Impact of Prescription Benefit Coverage Limits on Sevelamer Hydrochloride Adherence for Patients with ESRD

Abstract

Objective

To assess the impact of prescription benefit coverage on medication adherence in Medicare-eligible members diagnosed with end-stage renal disease taking sevelamer hydrochloride.

Methods

This pilot study involved a retrospective analysis of patients with end-stage renal disease taking sevelamer, with an annual cap on brand prescription drug spending compared with those without a cap. We compared sevelamer adherence and discontinuation proportions between the 2 groups of Medicare patients in 2003 and 2004. Medication adherence was calculated based on the proportion of available days covered in relationship to capped versus noncapped pharmacy benefit.

Results

Rate ratios showed that in 2003, the patients taking sevelamer under a capped benefit (N = 43) had 27% fewer days of drug use compared with those (N = 88) without a capped benefit (relative risk, 0.73; 95% CI, 0.58–0.93). Similarly, in 2004, those taking sevelamer under the capped benefit (N = 21) had 33% fewer days of drug use compared with those (N = 117) without a capped benefit (relative risk, 0.67; 95% CI, 0.46–0.96).

Conclusions

Medication adherence was significantly lower for patients with a capped brand-name drug benefit. These findings provide insight into potential drug utilization patterns, including for sevelamer, under the Medicare Part D benefit, where members could face significant out-of-pocket expenditures once coverage limits are reached.

Hyperphosphatemia is prevalent among patients with end-stage renal disease (ESRD) and continues to be an important and challenging area for drug therapy. The kidneys play a key role in maintaining normal serum phosphorus levels, which is impaired at an estimated glomerular filtration rate (GFR) of 50 mL/min to 60 mL/min.1,2 An adaptive process by the kidneys, which involves an increase in circulating parathyroid hormone (PTH) with normal calcitriol levels, helps to control phosphorus at this low GFR value.1,2 Once the GFR decreases to 25 mL/min to 30 mL/min, the PTH elevation and low levels of calcitriol cannot maintain a normal phosphorus level, leading to hyperphosphatemia.1,2 Hyperphosphatemia, in addition to elevated calcium phosphorus product and elevated PTH, is associated with significant morbidity and mortality.1

Failure to adequately control elevated serum phosphorus levels can result in cardiovascular events, including coronary artery disease (CAD), uncontrolled hyperparathyroidism, and fractures.1 These complications carry a significant economic burden and untoward personal costs associated with hospitalization, increased higher healthcare resource utilization, decreased quality of life, and premature death.3–7

The Kidney Disease Outcomes Quality Initiative (K/DOQI) clinical practice guidelines published by the National Kidney Foundation emphasize the importance of serum phosphorus control in patients with ESRD.8 The K/DOQI guidelines recommend the initiation of oral phosphate binders when serum phosphorus and intact PTH levels cannot be kept within the target range with dietary phosphorus restriction alone. Calcium-containing phosphate-binding medications are effective; however, they carry the risk of elevating serum calcium levels above the target range in patients with ESRD.8,9 Hypercalcemia has been linked to increased mortality risk in patients with ESRD.5,10–14

The K/DOQI guidelines suggest a preference for phosphate binders—which are noncalcium, nonmagnesium, and nonaluminum, such as sevelamer hydrochloride (Renagel; henceforth sevelamer)—in patients on dialysis with severe vascular and/or other soft-tissue calcifications. The guidelines also suggest the use of these agents in patients with hypercalcemia to control phosphorus levels.8,11 A recent study showed that 53% of patients with advanced chronic kidney disease fulfilled at least 1 criterion for the use of sevelamer.15

Sevelamer is a polymeric compound that binds phosphate within the intestinal lumen, limiting absorption of and decreasing serum phosphate concentrations without altering calcium, aluminum, or bicarbonate concentrations.16 It is not currently available as a generic product. For some patients, its use may be limited by its high cost.17 In addition, reduced adherence to sevelamer in patients with ESRD because of financial barriers can significantly increase utilization of high-cost drivers of healthcare, such as hospitalizations and outpatient visits.

