Care Quality Metrics in Medicare During COVID-19 Pandemic

ABSTRACT

Objective: To examine the impact on quality of care for individuals enrolled in Medicare Advantage (MA) plans or traditional Medicare (TM) during the COVID-19 pandemic.

Study Design: Retrospective cohort study.

Methods: The study examined beneficiaries enrolled in Medicare from 2017 through 2021. Beneficiaries were divided into 4 cohorts based on their enrollment in TM or MA and the year of enrollment in 2019, the year before the COVID-19 pandemic, or 2021, during the COVID-19 pandemic. For each cohort, 12 clinical quality measures were constructed, including 4 screening measures requiring in-person visits and 8 medication management and adherence measures.

Results: A total of 3,190,208 Medicare beneficiaries were included (58.4% female; mean age, 73.0 years). In both 2019 and 2021, the MA program performed significantly better than TM across the 12 clinical quality measures. Compared with the year before the pandemic, both programs experienced a decrease in screening measures that required in-person visits during the pandemic, with a slightly higher decrease for the MA plans. In contrast, measures of medication management and adherence improved for both programs, but especially for MA plans.

Conclusions: MA plans continued to outperform TM on clinical quality measures during the COVID-19 pandemic.

Am J Manag Care. 2026;32(2):In Press

Takeaway Points

This study examined the impact on quality of care for individuals enrolled in Medicare Advantage (MA) plans or traditional Medicare (TM) during the COVID-19 pandemic.

• MA plans outperformed TM on clinical quality care measures in 2019 and 2021.
• Both MA and TM experienced a decrease in screening measures that required in-person visits during the COVID-19 pandemic, with a slightly higher decrease for the MA plans.
• Measures of medication management and adherence improved during the COVID-19 pandemic for both programs, but especially for MA plans.

Medicare Advantage (MA), also known as Medicare Part C, has seen significant growth in recent years. In 2024, MA enrollment reached 33.8 million, accounting for half of the eligible Medicare population.¹ MA plans receive fixed, risk-adjusted payments from CMS, incentivizing them to control costs by using managed care tools to maximize use of high-value care, limit low-value care, and improve quality of care to avoid poor outcomes. MA plans are also incentivized to deliver high-quality care to attract enrollment and qualify for additional payments through the MA Star Ratings system.

Systematic reviews of studies published through 2020 showed that MA plans performed significantly better than traditional Medicare (TM) across various clinical quality metrics, including breast cancer screening, diabetes care, influenza and pneumonia vaccinations, cholesterol testing, β-blockers after heart attack, and antidepressant medication management.^2,3 However, many of these studies were based on data from before 2012.

More recent studies have shown that MA delivers a better quality of care than TM. MA plans performed significantly better than TM on preventive and chronic care clinical quality metrics, including breast cancer and diabetes screening, pharmacological treatment for chronic obstructive pulmonary disease (COPD) exacerbation, osteoporosis management, and monitoring patients on persistent medications.⁴ MA beneficiaries had lower rates of inpatient and emergency department (ED) admissions, avoidable ED visits, 30-day readmissions, admissions for stroke or myocardial infarction, and hospitalizations for COPD or asthma exacerbation.^5,6 MA beneficiaries received fewer low-value services than TM beneficiaries.⁷

However, current literature is largely based on data before 2020, the start of the COVID-19 pandemic. The COVID-19 pandemic led to major interruptions in health care delivery: Outpatient visit frequency decreased, whereas use of telemedicine increased.^8-10 Preventive and chronic care visits were especially affected.^11,12 In its March 2020 guidance, CMS recommended that health care delivery systems postpone nonemergent low-acuity treatment, including preventive care visits and screenings, in order to prioritize care for high-acuity patients.¹³ Despite the gradual reopening starting in April 2020,¹⁴ many patients and health care providers may have opted to postpone screenings for conditions such as cancer and diabetes to avoid potential exposure to COVID-19, especially given the heightened risk for older patients and patients with chronic conditions.¹⁵

In this analysis, we compared the trends in clinical quality of care measures in 2019 (the year before the start of the COVID-19 pandemic) and 2021 among MA and TM beneficiaries matched based on demographic, socioeconomic, and geographic variables. Because the quality-of-care measures in our study are typically based on 2 years of administrative claims data, 2021 is the first year in which the measures reflect the full impact of the pandemic.

