Medication Persistence and Its Impact on Type 2 Diabetes

ABSTRACT

Objectives: Medication persistence in type 2 diabetes (T2D) is a critical factor for preventing adverse clinical events. We assessed persistence among newly treated patients with T2D and documented the impact of persistence on clinical outcomes and costs.

Study Design: Retrospective study of Optum Clinformatics Data Mart commercial and Medicare Advantage enrollees from 2007 to 2020.

Methods: We identified adult patients who initiated antidiabetic treatments. Patients were required to have at least 1 treatment-free year prior to their first T2D prescription. Persistence was measured as the duration of continuous therapy until a 60-day gap in drug availability appeared in any antidiabetic therapy. Factors associated with duration were documented, focusing on the initial class(es) of T2D drugs. The impact of treatment duration on the risk of hospitalization and on total health care costs was also examined.

Results: A total of 673,265 patients were included, with a median follow-up of 7 years. Only 22% of patients maintained continuous treatment, of whom 10% added a second medication. A 1-month increase in duration was associated with reduced risk of hospitalization due to stroke by 0.54% (95% CI, 0.46%-0.60%), acute myocardial infarction by 0.51% (95% CI, 0.44%-0.57%), and all-cause hospitalization by 0.36% (95% CI, 0.34%-0.37%). A 1-month increase in duration was associated with a year-to-year decrease in medical costs of $51 (95% CI, –$54 to –$48) and an increase in year-to-year drug costs of $14 (95% CI, $13-$14).

Conclusions: Our findings show low persistence among patients with T2D and emphasize the importance of medication persistence, which is associated with cost savings and lower risk of hospitalizations.

Am J Manag Care. 2024;30(4):e124-e134. https://doi.org/10.37765/ajmc.2024.89534

Takeaway Points

We observed suboptimal medication persistence among newly treated patients with type 2 diabetes with a median follow-up of 7 years. Persistence is imperative in the management of chronic diseases.

• Only 22% of newly treated patients with type 2 diabetes were persistent with antidiabetic medications. The remaining 78% discontinued medications for at least 60 days; 25% never restarted antidiabetic medication, and 53% restarted treatment.
• Persistent patients were associated with reduced hospitalization rates, including for acute myocardial infarction and stroke.
• Persistent patients were associated with lower total health care costs because their increased pharmacy costs were offset by lower medical costs.
• Our findings highlight the importance of persistence in drug management among patients with type 2 diabetes.

Type 2 diabetes (T2D) is a complex, chronic, and progressive disease characterized by insufficient insulin production, insulin resistance, or both. In 2020, T2D affected more than 34 million adults (18 years or older) in the US and was ranked as the seventh leading cause of death.¹ Approximately 27% of US adults 65 years or older have received a diagnosis of T2D, and its prevalence is projected to increase, with an estimated 55 million US adults affected by 2030.^1,2 Uncontrolled T2D is associated with macrovascular and microvascular complications, including a 5-fold increase in the risk of developing stroke.³ Furthermore, T2D imposes a considerable economic burden, with estimated costs approximating $203 billion in direct medical expenses and a further $87 billion in lost productivity in 2017.⁴

The therapeutic strategy for T2D usually involves the use of antidiabetic medications. According to the 2022 American Diabetes Association guidelines, metformin was the preferred first-line therapy for T2D.⁵ However, in cases where metformin is contraindicated or not well tolerated, newly treated patients may be prescribed alternative medications such as sulfonylureas, thiazolidinediones, sodium-glucose cotransporter 2 (SGLT2) inhibitors, dipeptidyl-peptidase-4 (DPP-4) inhibitors, and glucagon-like peptide-1 receptor agonists (GLP-1 RAs) as first-line monotherapy or in conjunction with metformin.