In 2003, the average wholesale price (AWP) for sevelamer was $1.31/tablet or between $2832 and $8484 per year.18 In 2004, the AWP increased to $1.42/tablet or between $3072 and $9204 annually for an individual19 (based on a dosing frequency of 2–6 tablets/meal and 3 meals/day). Limitations with prescribing a phosphorus-restricted diet in addition to limitations in phosphorus removal by conventional dialysis, make the use of a noncalcium-based phosphate binder, such as sevelamer, the option to managing hyperphosphatemia without increasing the risk for hypercalcemia.5 Patients who are unable to comply with sevelamer therapy because of cost constraints are often left with the alternative of calcium-based phosphate binders and may be at risk for subsequent complications secondary to hypercalcemia.

In one study, patients treated with sevelamer had a 50% lower likelihood of hospitalization and had lower overall annual costs by more than $16,500 per patient than patients not receiving this drug.20 However, nonadherence to drug therapy has been a significant problem in the ESRD population.21–23 Nonadherence to hemodialysis treatment, as assessed by the number of missed dialysis sessions and hyperphosphatemia, has been associated with increased mortality.6 In that study, the relative risk for skipping 1 or more dialysis treatments per month was 1.30, and the mortality risk for serum phosphorus >6.5 mg/dL was 1.27 relative to a serum phosphorus of 2.4 mg/dL to 6.5 mg/dL.6,7 The direct cost-related nonadherence rate in the United States for patients with ESRD is 29%.24

KEY POINTS

• ▸Previous studies have shown that drug benefit coverage design can affect patient medication adherence.
• ▸Putting a cap on total drug spending has been associated with decreased medication adherence, resulting in poor clinical outcomes.
• ▸Results of this study showed that sevelamer hydrochloride adherence was significantly lower for Medicare patients with end-stage renal disease with a capped brand-name drug benefit compared with those without a capped benefit.
• ▸Reduced drug utilization may likely have been the result of greater patient cost-sharing. This utilization pattern has ramifications on biochemical outcomes in this patient population and potentially in other patients using brand-name medications.

Drug Benefit and Adherence

Caps on total drug spending have been associated with decreased medication adherence and reduced control of systolic blood pressure, low-density lipoprotein (LDL) cholesterol, and glycosylated hemoglobin in patients with hypertension, hyperlipidemia, and diabetes, respectively.25 Patients with drug coverage limits have increased emergency department visits and hospitalizations, which may have a direct impact on total healthcare spending.25–27

There is limited research examining the impact of sevelamer adherence within a capped drug benefit for patients with ESRD. An appropriate way to measure sevelamer adherence within a Medicare managed care population is by using the proportion of days covered (PDC).28,29 PDC is the number of days in the measurement period covered by prescription claims for the same medication or another in its therapeutic category as defined by the Pharmacy Quality Alliance.30 The PDC threshold is the level (typically 80%) above which the medication has a reasonable likelihood of achieving most of its potential clinical benefit.

Patients with ESRD are eligible for prescription drug coverage under Medicare Part D. The standard Part D benefit includes limits on total drug spending. For example, in 2006 the total drug spending limit was $2250, after which patients enter a coverage gap and are responsible for 100% of the cost of all medications until catastrophic coverage takes effect. The total drug spending limit for Medicare Part D increased to $2400 and $2510, respectively, in 2007 and 2008.31,32 Sevelamer is a tier 2 (preferred brand-name) drug in most Medicare Part D plans.33 Patients with ESRD taking sevelamer could be particularly vulnerable to spending limits in the Part D benefit, because they may be responsible for 100% of medication costs once they reach the coverage gap threshold.

