Use of Charity Financial Assistance for Novel Oral Anticancer Agents

Abstract

Novel oral targeted drugs are increasingly used for cancer therapy, but their extreme cost, often exceeding $10,000 per month, poses a significant barrier for patients and insurers alike, leading to the potential breakdown of traditional cost-sharing strategies. Insured patients’ routine use of charity assistance to supplement their coverage would indicate a major deficiency in the current health care policies. By using data from a specialty pharmacy affiliated with an academic center (1,557 prescriptions dispensed between January 2014 and March 2017), we examined sources of payment for novel oral anticancer agents, distinguishing contributions from health insurance, patients, and from charitable assistance organizations. Thirty-six percent of 211 patients received charity assistance, including 47% of patients who were 65 years old or older. Charity sources covered 4% of total drug costs and 64% of out-of-pocket expenditures. The proportion of patients receiving financial assistance ranged from 7% when the upfront out-of-pocket requirement was less than $100 to 67% when it exceeded $1,000. When patients’ out-of-pocket requirement exceeded $1,000, the median direct cash contribution paradoxically fell to $0 because of extensive use of charity support. Receipt of upfront charity assistance was associated with a longer time to filling the first prescription (median 9 v 7 days; P = .011) and with longer overall duration of therapy (median, 261 v 134 days; P = .014). These findings indicate that high out-of-pocket burden for expensive novel oral anticancer drugs leads to widespread use of charity support in the United States and that a significant financial barrier disparately affects older Medicare beneficiaries.

INTRODUCTION

Financial distress caused by the extreme cost of novel anticancer agents has been widely decried by clinicians, policy experts, and patient advocates.^1-3 Routinely priced at about $10,000 or more per month of therapy, these drugs (which include tyrosine kinase inhibitors, immunomodulatory drugs, and other targeted agents) often require prior authorizations, tiered copayments, or high coinsurance set by health insurance plans as cost-containment measures. Oral antineoplastic agents in the United States are typically covered under prescription benefits, which differ from coverage policies for injectable drugs administered in the office or hospital setting. In the case of Medicare Part D plans, which provide prescription coverage for a majority of older individuals in the United States, this involves a complex system of patient cost sharing, including a coverage gap translating into thousands of dollars in out-of-pocket expenses every year.^4,5 The variation in payments for parenteral and oral anticancer treatments contributes to disparities in use and adherence, as well as significant financial toxicity—a subject of active research, advocacy, and legislative efforts aimed at achieving equitable access to all treatment options.^6-8

Despite challenges, the clinical role of novel oral targeted agents in oncology has been steadily increasing.⁹ These agents now constitute essential standard of care for diseases such as chronic myelogenous leukemia; lung cancer with EGFR, ALK, or ROS1 mutations; renal cell carcinoma; or chronic lymphocytic leukemia with deletion of chromosome 17p. Many charity organizations (patient assistance programs [PAPs]), often partly funded by drug manufacturers, provide support that alleviates patients’ out-of-pocket expenses. Previous studies indicated that approximately 10% to 15% of patients who received any cancer therapy used PAPs to offset out-of-pocket costs, but the extent of this support for the novel ultra-expensive anticancer agents has not been investigated.^10-12 Studies that directly assessed the use of PAPs concluded that they cover only a small percentage of total cancer drug costs, but they can facilitate access for underinsured individuals.^12-14 Our objective was to examine sources of payment for patients who have been prescribed novel oral anticancer agents by using data from an oncology specialty pharmacy that is linked to patients’ clinical medical records and is affiliated with an academic cancer center.