Persistence with antidiabetic medications over extended periods is paramount for achieving optimal glycemic targets and preventing complications. Adherence and persistence are 2 distinct concepts used to assess the impact of medication-taking behaviors on health outcomes.⁶ Adherence refers to the extent to which a patient follows the health care provider’s recommendations for medication usage over a period of time, often measured as a proportion of days covered or a medication possession ratio. On the other hand, persistence measures the duration of time a patient continues taking a medication without interruption.⁶ It is calculated from when the initial prescription is filled to the termination of treatment or a significant gap in refills. Unlike adherence, persistence can clarify the temporal relationship between a patient’s treatment status and future clinical events. Using statistical methods such as Cox proportional hazards regression models, calculating persistence allows for the assessment of adverse outcomes without being limited to a fixed period.

Research has shown that among patients with T2D, nonpersistence with medication has been associated with poorer glycemic control, higher hospitalization rates, and increased clinical complications.^7-10 Systematic literature reviews have been published on persistence with antidiabetic medications. They showed that most studies focused primarily on estimating patients’ persistence over 12 to 24 months, with follow-up periods ranging from 6 months to 3 years.^8,11-13 The reported persistence rates varied from 16% to 94%.¹² This wide range of variability in persistence can be partly attributed to the length of the follow-up period. A meta-analysis of DPP-4 inhibitors showed that as the follow-up period increased, the persistence rate decreased from 75.6% at 1 year to 31.4% at 3 years.¹⁴ Given that previous studies provided only snapshots with relatively short follow-ups, we sought to assess the level of persistence among patients with T2D regarding antidiabetic medications over a longer period and to update our understanding of persistence in the current landscape (2007-2020).

In this study, we evaluated medication persistence achieved by patients with T2D starting from their initial episode of drug therapy and explored determinants associated with persistence in antidiabetic medications. In addition, we estimated the impact of treatment duration on the risk of hospitalization and on health care costs.

METHODS

Data Source

This retrospective study utilized Optum’s deidentified Clinformatics Data Mart database, which includes national data on more than 65 million commercial and Medicare Advantage enrollees, for January 1, 2007, to December 31, 2020. The data included member socioeconomic demographics, continuous insurance enrollment periods, medical claims, and pharmacy claims for all enrollees and outpatient laboratory results for a subset of patients (~ 35%). Baseline hemoglobin A_1c (HbA_1c) data were included in the analyses to comprehensively assess treatment persistence and its impact on outcomes. The use of the Optum data was approved by the University of Southern California Institutional Review Board.

Study Cohort

We identified adults 18 years or older who initiated therapy using 1 or more oral antidiabetic medications: metformin, sulfonylureas, thiazolidinediones, SGLT2 inhibitors, DPP-4 inhibitors, GLP-1 RAs, or meglitinides. Patients were included when they had at least 1 recorded diagnosis of T2D (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 250.x0, 250.x2; International Classification of Diseases, Tenth Revision, Clinical Modification [ICD-10-CM] code E11). In addition, patients were required to have continuous insurance enrollment for at least 1 year before and after the initiation of antidiabetic medication to focus this analysis on newly treated patients. Patients must not have been prescribed T2D drugs during the first year of their enrollment (treatment-free year). Patients with type 1 diabetes, gestational diabetes, pregnancy, or complications of pregnancy, childbirth, and puerperium were excluded.

Definition of Initial Treatment Regimen

We defined a patient’s initial treatment regimen as the class or classes of medications used within the first 6-week period after the patient’s earliest fill date for antidiabetic medication (index date). This time frame allows for adjustments in the initial drug treatment regimen, which are often made during the early weeks of therapy for newly treated patients. For our analysis, we categorized treatment regimens as monotherapy, which consists of a single class of antidiabetic medication, or combination therapy, which involves multiple classes. Patients with insulin as their initial therapy were excluded because they were assumed unlikely to be newly treated or had specific indications for initiating insulin as their first observed therapy.

Key Dependent and Independent Variables

This study of newly treated patients with T2D is divided into 2 parts. First, we examined factors associated with treatment duration, with a special interest in the drug classes used as initial therapy. The second set of analyses evaluated the impact of the treatment duration on patients’ risk of hospitalizations (both overall and cardiovascular related) and the change in their total health care costs in their first posttreatment year relative to their costs in the year prior to starting treatment.