The primary objective of this pilot study was to compare adherence and discontinuation proportions in patients with ESRD receiving sevelamer under a capped brand-name prescription benefit in Medicare plans and in retiree patients who have no caps in non-Medicare plans offered by Kaiser Permanente Colorado. Secondary end points included specific biochemical outcomes for patients with ESRD within a capped and a noncapped prescription benefit.

Methods

Kaiser Permanente Colorado is an integrated healthcare delivery system for more than 400,000 members in the Denver-Boulder metropolitan area. The physicians of the Colorado Permanente Medical Group contract exclusively with the Kaiser Foundation Health Plan to provide comprehensive healthcare services. Medicare-eligible members represent approximately 15% of Kaiser Permanente Colorado membership.

In 2003 and 2004, Medicare patients in a Medicare+Choice plan had a $30 or $50 copay for brand-name medications, with a $500 (2003 and 2004), $1000 (2004), or $1500 (2003) capped benefit. Members paid their copays until the prescription drug cap was met each year. At that point, they were responsible for 100% of the cost of all brand-name prescription drugs. The capped benefit differed slightly between 2003 and 2004. In 2004, patients with a Medicare+Choice plan also had a $300 deductible for brand-name medications, after which they faced a cap on brand-name drug spending.

The annual copay distribution for sevelamer in the capped group is shown in Table 1. A large proportion of patients in 2003 (95%) and 2004 (100%) had prescription copays of $50 for this medication.

Table 1.

Copays for Sevelamer in the Capped Group, 2003 and 2004

	2003		2004
Cap	$30 copay, N (%)	$50 copay, N (%)	$30 copay, N (%)	$50 copay N (%)
$500	2 (4.7)	14 (32.6)	NA	NA
$1000	NA	NA	NA	9 (42.9)
$1500	0	27 (62.8)	NA	12 (57.1)

The noncapped prescription benefit did not have any limits on prescription drug spending and represented a retiree Medicare population with employer-sponsored coverage. Retiree-based Medicare patients in the noncapped group typically paid a copay for brand-name medications ranging from $5 to $40. Table 2 shows the annual distribution of copays for sevelamer in the noncapped group. In 2003 and 2004, a large proportion of patients in the noncapped group had copays of $20 (27% and 20%, respectively) or $30 (15% and 20%, respectively).

Table 2.

Distribution of Copays for Sevelamer in the Noncapped Group, 2003 and 2004

Prescription drug copay, $	2003 N (%)	2004 N (%)
3	3 (3.4)	3 (2.6)
5	5 (5.7)	4 (3.4)
7	1 (1.1)	0
10	16 (18.2)	20 (17.1)
15	6 (6.8)	5 (4.3)
20	24 (27.3)	34 (29.1)
25	2 (2.3)	9 (7.7)
30	19 (21.6)	23 (19.7)
40	1 (1.1)	5 (4.3)
50	11 (12.5)	14 (12.0)

All patients with documented ESRD receiving sevelamer during the first 6 months of 2003 and 2004 were identified using pharmacy claims data. Inclusion criteria were continuous dialysis (peritoneal or hemodialysis) between January 1, 2003, and December 31, 2004, and continuous health plan membership with pharmacy benefits for this period. Because of a lack of awareness of the Medicare drug discount card (instituted in 2004), and the difficulty in identifying patients who had the card, the potential impact of the discount was excluded.34 The sevelamer index date was defined as the first sevelamer dispensed in each calendar year. All dispensing events were captured through the end of each calendar year. Patients were stratified into groups according to type of Medicare drug coverage (capped or noncapped). The 2 groups were kept separate, because the capped benefits changed from 2003 to 2004.

Sevelamer adherence was defined as the PDC during calendar years 2003 and 2004. The PDC was measured as a continuous variable and was calculated by dividing the total number of days supply of sevelamer dispensed by the total number of days between the index dispensing of sevelamer and the end of both calendar years.27–30

Patients were considered to have discontinued the use of sevelamer if they only had 1 prescription of the drug filled in each calendar year. Discontinuation proportions were calculated by dividing the number of patients who discontinued sevelamer by the total number of patients receiving a sevelamer prescription in the capped and noncapped group for each year.

