP costs for a 1-month supply of SGLT2 inhibitor and GLP-1 RA based on their health plan assignment. Data were analyzed from April 2021 to October 2022. Exposures: OOP cost for SGLT2 inhibitors and GLP-1 RA. Main Outcomes and Measures: The primary outcome was treatment intensification, defined as a new dispensing (ie, initiation) of either an SGLT2 inhibitor or GLP-1 RA, among patients with T2D previously treated with metformin monotherapy. For each drug class separately, Cox proportional hazards models were used to adjust for demographic, clinical, plan, clinician, and laboratory characteristics to estimate the hazard ratios of treatment intensification comparing the highest vs the lowest quartile of OOP costs. Results: Our cohort included 80 807 adult patients (mean [SD] age, 72 [9.5] years, 45 129 [55.8%] male; 71 128 [88%] were insured with Medicare Advantage) with T2D and established CVD on metformin monotherapy. Patients were followed for a median (IQR) of 1080 days (528 to 1337). The mean (SD) of OOP costs in the highest vs lowest quartile was $118 [32] vs $25 [12] for GLP-1 RA, and $91 [25] vs $23 [9] for SGLT2 inhibitors. Compared with patients in plans with the lowest quartile (Q1) of OOP costs, patients in plans with the highest quartile (Q4) of costs were less likely to initiate a GLP-1 RA (adjusted HR, 0.87 [95% CI, 0.78 to 0.97]) or an SGLT2 inhibitor (adjusted HR, 0.80 [95% CI, 0.73 to 0.88]). The median (IQR) number of days to initiating a GLP-1 RA was 481 (207-820) days in Q1 and 556 (237-917) days in Q4 of OOP costs and 520 (193-876) days in Q1 vs 685 (309-1017) days in Q4 for SGLT2 inhibitors. Conclusions and Relevance: In this cohort study of more than 80 000 older adults with T2D and established CVD covered by Medicare Advantage and commercial plans, those in the highest quartile of OOP cost were 13% and 20% less likely to initiate a GLP-1 RA or SGLT2 inhibitor, respectively, when compared with those in the lowest quartile of OOP costs.

AI snapshot

Written from the paper’s full text.

TL;DR

In a retrospective cohort of 80,807 adults with type 2 diabetes and established cardiovascular disease on metformin, higher out-of-pocket costs for SGLT2 inhibitors or GLP-1 RAs were associated with lower likelihood of treatment intensification. The highest vs lowest OOP quartile had adjusted hazard ratios of 0.87 for GLP-1 RA and 0.80 for SGLT2 inhibitor initiation.

Key findings

  • Among 80,807 adults (mean [SD] age, 72 [9.5] years; 45,129 [55.8%] male), the mean estimated 30-day OOP cost was $69.20 ($38.30) for a GLP-1 RA and $54.39 ($28.00) for an SGLT2 inhibitor.
  • Comparing highest vs lowest OOP quartile, unadjusted HRs for initiation were 0.70 (95% CI, 0.63-0.78) for GLP-1 RA and 0.68 (95% CI, 0.62-0.74) for SGLT2 inhibitor; fully adjusted HRs were 0.87 (95% CI, 0.78-0.97) and 0.80 (95% CI, 0.73-0.88), respectively.
  • Median days from index metformin dispensing to initiation for Q1 vs Q4 OOP costs were 481 vs 556 days for GLP-1 RA and 520 vs 685 days for SGLT2 inhibitor, corresponding to median delays of 75 days and 165 days.
  • In a statin falsification test, highest vs lowest OOP quartiles had adjusted HRs of 1.08 (95% CI, 1.01-1.16) for GLP-1 RA and 1.10 (95% CI, 1.03-1.18) for SGLT2 inhibitor, suggesting no strong negative association with statin initiation.

Methods

  • Retrospective cohort using Optum deidentified Clinformatics Data Mart Database, 2017-2021; eligible adults had T2D, at least 1 metformin fill between December 5, 2017, and December 31, 2020, continuous enrollment, established CVD, and no prior second-line glucose-lowering drug.
  • OOP costs were assigned using mean 30-day copay plus deductible from other Optum CDM members in the same health plan with 1 or more fills for an SGLT2 inhibitor or GLP-1 RA (n = 445,873); 3,631 members were excluded because OOP cost could not be determined.
  • Cox proportional hazards models compared highest vs lowest OOP quartile, sequentially adjusting for demographics, comorbidities, insurance type, clinician type, and presence of an HbA1c result; Kaplan-Meier cumulative incidence curves and a statin initiation falsification test were also used.