Treatment duration was measured as the total supply days from the index date to the last date of medication termination or the start of a significant medication gap. In this study, we defined a treatment gap as a period of 60 days, one of the most commonly used definitions for diabetes, because adverse events from discontinuing medications develop gradually over time.¹⁵ This definition allows for medication switches without considering them as breaks in therapy. Discontinuation was identified when a patient experienced a 60-day gap in treatment.

Diagnostic Histories

In this study, patients’ diagnostic histories were captured using ICD-9-CM/ICD-10-CM codes over 1 year prior to the index date to account for different baseline patient characteristics. These histories included conditions such as chronic kidney disease, hyperlipidemia, hypertension, congestive heart failure, and stroke (eAppendix Table 1 [eAppendix available at ajmc.com]).

Health Care Costs

Costs in this study reflected standardized prices for allowed payments across all provider services. The costs were categorized into medical costs and prescription drug costs. Medical costs measured spending from inpatient, outpatient, ancillary, emergency department, and professional services (eAppendix Table 2). Total health care costs were the sum of medical and prescription costs.

Prescription Drug Use

Prescription drug use during the 1-year pre–index date period was categorized based on either the American Hospital Formulary Service (AHFS) Clinical Drug Information or use of cardiovascular-related drugs. AHFS drug classifications included antihistamine, antineoplastic, cardiovascular, and anti-infective drugs (eAppendix Table 3). Our other approach factored in cardiovascular-related drugs prescribed during the 1-year baseline (α-blockers, angiotensin-converting enzyme inhibitors, β-blockers, calcium channel blockers, diuretics, angiotensin receptor blockers, statins, antiplatelets, and aldosterone).

Covariates

The covariates used in this study encompass sociodemographic and clinical characteristics: age, sex, health insurance type, race, time trends to account for changes in clinical practices, prior hospitalization during the 1-year preindex period, whether a baseline HbA_1c test was recorded to control for potential differences in providers and recording system, diagnostic history, total health care costs from the 1-year preindex period as a proxy of patient’s baseline disease severity, prescription of drug classes defined by AHFS Clinical Drug Information to control for different baseline health status, and/or use of cardiovascular-related drugs at baseline.

Statistical Analyses

First, we used Cox proportional hazards models to document the association between the initial antidiabetic medications and treatment discontinuation, controlling for the baseline sociodemographic and clinical characteristics. Two models were conducted to understand the impact of using the different definitions of prescription drug use. Model 1 controlled for age, sex, health insurance type, race, time trends, prior hospitalization, whether a baseline HbA_1c test was recorded, diagnostic history, total health care costs from the 1-year preindex period, and prescription of drug classes defined by AHFS. Model 2 controlled for the same covariates used in model 1 but replaced AHFS classification with the use of cardiovascular-related drugs at baseline.

Second, we examined the association between treatment duration and risk of all-cause hospitalization and hospitalizations specific to stroke and acute myocardial infarction (AMI) by using a time-varying multivariable Cox model, controlling for age, sex, health insurance type, race, time trends, prior hospitalization, whether a baseline HbA_1c test was recorded, diagnostic history, total health care costs from the 1-year preindex period, AHFS drug classifications, and use of cardiovascular-related drugs. To facilitate interpretation, we converted treatment duration from days to months by dividing its value by 30, allowing us to express the estimated coefficient from the Cox model as the estimated effect of an additional month of treatment.

Third, for economic analyses, we utilized multivariable ordinary least squares with a difference-in-difference design to assess the incremental costs between preindex and postindex years after the start of antidiabetic medication, controlling for age, sex, health insurance type, race, time trends, prior hospitalization, whether a baseline HbA_1c test was recorded, diagnostic history, and AHFS drug classifications.

Two sensitivity analyses were conducted: (1) controlling for baseline HbA_1c level instead of the binary variable of whether HbA_1c was recorded and (2) controlling for Charlson Comorbidity Index (CCI) score instead of diagnostic history. For the sensitivity analysis of HbA_1c level, initial episodes for which the baseline value of patients’ HbA_1c was not available were dropped. All analyses were performed using the SAS 9.3 for UNIX (SAS Institute Inc) and Stata 17.0 (StataCorp LLC).