Laboratory test results were obtained through a review process of each patient's charts. The last measured values for the relevant clinical markers were collected each month for each patient, and the average of these values was used as the outcome measure.

Medical and pharmacy data from administrative sources and patient chart reviews were used to identify patient age (based on the index date for a sevelamer dispensing); sex; comorbid conditions (major depressive disorder, atrial fibrillation, chronic heart failure [CHF], hypertension, CAD, diabetes mellitus); procedure for parathyroidectomy; pharmacy benefit structure (capped vs noncapped); concurrent drug use (calcimimetics; calcium-, aluminum-, or magnesium-containing phosphate binders; and vitamin D analog therapy); associated laboratory test results (corrected calcium-phosphorus product; and corrected serum calcium, serum phosphorus, and intact PTH).

For both calendar years, comorbid conditions of interest were identified by the codes of the International Classification of Diseases, 9th Revision, and procedures of interest were identified by the Current Procedural Terminology, 4th Edition. Although beneficiaries have the ability to change coverage during the year, pharmacy benefit structure at the index sevelamer prescription was assumed to continue during each calendar year. Concurrent drug use was primarily identified using National Drug Codes.

Statistical significance of differences for the descriptive variables noted above was tested using the Wilcoxon rank sum or chi-square tests for continuous and categorical variables, respectively. Associations between pharmacy benefit coverage and medication adherence and discontinuation were evaluated using Poisson regression modeling with the canonical loglink function and an offset parameter to accommodate the effect of unequal length of the individual follow-up period.35 The model was fit with a correction for overdispersion. Rate ratios were tabulated to compare the capped and noncapped groups. Logistic regression analysis was used to determine the relationship between pharmacy benefit coverage discontinuation proportions for sevelamer. Analyses were conducted using SAS 9.1 (SAS Institute; Cary, NC).

Results

Table 3 describes the characteristics of patients with ESRD receiving sevelamer during the first 6 months of 2003 and 2004. In 2003, 131 patients met the eligibility criteria: 43 (32.8%) patients were in the capped group and 88 (67.2%) in the noncapped group. In 2004, 138 patients met the eligibility criteria: 21 (15.2%) patients in the capped group and 117 (84.8%) in the noncapped group. The top 3 diagnoses for patients in the capped group in 2003 were diabetes mellitus, atrial fibrillation, and hypertension, and in 2004 diabetes mellitus, hypertension, and CAD. Similar trends were observed in the noncapped group.

Table 3.

Characteristics of Patients with Sevelamer Dispensing in the First 6 Months of 2003 and 2004

	2003 (N = 131)						2004 (N = 138)
	Capped (n = 43)		Noncapped (n = 88)			Capped (n = 21)		Noncapped (n = 117)
Characteristic	Median	(5th, 95th percentile)	Median	(5th, 95th percentile)	P	Median	(5th, 95th percentile)	Median	(5th, 95th percentile)	P
Age at first dispensing	62 yr	(33, 85 yr)	61 yr	(34, 78 yr)	.644	69 yr	(49, 88 yr)	62 yr	(27, 81 yr)	.014
Distinct oral medications, Na	6	(2, 15)	7	(2, 15)	.245	6	(3, 12)	7.5	(2, 16)	.134
Sex	N	%	N	%		N	%	N	%
Female	16	37.2	40	45.4		8	38.1	50	42.7
Male	27	62.8	48	54.6		13	61.9	67	57.3
Comorbid conditions
CAD	29	67.4	28	31.8	.001	11	52.4	47	40.2	.296
Diabetes mellitus	37	86.1	78	88.6	.670	20	95.2	100	85.5	.307
MDD	0	0.0	4	4.6	NA	0	0.0	2	2.3	NA
CHF	8	18.6	22	25.0	.413	7	33.3	24	20.5	.254
Atrial fibrillation	8	18.6	6	6.8	.067	6	28.6	8	6.8	.008
Hypertension	32	74.4	66	75.0	.942	15	71.4	92	78.6	.569
Parathyroidectomy	5	11.6	11	12.2	.921	1	4.8	8	6.7	.735

^aDistinct medications calculated by medications with a total of at least 90-day supply over the year summed up by the first 8 digits in the generic product identifier code (down to the drug name level).

