Baseline characteristics and secondary medication adherence among Medicare patients diagnosed with transthyretin amyloid cardiomyopathy and/or receiving tafamidis prescriptions: A retrospective analysis of a Medicare cohort

Abstract

BACKGROUND:

Transthyretin amyloid cardiomyopathy (ATTR-CM) is an underdiagnosed, life-threatening condition that mostly affects older persons. In May 2019, regulatory approval of tafamidis provided the first pharmacologic treatment of ATTR-CM. In the pivotal phase 3 Transthyretin Amyloidosis Cardiomyopathy Clinical Trial (ATTR-ACT), 97.2% of patients were classified as adherent (defined as taking ≥ 80% of scheduled doses). Given its recent approval, there is limited real-world evidence examining patient adherence to tafamidis.

OBJECTIVE:

To evaluate adherence patterns, demographics, and clinical characteristics of patients in the United States receiving tafamidis prescriptions through Medicare. Secondarily, we aimed to evaluate concomitant medications filled by this patient population.

METHODS:

We conducted a retrospective cohort study of US Medicare claims data, limited by the Health Insurance Portability and Accountability Act of 1996, in adult patients with an adjudicated pharmacy claim for tafamidis (tafamidis free acid 61-mg capsule once daily or tafamidis meglumine four 20-mg capsules once daily) between May 1, 2019, and June 30, 2021. Gaps in therapy were measured using day gaps between prescription refills and continuous measure of medication gaps. Implementation adherence was assessed through modified medication possession ratio (MPRm), medication refill adherence (MRA), and proportion of days covered (PDC). Patients were grouped based on Medicare coverage. Patients were analyzed by subgroups based on age and at the zip code level, via distressed communities index quartiles and rural-urban tiers.

RESULTS:

A total of 3,558 patients who received a prescription fill of a tafamidis formulation were identified using Medicare Fee-for-Service (FFS) and Medicare Advantage (MA) claims data from May 1, 2019, to June 30, 2021. The characteristics of this patient population were consistent with published literature, as 98.6% were older than 65 years, 53.4% were between 75 years and 84 years, and 81.5% were male. In the patient population receiving tafamidis refills, adherence was high across all 3 measures, with mean MPRm greater than 90% and mean MRA greater than 80%, across all age groups. Mean PDC adherence rates were 79% or more across all age groups. Concomitant medications were generally indicated for heart failure and thrombosis. Among monotherapy groups with similar demographic makeup, adherence was significantly higher among users of tafamidis free acid vs tafamidis meglumine (P < 0.0001 across all mean adherence measures).

CONCLUSIONS:

Our results demonstrate that real-world adherence to tafamidis in the Medicare population is high, regardless of age, zip code–level socioeconomic quartile, or geography. Adherence was higher among patients receiving tafamidis free acid, suggesting that the enhanced convenience of a single capsule once daily may positively contribute to adherence among patients with ATTR-CM.

Plain language summary

Tafamidis is the first drug approved to treat transthyretin amyloid cardiomyopathy (ATTR-CM). It gained approval by the US Food and Drug Administration in May 2019. ATTR-CM is a serious condition that mostly affects older persons. As of now, real-world adherence data for tafamidis are scarce. In this study, we included Medicare patients in the United States who filled tafamidis prescriptions. Adherence to tafamidis and patient traits, including age, sex, and race, were assessed. Here, we report that real-world adherence to tafamidis was high.

Implications for managed care pharmacy

Medication adherence in patients with cardiovascular disease has been associated with improved health outcomes, prevention of disease progression, and decreased health care costs; however, reported adherence in this population is low. Utilizing several adherence measures, our study demonstrated that real-world adherence to tafamidis in Medicare patients with ATTR-CM is consistently high regardless of age, sex, race, socioeconomic status, and geography. Therefore, real-world tafamidis adherence has the potential to support clinical and financial benefits.

