Morbidity and mortality that result from nonoptimized medication therapy were estimated to cost more than $500 billion annually in the United States in 2016,1 and are especially pervasive among adults aged ≥65 years.2 Common problems associated with medication therapy in older people include inappropriate or ineffective medication use, poor medication adherence, and adverse health effects.3 To reduce costs and improve pharmacotherapy outcomes, the Centers for Medicare & Medicaid Services (CMS) incorporated medication therapy management (MTM) in 2006 as a required component of the Medicare Prescription Drug Benefit Program (Part D).4
Although MTM services have been provided since the 1990s as a strategy to promote optimal medication use, the term MTM was officiated in the Medicare Prescription Drug, Improvement, and Modernization Act of 2003.5 Typically, these services include 5 core elements, including a medication therapy review, a personal medication record, a medication-related action plan, intervention and/or referral, and documentation and follow-up.6 The Medicare Part D MTM program was designed to target beneficiaries with multiple chronic conditions who are taking multiple drugs covered by Part D, and having high drug expenditures.7 However, the program has evolved over the years toward less restrictive eligibility criteria7 amid multiple, yet unsuccessful, attempts to address the issue of low enrollment.4
Policy scenario analyses revealed that the utilization-based criteria may be too restrictive for racial and ethnic minorities, who tend to use fewer medications and thus incur lower costs.8-10 Citing these findings, in 2015 CMS proposed to further relax the eligibility criteria, but the attempt was thwarted as a result of opposition to general Medicare reform from stakeholders.11
A crucial barrier in realizing MTM’s full potential is the lack of evidence about the effectiveness of the program, particularly its actual effects on minority populations. Among previous literature on the effects of MTM on patients with chronic conditions in the outpatient setting, there has been insufficient evidence for most of the evaluated outcomes because of heterogeneity in study populations and program interventions.12 Most studies did not examine patient characteristics other than age and sex.12
Furthermore, although MTM has been shown to have generally improved clinical outcomes, such as medication adherence, most studies have lacked strong designs and sufficient statistical power to produce conclusive evidence on the benefits of MTM.5 Limited data availability also did not allow for the evaluation of MTM’s effects on minorities,5 because race and ethnicity information is not included in commercial databases, and a national MTM database was not available to the research community until recent years.13 With Medicare Part D MTM data becoming available, it is now possible to examine the effects of MTM on racial and ethnic disparities.
Our study focused on the effects of comprehensive medication review (CMR), a required component of the Part D MTM program.14 As defined by CMS, CMR is “a systematic process of collecting patient-specific information, assessing medication therapies to identify medication-related problems, developing a prioritized list of medication-related problems, and creating a plan to resolve them with the patient, caregiver and/or prescriber.”14
In practice, CMR is required to be provided in real time either in person or over the phone.14 In 2016, the CMR completion rate, namely, the percentage of MTM enrollees who received a CMR in the reporting year, became a Star Rating quality measure for Part D health insurance plans.15 The measure was endorsed by the Pharmacy Quality Alliance, and CMS has adopted Pharmacy Quality Alliance measures in its Star Rating system that evaluates the quality of Medicare health plans.15
An emerging body of literature has explored factors associated with or strategies to improve the CMR completion rate.16-22 In addition, descriptive studies reported racial and ethnic disparities in CMR receipt among the general Medicare Part D beneficiary population3 and beneficiaries with mental health conditions.23 However, no published empirical assessment has been conducted of the impacts of CMR on racial and ethnic disparities in clinical outcomes.
The objective of this current study was to examine the effects of CMR on racial and ethnic disparities in medication utilization, measured by the likelihood of nonadherence to medications for diabetes, hypertension, or hyperlipidemia. We tested the hypothesis that CMR has reduced racial and ethnic disparities in the likelihood of nonadherence to medications to treat diabetes, hypertension, and hyperlipidemia. This study fills in the research gap, by using the latest Medicare Part D MTM data.