Limitations

  • The paper states results came from 1 national health insurance company with most patients covered under Medicare Advantage, so findings may not generalize to other commercial plans, Medicaid beneficiaries, or Medicare fee-for-service.
  • The paper states lack of formulary data precluded precise calculation of each member’s exact OOP cost; costs were estimated from other Optum CDM members with the same plan identification number and index year.
  • The paper states claims did not include samples or prescriptions paid with cash or reimbursed through secondary payers, possibly causing outcome misclassification; cohort entry required prior metformin, so results may not apply to front-line SGLT2 inhibitor or GLP-1 RA use, and residual confounding is possible. Evident from the design: this observational claims analysis cannot establish causality, and OOP costs for non-fillers were plan-level proxies rather than individual-level observed costs.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1001/jamanetworkopen.2023.17886

Cites

22 of 22 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

13 of 16 cited sources checked · 3 supported · 1 flagged · accuracy not yet measured

  • Premisecitation

    High out-of-pocket drug costs are identified as a factor that limits initiation of and adherence to prescribed glucose lowering therapies.

    “High out-of-pocket (OOP) drug costs are identified as a factor that limit initiation of and adherence to prescribed glucose lowering therapies.8,9,10,11,12”
    Quote in the paperSubstantial claim4 of 5 sources checked
    • cites [8] Kirkman, 2015, Determinants of adherence to diabetes medications: findings from a large pharmacy claims database., Diabetes Care source

      The Conclusions report that higher out-of-pocket costs were associated with lower odds of adherence to noninsulin antidiabetic medications. They also state that the study could not assess primary nonadherence—failure to fill an initial prescription—so it does not establish that costs limit initiation. (LLM, conf=0.90)

      Experimental — accuracy not yet measured
    • cites [9] Patel, 2016, Social determinants of health, cost-related non-adherence, and cost-reducing behaviors among adults with diabetes: findings from the National Health Interview Survey., Med Care source

      The Introduction identifies lack of affordability as a reason for nonadherence, and Methods > Measures defines cost-related nonadherence to include skipping doses, taking less medicine, and delaying a prescription fill to save money. The study concerns adults with diabetes, but these measures do not distinguish glucose-lowering drugs from other medications or establish that costs prevent initiation of glucose-lowering therapy. No page numbers are available in our copy. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites [10] Hawley, 2022, Three sides to the story: adherence trajectories during the first year of sodium–glucose cotransporter 2 inhibitor therapy among medicare beneficiaries., Diabetes Care source

      In Covariates Associated With SGLT2i Adherence (page unavailable), the paper reports that a higher SGLT2 inhibitor co-pay was positively associated with adherence, contrary to the claimed limiting effect of higher costs for this glucose-lowering drug. The study included patients who had already initiated therapy, so it does not establish an effect of costs on initiation. (LLM, conf=0.83)

      Checked by a reasoning model — may be wrong.

      Experimental — accuracy not yet measured
    • cites [11] Kang, 2018, Cost-related medication non-adherence among US adults with diabetes., Diabetes Res Clin Pract source

      The Introduction states that the high cost of diabetes care is associated with medication non-adherence and that the cost burden of prescribed medications can reduce adherence. It does not address whether out-of-pocket drug costs limit initiation of glucose-lowering therapy; a search for initiation found no supporting passage. (LLM, conf=0.87)

      Experimental — accuracy not yet measured
    • cites [12] Hepke, 2004, Costs and utilization associated with pharmaceutical adherence in a diabetic population., Am J Manag Care
  • Premisecitation

    Use of SGLT2 inhibitors and GLP-1 receptor agonists is paradoxically lower among patients with evidence of or increased risk factors for cardiovascular disease.