Transthyretin amyloid cardiomyopathy (ATTR-CM) is a life-threatening condition in which destabilizing mutations (hereditary) or spontaneous age-linked processes (wild-type) cause amyloid fibrils to form and accumulate in cardiac tissue.¹ Although diagnoses can be confirmed via nuclear scintigraphy and/or endomyocardial biopsy, lack of awareness for ATTR-CM and the heterogeneity of symptoms at presentation often lead to the disease going undiagnosed and untreated.^2-5 Across 2 autopsy studies, the incidence of wild-type transthyretin amyloidosis (ATTR) deposits was found to increase with age, with prevalence being as high as 20%-25% in patients aged 80-89 years and 37% in those older than 95 years.^1,6,7 Without treatment, transthyretin deposition leads to a thickening of left ventricular walls and diastolic dysfunction, eventually causing heart failure.¹ Severe, rapid disease progression diminishes quality of life and increases economic burden in untreated patients.¹ Ultimately, ATTR-CM is a fatal disease, with a median survival of 2-3.5 years if untreated.⁸ Historically, the lone treatment for ATTR-CM was heart or heart/liver transplantation, as no effective pharmacologic treatments were available.^5,9-12 The development and regulatory approval of tafamidis was a crucial landmark for the treatment landscape of ATTR-CM.^1,13

Vyndaqel (tafamidis meglumine) and Vyndamax (tafamidis free acid), herein collectively referred to as tafamidis, are bioequivalent, first-in-class therapies for wild-type or hereditary ATTR-CM approved by the US Food and Drug Administration (FDA).¹⁴ A single 61-mg tafamidis free acid capsule taken once daily is bioequivalent to four 20-mg tafamidis meglumine capsules taken once daily.¹⁵ Tafamidis free acid was developed for patient convenience.¹⁶ Tafamidis acts as a stabilizer of transthyretin and inhibits the rate-limiting step of the amyloidogenic process.^17,18 In the pivotal phase 3 Transthyretin Amyloidosis Cardiomyopathy Clinical Trial (ATTR-ACT), tafamidis provided significant reductions in all-cause mortality and cardiovascular-related health care utilization, alongside reducing the decline of functional capacity and quality of life, compared with placebo.¹⁹

It has been shown that in chronic conditions, and especially in cardiovascular disease, medication adherence is a key determinant in the optimization of patient outcomes.^20-22 Nonadherence to medications used for cardiovascular disease leads to a higher risk of experiencing cardiovascular events, hospitalization, and mortality.²³ Adherence was high in the ATTR-ACT, as 97.2% of patients took at least 80% of scheduled doses.^19,24 Secondary adherence, which measures whether patients receive refills as prescribed, is often used to evaluate real-world adherence.²⁵ For the remainder of the text, we utilize the term adherence to maintain consistency. Adherence rates are generally higher for both trial-related and nontrial pharmacotherapy while patients are participating in a clinical trial²²; however, in the real world, evidence suggests that adherence to cardiovascular medication is much lower than in the clinical trial setting.^21,23,26-28

At the time of this study, a limited amount of published data were available describing the patient characteristics and prescription refill patterns of tafamidis free acid and tafamidis meglumine in the US Medicare population.

The primary objectives of this study were to describe adherence patterns, demographics, and clinical characteristics of the Medicare patient population with evidence of tafamidis prescription fills and to evaluate concomitant medications filled by patients who had received tafamidis.

Methods

OVERVIEW

This was a noninterventional, observational, retrospective cohort study of US Medicare claims data, limited by the Health Insurance Portability and Accountability Act (HIPAA) of 1996, in patients with evidence of a prescription fill of tafamidis meglumine or tafamidis free acid. Both Medicare Fee-for-Service (FFS) and Medicare Advantage (MA) claims data spanning from January 1, 2010, to June 30, 2021, were accessed via the Centers for Medicare & Medicaid Services (CMS) Chronic Conditions Data Warehouse (CCW).

STUDY POPULATION

This study included adult patients (aged ≥ 18 years) with an adjudicated pharmacy claim reported in the CMS CCW for tafamidis meglumine or tafamidis free acid between May 1, 2019, to June 30, 2021, who had a positive days supply across all tafamidis claims and continuous Medicare Part D coverage 6 months prior to index and a minimum of 3 months post index. As the study used only commercially available HIPAA-limited secondary data sources, under US Department of Health and Human Services regulations 45 CFR 46.104, it was considered exempt from requirements for “human subjects research” in the United States, including institutional review board; therefore, informed consent was not sought/deemed necessary. This study was conducted in line with legal requirements, regulatory requirements, and generally accepted research practices presented in the Good Practices for Outcomes Research provided by the International Society for Pharmacoeconomics and Outcomes Research (ISPOR).^29,30

Within the study period, patients were indexed from their earliest tafamidis prescription date and observed until the end of follow-up, which was determined using 1 of the following: June 30, 2021 (end of data availability), loss of Part D coverage, or death. Patients had differing follow-up periods based on prescription fill dates.