RESULTS

Sample Selection and Baseline Characteristics

The study cohort consisted of a total of 673,265 patients (Figure), and the median follow-up period was 7 years. The most significant inclusion/exclusion criteria were the requirements for a minimum of 2 years of data around the index date and no history of antidiabetic medications within 1 year prior to the index date (Figure). The Figure provides a detailed breakdown of the study population according to their treatment patterns and treatment duration.

The mean (SD) age of the cohort was 59 (14.6) years, with 345,225 men (51%) and 328,040 women (49%).Metformin monotherapy was the most common treatment in the initial regimen, accounting for 81% to 92% of initial regimens (Table 1 [part A and part B]). Combination therapy accounted for 7% to 22% of all initial drug regimens and showed correlations with treatment duration and augmentation/switching (Table 1). We observed that 17% to 45% of patients with T2D were already on an angiotensin-converting enzyme (ACE) inhibitor, β-blocker, calcium channel blocker, and/or statin at the start of their enrollment or during the first year of the study data period (Table 1).

Persistence

Only 22% of all patients were persistent with initial treatment regimens; 12% stayed on the initial treatment for a median of 27 months over a median follow-up of 5.9 years, and another 10% of patients changed or augmented their initial therapy for a median duration of 39 months over a median follow-up of 7.1 years (Table 1). The remaining 78% of patients were not persistent with antidiabetic treatment. Of the total cohort, 25% discontinued their initial therapy with no follow-on treatment attempt, with a median treatment duration of 5 months over a median follow-up of 6.5 years (Table 1). Meanwhile, 29% stopped and then restarted their initial therapy after a break of more than 60 days, with a median treatment duration of 22 months over a median follow-up of 7.1 years (Table 1). Lastly, 24% changed to different regimens after a break in initial treatment, with a median treatment duration of 43 months over a median follow-up of 7.1 years (Table 1).

Factors Associated With Persistence

The results of the Cox analyses of T2D treatment duration are presented in Table 2 [part A and part B]. Sulfonylurea drugs, available prior to 2008, were found to significantly reduce the likelihood of treatment discontinuation during their initial treatment attempt compared with metformin. Patients using thiazolidinediones, SGLT2 inhibitors, or DPP-4 inhibitors were also significantly less likely to discontinue treatment compared with those using metformin. On the other hand, patients initiating treatment using a GLP-1 RA or meglitinide had a higher likelihood of discontinuation compared with those using metformin. Patients using combination therapy were associated with a significantly lower risk of discontinuation vs those using metformin.

Other factors associated with treatment discontinuation included age and health insurance type. Older patients and those enrolled in a health maintenance organization, fee-for-service insurance, or a preferred provider organization were less likely to discontinue treatment compared with younger patients and patients with point-of-service insurance. Patients with a prior hospitalization were associated with a significantly increased risk of discontinuation. Finally, all patient race categories other than White were more likely to discontinue therapy compared with White patients.

Risk of Hospitalizations

The results of the multivariable analyses of hospitalization risk, adjusting for other risk factors, are summarized in Table 3 [part A and part B]. A 1-month increase in treatment duration was associated with lower hazard rates of hospitalizations. The rate of all-cause hospital admissions after the start of antidiabetic regimens was approximately 31% in the study cohort. An additional month of continuous antidiabetic drug therapy was associated with a reduced risk of all-cause hospitalization by 0.36% (95% CI, 0.34%-0.37%). Hospitalization related to stroke was reported in 1.5% of the study cohort. A 1-month increase in treatment duration was associated with a lower incidence of stroke-related hospitalization by 0.54% (95% CI, 0.46%-0.60%). The hospitalization rate for AMI was 1.7%, and an additional month of continuous drug therapy was associated with a reduced risk of AMI-related hospitalization by 0.51% (95% CI, 0.44%-0.57%).

Change in Annual Costs

A 1-month increase in T2D treatment duration was associated with a significant reduction in the year-to-year difference in medical costs of $51 (Table 4 [part A and part B]). In contrast, an additional month of persistence was associated with a significant increase in the year-to-year difference in prescription drug costs of $14. Despite the increase in prescription drug costs, the overall effect on total health care costs was a reduction of $37 per month of continuous antidiabetic medication therapy. This suggests that longer treatment duration is associated with cost savings in terms of medical expenses, outweighing the increase in prescription drug costs.