CAD indicates coronary artery disease; CHF, chronic heart failure; MDD, major depressive disorder.

In 2003, the capped and noncapped groups were similar in most characteristics, except CAD, where the capped group had a higher percentage of members than the noncapped group (P = .001). In 2004, the noncapped group was younger than the capped group and had a lower percentage of members with atrial fibrillation.

The capped-group patients faced greater cost-sharing for sevelamer after they reached their cap (results not shown). In 2003, the mean copay for patients taking sevelamer in the capped group was $50 before the cap was reached and $395 after the cap was reached. In 2004, the mean copay was $160 before the cap was reached and $264 after. In 2003, the unadjusted median PDC was 40.0% for the capped group and 66.3% for the noncapped group; in 2004 the unadjusted median PDC for the capped group was 40.4% compared with 59.2% for the noncapped group. In 2003 and 2004, 60.5% of the patients in the capped group had PDCs <50% compared with 33% in the noncapped group. The unadjusted proportion of patients who discontinued sevelamer was 28% in 2003 and 19% in 2004 in the capped group and 15% in 2003 and 17% in 2004 in the noncapped group. None of these differences were significant.

Adjusted rate ratios are shown in Table 4. Adjusted analysis showed that factors influencing the PDC in 2003 included pharmacy benefit group and presence of the parathyroidectomy procedure. Adjusted rate ratios reveal that in 2003 patients with a capped benefit had 27% fewer days of sevelamer use compared with the noncapped benefit (odds ratio [OR], 0.73; 95% confidence interval [CI], 0.58–0.93). The 2004 adjusted rate ratios showed those with a capped benefit have 33% fewer days of sevelamer use compared with the noncapped benefit (OR, 0.67; 95% CI, 0.46–0.96).

Table 4.

Characteristics Associated with Medication Adherence for Sevelamer Users

	2003	2004
Variable characteristics	Adjusted rate ratios (95% CI)	Adjusted rate ratios (95% CI)
Noncapped benefit referent	0.73 (0.58–0.93)a	0.67 (0.46–0.96)a
No parathyroidectomy referent	0.50 (0.33–0.76)b	0.65 (0.37–1.17)
Continuous number of medications	0.99 (0.98–1.00)	1.00 (0.99–1.02)
Continuous age at first dispensing	1.00 (0.99–1.01)	1.00 (0.99–1.01)
Male referent	1.14 (0.94–1.39)	1.02 (0.82–1.27)
No depression referent	0.88 (0.50–1.55)	0.59 (0.21–1.67)
No atrial fibrillation referent	0.84 (0.59–1.19)	0.93 (0.62–1.40)
No hypertension referent	1.02 (0.81–1.29)	0.87 (0.66–1.14)
No CHF referent	1.14 (0.90–1.45)	NA
No CAD referent	NA	0.99 (0.79–1.25)
No metoprolol/atenolol use referent	1.05 (0.85–1.29)	1.20 (0.95–1.53)
No calcimimetics use referent	NA	0.98 (0.72–1.34)
Corrected Ca × P product ≤55 mg2/dL2 referent	1.16 (0.87–1.55)	1.01 (0.74–1.38)
No concurrent calcium-based phosphate binder referent	0.91 (0.72–1.15)	0.78 (0.58–1.05)

^aP <.05 (significant difference).

^bP <.01 (significant difference).

Ca × P indicates calcium-phosphorus product; CAD, coronary artery disease; CHF, chronic heart failure; CI, confidence interval.

None of the variables, including the type of pharmacy benefit, were associated with discontinuation of sevelamer in 2003 and 2004 (Table 5).

