Economic Burden of Nonadherence to Standards of Diabetes Care

ABSTRACT

Objectives: To evaluate the effect of nonadherence to American Diabetes Association (ADA) guidelines on health care expenditures for patients with type 2 diabetes (T2D).

Study Design: Retrospective cross-sectional cohort design, utilizing 2016-2018 Medical Expenditure Panel Survey data.

Methods: Patients with a diagnosis of T2D who completed the supplemental T2D care survey were included in the study. Participants were categorized based on adherence to the 10 processes in the ADA guidelines into adherent (≥ 9 processes) and nonadherent (≤ 6 processes) categories. Propensity score matching was employed using a logistic regression model. After matching, total annual health care expenditure change from baseline year was compared using a t test. Further, imbalanced variables were controlled for in a multivariable linear regression model.

Results: A total of 1619 patients representing 15,781,346 (SE = 438,832) individuals met the inclusion criteria, among whom 12.17% received nonadherent care. After propensity matching, those who received nonadherent care had $4031 higher total annual health care expenditures compared with their baseline year, whereas patients who received adherent care had $128 lower total annual health care expenditures compared with their baseline year. Further, multivariable linear regression adjusted for the imbalanced variables indicated that nonadherent care was associated with a mean (SE) $3470 ($1588) increase in the change from baseline health care expenditure.

Conclusions: Nonadherence to the ADA guidelines results in a significant increase in health care expenditures among patients with diabetes. The economic impact of nonadherent care for T2D is a significant and extensive issue that needs to be addressed. These findings emphasize the importance of providing care based on ADA guidelines.

Am J Manag Care. 2023;29(6):e176-e183. https://doi.org/10.37765/ajmc.2023.89376

Takeaway Points

Nonadherence to the standards of diabetes care provided by the American Diabetes Association (ADA) may result in increased health care expenditures among patients with type 2 diabetes.

• An estimated 12.17% of patients with diabetes in the United States during 2016-2018 received diabetes care nonadherent to the ADA guidelines.
• Patients with diabetes receiving nonadherent care had $4031 higher total annual health care expenditures compared with the prior year.
• Conversely, patients with diabetes who received adherent care had $128 lower total annual health care expenditures compared with the prior year.

Type 2 diabetes (T2D) is a chronic metabolic disease and is an ongoing epidemic in the United States. According to the CDC’s National Diabetes Statistics Report, an estimated 11.3% of adults in the United States have T2D, equating to 37.3 million patients.¹ In 2017, the American Diabetes Association (ADA) estimated the total nationwide cost of diagnosed T2D to be $327 billion.² T2D is a chronic disease that requires careful management and is associated with a severe decrease in life expectancy.³ Inadequate management of T2D can lead to increased risk of severe complications.^4-8 In this regard, both screening for and treating diabetes-related complications are essential components in the appropriate management of T2D.⁹

The ADA annually publishes guidelines for standards of medical care in T2D, detailing numerous processes of care for T2D. Adherence to ADA guidelines for T2D management requires appropriate prescription of antihyperglycemics, treatment of diabetes-associated complications, lifestyle modifications, preventive care measures, and annual screenings.¹⁰

Despite strong evidence supporting ADA guidelines, the prevalence of nonadherent care among patients with T2D remains high.¹¹ We demonstrated in our previous study that nonadherence to ADA standards of T2D care guidelines ranged from about 24% to 44% for 5 types of care: lifestyle management, immunization, pharmacologic therapies, physical examinations, and laboratory tests.¹¹

The economic burden of T2D has been steadily increasing over the years.^12-14 In 2017, the economic burden of T2D in the United States was estimated to be $327 billion annually, a 26% increase from 2012.¹⁴ Although the economic burden of T2D is well documented in the literature, the medical cost of nonadherence to guidelines in T2D remains unknown. There is a critical need to examine the health care expenditure and economic burden associated with nonadherent diabetes care. In this study, we aimed to evaluate the effect of receiving diabetes care nonadherent to ADA guidelines on medical costs for T2D.