Sensitivity Analyses

In the sensitivity analyses using patients with a baseline HbA1c value, we had a 63% reduction in the study sample (n = 250,003). Overall, the main outcomes were not sensitive to limiting the analysis to patients with a baseline HbA_1c value. The median baseline HbA_1c level was 6.7% for those with metformin as their initial regimen, 7.5% for sulfonylureas, 7.1% for thiazolidinediones, 7.5% for SGLT2 inhibitors, 7.3% for DPP-4 inhibitors, 6.2% for GLP-1 RAs, and 6.9% for meglitinides (eAppendix Table 4). However, starting therapy using 2 or more antidiabetic medications was associated with a significant reduction in the likelihood of discontinuation by 32% (eAppendix Table 5). This reduction may be attributed to a positive correlation between baseline HbA_1c and duration, indicating a reduced risk of terminating therapy. Hence, patients with higher baseline HbA_1c values are more likely to stay on therapy.

All sensitivity analyses confirmed that our conclusions remained robust. The impact of an additional month of uninterrupted therapy on patient risk and cost remained unchanged in the sensitivity analyses (eAppendix Table 6). Additionally, based on the sensitivity analyses with CCI score, the conclusions remained consistent whether we used the diagnostic history or CCI score as covariates to control for different patient characteristics (eAppendix Tables 7 and 8). Consistent with the main results, persistence with the initial therapy was associated with reduced risk of hospitalization and total health care costs.

DISCUSSION

With a median follow-up of 7 years, our study findings demonstrate that only 22% of patients with T2D maintained continuous treatment with antidiabetic medications and that 78% discontinued their use for at least 60 days. This finding aligns with previous research, which has reported suboptimal persistence and significant variability in persistence rates, ranging from 16.9% over 3 years to 94% over 6 months.^8,11-13,15 To the best of our knowledge, this study encompasses the longest follow-up period examining medication persistence among newly treated patients with T2D. Our study aimed to analyze the long-term persistence with established agents such as metformin, sulfonylureas, and thiazolidinediones as well as relatively newer agents including SGLT2 inhibitors, DPP-4 inhibitors, and GLP-1 RAs. The FDA approved the first agent in the GLP-1 RA class in 2005, the DPP-4 inhibitor class in 2006, and the SGLT2 inhibitor class in 2013. By utilizing data from 2007 to 2020, our study can provide insights into long-term persistence and its impact on hospitalization and costs. In addition, it is worth noting that our study adopted a conservative approach for defining nonpersistence as a gap in medication supply of 60 days, whereas other studies have utilized a threshold of more than 90 days. In our effort to understand persistence, we were able to track a patient’s persistence over time. Medication persistence is a crucial element for determining the effectiveness of pharmacological therapies for individuals with a chronic disease such as T2D. Unfortunately, our study found that 25% of newly treated patients with T2D stopped taking their antidiabetic treatments and never restarted; it also revealed that 53% discontinued and then restarted their medications. This nonpersistence can pose a challenge in disease management.

Our assessment of medication persistence among newly treated patients with T2D revealed significant differences in persistence for various antidiabetic medications. Sulfonylureas, thiazolidinediones, SGLT2 inhibitors, and DPP-4 inhibitors were associated with significantly longer persistence compared with metformin, whereas GLP-1 RAs and meglitinides were associated with significantly shorter persistence than metformin. Despite metformin being recommended as the preferred first-line therapy for T2D, our findings indicate low persistence with metformin among newly treated patients compared with other oral antidiabetic medications (both in the main analysis and the sensitivity analysis adjusting for baseline HbA_1c level). The low persistence with metformin could be attributed to the high occurrence of gastrointestinal adverse events, leading to intolerance and discontinuation.^15,16 Alternatively, those patients who started antidiabetic medications other than metformin may represent distinct populations with potentially higher risk factors, which might not be fully accounted for even after adjusting for heterogeneous baseline severity.