Baseline characteristic data were provided by CCW. Information regarding how they were measured was not available. Concomitant medication data were provided by CCW and reported based on frequency. To assess adherence, measures including modified medication possession ratio (MPRm), medication refill adherence (MRA), and proportion of days covered (PDC) were utilized. Patients were grouped into FFS or MA subgroups based on Medicare coverage. Baseline characteristics were assessed according to these cohorts. International Classification of Diseases, Tenth Revision, (ICD-10) diagnosis codes were derived from inpatient and outpatient FFS claims. Thus, FFS patients were stratified based on those with an ATTR-CM diagnosis (ICD-10 E85.82) and those with other diagnosis codes (ICD-10 codes E85.x, except E85.82). Clinical characteristics included comorbidities, Quan-Charlson Comorbidity Index (CCI) scores, and Hierarchical Condition Categories (HCC) scores. Patients were also classified at the zip code level, via geographic Rural-Urban Commuting Area (RUCA) tiers and socioeconomic Distressed Communities Index (DCI) quintiles.^31,32 Specific cohorts within the Medicare population that were evaluated for adherence included those younger than 65 years with end-stage renal disease (ESRD) or who were disabled (qualifying conditions such as amyotrophic lateral sclerosis), and those with dual Medicare FFS and Medicaid coverage. For inclusion in the adherence analysis, an additional attrition step was applied, in which patients must have had 2 fully adjudicated claims for tafamidis with a days supply greater than zero between May 2019 and June 30, 2021 (Figure 1).

FIGURE 1.

Study Design and Cohort Attrition

MEASURES OF MEDICATION DAY GAPS AND ADHERENCE

Follow-up intervals were calculated using time from first prescription claim of tafamidis (fill or diagnosis) until the end of follow-up, which was determined using 1 of the following: June 30, 2021 (end of data availability), loss of Part D coverage, or death. For MPRm, follow-up time was defined as days between first and last prescription fill plus days supply of the last fill.

Patients were grouped by the following minimal followup times: at least 3 months, at least 6 months, and at least 12 months. Patients may have been placed in multiple groups, depending on the totality of their individual follow-up. Adherence measures were assessed over each group that a patient was categorized into following the first tafamidis claim.

Gaps in therapy were measured using day gaps between prescription refills and continuous measure of medication gaps (CMGs). Day gaps between prescription refills were calculated as the days between prescription refills minus the days supply of the prescription. When time between fills exceeded the days supply of the previous fill, there was a “gap” in fills. Gaps were reported as categories (0 days, 1-30 days, 31-60 days, 61-90 days, and > 90 days).

Day Gaps Between Prescription Refills = Days between prescription refills-days’ supply

CMG was calculated using the equation below and was reported as a percentage.²⁵

$$CMG=\frac{\begin{array}{l}(\mathit{\text{Days from first dispensing to end of follow-up-sum of days supplied from first dispensing to end of follow-up}})\end{array}}{\begin{array}{l}(\mathit{\text{Days between first dispensing to end of follow-up}})\end{array}}\times 100$$

Adherence was measured using MPRm, PDC, and MRA. MPRm was calculated as days supply of tafamidis dispensed throughout the observation period, from first to last dispensing, divided by the number of days between first and last dispensing plus days supply of the last dispensing.²⁵ MPRm was capped at 1, multiplied by 100, and reported as a percentage. MPRm was the MPR variation used in this study because it included days supply dispensed with last dispensing.²⁵

$$MPRm=\frac{\begin{array}{l}(\mathit{\text{Sum of days supplied for all refills from first to last dispensing}})\end{array}}{\begin{array}{l}(\mathit{\text{Days between first dispensing to last dispensing plus days supplied in last dispensing}})\end{array}}\times 100$$