METHODS

Data Sources

This study had a retrospective cross-sectional cohort design and used 2016-2018 Medical Expenditure Panel Survey (MEPS) data. MEPS is a nationally representative collection of cross-sectional surveys in the United States that captures information on health care use and expenditures from households, providers, and insurers.¹⁵ Given the multicluster sample design that allows for representative sampling in MEPS, survey weights, strata, and primary sampling unit variables were applied to obtain national weighted frequencies.¹⁶ MEPS consists of several panel-based surveys, and individuals in each panel participate in 5 rounds of interviews over a 2-year period. The 2 most recent panels (2016-2017 and 2017-2018) were used for the current study.^17,18

For each panel, the index date was the first day of the panel (January 1, 2016, for the first panel and January 1, 2017, for the second panel). For both panels, baseline characteristics were obtained from the first year of the panel, and study outcomes were obtained from the second year of the panel. The cohort was developed by linking the following data files: a full-year consolidated data file, hospital inpatient stays file, emergency department visits file, outpatient visits file, office-based medical provider visits file, medical conditions file, and prescribed medicines file.¹⁹

Study Population

The sample included adults (≥ 18 years) with a diagnosis of T2D at the index date, based on the supplemental diabetes care survey.²⁰ This survey is given to all those surveyed by MEPS who were identified as having diabetes. Individuals with a diagnosis of T2D were defined as those who had a diagnosis code for T2D, had self-reported receiving a diagnosis of diabetes from a health practitioner, and had at least 1 prescription of an antihyperglycemic drug within 182 days of index date. Individuals were identified using theInternational Classification of Diseases, Tenth Revision, Clinical Modification code for T2D (E11).²¹ Individuals with a T2D diagnosis in the cohort were excluded if they were not prescribed antihyperglycemics within 6 months from the index date. Study sample attrition is demonstrated in Figure 1.

Antihyperglycemic drug classes included in this study were metformin, sulfonylureas, meglitinides, dipeptidyl peptidase-4 inhibitors, thiazolidinediones, glucagon-like peptide-1 receptor agonists, amylin analogues, sodium-glucose cotransporter 2 inhibitors, α-glucosidase inhibitors, and insulin. Drugs were identified using National Drug Codes, and Multum MediSource Lexicon names were used to confirm accuracy and completeness.^22,23

The ADA Standards of Medical Care in Diabetes guideline was used to define and identify 10 processes of T2D care (Table 1).⁹ Data regarding foot exams, eye exams, hemoglobin A_1c (HbA_1c) tests, diet modification, influenza vaccination, and cholesterol tests were obtained from the MEPS diabetes survey.²⁰ In addition, we included 3 processes of care based on medication adherence to antihyperglycemics, high-intensity statins, and medications for hypertension based on the ADA guideline. Medication adherence for antihyperglycemics was measured using the proportion of days covered (PDC) method. A measure of at least 80% was considered adherent. PDC was calculated by total number of days’ supply of filled antihyperglycemics of any type, divided by total days. Days began at first prescription fill of antihyperglycemics detected after the index date. In the case of using multiple antihyperglycemics, a fill of any antihyperglycemic was considered adherent. The next process of care was regarding utilization of high-intensity statins, defined as receiving prescription of atorvastatin 40 mg to 80 mg or rosuvastatin 20 mg to 40 mg during the first year of the study.⁹ Last, we evaluated hypertension treatment during the first year of the study, defined as receiving a prescription of angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, thiazide-like diuretics, or dihydropyridine calcium channel blockers.⁹

Adherent care was defined as adherence to 9 or more processes of care and nonadherent care was defined as adherence to 6 or fewer processes of care. The frequency distribution of process of care adherence was used a priori to help guide this decision. A cutoff of 9 or greater was chosen as adherent since approximately 50% of individuals met this threshold, whereas a cutoff of 6 was chosen as nonadherent, approximating 12% of the sample, representing individuals who are nonadherent outside of the gray area of those who met the criteria of 7 or 8 processes of care. Because this decision was made a priori to the analysis phase, additional analyses using separate cut-offs were not undertaken to avoid cherry-picking results.