    “Despite these recommendations, data from studies in the clinical setting show that use of these newer drug classes have been paradoxically lower among patients with evidence of, or increased risk factors for, CVD.3,4,5,6”
    Quote in the paperSubstantial claim4 of 4 sources checked
    • cites [3] McGurnaghan, 2019, The effect of dapagliflozin on glycaemic control and other cardiovascular disease risk factors in type 2 diabetes mellitus: a real-world observational study., Diabetologia source

      Table 1 (baseline characteristics, indexed under Discussion; page unavailable) shows that dapagliflozin users had less prior heart failure and stroke than never-users, supporting the claim for some evidence of CVD among users of one SGLT2 inhibitor. But users had higher BMI and systolic blood pressure, and this paper does not establish lower use of SGLT2 inhibitors as a class or of GLP-1 receptor agonists among patients with increased CVD risk factors. (LLM, conf=0.85)

      Experimental — accuracy not yet measured
    • cites [4] McCoy, 2021, Second-line therapy for type 2 diabetes management: the treatment/benefit paradox of cardiovascular and kidney comorbidities., Diabetes Care source

      The Results report that patients with prior myocardial infarction or cerebrovascular disease were less likely to start a GLP-1 receptor agonist than a DPP-4 inhibitor, and that patients with heart failure or nephropathy were less likely to start an SGLT2 inhibitor. This supports the treatment/benefit paradox for those specific conditions, but the study does not establish lower use of both drug classes across patients with cardiovascular risk factors generally. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites [5] McCoy, 2020, Paradox of glycemic management: multimorbidity, glycemic control, and high-risk medication use among adults with diabetes., BMJ Open Diabetes Res Care source

      The Discussion reports that patients with cardiovascular disease, heart failure, or CKD were less likely than younger, healthier patients to be prescribed SGLT2 inhibitors. It identifies cardiovascular benefits of both SGLT2 inhibitors and GLP-1 receptor agonists, but does not report similarly lower GLP-1 receptor agonist use among patients with cardiovascular disease or its risk factors. The study itself focused on insulin and sulfonylurea use. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites [6] McCoy, 2019, Adoption of new glucose-lowering medications in the US—the case of SGLT2 inhibitors: nationwide cohort study., Diabetes Technol Ther source

      The Discussion reports that patients with CVD or heart failure were among those least likely to initiate SGLT2 inhibitors, supporting the treatment-risk paradox for that class. This study examines SGLT2 inhibitor initiation, not whether GLP-1 receptor agonist use was lower in patients with CVD or its risk factors; it therefore does not support that part of the claim. (LLM, conf=0.89)

      Experimental — accuracy not yet measured
  • Premisecitation

    In prior work from a large integrated health system, one third of patients newly prescribed an SGLT2 inhibitor or GLP-1 receptor agonist did not fill their prescription within 30 days.

    “In prior work from a large integrated health system, we showed that one third of patients newly prescribed an SGLT2 inhibitor or GLP-1 RA did not fill their prescription within 30 days.15”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [15] Luo, 2022, Incidence and predictors of primary nonadherence to sodium glucose co-transporter 2 inhibitors and glucagon-like peptide 1 agonists in a large integrated healthcare system., J Gen Intern Med source
  • Study design and conductcitation

    Established cardiovascular disease was identified using validated ICD-10-CM diagnosis and procedure codes and Current Procedural Terminology codes.

    “who did not have evidence of established CVD, based on validated International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) diagnosis and procedure codes and Current Procedural Terminology (CPT) codes during the preindex period16”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [16] Fishman, 2018, Validation of a claims-based algorithm identifying eligible study subjects in the ADAPTABLE pragmatic clinical trial., Contemp Clin Trials Commun source

      The Abstract (Materials and methods) describes an ICD-9-CM/ICD-10 and CPT claims algorithm, and Case identification describes using codes for prior AMI, PCI, or CABG. But the Statistical analysis section validates the positive predictive value of identifying ADAPTABLE-eligible patients; it does not validate using an absence of codes during a preindex period to establish that someone has no CVD. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
  • Findingown data

    Among adults with type 2 diabetes and established cardiovascular disease treated with metformin, being in the highest quartile of out-of-pocket costs for SGLT2 inhibitors is associated with a lower likelihood of initiating an SGLT2 inhibitor compared with the lowest quartile, with an adjusted hazard ratio of 0.80 (95% CI, 0.73-0.88).