STUDY DATA AND METHODS

Administrative Data and Study Population

We used 2018-2021 CMS data sets containing medical and pharmacy claims data from a representative 5% sample of Medicare beneficiaries enrolled in either TM or MA. We constructed 4 cohorts based on the program (TM vs MA) and year of enrollment (2019 vs 2021). Each cohort included beneficiaries who were continuously enrolled for 2 years (the year of analysis and the previous year) in either TM (both Part A and Part B) or MA and had continuous prescription drug coverage (Part D). Beneficiaries who switched between programs or terminated their coverage during the study period were excluded from analysis. We further excluded Medicare beneficiaries who qualified for Medicare due to end-stage renal disease.

Covariates

We marked beneficiaries as enrolled in either TM or MA based on continuous enrollment in the same program for 12 months of the study year (2019 or 2021). We noted other demographic variables, including age, sex, race, ethnicity, and Medicare-Medicaid dual-eligibility status. We assessed a measure of comorbidity for each beneficiary by calculating the CMS Hierarchical Conditions Category (HCC) score using the algorithm provided by CMS.¹⁶ The HCC score is based on a predictive model of health care spending that accounts for a beneficiary’s sociodemographic characteristics and selected acute or chronic conditions identified from medical claims. To account for the differences in coding intensity between fee-for-service and MA, we applied a 5.9% HCC risk adjustment to MA beneficiaries to reflect the risk adjustment CMS applies when determining MA payments.¹⁷ For zip code–level data on socioeconomic indicators (eg, median household income, share of residents with less than a high school diploma), we linked data from the American Community Survey published by the US Census Bureau.¹⁸ We adjusted median household income for inflation to 2021 US$ using Consumer Price Index data published by the US Bureau of Labor Statistics.¹⁹ In addition, beneficiaries were identified as living in a rural area based on the rural-urban commuting area codes published by the US Department of Agriculture Economic Research Service.²⁰ A rural-urban commuting area code of 4 or greater was defined as rural. We also noted the census region in which beneficiaries resided.²¹

Clinical Quality Measures

For our analysis, we selected 12 Healthcare Effectiveness Data and Information Set (HEDIS) clinical quality measures developed by the National Committee for Quality Assurance.²² The measures focused on preventive care and chronic condition management. Only measures that applied to Medicare patients and could be constructed using administrative claims data were selected. Measures that relied on provider specialty identification were excluded because provider specialty was typically not reported for MA plans. To account for higher coding intensity in some MA plans^23,24 and ensure comparability between MA and TM cohorts, only the claims associated with performed procedures were used to construct HEDIS measures; chart reviews were excluded from the analysis.

Three of the measures—breast cancer screening, colon cancer screening, and hemoglobin A_1c (HbA_1c) testing for diabetes—require in-person visits. Another measure, osteoporosis management for women, can be fulfilled by either an in-person bone mineral density test or a prescription for an osteoporosis treatment. However, some osteoporosis medications, such as zoledronic acid or denosumab, must be administered by a health care professional or require laboratory tests prior to initiation and, therefore, may require in-person visits as well.²⁵

The remaining 8 measures were related to medication management and adherence, including β-blocker treatment after a heart attack, statin therapy for patients with cardiovascular diseases and patients with diabetes, disease-modifying antirheumatic drug (DMARD) therapy for rheumatoid arthritis, and corticosteroid and bronchodilator treatment for patients with COPD exacerbation.

Statistical Analysis

We used propensity score weighting to minimize the effects of potential confounders. In particular, we applied inverse probability of treatment weighting (IPTW) across the TM and MA populations for all study periods.^26-28 The advantage of this approach is that it adjusts for demographic differences between the TM and MA populations and compositional changes in these populations over time.