Among the drug classes used in the management of T2D, SGLT2 inhibitors and GLP-1 RAs have been gaining popularity in recent years due to their proven benefits in cardiovascular health, renal health, and weight loss. Annual prescriptions for SGLT2 inhibitors and GLP-1 RAs nearly tripled from 2015 to 2020.^17,18 In recent years, the American Diabetes Association has shifted its guideline recommendations to emphasize the use of SGLT2 inhibitors and GLP-1 RAs for cardiovascular and renal risk reduction in patients at high risk.⁵ Growing attention is being paid to persistence associated with these drug classes,^19,20 so our study compared persistence across different classes. Because persistence remains suboptimal, improving persistence with these drug classes could serve as an effective strategy for enhancing patient outcomes and health care efficiency in the long run.

This study’s results emphasize that nonpersistence with antidiabetic medications is associated with suboptimal health outcomes. We observed that improved persistence was associated with a reduction in hospitalizations, which is consistent with previous findings.^7-10,12 Additionally, our study’s results align with previous findings that demonstrated medication adherence is inversely associated with costs related to outpatient care, emergency department visits, and hospitalization.^12,21-23 Improved medication persistence would significantly affect long-term health care costs because each month of persistence is associated with reduced risk of major cardiovascular events and hospitalization. As T2D progresses and the risk of complications increases, the cost savings associated with reduced risk of complications would be expected to be even more significant.

The importance of treatment persistence becomes even more apparent when considering the prevention of complications. T2D is a progressive condition, often requiring a combination of medications over time.²⁴ This progression and the complexity of treatment protocols can challenge patients to stay on track, resulting in discontinuation. Findings from the UK Prospective Diabetes Study, a long-term observational study, suggested that early glycemic control may have lasting effects in reducing the risk of T2D-related complications.²⁵ Recent research on patients with newly diagnosed T2D also demonstrated that tight glycemic control during the first year after diagnosis was significantly associated with a decrease in future risk of diabetic complications and mortality.²⁶ These findings underscore the importance of intensively treating patients with T2D at an earlier stage to achieve better long-term improvements in patient health. Therefore, establishing and maintaining treatment persistence from the beginning can greatly improve overall health outcomes.

Limitations

Our study is subject to several limitations. First, our study may have treatment selection bias due to the nature of a retrospective study design. There may still exist inherent and unobservable differences between persistent and nonpersistent patients (eg, self-care behaviors) that could affect our results. Second, some antidiabetic medications are generically available at a low price, so a patient may elect to fill their prescriptions for these medications without submitting the prescription for insurance coverage if the co-pay exceeds the direct price at the pharmacy. Although the extent of direct purchase is unknowable, this practice might be uncommon given that a zero or very low co-pay is specified by most drug plans for many generic medications. Moreover, direct purchase must cover both the ingredient cost and the pharmacy’s dispensing fee. Third, we cannot track adverse events or unrecorded health outcomes (eg, reduced symptom severity) from the data. Fourth, cardiovascular-related mortality is an important outcome for patients with T2D, but mortality information was not available. Moreover, all patients were required to have survived at least 1 year following the initiation of drug therapy, which may introduce survivorship bias. Finally, this study is limited in its direct comparability with other studies, especially those that utilized severity indices such as the Diabetes Complications Severity Index (DCSI). For example, we controlled for diagnostic histories of diabetic retinopathy, nephropathy, neuropathy, and cerebrovascular, cardiovascular, and peripheral vascular diseases but were unable to control for severity levels. Also, laboratory test results listed in the DCSI were not utilized in our study due to limited access to laboratory data. Our approach may pose a challenge when comparing our findings with those of studies that used the complete severity index. Nonetheless, given that this study focuses on newly treated patients with T2D and excludes those with insulin as their first therapy, we expect that our predictions would perform well.

CONCLUSIONS

Persistence with antidiabetic medications among newly treated patients with T2D can significantly reduce the risk of hospitalizations as well as total health care costs. These results underscore the importance of promoting medication persistence in T2D management and implementing an assistance program that would support patients to remain persistent with antidiabetic medications.