Table 5.

Characteristics Associated with Discontinuation of Sevelamer

	2003	2004
Variable characteristics	Adjusted odds ratios (95% CI)	Adjusted odds ratios (95% CI)
Noncapped benefit referent	0.50 (0.20–1.27)	1.31 (0.41–4.16)
No parathyroidectomy referent	1.83 (0.54–6.18)	0.48 (0.07–3.16)
Continuous number of medications	0.98 (0.92–1.04)	1.00 (0.95–1.05)
Continuous age at first dispensing	0.99 (0.96–1.02)	1.01 (0.99–1.04)
Male referent	0.99 (0.44–2.19)	1.21 (0.55–2.65)
No atrial fibrillation referent	2.36 (0.61–9.14)	0.73 (0.18–2.88)
No hypertension referent	1.13 (0.43–2.96)	1.16 (0.42–3.21)
No CHF referent	0.79 (0.29–2.18)	NA
No CAD referent	NA	0.69 (0.31–1.55)
No metoprolol or atenolol use referent	1.76 (0.75–4.10)	1.26 (0.54–2.94)
No calcimimetics use referent	NA	3.11 (1.09–8.86)
Corrected Ca × P product ≤55 mg2/dL2 referent	0.39 (0.14–1.08)	2.56 (0.90–7.24)
No concurrent calcium-based phosphate binder referent	1.63 (0.66–4.02)	1.54 (0.60–3.96)

Ca × P indicates calcium-phosphorus product; CAD, coronary artery disease; CHF, chronic heart failure; CI, confidence interval.

In 2003, there was a significant difference in the corrected serum calcium level, serum phosphorus level, and the corrected calcium-phosphorus product in patients with ESRD in the capped group compared with those in the noncapped group (Table 6). The intact PTH values did not differ significantly between the capped and noncapped groups.

Table 6.

Clinical Marker Differences between Capped/Noncapped Groups

	2003						2004
	Capped (n = 32)		Noncapped (n = 60)			Capped (n = 20)		Noncapped (n = 96)
Characteristic	Median	(5th, 95th percentile)	Median	(5th, 95th percentile)	P	Median	(5th, 95th percentile)	Median	(5th, 95th percentile)	P
Corrected calcium values	9.7	(8.5, 10.8)	9.6	(8.0, 10.7)	.022	9.7	(8.4, 10.7)	9.5	(8.2, 10.5)	.003
Serum phosphorus values	5.5	(3.3, 7.8)	5.1	(3.4, 8.1)	.015	4.8	(3.0, 7.5)	4.9	(3.2, 7.9)	.097
Corrected calcium-phosphorus product	53.5	(30.9, 74.9)	48.8	(31.9, 74.7)	<.001	45.6	(28.8, 73.7)	46.5	(30.6, 74.0)	.433
Intact PTH values	299.5	(53.5, 1603.5)	300	(90.0, 1003.0)	.089	391	(100, 2186)	421	(98, 1449)	.369

PTH indicates parathyroid hormone.

Patients with ESRD in the noncapped group had slightly lower corrected serum calcium and serum phosphorus values but much lower corrected calcium-phosphorus product values (a difference of 4.7 mg/dL). In 2004, results were similar to 2003, except that the corrected calcium-phosphorus product was not significantly different between the 2 groups.