    “Similarly, for SGLT2 inhibitors, those in the highest vs the lowest quartile of OOP costs had an unadjusted HR of 0.68 (95% CI, 0.62-0.74) for SGLT2 inhibitor initiation (Table 2 and Figure 2).”
    • Also stated in: § Likelihood of Treatment Intensification by Quartile of OOP Costs · plain
      “In the most comprehensive model, adjusting for demographic, comorbidities, insurance type and presence of HbA1c, the HR was 0.80 (95% CI, 0.73-0.88).”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 4
  • The quote only partly states this claim. 2
  • May combine two claims. 3

  • Classified as probably not a substantial claim.

    Premisecitation

    Professional society guidelines recommend SGLT2 inhibitors and GLP-1 receptor agonists for patients with type 2 diabetes and established cardiovascular disease.

    “Based on evidence from large cardiovascular outcomes trials, the latest professional society guidelines continue to recommend sodium-glucose cotransporter 2 (SGLT2) inhibitors and glucagon-like peptide-1 receptor agonists (GLP-1 RAs) for patients with type 2 diabetes (T2D) and established cardiovascular disease (CVD).1,2”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [1] Kelsey, 2022, Guidelines for cardiovascular risk reduction in patients with type 2 diabetes: JACC guideline comparison., J Am Coll Cardiol source
    • cites [2] Draznin, 2022, 9. Pharmacologic approaches to glycemic treatment: standards of medical care in diabetes-2022., Diabetes Care source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Patients with type 2 diabetes and established cardiovascular disease tend to be older and have comorbidities such as hypertension, heart failure, and chronic kidney disease.

    “Patients with T2D and established CVD tend to be older and have comorbidities, such as hypertension, heart failure, and chronic kidney disease.7”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [7] Korytkowski, 2016, Use of an electronic health record to identify prevalent and incident cardiovascular disease in type 2 diabetes according to treatment strategy., BMJ Open Diabetes Res Care source

      The Results report hypertensive heart disease and heart failure among the diagnoses in the study’s CVD category. However, Table 2 reports age by treatment strategy, not by whether patients had established CVD, and the paper does not establish chronic kidney disease as a comorbidity of those patients. It therefore supports some examples of coexisting conditions, but not the full characterization. (LLM, conf=0.85)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Cost-related nonadherence affects up to 1 in 4 insulin users.

    “For example, studies have shown that cost-related nonadherence affects up to 1 in 4 insulin users.13,14”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites [13] Herkert, 2019, Cost-related insulin underuse among patients with diabetes., JAMA Intern Med source
    • cites [14] Bakkila, 2022, Catastrophic spending on insulin in the United States, 2017–18., Health Affairs source

  • The quote only partly states this claim.

    Findingown data

    Among adults with type 2 diabetes and established cardiovascular disease treated with metformin, patients in the highest out-of-pocket cost quartile experienced a median delay in treatment intensification of 75 days for GLP-1 receptor agonists compared with the lowest out-of-pocket cost quartile.

    “In addition, patients in the highest OOP 2 category experienced a median delay in treatment intensification of 75 days for GLP-1 RAs and 165 days for SGLT2 inhibitors when compared to patients in the lowest OOP cost category.”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    Among adults with type 2 diabetes and established cardiovascular disease treated with metformin, patients in the highest out-of-pocket cost quartile experienced a median delay in treatment intensification of 165 days for SGLT2 inhibitors compared with the lowest out-of-pocket cost quartile.

    “In addition, patients in the highest OOP 2 category experienced a median delay in treatment intensification of 75 days for GLP-1 RAs and 165 days for SGLT2 inhibitors when compared to patients in the lowest OOP cost category.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Baseline characteristics considered clinically related to treatment intensification were selected a priori.

    “Baseline characteristics that were considered to be clinically related to treatment intensification were selected a priori.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    The study used a large administrative claims database, large sample size, and focused on a specific population of patients with strong indications for use of newer glucose lowering medications.

    “Strengths of our study include its use of a large administrative claims database, large sample size, and focus on a specific population of patients with strong indications for use of newer glucose lowering medications (ie, those with established or preexisting CVD).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    The study's results would not likely be biased by differential use of manufacturer-sponsored drug coupons because such discounts are not allowed in Medicare plans.