The most recent ADA clinical practice guideline was used to conceptualize the processes of care in patients with T2D.⁹ The rationale for the processes of care used for this study and detailed operationalization for processes of care are defined in detail in a prior publication.¹¹ Some diabetes care survey responses contained missing values for some variables, such as number of HbA_1c tests. Individuals with 3 or more missing values for processes of care were excluded, whereas multiple imputation was used to fill missing data points for any individuals with fewer than 3 missing values, using imputation methodology detailed in our previous publication.¹¹

Study Outcomes

Andersen’s behavior model of health care utilization was adapted and used as a conceptual framework to guide the creation of predictor models (eAppendix Figure [eAppendix available at ajmc.com]).^4,24 The outcome measured in the study was the total health care expenditure during the follow-up year compared with the baseline year. Health care expenditure included all health care–related costs, including office-based, emergency department, hospital outpatient, inpatient, prescribed medications, dental, home health, and medical supplies/equipment. All dollars were converted to 2018 US$ based on the Consumer Price Index for medical care.²⁵

Statistical Analysis

A propensity score (PS) model (1:1) was developed using patient, system, and physician variables to match patients with T2D receiving nonadherent care with patients receiving adherent care. PS matching is the optimal approach for this study because this method enables a clear and easily interpretable direct comparison between groups while using most of the nonadherent sample. This methodology also excludes individuals within each group who may have a combination of characteristics that a traditional model could not adjust for if the full sample was included. PS models were employed with the application of the caliper method. The PS model was developed using a logistic regression model for the probability of receiving adherent care vs nonadherent care. Model covariates included age, sex, educational attainment, race, income, smoking status, source of care, geographic region, insurance type, diabetes severity, and comorbidities based on Elixhauser Comorbidity Index (ECI) score, which are described in our prior publication.¹¹ Diabetes severity was identified by using the Diabetes Complications Severity Index (DCSI). The DCSI is a 14-level metric that uses scores for diabetes complications—including cardiovascular, cerebrovascular, and metabolic complications, as well as retinopathy, peripheral vascular disease, neuropathy, and nephropathy—to quantify diabetes severity.²⁶

After matching, a t test was used for comparing the health care expenditure change from baseline during the second year for both cohorts. Further, a multivariable linear regression controlled for variables that were not balanced after matching was performed. Imbalanced covariates were variables with a standardized difference above 0.1 after matching.²⁷ All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc), with a significance level of P < .05. The University of Houston Institutional Review Board approved the project under the exempt category for secondary data analysis.

RESULTS

Population Characteristics

There were 1619 adults with T2D who met the study inclusion criteria, representing a weighted frequency (WF) of 15,781,346 (SE = 438,832) individuals. Among the patients who met the inclusion criteria, 197 (12.17%) (WF = 1,650,297; SE = 133,126) were categorized as receiving nonadherent care and 826 (51.02%) (WF = 8,359,289; SE = 285,406) were categorized as receiving adherent care (Figure 1). Individuals who were adherent to 7 or 8 processes of care did not meet the inclusion criteria and were excluded (n = 596).Those who received nonadherent care were compared with those who received adherent care (eAppendix Table 1). Adherence to annual HbA_1c test was 30.51% and adherence to cholesterol test was 21.93% during the first year (eAppendix Table 2). The frequency of adherence to each component of diabetes care process is provided in eAppendix Table 2. After PS matching, there was a similar number of men and women in each cohort and most individuals in each group were White (47.42% of nonadherent, 42.66% of adherent). More detail on characteristics of cohorts is provided in Table 2 [part A and part B]. Standardized difference was used to identify characteristics that differed significantly between the cohorts.

PS Model

Individuals who received nonadherent care were matched to similar individuals who received adherent care using PS 1:1 matching. Unmatched individuals were excluded (n = 633). Table 2 compares the postmatch characteristics between the cohorts. After matching, some characteristics remained significantly unbalanced between the cohorts including race, education, income level, usual diabetes care facility, physician specialty, insurance type, and ECI score.