Author Affiliations: Department of Pharmaceutical and Health Economics (JHC, SX, JMc) and Department of Clinical Pharmacy (REK), USC Alfred E. Mann School of Pharmacy and Pharmaceutical Sciences (AG, JMa), University of Southern California, Los Angeles, CA.

Source of Funding: None.

Author Disclosures: Ms Choe reports receiving grants from the Department of Pharmaceutical and Health Economics in the USC Alfred E. Mann School of Pharmacy and Pharmaceutical Sciences at the University of Southern California. The department also provided access to the Optum Clinformatics Data Mart for graduate student training and research. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (JHC, SX, JMc, REK); acquisition of data (JHC, JMc); analysis and interpretation of data (JHC, SX, AG, JMa, JMc, REK); drafting of the manuscript (JHC, SX, JMc); critical revision of the manuscript for important intellectual content (JHC, AG, JMa, JMc, REK); statistical analysis (JHC, JMc); administrative, technical, or logistic support (AG, JMa); and supervision (JMc, REK).

Address Correspondence to: Jee H. Choe, MS, Department of Pharmaceutical and Health Economics, USC Alfred E. Mann School of Pharmacy and Pharmaceutical Sciences, University of Southern California, 635 Downey Way, VPD 414C, Los Angeles, CA 90089-3333. Email: jhchoe@usc.edu.

REFERENCES

1. National Diabetes Statistics Report, 2020: Estimates of Diabetes and Its Burden in the United States. CDC; 2020. Accessed October 12, 2022. https://www.cdc.gov/diabetes/pdfs/data/statistics/national-diabetes-statistics-report.pdf

2. Rowley WR, Bezold C, Arikan Y, Byrne E, Krohe S. Diabetes 2030: insights from yesterday, today, and future trends. Popul Health Manag. 2017;20(1):6-12. doi:10.1089/pop.2015.0181

3. Tun NN, Arunagirinathan G, Munshi SK, Pappachan JM. Diabetes mellitus and stroke: a clinical update. World J Diabetes. 2017;8(6):235-248. doi:10.4239/wjd.v8.i6.235

4. American Diabetes Association. Economic costs of diabetes in the U.S. in 2017. Diabetes Care. 2018;41(5):917-928. doi:10.2337/dci18-0007

5. American Diabetes Association Professional Practice Committee. 9. pharmacologic approaches to glycemic treatment: Standards of Medical Care in Diabetes—2022. Diabetes Care. 2022;45(suppl 1):S125-S143. doi:10.2337/dc22-S009

6. Cramer JA, Roy A, Burrell A, et al. Medication compliance and persistence: terminology and definitions. Value Health. 2008;11(1):44-47. doi:10.1111/j.1524-4733.2007.00213.x

7. Ho PM, Rumsfeld JS, Masoudi FA, et al. Effect of medication nonadherence on hospitalization and mortality among patients with diabetes mellitus. Arch Intern Med. 2006;166(17):1836-1841. doi:10.1001/archinte.166.17.1836

8. Asche C, LaFleur J, Conner C. A review of diabetes treatment adherence and the association with clinical and economic outcomes. Clin Ther. 2011;33(1):74-109. doi:10.1016/j.clinthera.2011.01.019

9. Yu AP, Yu YF, Nichol MB. Estimating the effect of medication adherence on health outcomes among patients with type 2 diabetes—an application of marginal structural models. Value Health. 2010;13(8):1038-1045. doi:10.1111/j.1524-4733.2010.00787.x

10. Stuart BC, Simoni-Wastila L, Zhao L, Lloyd JT, Doshi JA. Increased persistency in medication use by U.S. Medicare beneficiaries with diabetes is associated with lower hospitalization rates and cost savings. Diabetes Care. 2009;32(4):647-649. doi:10.2337/dc08-1311

11. Cramer JA. A systematic review of adherence with medications for diabetes. Diabetes Care. 2004;27(5):1218-1224. doi:10.2337/diacare.27.5.1218