To calculate the propensity score for each beneficiary in each cohort, we fitted a logistic regression model of the probability of the beneficiary’s enrollment in the MA 2019 cohort as a function of the following demographic variables: age, sex, race and ethnicity, Medicare-Medicaid dual-eligibility status, HCC score, residence in rural area, census region, share of residents with less than a high school diploma, and median household income. We used the inverse of the propensity score to obtain IPTW and then assigned these weights to each beneficiary. After weighting, all cohorts exhibited similar demographic characteristics (eAppendix Table 1 [eAppendix available at ajmc.com]). Using standardized differences, we tested the balance of covariates before and after IPTW adjustment. A difference of 10% or higher was treated as problematic. The cohorts appeared balanced after IPTW adjustment (eAppendix Table 2).

We fitted propensity score–weighted logistic regression models to adjust for differences between TM and MA beneficiaries and to estimate adjusted disparities in clinical quality of care outcomes in TM and MA. The analyses were performed separately for each year. We compared the impact of the COVID-19 pandemic on clinical quality of care outcomes within each program by using generalized estimating equation models with binomial distributions and the logit link function, accounting for patient-level clustering.

We repeated the analysis with beneficiaries who switched between programs to test the sensitivity of our results to data issues and research design choices. We also tested the sensitivity of our model to coding intensity adjustment by applying a 10% risk adjustment to HCC scores for MA beneficiaries. Finally, we tested the impact of excluding propensity score weights.

RESULTS

Patient Characteristics

More than 3 million Medicare beneficiaries met eligibility criteria and were included in the analysis. Table 1 shows unadjusted descriptive statistics for beneficiaries enrolled in TM or MA. Compared with MA, TM beneficiaries were slightly older, more likely to be White and to live in rural areas, and less likely to reside in the Western census region. On average, TM beneficiaries had higher household incomes and were likelier to have completed high school. The share of beneficiaries dually enrolled in Medicare and Medicaid was higher in TM in 2019 but was higher in MA in 2021.

Measures

Table 2 shows IPTW-adjusted measures for all cohorts. In both 2019 and 2021, the MA program performed significantly better compared with TM across the 12 HEDIS measures. In 2021, the first year in which the measures capture the impact of the pandemic, the most notable differences were in osteoporosis management in women, with a difference in HEDIS performance of 16.7%; access to recommended medications after a COPD exacerbation, with a difference in performance of 4.9% for corticosteroid treatment and 5.9% for bronchodilator treatment; and persistence of β-blocker treatment, with a difference of 6.5%. Table 3 shows the ORs for the association between enrollment in MA vs TM and HEDIS quality-of-care outcomes. Beneficiaries enrolled in MA had higher odds of receiving recommended cancer screening, diabetes screening, and bone mineral density testing, and of adhering to the prescribed course of preventive pharmacological therapy.

The impact of COVID-19 on HEDIS metrics varied. Table 4 shows the ORs for the association between cohort year and HEDIS quality-of-care outcomes. For both MA and TM beneficiaries, the odds of receiving preventive care that required in-person office visits (ie, cancer screening, diabetes screening, or bone mineral density test) were lower in 2021 compared with 2019, although findings are not statistically significant for osteoporosis management in women in TM. In contrast, the odds of MA and TM beneficiaries receiving pharmacological treatment, such as statin or β-blocker therapy, were higher in 2021 compared with 2019, with the exception of DMARD therapy in TM (OR, 0.98; 95% CI, 0.91-1.06) and MA (OR, 0.89; 95% CI, 0.82-0.97) and β-blocker treatment in TM (OR, 0.91; 95% CI, 0.74-1.13), which saw a decrease in odds of treatment, although these findings were not statistically significant for TM.

Despite better performance across the 12 HEDIS measures in both years, MA experienced a greater decline in 2021 compared with TM on measures requiring in-person visits, especially on breast cancer screening and osteoporosis management in women. In contrast, MA experienced larger increases than TM on measures of medication adherence, including bronchodilator and corticosteroid treatment in COPD exacerbation and persistence of β-blocker treatment.