Health Expenditure Outcomes

Post matching, those who received nonadherent care had a significant increase from baseline in terms of total health care expenditure (mean = $4031; SE = $600) whereas those who received adherent care had a decrease in health care expenditure (mean = –$128; SE = $987) (P < .001) (Figure 2).

After adjusting for variables that remained imbalanced in PS matching (education, race, poverty level, region, usual diabetes care facility, physician specialty, insurance type, and ECI score), individuals with nonadherent care on average had a significantly higher increase from baseline in health care expenditure compared with individuals who received adherent care (mean [SE] difference, $3470 [$1588]) (P = .028) (Table 3).

DISCUSSION

It is well-known that nonadherence to ADA guidelines results in poorer diabetes outcomes.^4-8 Several studies have shown the high burden of T2D complications.^26,28-30 However, this study is the first to examine the effect of nonadherent diabetes care on health care expenditures using factors related to patient, physician, and systems as covariates. In this real-world retrospective study, we examined the effect of nonadherent diabetes care, as defined by ADA guidelines, on total annual health care expenditures. In this study, we utilized a theoretical model including patient, physician, and system factors to evaluate the impact of nonadherent care on the economic burden of T2D. Our study using a nationally representative sample of the US population indicated that nonadherent T2D care leads to $3470 higher health care expenditures compared with patients receiving adherent care. Considering the prevalence of T2D in the United States (37.3 million)¹ and the prevalence of nonadherent care (12.17%) from our study, the annual burden of nonadherent diabetes care could be as large as $16 billion. Our finding regarding the prevalence of nonadherent care is consistent with those of prior studies that reported adherence to ADA guidelines.^5,31

After PS matching and controlling for unbalanced covariates, nonadherent care was associated with higher change from baseline in terms of the total health care expenditure. To the best of our knowledge, no studies have evaluated the economic burden of nonadherent care based on all components of the ADA guideline. However, there are prior studies that used economic modeling to demonstrate the burden of poor glycemic control in the United Kingdom and United States.^32,33 Bain et al in the United Kingdom used a validated diabetes model (IQVIA CORE) to evaluate the economic burden of poor glycemic control among patients with diabetes at different levels of HbA_1c.³² They found that poor glycemic control (HbA_1c of 8.2% vs 7.0%) even for a brief period (1 year) may result in a population-level economic burden of £85 million ($106.15 million) on a 3-year horizon.³² Another study by Ali et al in the United States using the same diabetes model that was used in the Bain et al study reported an incremental population-level burden of $1.7 billion for 1 year of nonadherence to the recommended HbA_1c values (HbA_1c of 9.0% vs 7.0%).³³ Although our study’s findings are aligned with these previous findings, comparing the findings should take into consideration the difference between the population sizes in each study and the differences among the comparison groups. In addition, it should be noted that our study was based on a nationally representative sample of the US population whereas the aforementioned studies are both economic models with particular assumptions.^32,33

There are several implications of these study findings. Whereas our previous study supported the vital role of ADA guidelines in lowering the risk of T2D complications,⁴ the current study demonstrated the economic impact of low adherence to ADA guidelines. Further, it showed the immediate effect, presented as higher health expenditure in the year following low adherence. Health care cost is a major concern for patients, clinicians, and payers, and our study illustrates the potential cost-saving aspect of adherence to T2D guidelines. Our findings can be used as a basis for implementation of cost-saving managed care programs to improve long-term adherence to T2D guidelines. In a recent study, Benedict et al designed a team-based approach for providing comprehensive care for patients with T2D.³⁴ In the intervention, clinical pharmacists were added to an integrated health care team. The clinical pharmacists were responsible for managing therapy for T2D, as well as for hypertension and hypercholesterolemia, to reach optimal clinical goals. Pharmacists were allowed to order laboratory tests for monitoring the drug therapy and to make modifications to the treatment regimen.³⁴ In addition, pharmacists could address other care gaps including medication adherence, refills, vaccinations, screenings, and lifestyle modifications. Through their intervention, on average, patients with HbA_1c greater than 8.0% saw a 1-unit decrease in their HbA_1c level in 3.4 months.³⁴ Another study evaluated the effect of physician-pharmacist collaboration for management of diabetes.³⁵ The decrease in HbA_1c level among patients who benefited from physician-pharmacist collaboration was significantly higher than among patients who received care solely from a physician (1.75% vs 0.16%).³⁵