This study was a retrospective analysis using the 2017 Medicare Part D database linked to Medicare Parts A and B claims, the Master Beneficiary Summary File (MBSF) base segment, and the Area Health Resources Files (AHRF). The study period covered the entire year of 2017 from January 1 to December 31. Specifically for the Medicare Part D database, the Part D Drug Event File and MTM Data File were used, with the former containing all prescription drug dispensing records submitted by Part D health plans to CMS and the latter supplying information on MTM participants, such as their dates of MTM enrollment and CMR receipt. The MBSF base segment provides demographic information, such as age, sex, and race.24 All Medicare Parts A, B, and D and MBSF data were available at the beneficiary level. The county-level data were obtained from AHRF, which is managed by the Health Resources and Services Administration and supplies information such as population characteristics and health service availability.25
The study sample was restricted to beneficiaries who met the following 3 inclusion criteria: (1) age ≥65 years at the beginning of 2017, (2) having continuous Parts A, B, and D coverage throughout the year, and (3) being enrolled in the MTM program. Five racial and ethnic groups were analyzed, including non-Hispanic whites (hereafter referred to as “whites”), non-Hispanic blacks (hereafter, “blacks”), Hispanics, Asians and Pacific Islanders (hereafter, “Asians”), and “other,” which includes all other races and ethnicities. Although American Indians/Alaska Natives are a major racial and ethnic group that warrants research attention, the group is underidentified based on the current race and ethnicity variables in the MBSF,26 and was therefore combined with other races and ethnicities in the “other” category.
Medication utilization was measured by the likelihood of nonadherence to medications for diabetes, hypertension, and hyperlipidemia. Medication adherence measures for these 3 diseases were developed by the Pharmacy Quality Alliance and were adopted by CMS as core Star Rating measures.15 Specifically, the measures consider reaching a proportion of days covered threshold of 80% as adherence.27
We calculated the proportion of days covered as the proportion of days covered by at least 1 prescription drug in the drug class in 2017. The outcome variables included 3 dummy variables, each constructed for 1 of the 3 diseases of interest. The value of 1 of the dummy variables indicated a beneficiary had a proportion of days covered of less than 80%, and therefore the patient was nonadherent to the study medication; the value of 0 indicated otherwise.
The theoretical framework for our study was the Gelberg-Andersen Behavioral Model for Vulnerable Populations, which suggested that healthcare utilization is determined by predisposing, enabling, and need factors at the individual and community levels.28 Variables potentially associated with medication nonadherence were selected as covariates based on these factors. The predisposing factors, which include demographic and social structure characteristics, were measured by age, sex (male and female), race or ethnicity (whites, blacks, Hispanics, Asians, and other), and the following county-level variables: education (percentage of people with education level of high school or higher), married family (percentage of married-couple families), income per capita, and health insurance status (percentage of people without health insurance).
The enabling factors, which encompass personal and community resources that facilitate medication utilization, were measured by county-level variables including metropolitan statistical area, census regions (Northeast, Midwest, South, and West), and health professional shortage area. The need factors, which include perceived and evaluated risks in health, were measured by a risk adjustment summary score calculated using beneficiary diagnostic and demographic information. A higher score represented a greater likelihood of incurring increased medical care expenditures.29 The score was a proxy for health status.
Propensity Score Matching
To reduce the observed bias in intervention effects estimated from observational data, we used propensity score matching30 to construct 2 comparable groups, with one group consisting of CMR recipients and the other of CMR nonrecipients. The 2 groups were mutually exclusive. A propensity score for each beneficiary was estimated from a logistic regression model that predicted the probability of receiving a CMR.
The model included all individual and community characteristics in the Gelberg-Andersen model. CMR nonrecipients were then matched to CMR recipients in a 1:1 ratio. A higher ratio was not used, because the percentage of CMR receipt in the study sample was 43.1%. Any ratio higher than 1:1 would have required a larger number of nonrecipients. A nearest neighbor algorithm without replacement was used to create the most balanced comparison groups among all matching algorithms considered.30
To examine the differences in beneficiary characteristics between the CMR recipients and nonrecipients, chi-square tests and t-tests were conducted for categorical and continuous variables, respectively. To control for any remaining unbalanced beneficiary characteristics after the propensity score matching, a multivariate difference-in-differences logistic regression model was used to evaluate the effects of CMR on racial and ethnic disparities in medication nonadherence.