PDC was calculated by taking the number of days covered in the follow-up period divided by the number of days in that follow-up period, multiplied by 100, and was reported as a percentage.²⁵

$$PDC=\frac{\begin{array}{l}(\mathit{\text{Sum of days covered over the follow-up period}})\end{array}}{\begin{array}{l}(\mathit{\text{Days between first dispensing to end of follow-up}})\end{array}}\times 100$$

MRA was calculated by taking total days supply divided by number of days in the observation period.25 MRA was capped at 1, multiplied by 100, and reported as a percentage.

$$MRA=\frac{\begin{array}{l}(\mathit{\text{Sum of days supplied for all refills from first to end of follow-up}})\end{array}}{\begin{array}{l}(\mathit{\text{Days between first dispensing to end of follow-up}})\end{array}}\times 100$$

For this exploratory analysis, multiple thresholds were assessed to accommodate the various thresholds reported in literature and relatively recent approval of tafamidis.^27,33 An adherence rate of 80% is a commonly applied threshold in the literature and was the defined adherence cutoff in the ATTR-ACT.^19,33-35 Thus, patients were categorically classified as adherent (yes/no) based on an MPRm of at least 70%, 75%, and 80%. Similarly, patients were categorically classified as adherent (yes/no) based on an MRA and PDC of at least 70%, 75%, and 80%.

MEASURES OF THERAPY DURATION

Duration of therapy was assessed using 2 different methods. The first method utilized dates of prescription fills observed, whereas the second method used PDC. The first method calculated duration of therapy by using days between first and last index claim of tafamidis plus days supply of that last claim. For example, if the first index claim occurred on January 1, 2020, and the last claim was on September 1, 2020, for a 30-day fill, then the duration of therapy would be 273 days. The second method calculated duration of therapy by using total number of days in the follow-up period multiplied by PDC. Using this method, the days of follow-up a patient was covered for treatment could be determined, but it did not consider any permissible day gaps between fills. For example, if a patient was followed for 365 days and had a PDC of 90%, then their duration of therapy would have been 328.5 days. Both methods were used to analyze the data, and duration of therapy was reported using the conventions of continuous variables, as described in the statistical methods section.

STATISTICAL METHODS AND DATA ANALYSES

Data were analyzed using SAS version 9.4 (SAS Institute Inc, Cary, NC). Patient and treatment characteristics were summarized using descriptive statistics. Categorical variables were summarized by number of available observations, frequency, percentage, and 95% CIs. Continuous variables were summarized by number of available observations, mean, SD, 95% CIs, median, quartiles, minimum, and maximum. Missing categorical data were reported as a separate “missing” category. Missing continuous data were not included in summaries and analyses. No imputations were performed. As this study did not specify hypotheses a priori, all statistical analyses are descriptive and exploratory. Adherence assessments were stratified by the following variables: minimum length of follow-up (3 months, 6 months, and 12 months), enrollment type (FFS vs MA), age, sex, race, presence of an ATTR-CM diagnosis code, tafamidis formulation (tafamidis meglumine or tafamidis free acid), and whether a patient was disabled or had ESRD. Among monotherapy groups with similar demographic makeup, analyses comparing adherence rates between tafamidis formulations were explored. RUCA and DCI were also considered as stratifying variables. An α = 0.05 was utilized to determine statistical significance in analysis of variance tests that were completed.

Results

Study cohort attrition is presented in Figure 1. From the Medicare claims database, there were a total of 4,513 eligible patients with at least 1 adjudicated pharmacy claim for a tafamidis product between May 1, 2019, to June 30, 2021. Of 4,016 patients who had at least 2 fully adjudicated claims for tafamidis with a positive days supply greater than zero, 3,848 patients had continuous Medicare Part D coverage 6 months prior and 3 months after first prescription. Patients were divided into specific subgroups based on amount of follow-up time and then evaluated using the measures of adherence. There were 3,309 patients who had at least 1 gap in fills.

PATIENT CHARACTERISTICS

Overall, the highest proportion of patients receiving tafamidis were aged between 75 years and 84 years (53.4%), and there was a higher proportion of male patients than female patients (81.5% vs 18.5%, respectively). Patients were variably distributed, with the largest discrepancy occurring between the South (40.0%) and West (8.3%), most were non-Hispanic White (71.3%), and the majority of patients were enrolled in FFS (57.0%) (Table 1).