Discussion

Although the ESRD population accounts for less than 1% of the entire US population, it continues to increase at a rate of 3% annually across all races and age-groups.36 Total Medicare spending in 2006 was nearly $355 billion; ESRD costs rose to $23 billion (6.4% of the total Medicare budget).36 Patients with ESRD have many comorbidities, and a majority of the older patients take 5 or more medications.36 The leading cause of ESRD is diabetes (44.8%), followed by hypertension (26.8%) and glomerulonephritis (12.8%).36

Drug therapy for patients with ESRD is complex, requiring many oral and injectable medications, some of which require multiple doses each day. Attempts to use behavioral models to predict medication nonadherence have suggested complex interrelationships between psychologic factors and adherence in this patient population.21 A recent review suggests health beliefs (eg, perceptions of self-efficacy with regard to taking medication), social support (eg, support of friends, family, and renal staff), family dynamics (eg, family problems caused by the patient's illness), and personality traits (eg, low conscientiousness) as significant predictors of nonadherence to phosphate-binding medications in these patients.37

Results of our analysis show that adherence to sevelamer treatment was 27% lower in the capped group in 2003 and 33% lower in 2004 compared with the noncapped group. Patients with ESRD were taking an average of 10 to 12 medications in both years. The total annual out-of-pocket (OOP) expenses for sevelamer were 26% and 57% higher for patients in the capped group compared with patients in the noncapped group in 2003 and 2004, respectively.

These rates of adherence are higher than those reported in a previous study, although that study only examined patient self-reports of cost-related nonadherence, which are typically lower than what is examined in claims data, because of recall bias issues.24 In addition, our small sample size did not allow us to examine whether the use of sevelamer was influenced by members reaching the capped portion of their prescription benefit. However, despite the generosity of the noncapped benefit, median adherence rates were only 40.4% in 2003 and 59.2% in 2004. Thus, although cost is a factor that may affect adherence, other factors not examined in this pilot study may affect adherence as well.

Our findings also suggest that lower adherence to sevelamer treatment may have been associated with less desirable biochemical outcomes. The K/DOQI guidelines recommend a serum phosphorus level of 3.5 mg/dL to 5.5 mg/dL for patients with ESRD, a corrected serum calcium level of 8.4 mg/dL to 9.5 mg/dL, and a goal for corrected calcium-phosphorus product of 55 mg²/dL².8 The corrected serum calcium values were above K/DOQI targets in the capped group in both years. The serum phosphorus values and corrected calcium-phosphorus product were at the high end of the target goal range in the capped group in 2003.

Although this pilot study was not designed to evaluate the clinical significance of the biochemical values, results from previous studies have identified significant morbidity and mortality risks associated with hyperphosphatemia and hypercalcemia.3–6,10–14 Because there are no generic alternatives to sevelamer hydrochloride, its use in this population and the impact of cost-sharing on adherence could have implications in the current healthcare environment.

The impact of sevelamer adherence on health plan resource utilization and total healthcare costs is an area for further research. Medicare+Choice beneficiaries with a capped annual drug benefit were found to have higher odds of nonadherence to antihypertensive, lipid-lowering, and diabetes medications (30%, 27%, 33%, respectively).25 Furthermore, patients with a capped benefit were more likely to have elevated systolic blood pressure (OR, 1.05; 95% CI, 1.00–1.09), LDL cholesterol (OR, 1.13; 95% CI, 1.03–1.25), and glycosylated hemoglobin (OR, 1.23; 95% CI, 1.03–1.46).25 Finally, patients with a capped benefit had 28% lower pharmacy costs, 4% lower office visit costs, but 13% higher hospital costs and 9% higher emergency department costs.25

Sokol and colleagues examined the relationship between medication adherence for diabetes, hypertension, dyslipidemia, and CHF on the risk of hospitalizations and total healthcare expenditures.38 Using diabetes as an example, they found that diabetic patients with adherence levels between 80% and 100% had a 13% risk of hospitalizations compared with 26% risk for those with adherence levels <40%. In patients with schizophrenia, 17% of adherent patients (>80% adherence) had a psychiatric hospitalization compared with 30% of nonadherent patients (<80% adherence).39 In another study of patients with type 2 diabetes, adherence remained the strongest predictor for annual healthcare costs, where a decrease of between 8.6% and 28.9% was observed for every 10% increase in adherence.40 Therefore, nonadherence has been associated with greater utilization of high-cost drivers of healthcare expenditures, such as inpatient visits and emergency department visits, which ultimately results in increased medical costs and the subsequent economic implications of medication nonadherence.