In the sensitivity analysis, the results for the models with unadjusted cohorts, IPTW-adjusted cohorts with an alternative HCC deflator, and IPTW-adjusted cohorts including beneficiaries who switched programs were consistent with our main findings (eAppendix Tables 3-11). In all 3 models, MA outperformed TM across the 12 HEDIS measures in 2019 and 2021. Similarly, all 3 models show the same pattern of lower odds of beneficiaries receiving cancer and diabetes screening and bone density testing, and higher odds of beneficiaries receiving pharmacological treatment in 2021 compared with 2019.

DISCUSSION

Delayed Screenings

This study’s results align with previous research showing postponed care and screenings due to the COVID-19 pandemic. For example, studies have documented a decline in the rate of patients with diabetes receiving HbA_1c tests during the pandemic.¹⁵ In the case of screenable cancers, including breast and colon cancer, another study showed that postponed screening resulted in 13.9% lower cancer incidence from March to December 2020 compared with prepandemic levels, which translated into 52,000 potentially undiagnosed cases.²⁹

The disruption in care during the pandemic can have major negative health implications for Medicare beneficiaries. Previous studies have shown that adherence to screening guidelines for diabetes lowers the risk of cardiovascular and kidney complications, amputation, and death.^30,31 Delays in diagnosing cancer can allow the disease to progress to a more advanced stage and result in higher mortality rates.^32,33 Untreated osteoporosis can lead to greater need for subsequent health care. Compared with patients with osteoporosis who have high medication adherence, patients with low medication adherence experienced higher costs of inpatient and outpatient care.³⁴

This study found that preventive screenings declined during the pandemic for MA and TM enrollees. Our analysis further showed a higher decline in the performance on HEDIS measures requiring in-person visits in MA compared with TM during the COVID-19 pandemic, although MA enrollees continued to have higher rates of preventive cancer and diabetes screenings than TM in 2021. MA plans disproportionately serve populations more likely to face challenges adhering to cancer screenings during the COVID-19 pandemic (eg, non-White individuals, those with low income, those with low educational attainment).³⁵ Although the study controlled for some confounders, other factors, such as smoking, morbid obesity, uncontrolled diabetes, or family history of cancer, were shown to influence patients’ adherence to screenings.^36,37 The study could not account for these factors because they were unavailable in the data source.

The findings of this study highlight the need for stakeholders across the health care system, including public health officials, academics, hospitals, providers, and health plans, to collaboratively develop policies that would allow patients to receive critical care, including preventive care, in case of major health care disruptions. Such policies may include the need for hospital disaster plans to prioritize care, expedited publication of research to facilitate rapid sharing of best practices, and greater use of telemedicine and artificial intelligence to supplement in-person care.³⁸

Improved Medication Adherence

In our analysis, statin adherence improved for patients with diabetes and patients with recent cardiovascular events, which aligns with previous trends.³⁹ Previous research showed that higher adherence to statins reduced the risk of hospitalizations for cardiovascular diseases and the risk of mortality.^40,41 Furthermore, patients already on statin therapy had a lower risk of severe COVID-19 infection and lower mortality from COVID-19.⁴²

In this study, adherence to β-blocker therapy after a cardiovascular event improved for MA beneficiaries but decreased modestly for TM beneficiaries. Another study found modest decreases in adherence to antihypertensive drugs during the pandemic, which was partially mitigated by greater use of mail-order prescriptions and a greater share of prescriptions with more than 90 days’ supply.⁴³

In both MA and TM, a higher share of patients with COPD were prescribed and filled the recommended corticosteroid and bronchodilator treatment following respiratory exacerbation. Other studies similarly found that medication adherence to daily controller medication among patients with asthma and COPD improved during the pandemic, likely due to heightened concerns over contracting SARS-CoV-2.⁴⁴ Lockdowns and other measures to contain the spread of SARS-CoV-2 also decreased the share of COPD-related hospitalizations, because viral infections account for 40% to 50% of COPD exacerbations.⁴⁵