TABLE 1.

Patient Characteristics

Baseline characteristics	n	%
Age, y
< 65	50	1.4
65-74	834	23.4
75-84	1,899	53.4
> 85	775	21.8
Sex
Female	657	18.5
Male	2,901	81.5
Region
Midwest	661	21.2
Northeast	950	30.5
South	1,248	40.0
West	260	8.3
Race
African American	855	24.0
Hispanic	39	1.1
Non-Hispanic White	2,538	71.3
Other	126	3.5
Enrollment type
FFS	2,027	57.0
MA	1,531	43.0
Clinical characteristics	n	%
CCIb
Mild (1-2)	547	26.6
Moderate (3-4)	599	29.2
Severe (5+)	906	44.1
	n	Mean HCC Score
Racea,b,c
African American	355	2.15
Hispanic	14	2.89
Non-Hispanic White	1,717	2.00
Other	84	2.20
Diagnosisb
Patients with E85.82 diagnosis	1,189	2.13
Patients without E85.82 diagnosis	981	1.92

^aDenotes statistical significance.

^bClinical characteristics were evaluated for Medicare FFS patients only.

^cAssessed by analysis of variance, HCC: P = 0.0067.

CCI = Quan-Charlson Comorbidity Index; FFS = Medicare Fee-for-Service;

HCC = Hierarchical Condition Categories; MA = Medicare Advantage; y = years.

The most common comorbidities for the FFS population were congestive heart failure (90.3%), hypertension (75.7%), ischemic heart disease (73.5%), hyperlipidemia (66.9%), chronic kidney disease (60.3%), rheumatoid arthritis/osteoarthritis (54.4%), and atrial fibrillation (43.6%). CCI severity varied, with the highest proportion of patients having a severe score (44.1%). Mean HCC scores across all races varied significantly (P = 0.0067), with the highest mean score being reported in Hispanic patients (2.89) and the lowest mean score reported in non-Hispanic White patients (2.00). Mean HCC scores were similar between patients with and without an E85.82 diagnosis code at 2.13 and 1.92, respectively (Table 1).

PRESCRIPTION CHARACTERISTICS

Among patients analyzed, the majority of patients stayed on the same tafamidis product (n = 3,170) for the duration of the follow-up period. A larger proportion of patients stayed on tafamidis free acid (n = 2,000) for the duration of the follow-period than those who stayed on tafamidis meglumine (n = 1,170). A total of 675 patients switched from tafamidis meglumine to tafamidis free acid, whereas less than 10 patients switched from tafamidis free acid to tafamidis meglumine. The most prominent specialties reported among prescribers in this patient population were cardiology and advanced heart failure and transplants.

PRESCRIPTION CHARACTERISTICS OVER FOLLOW-UP

Generally, concomitant medications in this patient population were indicated for use in heart failure and thrombosis. Most commonly prescribed medications included apixaban, torsemide, furosemide, and rivaroxaban. Most patients did not have any gaps in refills, and those who did were likely to have a gap of less than 30 days. The majority of patients with gaps had between 1 and 4 gaps in fills (60.7%).

ADHERENCE TO TAFAMIDIS

Across all FFS and MA cohorts, more than 87% of patients had an MPRm greater than 80%, at least 62% of patients had a PDC greater than 80%, and more than 65% of patients had an MRA greater than 80% (Table 2). Adherence results by subgroup are presented in Table 3. The mean MPRm across all follow-up groups for both FFS and MA claims decreased as follow-up time increased. This trend was also seen for mean and median values for PDC and MRA. Across all age groups, mean MPRm was high, as evidenced by rates greater than 90%. All age groups had a mean MRA greater than 80%. Mean PDC adherence rates were 79% or more across all age groups. Increases in age corresponded to an increase in CMGs with the age group older than 85 years having the highest median CMG (2.2%). The mean MPRm, PDC, and MRA were similar between sexes. Median MPRm, PDC, and MRA values were 89.9% or more for both sexes. Median CMG was higher in the female cohort (4.2% vs 1.4%). All cohorts by race had a mean MPRm greater than 91% as well as mean PDC and MRA greater than 77%. Median CMG was highest in the Hispanic cohort (7.7%).