12. Evans M, Engberg S, Faurby M, Fernandes JDDR, Hudson P, Polonsky W. Adherence to and persistence with antidiabetic medications and associations with clinical and economic outcomes in people with type 2 diabetes mellitus: a systematic literature review. Diabetes Obes Metab. 2022;24(3):377-390. doi:10.1111/dom.14603

13. Iglay K, Cartier SE, Rosen VM, et al. Meta-analysis of studies examining medication adherence, persistence, and discontinuation of oral antihyperglycemic agents in type 2 diabetes. Curr Med Res Opin. 2015;31(7):1283-1296. doi:10.1185/03007995.2015.1053048

14. Ogundipe O, Mazidi M, Chin KL, et al. Real-world adherence, persistence, and in-class switching during use of dipeptidyl peptidase-4 inhibitors: a systematic review and meta-analysis involving 594,138 patients with type 2 diabetes. Acta Diabetol. 2021;58(1):39-46. doi:10.1007/s00592-020-01590-w

15. McGovern A, Tippu Z, Hinton W, Munro N, Whyte M, de Lusignan S. Comparison of medication adherence and persistence in type 2 diabetes: a systematic review and meta-analysis. Diabetes Obes Metab. 2018;20(4):1040-1043. doi:10.1111/dom.13160

16. Bonnet F, Scheen A. Understanding and overcoming metformin gastrointestinal intolerance. Diabetes Obes Metab. 2017;19(4):473-481. doi:10.1111/dom.12854

17. McGuire DK, Shih WJ, Cosentino F, et al. Association of SGLT2 inhibitors with cardiovascular and kidney outcomes in patients with type 2 diabetes: a meta-analysis. JAMA Cardiol. 2021;6(2):148-158. doi:10.1001/jamacardio.2020.4511

18. Del Olmo-Garcia MI, Merino-Torres JF. GLP-1 receptor agonists and cardiovascular disease in patients with type 2 diabetes. J Diabetes Res. 2018;2018:4020492. doi:10.1155/2018/4020492

19. Uzoigwe C, Liang Y, Whitmire S, Paprocki Y. Semaglutide once-weekly persistence and adherence versus other GLP-1 RAs in patients with type 2 diabetes in a US real-world setting. Diabetes Ther. 2021;12(5):1475-1489. doi:10.1007/s13300-021-01053-7

20. Ofori-Asenso R, Liew D, Lalic S, et al. Adherence, persistence, and switching among people prescribed sodium glucose co-transporter 2 inhibitors: a nationwide retrospective cohort study. Adv Ther. 2019;36(11):3265-3278. doi:10.1007/s12325-019-01077-3

21. Nasseh K, Frazee SG, Visaria J, Vlahiotis A, Tian Y. Cost of medication nonadherence associated with diabetes, hypertension, and dyslipidemia. Am J Pharm Benefits. 2012;4(2):e41-e47.

22. Roebuck MC, Liberman JN, Gemmill-Toyama M, Brennan TA. Medication adherence leads to lower health care use and costs despite increased drug spending. Health Aff (Millwood). 2011;30(1):91-99. doi:10.1377/hlthaff.2009.1087

23. Breitscheidel L, Stamenitis S, Dippel FW, Schöffski O. Economic impact of compliance to treatment with antidiabetes medication in type 2 diabetes mellitus: a review paper. J Med Econ. 2010;13(1):8-15. doi:10.3111/13696990903479199

24. Higgins V, Piercy J, Roughley A, et al. Trends in medication use in patients with type 2 diabetes mellitus: a long-term view of real-world treatment between 2000 and 2015. Diabetes Metab Syndr Obes. 2016;9:371-380. doi:10.2147/DMSO.S120101

25. Holman RR, Paul SK, Bethel MA, Matthews DR, Neil HA. 10-year follow-up of intensive glucose control in type 2 diabetes. N Engl J Med. 2008;359(15):1577-1589. doi:10.1056/NEJMoa0806470

26. Laiteerapong N, Ham SA, Gao Y, et al. The legacy effect in type 2 diabetes: impact of early glycemic control on future complications (the Diabetes & Aging Study). Diabetes Care. 2019;42(3):416-426. doi:10.2337/dc17-1144