PATIENTS AND METHODS

This retrospective cohort study was approved by the Institutional Review Board at Rhode Island Hospital. The setting of the study was Lifespan Cancer Institute, the largest program that provids cancer care to the residents of Rhode Island. It encompasses an academic research center (Rhode Island Hospital) as well as several community centers in affiliated regional hospitals and stand-alone facilities. Lifespan Pharmacy is an affiliated specialty pharmacy that fills all prescriptions for oral anticancer drugs for patients treated within the program, except when the use of an outside specialty pharmacy is required by an insurance plan or drug manufacturer. The pharmacy initially processes all prescriptions for oral chemotherapy generated by prescribers in the cancer program and offers an integrated continuum of care practice model with direct access to providers, electronic medical record notes with outlined plans for chemotherapy and monitoring, and interaction checking with a complete medication profile. Financial assistance is provided with the help of dedicated pharmacy staff and social workers in the oncology office who reach out to both industry-sponsored PAPs and charitable organizations when patients in need are identified.

The list of novel oral anticancer drugs of interest was compiled by a clinical expert (A.J.O.), and it comprised all approved tyrosine kinase inhibitors, immunomodulatory agents, novel antiandrogens, proteasome inhibitors, histone deacetylase inhibitors, and other targeted agents (Appendix Table A1, online only). By using pharmacy records, we classified payment contributions for each filled prescription as originating from the insurance plan (primary or supplemental), patient (cash), or a charity assistance organization. Patients’ primary insurance was classified as commercial/employer-provided, Medicare (Part D plans available for patients older than age 65 years or those with a disability), or Medicaid (state-sponsored plans available to patients who meet low-income criteria). Assistance provided by point-of-sale copayment cards or coupons was not identified because it was indistinguishable from out-of-pocket contributions in pharmacy records. Such discounts, offered directly by drug manufacturers to commercially insured patients only, and largely independent of income, were thus included with patients’ other out-of-pocket contributions. Sex and self-reported race/ethnicity were obtained from linked electronic medical records. Time from the initial prescription to start of therapy was calculated from the day the provider sent the prescription to the pharmacy to the day of drug dispensation. Observed duration of therapy was defined as the time from the first drug dispensation to the last day covered by the prescriptions dispensed. We compared categorical variables by two-sided Fisher’s exact test, ordinal variables by χ² test for trend, and continuous variables by Somers’ D statistic.¹⁵ Because of contractual confidentiality, we replaced the exact dollar amounts for pharmacy charges with average wholesale prices from a national database,¹⁶ but all proportional contributions (percentages) were based on actual payments.

RESULTS

The data encompassed 1,557 prescriptions dispensed for 211 patients between January 2014 and March 2017, capturing an estimated 68% of all prescriptions issued for the drugs of interest in our cancer program. Median patient age was 70 years (range, 7 to 94 years), and 39% were women. The distribution of race/ethnicity was as follows: white non-Hispanic, 83%; white Hispanic, 12%; black, 3%; and Asian, 2%. Most patients were insured by Medicare Part D (66%), followed by commercial/employer coverage (17%), and Medicaid (14%); 3% had no insurance. The most frequently dispensed drugs included ibrutinib (19%), enzalutamide (15%), abiraterone (13%), dasatinib (11%), erlotinib (9%), and imatinib (5%), at a median average wholesale price of $11,275 (interquartile range [IQR], $9,061 to $13,324). Median time from issuing the first prescription to dispensation was 8 days (IQR, 3 to 15 days), median number of prescriptions was five (IQR, two to 11), and median observed duration of therapy was 151 days (IQR, 54 to 347 days).

Altogether, 36% of patients received charity financial assistance, and 32% received it at the time of the initial prescription. Charity sources covered 4% of all drug costs, whereas insurance contributed 94% and patients 2%. Consequently, charity covered 64% of all required out-of-pocket expenses. Receipt of assistance was strongly associated with age (median, 76 years for recipients and 66 for nonrecipients; P < .001), with 47% of patients age 65 years or older obtaining assistance (Fig 1A). There was no difference by sex (P = .24). Charitable support was more frequent among white non-Hispanic patients than other groups (39% v 13%; P = .004). Patients who were white non-Hispanic were less likely to have Medicaid as primary insurance (10% compared with 39% for other groups; P < .001). Assistance was most frequent (47%) among patients covered by Medicare (Fig 1B). Eighty-eight percent of all recipients of charitable support were Medicare beneficiaries, and 79% had out-of-pocket responsibility for the first prescription exceeding $1,000 (median, $3,469; IQR, $2,225 to $3,804). Among patients with commercial insurance or Medicare, upfront use of charity assistance was (nonsignificantly) more common in the first quarter, possibly as a result of the higher likelihood of facing deductibles or the Medicare coverage gap, or possibly as a result of depletion of available philanthropic funds later in the year (Fig 1C). Age, Medicare coverage, and high out-of-pocket requirements (related to Part D coverage gap) were all highly correlated, precluding separation of their associations in a multivariable model. Charity assistance for the initial prescription was associated with a longer time to drug dispensation (median, 9 v 7 days; P = .011) and longer observed duration of therapy (median, 261 v 134 days; P = .014).