TABLE 2.

Percentage of Patients With More Than 80% Adherence

Minimum follow-up	Enrollment type	n	Percentage of patients with > 80% adherence
			MPRm	PDC	MRA
3 months	FFS	2,027	92.7	69.6	73.7
6 months		1,657	91.9	67.5	71.0
12 months		1,212	90.5	63.8	67.2
3 months	MA	1,531	90.8	68.3	71.8
6 months		1,243	89.1	64.8	67.9
12 months		887	87.6	62.0	65.1

FFS = Medicare Fee-for-Service; MA = Medicare Advantage; MPRm = modified medication possession ratio; MRA = medication refill adherence;

PDC = proportion of days covered.

TABLE 3.

Adherence Rates by Subgroup

	Enrollment type	n	Mean adherence (%)			Median adherence (%)			Median CMG (%)
			MPRm	PDC	MRA	MPRm	PDC	MRA
Minimum follow-up
3 months	FFS	2,027	95.2	79.3	83.3	100.0	92.0	98.3	—
6 months		1,657	94.7	77.5	81.6	100.0	91.5	97.8	—
12 months		1,212	93.9	74.8	79.0	100.0	90.3	96.8	—
3 months	MA	1,531	94.3	79.0	82.9	100.0	91.7	98.2	—
6 months		1,243	93.5	76.6	80.6	100.0	90.7	97.2	—
12 months		887	92.6	74.2	78.4	100.0	89.5	96.2	—
Age, y
< 65	—	50	95.5	79.4	82.4	100.0	94.9	99.7	0.3
65-74	—	834	94.4	79.3	83.1	100.0	92.5	98.5	1.5
75-84	—	1,899	94.9	79.2	83.2	100.0	91.8	98.1	1.9
> 85	—	775	94.8	79.0	83.0	100.0	91.1	97.8	2.2
Sex
Female	—	657	93.7	78.0	82.0	100.0	89.9	95.8	4.2
Male	—	2,901	95.1	79.4	83.4	100.0	92.2	98.6	1.4
Race
African American	—	950	92.6	78.1	82.1	100.0	89.5	96.2	3.7
Hispanic	—	661	91.6	77.8	81.6	99.2	88.0	92.3	7.7
Non-Hispanic White	—	260	95.5	79.4	83.4	100.0	92.3	98.7	1.3
Other	—	1,248	96.9	82.2	86.1	100.0	93.1	98.7	1.3
Diagnosisa
Patients with E85.82 diagnosis	—	1,189	95.0	78.0	82.0	100.0	91.0	98.0	—
Patients without E85.82 diagnosis	—	981	96.0	84.0	87.0	100.0	94.0	100.0	—
Tafamidis formulationb
Tafamidis free acid		2,000	97.0	87.0	90.0	100.0	95.0	100.0	—
Tafamidis meglumine		1,170	95.0	69.0	73.0	100.0	85.0	92.0	—
RUCA tier
1 Most urban	—	3,206	—	80.6	—	—	—	—	—
2	—	373	—	80.5	—	—	—	—	—
3	—	118	—	81.3	—	—	—	—	—
4 Most rural	—	140	—	75.1	—	—	—	—	—
DCI quintile
1 Prosperous	—	1,286	—	81.4	—	—	—	—	—
2	—	952	—	81.0	—	—	—	—	—
3	—	588	—	80.8	—	—	—	—	—
4	—	476	—	76.8	—	—	—	—	—
5 Distressed	—	460	—	79.6	—	—	—	—	—
Medicare cohort
Disabled	—	57	93.0	77.0	80.0	100.0	89.0	99.0	—
Dualc	—	375	94.0	83.0	88.0	99.0	92.0	97.0	—
ESRD	—	15	97.0	78.0	81.0	100.0	91.0	93.0	—

^aOnly evaluated FFS patients.

^bDenotes P < 0.0001 assessed by analysis of variance for mean adherence (MPRm, PDC, and MRA).

^cPertains to patients enrolled in both Medicare FFS and Medicaid.

— = not applicable; CMG = continuous measure of medication gaps; DCI = Distressed Communities Index; ESRD = end-stage renal disease; FFS = Medicare Fee-for-Service; MA = Medicare Advantage; MPRm = modified medication possession ratio; MRA = medication refill adherence; PDC = proportion of days covered; RUCA = Rural-Urban Commuting Area; y = years.