Our study had several features that provided exceptional strength. First, we applied strict inclusion criteria for identification of individuals with a T2D diagnosis. In addition, we pulled our cohorts from a nationally representative sample, providing a unique glimpse of T2D care throughout the United States. Further, we utilized PS matching to enhance the comparability between our cohorts. In our matching process we used the DCSI (a metric that covers a wide range of diabetes complications) and the ECI, as well as diabetes drug classification, to control for disease severity. Finally, we applied a difference-in-difference methodology to examine the change from baseline in health care expenditure. For this purpose, we compared the health care expenditure during the follow-up year with that of the initial year. The calculated difference was then used for comparison between the cohorts. This method provided a better control for factors and differences within the cohorts, which may have not been accounted for by other techniques.

Limitations

The first limitation of our study was the small sample size of our cohorts, which might have affected the precision of our estimates. Although utilization of PS matching strengthened our study, the generalizability of our findings might be limited due to the exclusion of some individuals during PS matching. Further, after the PS matching, several covariates remained imbalanced; however, we used a regression model following the PS matching to adjust for any covariates that remained imbalanced. Some variables, such as HbA_1c tests, were missing for some individuals in the data set, and imputation was performed that could result in some bias. Additionally, MEPS relies on self-reported information for some variables, which might result in biases, including reporting bias, recall bias, and response bias. Another important note is that the current study relied on process-of-care definitions based on the ADA guidelines for diabetes and cut-off points based on the adherence distribution among the cohort. Any modification in the definition of nonadherent care may result in deviations in our findings. In consistency with ADA guidelines, we limited antihypertensive agents to specific classes and the anticholesterol agents to high-intensity agents only.³⁶ Another important factor is that we excluded individuals who did not fill their antihyperglycemic medication during the first 6 months after index date, which may have resulted in exclusion of individuals who were severely nonadherent to T2D care. Exclusion of these individuals from the study might have resulted in underestimation of the prevalence of nonadherent care and the effects of nonadherent care on health outcomes. Another limitation of the study might be the reliability of MEPS in identifying patients with T2D, as this function of MEPS has not been checked for validity in previous literature. Finally, even though we adjusted the cohorts for comorbidities, it should be noted that we evaluated total health care expenditures, which may be influenced by factors other than T2D.

CONCLUSIONS

Patients who did not receive adherent care based on the ADA guidelines for diabetes had a significant increase in their expenditures in the second year compared with the baseline year. Insufficient care provided for T2D is an extensive issue with significant economic burden. These findings emphasize the importance of providing care according to the ADA guidelines in the management of T2D.

Author Affiliations: Department of Pharmaceutical Health Outcomes and Policy, College of Pharmacy, University of Houston (MZ, BDL, SSS), Houston, TX; Department of Practice, Sciences, and Health Outcomes Research, School of Pharmacy, University of Maryland Baltimore (JC), Baltimore, MD.

Source of Funding: None.

Author Disclosures: The 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 (MZ, BDL, JC, SSS); acquisition of data (BDL); analysis and interpretation of data (BDL); drafting of the manuscript (MZ, BDL, JC, SSS); critical revision of the manuscript for important intellectual content (MZ, BDL, JC, SSS); statistical analysis (BDL); and supervision (SSS).

Address Correspondence to: Sujit S. Sansgiry, PhD, MS, Department of Pharmaceutical Health Outcomes and Policy, College of Pharmacy, University of Houston, 4849 Calhoun Rd, Health and Biomedical Sciences Building 2 – Office 4055, Houston, TX 77204. Email: ssansgiry@uh.edu.