Adherence to DMARD therapy is our study’s only medication adherence measure that showed a statistically significant decrease during the pandemic. Patients with rheumatoid arthritis prescribed a DMARD therapy faced an additional challenge in the early days of the pandemic.⁴⁶ Hydroxychloroquine, a commonly used conventional DMARD therapy, was initially believed to have some benefits in treating COVID-19. Although the drug ultimately proved ineffective against COVID-19, patients who were prescribed hydroxychloroquine for other conditions, including rheumatoid arthritis, faced shortages of the drug.⁴⁷

Study findings show that greater medication adherence improves health and reduces health care utilization.⁴⁸ A systematic review of the health impact of medication adherence on adults 50 years and older found a 17% decrease in all-cause hospitalization and a 21% decrease in mortality among adherent individuals compared with those considered nonadherent.⁴⁹ Drug-related treatment failure, including poor medication adherence, results in $528 billion in potentially avoidable costs annually, including hospitalizations, ED visits, and long-term care admissions.⁵⁰

Limitations

Our study had several limitations. First, it relied exclusively on administrative data to calculate rates, which can result in underreporting in some cases due to incomplete data. In contrast, health plans rely on a mix of administrative data and medical records for a more comprehensive view. Second, several diabetes care measures (eg, retinal exam, monitoring for nephropathy, glucose and blood pressure control) could not be constructed due to data limitations. Third, to isolate the impact of the COVID-19 pandemic, the study used a shorter lookback period for some measures (eg, 2-year lookback instead of the required 10-year period in colorectal cancer screening). Fourth, the study did not account for heterogeneity within MA plans (eg, health maintenance organization vs preferred provider organization, at-risk vs fee-for-service contracts) or TM plans (accountable care organization vs not). Fifth, by using uniformly deflated HCC scores, the study may not fully account for differences in coding intensity between MA and TM.

CONCLUSIONS

In line with previous literature, our study finds that MA plans continue to outperform TM across various preventive and chronic condition care measures both before and during the COVID-19 pandemic. These findings demonstrate that MA plans continued to provide greater value than TM even during the public health emergency.

Given CMS recommendations to postpone nonemergent treatment—such as preventive care visits and screenings—and the continued fears of COVID-19 exposure among older patients, both MA and TM experienced fewer screenings that required in-person visits during the pandemic. Although that decrease was slightly greater for MA than for TM enrollees, overall screening rates remained higher for those in MA plans compared with those in TM. Additionally, medication adherence improved for enrollees of both MA and TM, but the degree of improvement was more pronounced for those in MA.

Because timely screening and medication adherence are associated with lower mortality, reduced health care utilization, and lower costs, future research is needed to support policies that can improve adherence to cancer and diabetes screening among all Medicare enrollees during major health care disruptions.

Author Affiliations: AHIP (JT, MH, LN, SA, GV, SS), Washington, DC.

Source of Funding: None.

Author Disclosures: The authors are employed by AHIP, which counts Medicare Advantage plans as its members.

Authorship Information: Concept and design (JT, MH, LN, SA, GV, SS); acquisition of data (JT, LN, SA, GV); analysis and interpretation of data (MH, LN, SA, GV); drafting of the manuscript (SA, GV); critical revision of the manuscript for important intellectual content (JT, MH, LN, SA, GV, SS); statistical analysis (SA, GV); administrative, technical, or logistic support (JT, SA, GV); responding to reviewer feedback (SA, SS); and supervision (JT, SS).

Address Correspondence to: Sherzod Abdukadirov, PhD, AHIP, 601 Pennsylvania Ave NW, South Building, Ste 500, Washington, DC 20004. Email: sabdukadirov@ahip.org.