Specifically, interaction terms between dummy variables for CMR receipt and each minority race or ethnicity (in total, 4 interaction terms) were added to the logistic regression. If the coefficient on the interaction term for a minority group was negative and significant, that would indicate that CMR was associated with disparity reduction for that minority group. Since the model included county-level measures from AHRF, standard errors were clustered within a county to account for intracounty correlation.
To explore whether CMR had any long-term effects, sensitivity analyses were conducted using 3-, 6-, 9-, and 11-month lags. A lag of more than 11 months was not used because MTM services are provided annually. All data analyses were conducted using SAS Enterprise Guide 7.1 (SAS Institute; Cary, NC) at the CMS Virtual Research Data Center. The statistical significance level was set a priori at 0.05. This study was approved by the Institutional Review Board at the corresponding author’s institution.
After propensity score matching, the patient sample included 1,962,478 Medicare Part D beneficiaries, of whom CMR recipients and nonrecipients each accounted for 50%. The beneficiaries’ demographic and socioeconomic characteristics are presented in Table 1.
The 2 groups had the same racial and ethnic composition, which included 716,694 (73.04%) whites, 106,831 (10.89%) blacks, 102,849 (10.48%) Hispanics, 34,759 (3.54%) Asians, and 20,106 (2.05%) other. All other characteristics were balanced between the 2 groups, but age, income per capita, and the risk adjustment summary score were significantly different (P <.05). Compared with CMR nonrecipients, CMR recipients were more likely to be younger, have lower income per capita, and have lower risk adjustment summary scores.
Table 2 shows the numbers and rates of beneficiaries by outcome measure and by race or ethnicity for CMR recipients and nonrecipients. Compared with CMR nonrecipients, CMR recipients had lower rates of medication nonadherence across all medications and racial and ethnic groups. For example, the rates of beneficiaries with diabetes medication nonadherence in the CMR recipients and nonrecipients, respectively, were 7.13% and 8.70% among whites, 10.48% and 13.18% among blacks, 6.63% and 9.62% among Hispanics, 4.56% and 6.81% among Asians, and 5.35% and 7.86% among other.
Regardless of their CMR receipt status, blacks and Hispanics generally had higher rates of medication nonadherence than whites. For example, among CMR recipients, the rates of nonadherent beneficiaries among whites and blacks, respectively, were 7.13% and 10.48% for diabetes, 9.11% and 11.61% for hypertension, and 9.25% and 13.74% for hyperlipidemia; among CMR nonrecipients, the rates were 8.70% and 13.18% for diabetes, 11.55% and 15.89% for hypertension, and 12.35% and 19.36% for hyperlipidemia. By comparison, Asians tended to have lower rates of medication nonadherence compared with whites in each of the 2 study groups, namely, CMR recipients and nonrecipients.
Although racial and ethnic disparities in each outcome measure were observed in the CMR recipient and nonrecipient groups, the difference was lower among CMR recipients between whites and blacks and between whites and Hispanics compared with CMR nonrecipients. By contrast, this gap was greater among CMR recipients between whites and Asians across all the outcome measures compared with the whites–Asians differences among CMR nonrecipients.
For example, for nonadherence to diabetes medication, the difference between whites and blacks was 4.48% (8.70% vs 13.18%, respectively) for CMR nonrecipients compared with a difference of 3.35% (7.13% vs 10.48%, respectively) for the CMR recipients. For the same outcome measure, the difference between whites and Asians was –1.89% (8.70% vs 6.81%, respectively) for CMR nonrecipients compared with –2.57% (7.13% vs 4.56%, respectively) for CMR recipients.