    “Notably, our results would not likely be biased by differential use of manufacturer-sponsored drug coupons since the use of these types of discounts are not allowed in Medicare plans.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Manufacturer-sponsored drug coupons are not allowed in Medicare plans.

  • May combine two claims.

    Study design and conductown data

    Optum CDM claims did not include information on samples or prescriptions paid with cash or reimbursed through secondary payers, which may have led to outcome misclassification.

    “Third, Optum CDM claims did not include information on samples or prescriptions paid with cash or reimbursed through secondary payers, which may have led to outcome misclassification.”
    Quote in the paperNot a substantial claim
9 of 9

  • The paper relies on this without support.

    Manufacturer-sponsored drug coupons are not allowed in Medicare plans.

    “Notably, our results would not likely be biased by differential use of manufacturer-sponsored drug coupons since the use of these types of discounts are not allowed in Medicare plans.”

  • The paper relies on this without support.

    The cohort of patients on metformin monotherapy is representative of type 2 diabetes patients on metformin monotherapy.

    “This ensured that the sample was representative of T2D patients on metformin monotherapy.”
2 of 2

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • In the United States, the cost of GLP-1 RAs can range from $800 to $1300 per month without insurance.

    “In the United States, the cost of GLP-1 RAs can range from $800 to $1300 per month without insurance, which poses a substantial barrier to access and adherence for many patients.4,5”

    From The expanding role of GLP-1 receptor agonists: a narrative review of current evidence and future directions (2025) · cites this paper as [5]

    The Results (Study Participants; page unavailable) report an estimated mean of $69.20 for a 30-day GLP-1 RA supply, but that figure is an insured cohort's out-of-pocket cost. The Methods (Determining OOP Drug Costs; page unavailable) explain that the estimate uses health-plan copays and deductibles. Searches for the $800–$1300 range and uninsured or cash prices found no passage addressing the claimed monthly cost without insurance. (LLM, conf=0.76)

    Experimental — accuracy not yet measured

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Luo, J., Feldman, R., Callaway Kim, K., Rothenberger, S., Korytkowski, M., Hernandez, I., & Gellad, W. F.
Affiliations
  • University of Pittsburgh (US)
  • Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania
  • University of California San Diego (US)
Corresponding author
Luo, Jing
ORCID
Venue
JAMA Network Open
Publisher
American Medical Association (AMA)
Published
2023-06-12
Volume
6
Issue
6
Pages
e2317886
ISSN
2574-3805 (ISSN-L)
DOI
10.1001/jamanetworkopen.2023.17886
PMC
PMC10261995
OpenAlex
W4380291064
Cited by
75
References
22
Citations by year
2026: 23 · 2025: 24 · 2024: 23 · 2023: 5
FWCI
12.39
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • JAMA Network Open (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (submittedVersion, cc-by)
Funders
  • Pfizer
  • Leona M. and Harry B. Helmsley Charitable Trust
  • National Institutes of Health (5kl2tr001856-10)
  • Agency for Healthcare Research and Quality
  • National Heart, Lung, and Blood Institute (K01HL142847)
  • National Institute of Diabetes and Digestive and Kidney Diseases (K23DK120956)
  • National Center for Advancing Translational Sciences (KL2?TR001856)
Topics
  • Health Sciences › Medicine › Endocrinology, Diabetes and Metabolism › Diabetes Treatment and Management
  • Social Sciences › Economics, Econometrics and Finance › Economics and Econometrics › Pharmaceutical Economics and Policy
  • Social Sciences › Economics, Econometrics and Finance › Economics and Econometrics › Health Systems, Economic Evaluations, Quality of Life
Keywords
out-of-pocket costs, SGLT2 inhibitors, GLP-1 receptor agonists, type 2 diabetes, cardiovascular disease, treatment intensification, metformin monotherapy, medication adherence, health insurance coverage, Medicare Advantage, financial barriers to care, glycemic control, HbA1c, LDL cholesterol, blood pressure, cardiovascular risk reduction
MeSH terms
Sodium-Glucose Transporter 2 Inhibitors, Aged, Cardiovascular Diseases, Diabetes Mellitus, Type 2, Health Expenditures, Female, Medicare, Humans, Male, Metformin, Retrospective Studies, United States, Cohort Studies, Glucagon-Like Peptide 1

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-16.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-16.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review