Among FFS patients, adherence rates were similar for those with and without an ATTR-CM diagnosis (ICD-10 E85.82). There were 1,189 patients who had evidence of an E85.82 diagnosis code and 981 who had evidence of other amyloidosis diagnosis codes (ICD-10 codes E85.x, except E85.82).

Among groups with similar demographic makeup, there was a significant difference (P < 0.0001) across all mean adherence measures between tafamidis meglumine and tafamidis free acid monotherapy groups. The tafamidis meglumine monotherapy group had mean MPRm, PDC, and MRA values of 95%, 69%, and 73%, whereas tafamidis free acid monotherapy users had values of 97%, 87%, and 90%, respectively. All median MPRm, PDC, and MRA values were 85% or more across users of both tafamidis products. The highest proportion of patients were classified as being in the “most urban” RUCA tier and “prosperous” DCI quintile. Mean PDC were similar across all DCI quintiles and RUCA tiers. Across the specific Medicare cohorts (ESRD, disabled, and dual coverage), adherence was high, with mean MPRm greater than 90% as well as PDC and MRA values 77% or more. Median MPRm, PDC, and MRA were also 89% or more in these cohorts.

A histogram of patients’ adherence rate to tafamidis, bucketed by increments of 0.05, showed that roughly 69% of patients had a PDC-measured adherence greater than 0.80, categorizing patients as adherent to their medication (Supplementary Figure 1 ^{(79.8KB, pdf)} , available in online article). More than half (50.93%) of patients with a minimum of 12 months of follow-up had 12 or more fills/refills of tafamidis within the study period and more than 70% received 7 or more prescriptions over the study period.

Discussion

As described in a medication adherence report published by the World Health Organization, average adherence is 50% in patients with a chronic condition.²⁸ Nonadherence to pharmacologic treatment prescribed for chronic conditions such as cardiovascular disease greatly reduce the effectiveness of treatment and, over time, increase the likelihood of poor patient outcomes, lower quality of life, and heightened economic burden.^20-22 In patients diagnosed with ATTR-CM who go untreated, severe disease progression leads to a lower quality of life and relatively short survival.⁸ As a transthyretin stabilizer, it is imperative that effective tafamidis concentration levels are maintained throughout the duration of treatment to slow progression of ATTR-CM and minimize complications associated with it. Although the ATTR-ACT trial demonstrated that adherence was high in participants, this study constitutes one of the first demonstrations of high real-world adherence rates to tafamidis meglumine and tafamidis free acid in Medicare beneficiaries.

Because Medicare primarily covers individuals older than 65 years, the FFS and MA database provided access to a population in which incidence of ATTR-CM is highest.^2-4 Nearly all patients (98.6%) were older than 65 years, and the highest proportion of patients were aged between 75 years and 84 years (53.4%). We found that 81.5% of patients were male, similar to proportions (83% and 86.9%) recently reported in literature.^36,37 Others have reported much higher proportions of male patients, with ratios between 10:1 and 20:1.^38-43 Patients were variably distributed across the United States, with the largest cohort being in the South (40.0%). The majority of patients were non-Hispanic White (71.3%), and more than half were enrolled in FFS (57.0%).

Comorbidities and concomitant medications were as expected for this patient population based on age and potential complications of ATTR-CM. The most common comorbidities included congestive heart failure (90.3%), hypertension (75.7%), and ischemic heart disease (73.5%). The significant difference seen in HCC scores between races provides insight into potential confounders that warrant future investigation.⁴⁴

Throughout the study period, most patients did not have a gap between refills and most gaps in fills were less than 30 days, suggesting good refill adherence. Adherence was high and well over the industry standard of 80%, as evidenced by median MPRm, PDC, and MRA values of 85% or more. Mean MPRm across all evaluated subgroups were between 91.6% and 97.0%. Patients who received tafamidis free acid had higher adherence across all measures. The significant difference between the monotherapy groups may be related to the enhanced convenience associated with tafamidis free acid, which is taken as 1 capsule once daily, whereas tafamidis meglumine is taken as 4 capsules once daily.⁴⁵ Although most patients stayed on the same tafamidis product for the duration of the study period, the majority of patients who switched went from tafamidis meglumine to tafamidis free acid. Future investigation into whether the enhanced convenience of tafamidis free acid is associated with increased adherence may provide further insight. Moreover, there were no substantial variations in adherence when patients were classified based on their socioeconomic (as measured by DCI) or geographical (as measured by RUCA) characteristics at the zip code level. Overall, our findings demonstrate that adherence to tafamidis is high within the Medicare patient population.