REFERENCES

1. Monthly contract and enrollment summary report. CMS. Updated September 10, 2024. Accessed October 10, 2024. https://www.cms.gov/data-research/statistics-trends-and-reports/medicare-advantagepart-d-contract-
and-enrollment-data/monthly-contract-and-enrollment-summary-report

2. Agarwal R, Connolly J, Gupta S, Navathe AS. Comparing Medicare Advantage and traditional Medicare: a systematic review. Health Aff (Millwood). 2021;40(6):937-944. doi:10.1377/hlthaff.2020.02149

3. DuGoff E, Tabak R, Diduch T, Garth V. Quality, health, and spending in Medicare Advantage and traditional Medicare. Am J Manag Care. 2021;27(9):395-400. doi:10.37765/ajmc.2021.88641

4. Landon BE, Zaslavsky AM, Anderson TS, Souza J, Curto V, Ayanian JZ. Differences in use of services and quality of care in Medicare Advantage and traditional Medicare, 2010 and 2017. Health Aff (Millwood). 2023;42(4):459-469. doi:10.1377/hlthaff.2022.00891

5. Cohen K, Ameli O, Chaisson CE, et al. Comparison of care quality metrics in 2-sided risk Medicare Advantage vs fee-for-service Medicare programs. JAMA Netw Open. 2022;5(12):e2246064. doi:10.1001/jamanetworkopen.2022.46064

6. Jung J, Carlin C, Feldman R. Medicare Advantage has lower resource use and better quality of care than traditional Medicare. Am J Health Econ. Published online March 14, 2024. doi:10.1086/730436

7. Boudreau E, Schwartz R, Schwartz AL, et al. Comparison of low-value services among Medicare Advantage and traditional Medicare beneficiaries. JAMA Health Forum. 2022;3(9):e222935. doi:10.1001/jamahealthforum.2022.2935

8. Betancourt JA, Rosenberg MA, Zevallos A, Brown JR, Mileski M. The impact of COVID-19 on telemedicine utilization across multiple service lines in the United States. Healthcare (Basel). 2020;8(4):380. doi:10.3390/healthcare8040380

9. Papautsky EL, Rice DR, Ghoneima H, et al. Characterizing health care delays and interruptions in the United States during the COVID-19 pandemic: internet-based, cross-sectional survey study. J Med Internet Res. 2021;23(5):e25446. doi:10.2196/25446

10. Roy CM, Bollman EB, Carson LM, Northrop AJ, Jackson EF, Moresky RT. Assessing the indirect effects of COVID-19 on healthcare delivery, utilization and health outcomes: a scoping review. Eur J Public Health. 2021;31(3):634-640. doi:10.1093/eurpub/ckab047

11. Chang AY, Cullen MR, Harrington RA, Barry M. The impact of novel coronavirus COVID-19 on noncommunicable disease patients and health systems: a review. J Intern Med. 2021;289(4):450-462. doi:10.1111/joim.13184

12. Sabetkish N, Rahmani A. The overall impact of COVID-19 on healthcare during the pandemic: a multidisciplinary point of view. Health Sci Rep. 2021;4(4):e386. doi:10.1002/hsr2.386

13. Non-emergent, elective medical services, and treatment recommendations. CMS. April 7, 2020. Accessed August 16, 2024. https://www.cms.gov/files/document/cms-non-emergent-elective-medical-recommendations.pdf

14. Re-opening facilities to provide non-emergent non-COVID-19 healthcare: phase I. CMS. April 19, 2020. Accessed August 16, 2024. https://www.cms.gov/files/document/covid-flexibility-reopen-essential-non-covid-services.pdf

15. Bancks MP, Lin MY, Bertoni A, et al. Impact of the COVID-19 pandemic on diabetes care among a North Carolina patient population. Clin Diabetes. 2022;40(4):467-476. doi:10.2337/cd21-0136

16. Risk adjustment. CMS. Updated October 24, 2024. Accessed March 3, 2025. https://www.cms.gov/medicare/payment/medicare-advantage-rates-statistics/risk-adjustment

17. Medicare Payment Advisory Commission. Report to the Congress: Medicare Payment Policy. Medicare Payment Advisory Commission; 2021. Accessed February 28, 2025. https://www.medpac.gov/document/march-2021-report-to-the-congress-medicare-payment-policy/