Table 3 presents the results from adjusted difference-in-differences logistic regression analyses with standard errors clustered at the county level. Compared with CMR nonrecipients, CMR recipients had significantly lower racial and ethnic disparities across all 3 outcome measures, with the exception of the difference between whites and blacks in nonadherence to diabetes medications.
For example, among patients receiving hypertension medications, the odds ratios (ORs) for the interaction terms between the dummy variables for CMR receipt and the minority groups were 0.92 for blacks (95% confidence interval [CI], 0.88-0.96), 0.82 for Hispanics (95% CI, 0.78-0.86), 0.84 for Asians (95% CI, 0.77-0.91), and 0.91 for other (95% CI, 0.85-0.98), respectively. This means that compared with CMR nonrecipients, the difference in the odds of hypertension medication nonadherence was reduced by 8% between whites and blacks, by 18% between whites and Hispanics, by 16% between whites and Asians, and by 9% between whites and the other minority group among CMR recipients. Similar significant disparity reductions were detected for nonadherence to hyperlipidemia medications between whites and each minority group.
For nonadherence to diabetes medications, although the disparity reductions between whites and Hispanics, whites and Asians, and whites and the “other” group were significant, the reduction between whites and blacks was not significant. Of note, the ORs for the interaction terms between the dummy variables for CMR receipt and the minority groups were 0.84 for Hispanics (95% CI, 0.79-0.90), 0.81 for Asians (95% CI, 0.73-0.90), and 0.82 for other (95% CI, 0.70-0.97). By contrast, the OR for the interaction term between the dummy variables for CMR receipt and blacks was 0.97 (95% CI, 0.92-1.03).
In addition to the main results, Table 3 shows that some individual and community characteristics also had significant and consistent associations with nonadherence to medications for diabetes, hypertension, and hyperlipidemia. These characteristics included male sex, married family, South region, and risk adjustment summary score.
Specifically, the male sex and married family characteristics were negatively associated with the likelihood of nonadherence to all 3 medications. For example, the OR for male sex with the 3 outcome measures were 0.80 for diabetes medications (95% CI, 0.78-0.82), 0.97 for hypertension medications (95% CI, 0.96-0.99), and 0.84 for hyperlipidemia medications (95% CI, 0.83-0.86). By comparison, South region and risk adjustment summary score were positively associated with all 3 outcome measures.
The results from the sensitivity analyses were consistent with the results of the main analysis and are not presented here.
Our study is among the first that used the latest Medicare Part D MTM data to evaluate the effects of CMR on racial and ethnic disparities in medication use among Medicare beneficiaries aged ≥65 years who were enrolled in MTM in 2017. The adjusted multivariate logistic regression results indicated that compared with CMR nonrecipients, CMR recipients generally had significantly lower racial and ethnic disparities across all 3 outcome measures. The only exception is the whites–blacks disparities in nonadherence to diabetes medications, which were not significantly different between the 2 study groups (ie, CMR recipients and nonrecipients). These findings therefore provide support to the hypothesis that CMR has reduced racial and ethnic disparities in the likelihood of nonadherence to medications for diabetes, hypertension, and hyperlipidemia.
This study provides critical empirical evidence that may inform the future development of the Medicare Part D MTM program. As the largest health insurance provider in this country,31 Medicare covered 59.9 million Americans in 2018, of whom approximately 85% were aged ≥65 years.32 Among the older beneficiaries, 65% had multiple chronic conditions.33 The MTM program is therefore of tremendous value for improving pharmacotherapy outcomes for this population.
However, low enrollment has kept the MTM program from reaching its full potential, and CMS has attempted to address the issue with measures such as requiring opt out only (ie, eligible beneficiaries are automatically enrolled in MTM unless they opt out) and lowering the eligibility criteria thresholds.7 For example, starting from 2010, CMS specified that the number of Part D drugs that beneficiaries are required to be taking to be qualified for MTM services cannot be greater than 8, but this number was as high as 15 in 2008.7 Most of the plans opted for using the ceiling instead of the floor in their eligibility requirements, as is evident in that more than 70% of the plans required 8 Part D drugs in 2018.34
Although policy scenario analyses noted that the existing utilization-based criteria may still be too restrictive for minorities,8,9 the lack of evidence of the program’s actual effects has hindered further reform efforts. Findings from this current study show the effects of MTM in reducing health disparities and suggest that such effects may be further expanded if more minority populations become eligible for the program. Because Part D plans tend to align their eligibility requirements with the CMS ceiling levels, CMS may want to consider further lowering the existing ceilings so that more minority beneficiaries may be eligible for MTM.