REFERENCES

1. National Diabetes Statistics Report. CDC. Accessed June 4, 2021. https://www.cdc.gov/diabetes/data/statistics-report/index.html

2. The cost of diabetes. American Diabetes Association. Accessed May 12, 2022. https://www.diabetes.org/resources/statistics/cost-diabetes

3. Tachkov K, Mitov K, Koleva Y, et al. Life expectancy and survival analysis of patients with diabetes compared to the non diabetic population in Bulgaria. PloS One. 2020;15(5):e0232815. doi:10.1371/journal.pone.0232815

4. Lewing B, Sansgiry S, Abughosh SM, Lal LS, Essien EJ, Sansgiry SS. Effect of inadequate care on diabetes complications and healthcare resource utilization during management of type 2 diabetes in the United States. Postgrad Med. 2022;134(5):494-506. doi:10.1080/00325481.2022.2061260

5. Lian J, Liang Y. Diabetes management in the real world and the impact of adherence to guideline recommendations. Curr Med Res Opin. 2014;30(11):2233-2240. doi:10.1185/03007995.2014.952716

6. Suckow BD, Newhall KA, Bekelis K, et al. Hemoglobin A1c testing and amputation rates in Black, Hispanic, and White Medicare patients. Ann Vasc Surg. 2016;36:208-217. doi:10.1016/j.avsg.2016.03.035

7. Goodney PP, Newhall KA, Bekelis K, et al. Consistency of hemoglobin A1c testing and cardiovascular outcomes in Medicare patients with diabetes. J Am Heart Assoc. 2016;5(8):e003566. doi:10.1161/JAHA.116.003566

8. Knighton AJ, McLaughlin M, Blackburn R, et al. Increasing adherence to evidence-based clinical practice. Qual Manag Health Care. 2019;28(1):65-67. doi:10.1097/QMH.0000000000000195

9. Introduction: Standards of Medical Care in Diabetes—2018. Diabetes Care. 2018;41(suppl 1):S1-S2. doi:10.2337/dc18-Sint01

10. American Diabetes Association. Standards of Medical Care in Diabetes—2018 abridged for primary care providers. Clin Diabetes. 2018;36(1):14-37. doi:10.2337/cd17-0119

11. Lewing B, Abughosh SM, Lal LS, Essien EJ, Sansgiry S, Sansgiry SS. Patient, physician, and health system factors associated with five types of inadequate care during management of type 2 diabetes mellitus in the United States. Diabet Epidemiol Manag. 2022;6:100046. doi:10.1016/j.deman.2021.100046

12. McAdam Marx C. Economic implications of type 2 diabetes management. Am J Manag Care. 2013;19(suppl 8):S143-148.

13. Fonseca V, Chou E, Chung HW, Gerrits C. Economic burden of hypoglycemia with basal insulin in type 2 diabetes. Am J Manag Care. 2017;23(2):114-122.

14. 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

15. Cohen JW, Cohen SB, Banthin JS. The Medical Expenditure Panel Survey: a national information resource to support healthcare cost research and inform policy and practice. Med Care. 2009;47(7 suppl 1):S44-S50. doi:10.1097/MLR.0b013e3181a23e3a

16. Download data files — codebook. Agency for Healthcare Research and Quality. March 2, 2020. Accessed May 13, 2022. https://meps.ahrq.gov/mepsweb/data_stats/download_data_files_codebook.jsp?PUFId=H201

17. MEPS HC-210: MEPS panel 22 longitudinal data file. Agency for Healthcare Research and Quality. September 2020. Accessed May 13, 2022. https://www.meps.ahrq.gov/mepsweb/data_stats/download_data_files_detail.jsp?cboPufNumber=HC-210

18. MEPS HC-217: MEPS panel 23 longitudinal data file. Agency for Healthcare Research and Quality. September 2021. Accessed May 13, 2022. https://www.meps.ahrq.gov/mepsweb/data_stats/download_data_files_detail.jsp?cboPufNumber=HC-217