LIMITATIONS

The limitations of our study are consistent with those of others utilizing pharmacy claims data. PDC relies on the assumption that a patient actually receives prescriptions for refill for the duration they are enrolled and eligible to receive a prescription. As such, we included multiple measures of adherence, including MPRm, which does not rely on this assumption. Retrospective analyses have the potential to include discrepancies in input data, to fail to identify partial adherence or capture prescriptions not filled through Medicare, and to provide incomplete records that do not include reasons for discontinuation.⁴⁶ To prevent errors in data entry from being included, only fully adjudicated claims were utilized. As claims were not accompanied by medical charts or electronic health records, we were unable to establish factors contributing to the prescription trends observed. This study also did not take into consideration the cost of tafamidis.

The adherence realized in this study is only considered a proxy of real-world adherence in patients who received tafamidis. For all adherence measures used in this study, we were only able to ascertain specific time points, such as when claims were adjudicated, prescriptions were filled, and when prescriptions were received or picked up by patients. Thus, the assumption was made that patients who received medication administered it correctly.⁴⁶ Identified nonadherence behaviors include noninitiation (no initial fill or administration of medication), incorrect implementation (late refills, missed doses, and improper timing or administration of dosing), and nonpersistence (early discontinuation).⁴⁷ Like most research measuring secondary adherence, our analysis provides information regarding the late-refills aspect of the 3 nonadherence behaviors.

The adherence terminology used in this study aligns with the implementation phase, as defined by the Ascertaining Barriers to Compliance project team and ESPACOMP Medication Adherence Reporting Guideline, but does not use the adherence phases of initiation and discontinuation.^48,49 The adherence terminology provided by ISPOR was used in this study, as it places a specific focus on prescription refill data.⁴⁹

The first notable strength of this study was the utilization of several adherence measures, which allowed us to assess adherence to tafamidis from multiple perspectives in this exploratory analysis. Second, the patient pool of the Medicare database allowed for the identification of patients with ATTR-CM within the United States. With the Medicare database, we were able to focus mainly on an older age group, of which individuals are particularly affected by ATTR-CM.^1,6,7 ICD-10 codes allowed us to evaluate patients who received tafamidis and had a specific diagnosis of ATTR-CM. Although data, such as ICD-10 codes, from FFS claims were available through June 30, 2021, MA claims in this data set were not available after 2019, and thus could not be used in the same capacity.

Other recent studies have utilized the Medicare database to provide data regarding real-world treatment patterns of a specific pharmacotherapy; reported results demonstrated higher adherence rates corresponding to lower health care resource utilization and identified at-risk groups that can be assisted to increase adherence.^50,51 Moving forward, areas of investigation include further examination into nonadherence behaviors, such as frequency of missed doses, dose timing, administration of dosing, and nonpersistence. Future studies may also assess other potential factors that may lead to lower adherence, such as to help inform efforts to support patients’ access and adherence to their tafamidis regimen to maximize their outcomes.

Conclusions

This study is one of the first to report real-world adherence of tafamidis in patients with ATTR-CM covered by Medicare in the United States. This examination showed that the characteristics of this patient population were consistent with published literature, as they were predominantly older and male. Patients generally used the same tafamidis product for the duration of the observation period and commonly received concomitant medications indicated for heart failure and thrombosis. Adherence to tafamidis was high across multiple subgroups, such as age, sex, and race. Adherence is higher in patients who received tafamidis free acid, which is taken as 1 capsule once daily. Most patients had no gaps in fills for tafamidis, and among those who did, gaps were primarily less than 30 days. Optimal adherence to pharmacologic treatment, especially for progressive conditions such as ATTR-CM, is imperative to ensuring that maximal effectiveness and patient outcomes are achieved.