18. American Community Survey (ACS). US Census Bureau. Accessed August 16, 2024. https://www.census.gov/programs-surveys/acs

19. Consumer Price Index. US Bureau of Labor Statistics. Accessed August 19, 2024. https://www.bls.gov/cpi/

20. Rural-urban commuting area codes. US Department of Agriculture Economic Research Service. Accessed August 16, 2024. https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes/

21. Geographic levels. US Census Bureau. Revised October 8, 2021. Accessed August 19, 2024. https://www.census.gov/programs-surveys/economic-census/guidance-geographies/levels.html

22. HEDIS and performance measurement. National Committee for Quality Assurance. Accessed August 16, 2024. https://www.ncqa.org/hedis/

23. Meyers DJ, Trivedi AN. Medicare Advantage chart reviews are associated with billions in additional payments for some plans. Med Care. 2021;59(2):96-100. doi:10.1097/MLR.0000000000001412

24. Carlin CS, Feldman R, Jung J. The mechanics of risk adjustment and incentives for coding intensity in Medicare. Health Serv Res. 2024;59(3):e14272. doi:10.1111/1475-6773.14272

25. Yu EW, Tsourdi E, Clarke BL, Bauer DC, Drake MT. Osteoporosis management in the era of COVID-19. J Bone Miner Res. 2020;35(6):1009-1013. doi:10.1002/jbmr.4049

26. McCaffrey DF, Griffin BA, Almirall D, Slaughter ME, Ramchand R, Burgette LF. A tutorial on propensity score estimation for multiple treatments using generalized boosted models. Stat Med. 2013;32(19):3388-3414. doi:10.1002/sim.5753

27. Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med. 2015;34(28):3661-3679. doi:10.1002/sim.6607

28. Stuart EA, Huskamp HA, Duckworth K, et al. Using propensity scores in difference-in-differences models to estimate the effects of a policy change. Health Serv Outcomes Res Methodol. 2014;14(4):166-182. doi:10.1007/s10742-014-0123-z

29. Burus T, Lei F, Huang B, et al. Undiagnosed cancer cases in the US during the first 10 months of the COVID-19 pandemic. JAMA Oncol. 2024;10(4):500-507. doi:10.1001/jamaoncol.2023.6969

30. Chen Y, Sloan FA, Yashkin AP. Adherence to diabetes guidelines for screening, physical activity and medication and onset of complications and death. J Diabetes Complications. 2015;29(8):1228-1233. doi:10.1016/j.jdiacomp.2015.07.005

31. Bayer FJ, Galusha D, Slade M, Chu IM, Taiwo O, Cullen MR. Process of care compliance is associated with fewer diabetes complications. Am J Manag Care. 2014;20(1):41-52.

32. Negoita S, Chen HS, Sanchez PV, et al. Annual Report to the Nation on the Status of Cancer, part 2: early assessment of the COVID-19 pandemic’s impact on cancer diagnosis. Cancer. 2024;130(1):117-127. doi:10.1002/cncr.35026

33. Nascimento de Lima P, van den Puttelaar R, Hahn AI, et al. Projected long-term effects of colorectal cancer screening disruptions following the COVID-19 pandemic. Elife. 2023;12:e85264. doi:10.7554/eLife.85264

34. Zhao Y, Johnston SS, Smith DM, McMorrow D, Krege J, Krohn K. Association between teriparatide adherence and healthcare utilization and costs among hip fracture patients in the United States. Bone. 2014;60:221-226. doi:10.1016/j.bone.2013.12.016

35. Sabatino SA. Up-to-date breast, cervical, and colorectal cancer screening test use in the United States, 2021. Prev Chronic Dis. 2023;20:E94. doi:10.5888/pcd20.230071

36. Gangcuangco LMA, Rivas T, Basnet A, et al. Factors associated with colorectal cancer screening adherence and the impact of COVID-19 on screening patterns in Connecticut, USA. Intern Emerg Med. 2022;17(8):2229-2235. doi:10.1007/s11739-022-03053-2