What are the implications of the lower adherence for sevelamer we found in our study on overall healthcare utilization and costs? It was beyond the scope of our study to directly examine this impact, but we conducted an estimation of the impact of lower adherence on total healthcare costs for patients with ESRD using published literature estimates for patients with diabetes who could be at risk for chronic kidney disease and ESRD.

The recent US Renal Data System (USRDS) reported an overall 50% change in new ESRD cases due to diabetes between 1996 and 2006.36 Balkrishnan and colleagues found that a 10% decrease in medication adherence for diabetes can lead to an 8.6% to 28.9% increase in total healthcare costs.40 We found that in 2003, patients with ESRD with a capped benefit had 27% lower adherence compared with those without a noncapped benefit. In 2004, those with a capped benefit had 33% lower adherence compared with those with a noncapped benefit. The USRDS 2008 annual data report for ESRD show that the per-member per-month (PMPM) cost of treating a patient with ESRD under Medicare in 2006 was $6266.36 Moreover, Medicare expenditures were approaching $72,000 per patient annually in 2006 and inpatient/outpatient costs of $4300 PMPM for patients with ESRD.36

Using the conservative estimate of 8.6%40 to extrapolate the impact of reduced adherence for sevelamer on healthcare costs, we can estimate that in 2003, patients with a capped benefit taking sevelamer potentially could have had 2.7-fold greater healthcare costs or a potential increase of $1454 PMPM in healthcare costs compared with patients with a noncapped benefit in 2003. Using the high-end estimate of 28.9%, we can estimate that patients taking sevelamer with a capped benefit could potentially experience an increase of $4889 PMPM in healthcare costs compared with patients with a noncapped benefit. Similarly, in 2004 the conservative estimate can potentially yield an increase of $1778 PMPM in total healthcare costs for patients who were taking sevelamer with a capped benefit compared with those with a noncapped benefit. The high-end estimate can potentially yield an increase of $5975 PMPM in total healthcare costs in 2004 for patients taking sevelamer with a capped benefit compared with a noncapped benefit.

What are the implications of our findings for the Medicare Part D benefit? Using the Medicare Current Beneficiary Survey data between 1997 and 2001, Patel and Davis estimate that under the current Part D benefit, patients with ESRD will have mean annual OOP drug costs that are twice those of non-ESRD Medicare patients ($2329 vs $1331, respectively).41 Approximately 54% of the patients with ESRD were taking more than 10 medications, and the majority of patients with ESRD reached the benefit gap (70% vs 43%, respectively) and the catastrophic coverage (39% vs 14%, respectively).41 Patients with ESRD approached the benefit gap and catastrophic coverage earlier (June vs July and July vs September, respectively).41 With a majority of patients with ESRD expected to reach prescription drug coverage limits by midyear and have significant OOP costs, our findings provide evidence of a negative impact of increased cost-sharing on sevelamer adherence.

Limitations

The small sample size was a limitation of this study, which may limit the generalizability of the findings, although the characteristics of the patients examined are similar to national estimates.36 We compared retirees to Medicare members in this study and were unable to measure important variables, such as income, that could affect the ability to purchase prescription medications. It is possible that the retiree population in the noncapped group had higher income levels than the Medicare beneficiaries in the capped group, which might have influenced their adherence behavior as well.

We were also not able to measure whether the impact of a potential selection bias of patients with ESRD in a Medicare+Choice plan versus retirees in an employer-sponsored plan influenced the use of sevelamer.

Conclusion

Our pilot study revealed that patients with ESRD taking sevelamer under a capped benefit for brand-name coverage had lower levels of adherence compared with those who did not face such a restriction. These results provide preliminary evidence of how ESRD beneficiaries may fare under the current Part D benefit. With escalating inpatient and outpatient costs for patients with ESRD, research on medication adherence and its impact on healthcare costs will be beneficial.

Biography

Bharati Bhardwaja