19. MEPS HC-036: MEPS 1996-2020 pooled linkage file for common variance structure. Agency for Healthcare Research and Quality. August 2022. Accessed May 6, 2023. https://meps.ahrq.gov/mepsweb/data_stats/download_data_files_detail.jsp?cboPufNumber=HC-036

20. A survey about your diabetes care. Agency for Healthcare Research and Quality. 2018. Accessed February 5, 2022. https://meps.ahrq.gov/survey_comp/hc_survey/paper_quest/2018/2018_DCS_ENG.shtml

21. ICD-10. CMS. December 1, 2021. Accessed March 1, 2022. https://www.cms.gov/Medicare/Coding/ICD10

22. Lexicomp. Wolters Kluwer. Accessed February 5, 2022. https://www.wolterskluwer.com/en/solutions/lexicomp

23. National Drug Code directory. FDA. July 22, 2022. Accessed May 6, 2023. https://www.fda.gov/drugs/drug-approvals-and-databases/national-drug-code-directory

24. Andersen RM. Revisiting the behavioral model and access to medical care: does it matter? J Health Soc Behav. 1995;36(1):1-10.

25. Consumer price index for all urban consumers: medical care in U.S. city average (CPIMEDSL). Federal Reserve Bank of St. Louis. Accessed March 5, 2022. https://fred.stlouisfed.org/series/CPIMEDSL#

26. Glasheen WP, Renda A, Dong Y. Diabetes Complications Severity Index (DCSI)—update and ICD-10 translation. J Diabetes Complications. 2017;31(6):1007-1013. doi:10.1016/j.jdiacomp.2017.02.018

27. Austin PC. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Commun Stat Simul Comput. 2009;38(6):1228-1234. doi:10.1080/03610910902859574

28. Caro JJ, Ward AJ, O’Brien JA. Lifetime costs of complications resulting from type 2 diabetes in the U.S. Diabetes Care. 2002;25(3):476-481. doi:10.2337/diacare.25.3.476

29. Andersson E, Persson S, Hallén N, et al. Costs of diabetes complications: hospital-based care and absence from work for 392,200 people with type 2 diabetes and matched control participants in Sweden. Diabetologia. 2020;63(12):2582-2594. doi:10.1007/s00125-020-05277-3

30. Lesniowska J, Schubert A, Wojna M, Skrzekowska-Baran I, Fedyna M. Costs of diabetes and its complications in Poland. Eur J Health Econ. 2014;15(6):653-660. doi:10.1007/s10198-013-0513-0

31. Delevry D, Ho A, Le QA. Association between processes of diabetes care and health care utilization in patients with diabetes: evidence from a nationally representative US sample. J Diabetes. 2021;13(1):78-88. doi:10.1111/1753-0407.13109

32. Bain SC, Bekker Hansen B, Hunt B, Chubb B, Valentine WJ. Evaluating the burden of poor glycemic control associated with therapeutic inertia in patients with type 2 diabetes in the UK. J Med Econ. 2020;23(1):98-105. doi:10.1080/13696998.2019.1645018

33. Ali SN, Dang-Tan T, Valentine WJ, Bekker Hansen B. Evaluation of the clinical and economic burden of poor glycemic control associated with therapeutic inertia in patients with type 2 diabetes in the United States. Adv Ther. 2020;37(2):869-882. doi:10.1007/s12325-019-01199-8

34. Benedict AW, Spence MM, Sie JL, et al. Evaluation of a pharmacist-managed diabetes program in a primary care setting within an integrated health care system. J Manag Care Spec Pharm. 2018;24(2):114-122. doi:10.18553/jmcp.2018.24.2.114

35. Norton MC, Haftman ME, Buzzard LN. Impact of physician-pharmacist collaboration on diabetes outcomes and health care use. J Am Board Fam Med. 2020;33(5):745-753. doi:10.3122/jabfm.2020.05.200044

36. Prevalence of both diagnosed and undiagnosed diabetes. CDC. September 30, 2022. Accessed May 6, 2023. https://www.cdc.gov/diabetes/data/statistics-report/diagnosed-undiagnosed-diabetes.html