Fig 1.

Percentage of patients receiving charity financial support for novel oral anticancer agents, stratified by (A) age group,(B) primary insurance, (C) month of starting therapy (first prescription only, patients with Medicare and commercial insurance only), or (D) out-of-pocket financial responsibility for the first prescription. Whisker bars indicate binomial 95% CIs; two-sided P values are derived from χ² test for trend (A, C, D) or Fisher’s exact test (B).

The largest differences in the use of charity were evident when patients were stratified by the amount of out-of-pocket requirement for the first prescription (Fig 1D; Table 1). Among patients facing out-of-pocket expenses of $100 to $1,000, 55% used assistance, and this proportion rose to 67% when cost sharing exceeded $1,000. Most Hispanic, black, or Asian patients had low or no out-of-pocket requirements, thanks to primary or secondary coverage from Medicaid, which may explain their lower use of assistance. Paradoxically, when the required cost sharing exceeded $1,000, median actual cash contribution by patients decreased to $0 because charity sources completely covered most out-of-pocket expenses. Median initial out-of-pocket requirement was $2,751 (IQR, $9 to $3,593) for Medicare, $28 (IQR, $0 to $54) for commercial insurance, and $0 (IQR, $0 to $0) for Medicaid.

Table 1.

Sources of Payment for Novel Oral Anticancer Agents, Stratified by Out-of-Pocket Requirement for the First Prescription, January 2014 to March 2017

DISCUSSION

In this retrospective study on the basis of records from an integrated regional cancer program that encompassed academic and community sites, we investigated the prevalent use of charity assistance during treatment with novel oral anticancer drugs. We found that more than a third of patients and nearly half of those age 65 years or older received charity funds to help offset the out-of-pocket costs. Moreover, the use of assistance was strongly tied to insurance-imposed cost sharing, and Medicare plans were associated with the highest financial burden for patients and the highest use of charitable sources.

Our results illustrate several phenomena caused by the interlinking effects of cancer drug prices, widespread application of high out-of-pocket requirements, and the complex operation of PAPs that channel donations from drug manufacturers to facilitate use of expensive medications, thus effectively creating an indirect kickback mechanism.^10,17 In an earlier study by Zullig et al,¹² who investigated the use of PAPs for all anticancer medications dispensed by an academic oncology pharmacy, the median drug price was $2,762, and 12% of prescriptions were paid for with financial assistance. Similarly, Mitchell et al¹¹ reported on the use of PAP assistance among 10.6% of patients receiving any oral anticancer agents in a similar setting. In that study, 40% of recipients of charitable support were uninsured, and only 20% had Medicare coverage. In contrast, our study revealed that more than 35% of patients treated with novel targeted agents received charitable support, and that such assistance paid for nearly two thirds of out-of-pocket expenses for this class of drugs—an estimate that notably did not include direct discounts offered by drug manufacturers for commercially insured patients. Assistance needs are disparately higher among older Medicare beneficiaries, nearly half of whom (and a majority of those age 75 years or older) used charity programs. This confirms prior observations that Medicare beneficiaries who are age 75 years old or older are particularly sensitive to financial toxicity imposed by the Part D coverage gap.^5,8 According to the pharmaco-economic analysis by Dusetzina and Keating,⁴ gradual elimination of the gap according to provisions of the Affordable Care Act will not resolve this disparity because of residual cost-sharing requirements.