In addition to relaxing the MTM eligibility criteria, another policy implication from this study’s findings is that effective measures are needed to increase the CMR receipt rate, which until recently remained below 20%. For example, in the years 2013 and 2014, the percentages of Medicare MTM enrollees who were offered a CMR were 81.1% and 96.5%, respectively.13 By comparison, only 12.3% and 16.9% of enrollees actually received a CMR in the same 2 years.13
Previous studies identified the factors that influence beneficiaries’ CMR acceptance decisions, including their health status, plan types, and perception of the time and cost associated with getting a CMR.17,18,35,36 Research attention has also been devoted to strategies that can improve CMR acceptance, such as using a standardized recruitment script18 and different outreach models.19
Because the CMR completion rate is a Star Rating quality measure, which directly affects a drug plan’s contract status with CMS,15 Part D plans are incentivized to increase their completion rate, and yet, the majority of those plans struggle to accomplish that goal. The Star Ratings are on a scale of 1 to 5, with 5 representing the highest quality; before 2020, Part D plans on average scored less than 2.8 on this particular measure.37,38
Currently, Part D plans are required to use a designated standardized format for their CMR written summaries. However, studies have shown that less than half of the surveyed Medicare beneficiaries considered the standardized format as a good tool for their medication management,16 and that the standardized format had elements that could be modified to optimize CMR delivery.21 A more consumer-oriented standardized format, and CMR at large, may therefore be needed to increase the CMR completion rate effectively.
Efforts to orient CMR around the medication management needs of beneficiaries can be facilitated by a better understanding of beneficiary and community characteristics that influence medication adherence. The findings of our study revealed that being male and living in a county with a higher percentage of “married-couple” families were associated with a lower likelihood of medication nonadherence. By contrast, living in the South and having a higher risk adjustment summary score were associated with a higher likelihood of medication nonadherence. These results were generally consistent with findings in previous studies. Relative to male sex, female sex is associated with higher medication nonadherence, likely because women tend to prioritize the care needs of others over their own needs.39,40
The percentage of married-couple families in the beneficiary’s county of residence may be considered as a proxy for social support, which has a positive effect on medication adherence among black patients with hypertension.41,42 Compared with other regions, the South was associated with higher rates of nonadherence to medications for diabetes, hypertension, and hyperlipidemia.43 This was likely a result of the region being generally more likely to have adverse health outcomes,44 which have been identified as predictors of increased risks for medication nonadherence.45,46 Similarly, having a higher risk adjustment summary score was a proxy for a higher disease burden that has been documented as a determinant of medication nonadherence.45,46
Overall, this study has expanded the existing knowledge of the effects of the Medicare Part D MTM program on racial and ethnic disparities. It is among the first that used the most recent Part D MTM data to evaluate empirically the effects of CMR, which is a core component of MTM. By including a control group, identified with propensity score matching, and by controlling for characteristics at the individual and community levels, our study addresses the shortcomings that are inherent in administrative data to the extent possible and thus contributes to the rigor of research in this subject area.
This study has several limitations. First, Medicare data are administrative data; thus, the beneficiary-level characteristics are limited by the availability of variables. For example, it would have been helpful to have data on beneficiaries’ education, income, and insurance coverage. Although the study included corresponding county-level measures, they may not be optimal proxies because of their aggregate nature.
In addition, the study was cross-sectional only. It therefore did not capture any temporal variation in the racial or ethnic disparities in medication utilization.