37. Harris A, Bather JR, Kawamoto K, et al. Determinants of breast cancer screening adherence during the COVID-19 pandemic in a cohort at increased inherited cancer risk in the United States. Cancer Control. 2024;31:10732748241272727. doi:10.1177/10732748241272727

38. Akbulut S, Hargura AS, Garzali IU, Aloun A, Colak C. Clinical presentation, management, screening and surveillance for colorectal cancer during the COVID-19 pandemic. World J Clin Cases. 2022;10(26):9228-9240. doi:10.12998/wjcc.v10.i26.9228

39. Statin therapy for patients with cardiovascular disease and diabetes. National Committee for Quality Assurance. Accessed October 10, 2024. https://web.archive.org/web/20241010065010/https://www.ncqa.org/hedis/measures/statin-therapy-for-patients-with-cardiovascular-disease-and-diabetes/

40. Rodriguez F, Maron DJ, Knowles JW, Virani SS, Lin S, Heidenreich PA. Association of statin adherence with mortality in patients with atherosclerotic cardiovascular disease. JAMA Cardiol. 2019;4(3):206-213. doi:10.1001/jamacardio.2018.4936

41. Kengne AP, Brière JB, Zhu L, et al. Impact of poor medication adherence on clinical outcomes and health resource utilization in patients with hypertension and/or dyslipidemia: systematic review. Expert Rev Pharmacoecon Outcomes Res. 2024;24(1):143-154. doi:10.1080/14737167.2023.2266135

42. Yetmar ZA, Chesdachai S, Kashour T, et al. Prior statin use and risk of mortality and severe disease from coronavirus disease 2019: a systematic review and meta-analysis. Open Forum Infect Dis. 2021;8(7):ofab284. doi:10.1093/ofid/ofab284

43. Steiner JF, Powers JD, Malone A, et al. Hypertension care during the COVID-19 pandemic in an integrated health care system. J Clin Hypertens (Greenwich). 2023;25(4):315-325. doi:10.1111/jch.14641

44. Kaye L, Theye B, Smeenk I, Gondalia R, Barrett MA, Stempel DA. Changes in medication adherence among patients with asthma and COPD during the COVID-19 pandemic. J Allergy Clin Immunol Pract. 2020;8(7):2384-2385. doi:10.1016/j.jaip.2020.04.053

45. So JY, O’Hara NN, Kenaa B, et al. Population decline in COPD admissions during the COVID-19 pandemic associated with lower burden of community respiratory viral infections. Am J Med. 2021;134(10):1252-1259.e3. doi:10.1016/j.amjmed.2021.05.008

46. Schwartz IS, Boulware DR, Lee TC. Hydroxychloroquine for COVID19: the curtains close on a comedy of errors. Lancet Reg Health Am. 2022;11:100268. doi:10.1016/j.lana.2022.100268

47. Michaud K, Wipfler K, Shaw Y, et al. Experiences of patients with rheumatic diseases in the United States during early days of the COVID-19 pandemic. ACR Open Rheumatol. 2020;2(6):335-343. doi:10.1002/acr2.11148

48. Cutler RL, Fernandez-Llimos F, Frommer M, Benrimoj C, Garcia-Cardenas V. Economic impact of medication non-adherence by disease groups: a systematic review. BMJ Open. 2018;8(1):e016982. doi:10.1136/bmjopen-2017-016982

49. Walsh CA, Cahir C, Tecklenborg S, Byrne C, Culbertson MA, Bennett KE. The association between medication non-adherence and adverse health outcomes in ageing populations: a systematic review and meta-analysis. Br J Clin Pharmacol. 2019;85(11):2464-2478. doi:10.1111/bcp.14075

50. Watanabe JH, McInnis T, Hirsch JD. Cost of prescription drug-related morbidity and mortality. Ann Pharmacother. 2018;52(9):829-837. doi:10.1177/1060028018765159