From the payer’s perspective, we have observed that increasing a patient’s initial financial responsibility beyond $1,000 creates a severe access barrier, forcing most to request charity assistance and paradoxically decreasing actual patient contribution to the median of zero. Earlier research also showed that increasing out-of-pocket requirements affects adherence, with a continuous relationship between patients’ financial burden and outcomes such as delays in treatment, consistency of refills, or treatment discontinuation.^5,18,19 The lower median out-of-pocket requirement for commercially insured patients in our data may result from the prevalent use of fixed copayments rather than percentage-based coinsurance in this group. However, we cannot exclude potential lack of generalizability resulting from differential use of alternative specialty pharmacies by plans with higher out-of-pocket requirements. It is unclear whether the association between receipt of charity funds and longer duration of therapy in our study is causal because we could not discern reasons for discontinuation (including a switch to a different pharmacy) or adjust for clinical covariates such as treatment toxicity and loss of efficacy.

From the point of view of drug manufacturers, supporting the system of drug-agnostic charity programs ultimately generates profits, because more than 90% of payments are still provided by the insurance plans.²⁰ Extensive use of PAPs increases the actual costs of therapy by driving inflation or by shifting the policy and advocacy focus away from the escalating drug prices.²¹ We note that our study could not characterize the most common type of assistance among patients with commercial/employer insurance: point-of-sale copayment cards provided by drug manufacturers. Such cards and coupons directly undermine the intended effect of out-of-pocket requirements, and their use is prohibited for beneficiaries of federal Medicare and Medicaid programs.^20-22 However, some charity organizations effectively serve as one additional intermediary in the same mechanism, as evidenced by instances of their alleged financing by the pharmaceutical industry.^23,24

Limitations of our analysis include potential missing data related to the use of outside specialty pharmacies, mandated by some insurance plans or by certain drug manufacturers. These restrictions may explain the lack of several specific drugs (eg, lenalidomide) in our data set and may distort the proportions of patients covered by different types of health insurance. The data set contained only a few uninsured individuals because they may receive drugs directly from manufacturers through free drug programs that deliver drugs directly to the consumer, bypassing the pharmacy. Furthermore, we lacked information about patients’ socioeconomic status or about applications for charity assistance that were rejected. We also note that the prevalence of charitable support use may differ geographically because of variable population and insurance plan characteristics. For example, in contrast to our analysis, one study from Texas found that indigent and younger patients used PAPs more frequently, partly because of restrictive local Medicaid policies.¹⁴

In conclusion, our study provides empirical evidence of a widespread use of charity funds to access novel oral anticancer agents in the United States, particularly among older patients insured by Medicare. The fact that charity assistance has become an indispensable component of cancer treatment exposes inadequacy of current coverage policies. These in turn reflect the inability of health care systems to effectively address the exorbitant cost of novel anticancer agents—an area that requires urgent attention in view of the rapidly evolving landscape of oncologic therapies. Because access to the charity funds by itself is unequal, depending on patients’ assertiveness, external support from social workers (which may vary by setting of care), and even on the calendar month, the current system is prone to creating disparities and undermines the principles of equity in health care.

Appendix

Table A1.

Novel Oral Anticancer Drugs Included in the Study

AUTHOR CONTRIBUTIONS

Conception and design: Adam J. Olszewski, Andrew R. Zullo, Justin P. Huynh

Collection and assembly of data: Andrew R. Zullo, Christopher R. Nering,

Data analysis and interpretation: All authors

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Use of Charity Financial Assistance for Novel Oral Anticancer Agents

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jop/site/ifc/journal-policies.html.

Adam J. Olszewski

Consulting or Advisory Role: Spectrum Pharmaceuticals

Research Funding: Spectrum Pharmaceuticals (Inst), Genentech/Roche (Inst), and TG Therapeutics (Inst)

Andrew R. Zullo

No relationship to disclose

Christopher R. Nering

No relationship to disclose

Justin P. Huynh

No relationship to disclose