Third, our study only examined 3 measures in the drug safety domain of the Star Rating system.47 Thus, the identified MTM effects were not generalizable to other measures in the same domain, such as high-risk medication, that were measured by the percentage of older Medicare beneficiaries who received prescription drugs that have a high risk of side effects.47
CMR has reduced the racial and ethnic disparities in adherence to diabetes, hypertension, and hyperlipidemia medications. The findings of this study contribute to the existing literature by providing critical empirical evidence that may inform the future design of the Medicare Part D MTM program, which is valuable for improving pharmacotherapy outcomes, and could further realize its potential when greater numbers of racial and ethnic minority beneficiaries are enrolled in the program.
Further relaxing the current MTM eligibility criteria could improve medication adherence among racial and ethnic minority populations in Medicare. In the meantime, strategies, such as a consumer-oriented approach, may be implemented to maximize the benefit of CMR for the Medicare population.
As more data become available, future studies can explore temporal variation by including multiple years. Another area that warrants further investigation is other Star Rating measures related to drug safety, such as high-risk medications. Finally, the evaluation of Part D MTM program effects can be extended to other outcomes, such as health service utilization and costs.
The authors would like to acknowledge the assistance of Jamie Browning, a PharmD/MBA student at the College of Pharmacy, University of Tennessee Health Science Center, in this article.
Author Disclosure Statement
Dr Dong, Dr Tsang, Dr Zhao, Dr Wan, Dr Chisholm-Burns, and Dr Hines have no conflicts of interest to report; Dr Shih served on a Grant Review Panel for Pfizer and on an Advisory Board for AstraZeneca; Dr Dagogo-Jack is a Consultant to AstraZeneca, Bayer, Janssen, Merck, and Sanofi; Dr Cushman received an institutional grant from Eli Lilly; Dr Wang received grants from AbbVie, Curo, Bristol Myers Squibb, Pfizer, and Pharmaceutical Research and Manufacturers of America (PhRMA), and serves on the Health Outcomes Research Advisory Committee of the PhRMA Foundation.
- Watanabe JH, McInnis T, Hirsch JD. Cost of prescription drug–related morbidity and mortality. Ann Pharmacother. 2018;52:829-837.
- Petrone K, Katz P. Approaches to appropriate drug prescribing for the older adult. Prim Care. 2005;32:755-775.
- Coe AB, Adeoye-Olatunde OA, Pestka DL, et al. Patterns and predictors of older adult Medicare Part D beneficiaries’ receipt of medication therapy management. Res Social Adm Pharm. 2020;16:1208-1214.
- Centers for Medicare & Medicaid Services, HHS. Medicare program; contract year 2015 policy and technical changes to the Medicare Advantage and the Medicare prescription drug benefit programs. Proposed rule. Fed Regist. 2014;79:1917-2073.
- Ai AL, Carretta H, Beitsch LM, et al. Medication therapy management programs: promises and pitfalls. J Manag Care Spec Pharm. 2014;20:1162-1182.
- American Pharmacists Association; National Association of Chain Drug Stores Foundation. Medication therapy management in pharmacy practice: core elements of an MTM service model (version 2.0). J Am Pharm Assoc (2003). 2008;48:341-353.
- Gray C, Cooke CE, Brandt N. Evolution of the Medicare Part D medication therapy management program from inception in 2006 to the present. Am Health Drug Benefits. 2019;12(5):243-251.
- Wang J, Mullins CD, Brown LM, et al. Disparity implications of Medicare eligibility criteria for medication therapy management services. Health Serv Res. 2010;45:1061-1082.
- Munshi KD, Shih YCT, Brown LM, et al. Disparity implications of the Medicare medication therapy management eligibility criteria: a literature review. Expert Rev Pharmacoecon Outcomes Res. 2013;13:201-216.
- Briesacher B, Limcangco R, Gaskin D. Racial and ethnic disparities in prescription coverage and medication use. Health Care Financ Rev. 2003;25:63-76.
- Wang J, Qiao Y, Spivey CA, et al. Disparity implications of proposed 2015 Medicare eligibility criteria for medication therapy management services. J Pharm Health Serv Res. 2016;7:209-215.
- Viswanathan M, Kahwati LC, Golin CE, et al. Medication therapy management interventions in outpatient settings: a systematic review and meta-analysis. JAMA Intern Med. 2015;175:76-87.
- Pestka DL, Zillich AJ, Coe AB, et al. Nationwide estimates of medication therapy management delivery under the Medicare prescription drug benefit. J Am Pharm Assoc (2003). 2020;60:456-461.
- Chavez-Valdez AL; Centers for Medicare & Medicaid Services. CY 2019 Medication Therapy Management program guidance and submission instructions. April 6, 2018. www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovContra/Downloads/Memo-Contract-Year-2019-Medication-Therapy-Management-MTM-Program-Submission-v-040618.pdf. Accessed November 4, 2020.
- Centers for Medicare & Medicaid Services. Part C and D performance data. 2016 Star Ratings and display measures. Request for comments: enhancements to the star ratings for 2016 and beyond. www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovGenIn/PerformanceData. Accessed November 4, 2020.
- Brandt NJ, Cooke CE, Sharma K, et al. Findings from a national survey of Medicare beneficiary perspectives on the Medicare Part D Medication Therapy Management Standardized Format. J Manag Care Spec Pharm. 2019;25:366-391.
- Doucette WR, Zhang Y, Chrischilles EA, et al. Factors affecting Medicare Part D beneficiaries’ decision to receive comprehensive medication reviews. J Am Pharm Assoc (2003). 2013;53:482-487.
- Miguel A, Hall A, Liu W, et al. Improving comprehensive medication review acceptance by using a standardized recruitment script: a randomized control trial. J Manag Care Spec Pharm. 2017;23:13-21.
- Miller DE, Roane TE, Salo JA, Hardin HC. Evaluation of comprehensive medication review completion rates using 3 patient outreach models. J Manag Care Spec Pharm. 2016;22:796-800.
- Murray ME, Barner JC, Pope ND, Comfort MD. Impact and feasibility of implementing a systematic approach for medication therapy management in the community pharmacy setting: a pilot study. J Pharm Pract. 2019;32:664-670.
- Sharma K, Cooke CE, Howard AK, et al. Beneficiary, caregiver, and case manager perspectives on the Medication Therapy Management Program Standardized Format. J Gerontol Nurs. 2019;45:7-13.
- Snyder ME, Jaynes HA, Gernant SA, et al. Factors associated with comprehensive medication review completion rates: a national survey of community pharmacists. Res Social Adm Pharm. 2020;16:673-680.
- Murugappan MN, Seifert RD, Farley JF. Examining Medicare Part D medication therapy management program in the context of mental health. J Am Pharm Assoc (2003). 2020;60:571-579.e1.
- Research Data Assistance Center. Data file search. www.resdac.org/cms-data?tid_1%5B%5D=1. Accessed November 4, 2020.
- Health Resources and Services Administration. Area Health Resources Files. https://data.hrsa.gov/topics/health-workforce/ahrf. Accessed November 4, 2020.
- Jarrín OF, Nyandege AN, Grafova IB, et al. Validity of race and ethnicity codes in Medicare administrative data compared with gold-standard self-reported race collected during routine home health care visits. Med Care. 2020;58:e1-e8.
- Pharmacy Quality Alliance. Adherence: PQA adherence measures. www.pqaalliance.org/adherence-measures#:~:text=PQA%20Adherence%20Measures&text=Adherence%20measures%20assess%20the%20percentage,class%2C%20within%20the%20measurement%20year. Accessed November 4, 2020.
- Gelberg L, Andersen RM, Leake BD. The Behavioral Model for Vulnerable Populations: application to medical care use and outcomes for homeless people. Health Serv Res. 2000;34:1273-1302.
- Centers for Medicare & Medicaid Services. Part D payment and risk adjustment. www.cms.gov/Medicare/Prescription-Drug-Coverage/DrugCoverageClaimsData/RxClaims_PaymentRiskAdjustment. Accessed November 4, 2020.
- Austin PC. A comparison of 12 algorithms for matching on the propensity score. Stat Med. 2014;33:1057-1069.
- Davis K, Schoen C, Bandeali F. Medicare: 50 Years of Ensuring Coverage and Care. The Commonwealth Fund; April 2015. www.commonwealthfund.org/sites/default/files/documents/___media_files_publications_fund_report_2015_apr_1812_davis_medicare_50_years_coverage_care.pdf. Accessed November 4, 2020.
- Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds. 2019 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds. April 22, 2019. www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/ReportsTrustFunds/Downloads/TR2019.pdf. Accessed November 4, 2020.
- Wolff JL, Starfield B, Anderson G. Prevalence, expenditures, and complications of multiple chronic conditions in the elderly. Arch Intern Med. 2002;162:2269-2276.
- Centers for Medicare & Medicaid Services. 2018 Medicare Part D Medication Therapy Management (MTM) programs: fact sheet: summary of 2018 MTM programs. August 20, 2018. www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovContra/Downloads/CY2018-MTM-Fact-Sheet.pdf. Accessed November 4, 2020.
- Garcia GM, Snyder ME, Harriman McGrath S, et al. Generating demand for pharmacist-provided medication therapy management: identifying patient-preferred marketing strategies. J Am Pharm Assoc (2003). 2009;49:611-616.
- Schultz H, Westberg SM, Ramalho de Oliveira D, Brummel A. Patient-perceived value of Medication Therapy Management (MTM) services: a series of focus groups. Innov Pharm. 2012;3:Article 96.
- Centers for Medicare & Medicaid Services. Fact sheet-2017 Star Ratings. www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovGenIn/Downloads/2017_Star_Ratings_and_Display_Measures.zip. Accessed November 4, 2020.
- Centers for Medicare & Medicaid Services. Fact sheet-2020 Part C and D Star Ratings. www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovGenIn/Downloads/2020-Star-Ratings-Fact-Sheet-.pdf. Accessed November 4, 2020.
- Puskas CM, Forrest JI, Parashar S, et al. Women and vulnerability to HAART non-adherence: a literature review of treatment adherence by gender from 2000 to 2011. Curr HIV/AIDS Rep. 2011;8:277-287.
- Vimalananda VG, Miller DR, Hofer TP, et al. Accounting for clinical action reduces estimates of gender disparities in lipid management for diabetic veterans. J Gen Intern Med. 2013;28(suppl 2):S529-S535.
- Lewis LM, Askie P, Randleman S, Shelton-Dunston B. Medication adherence beliefs of community-dwelling hypertensive African Americans. J Cardiovasc Nurs. 2010;25:199-206.
- Ogedegbe G, Harrison M, Robbins L, et al. Barriers and facilitators of medication adherence in hypertensive African Americans: a qualitative study. Ethn Dis. 2004;14:3-12.
- Couto JE, Panchal JM, Lal LS, et al. Geographic variation in medication adherence in commercial and Medicare Part D populations. J Manag Care Spec Pharm. 2014;20:834-842.
- Artiga S, Damico A. Health and health coverage in the South: a data update. Issue brief. Kaiser Family Foundation; February 10, 2016. www.kff.org/racial-equity-and-health-policy/issue-brief/health-and-health-coverage-in-the-south-a-data-update/. Accessed November 4, 2020.
- Mojtabai R, Olfson M. Medication costs, adherence, and health outcomes among Medicare beneficiaries. Health Aff (Millwood). 2003;22:220-229.
- Soumerai SB, Pierre-Jacques M, Zhang F, et al. Cost-related medication nonadherence among elderly and disabled Medicare beneficiaries: a national survey 1 year before the Medicare drug benefit. Arch Intern Med. 2006;166:1829-1835.
- Centers for Medicare & Medicaid Services. Medicare 2017 Part C & D Star Rating technical notes. Updated September 26, 2016. www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovGenIn/Downloads/2017_Star_Ratings_and_Display_Measures.zip. Accessed November 4, 2020.