Medicare Part D Coverage Restrictions and Patient Cost-Sharing for Opioids Commonly Used for Cancer Pain, 2015-2021

PURPOSE:

Nation-wide rapid declines in prescription opioid dispensing gave rise to concerns regarding restricted access to effective pain management for patients with cancer-related pain. One important mechanism for such restrictions could be through more restrictive insurance coverage for opioids. This study aims to assess recent changes in Medicare Part D formulary designs for opioids commonly used for cancer-related pain.

METHODS:

We used data from the 2015-2021 Medicare Prescription Drug Plan (PDP) Formulary Files to assess formulary changes for six opioid-dose combinations commonly used for cancer-related pain. We estimated % of PDPs adopting prior authorization, quantity limits (and limits adopted), and a higher cost-sharing tier for each opioid-dose combination. We further estimated median and mean out-of-pocket (OOP) costs across all PDPs for a 30-day supply of the drug. Trends in proportions were tested using the Cochrane-Armitage test; trends in continuous measures were tested using the Jonckheere-Terpstra test.

RESULTS:

Proportion of PDPs adopting prior authorization increased from close to 0% to about 50% for two long-acting opioids (P < .001). Distribution of quantity limits across PDPs shifted over time to being more restrictive for all opioids considered (P < .001). For four of the six opioids, the proportion of PDPs adopting tier 3 or above increased from below or about 50% to well over 70% (P < .001). For the same four opioids, median OOP costs doubled to quadrupled (P < .001).

CONCLUSION:

Medicare PDP coverage has become increasingly restrictive for opioids commonly used for cancer-related pain, with multifold increases in patient OOP costs over the past 7 years. These changes pose concerns for patients with cancer needing opioid therapies for pain control and call for strategies to effectively exempt cancer-related pain from insurance and pharmacy rules intended to apply to opioids for noncancer chronic pain.

INTRODUCTION

Dispensing of prescription opioids in the United States declined from a peak of 81.3 in 2012 to 46.7 prescriptions per 100 persons in 2019.¹ Such rapid declines raised concerns for restricted access to effective pain management for patients with cancer. For this population, international consensus supports the use of opioids for moderate-to-severe pain associated with cancer and its treatment.² Adding to the concerns are recent studies revealing 20%-40% declines in opioid prescriptions dispensed to patients with cancer with poor prognosis and near the end of life³ or patients receiving cancer-related palliative care.^4,5

One important mechanism by which access could be restricted is through increasingly restrictive prescription formulary designs for opioids.⁶ Coverage exclusions, prior authorization, quantity limits, and higher cost-sharing tiers are common utilization management strategies shown to be associated with changes in opioid prescriptions in the United States^7,8 and Canada.^9,10 Because formulary design is not delineated along cancer versus noncancer lines, shifts toward more restrictive coverage terms for opioids may have inadvertently restricted access to opioids for patients with cancer-related chronic pain, especially in absence of deliberate policies or procedures to exempt patients with cancer pain.

Although existing literature reported rapid shifts to restrictive formulary designs for opioids by 2015,¹¹ such shifts may have accelerated in years since 2015. In this study, we examined changes in Medicare prescription drug plan (PDP) coverage restrictions and patient out-of-pocket (OOP) costs for opioids commonly used for cancer-related pain over 2015-2021.

METHODS

We used data from the Medicare PDP Formulary Files¹² pertaining to the first quarters of 2015-2021, providing detailed coverage information at the National Drug Code (NDC) level for both stand-alone and Medicare Advantage PDPs. We excluded PDPs operating exclusively in US territories. The number of PDPs included in analysis increased from 3,237 in 2015 to 5,132 in 2021.

We grouped NDCs of schedule II-V opioids¹³ by unique generic name-strength/dose combinations. The clinician researchers on our team (one oncology palliative care specialist and one anesthesiologist/pain specialist) identified six opioid-dose combinations (three short-acting, three long-acting) that are commonly used to treat cancer-related pain. We grouped short-acting oxycodone 5 mg and oxycodone-acetaminophen 5 mg/325 mg because they are used interchangeably in clinical care. To provide a context for concurrent changes in coverage and OOP costs for drugs commonly used to alleviate side effects of cancer treatment, we additionally examined coverage restrictions and OOP costs of ondansetron HCL 4 mg, an antiemetic commonly used for chemotherapy-related nausea.

For each drug-dose combination, we determined, in a given year, % of PDPs that covered the drug and, among PDPs covering the drug, % adopting prior authorization and % adopting a quantity limit and, among those with a quantity limit, the median and interquartile range of the quantity limit in terms of tablets/capsules or patches (we did not examine step therapy because no PDP in the study years adopted step therapy for the opioids considered). We additionally determined % of PDPs that placed the opioid-dose combination on a higher cost-sharing tier (tier 3 or higher, where tier 3 is typically associated with a nonpreferred generics or preferred brands designation) and estimated patient OOP cost for a 30-day supply of the opioid-dose combination. The OOP cost was based on either co-pay (with amount directly tied to a tier) or coinsurance (calculated as the product of the coinsurance rate and average retail price for a 30-day supply of the drug at pharmacies in the PDP's service area).

Typically, multiple NDCs for the same drug-dose combination were included in a formulary adopted by a PDP. Although prior authorization and quantity limit were almost always the same across the multiple NDCs, cost-sharing could differ. We assigned, to a PDP, the value representing the least restrictive coverage, for example, the lowest tier and lowest OOP cost across all NDCs. Our estimates thus reflect the lower bound of coverage restrictions faced by patients.

We derived estimates unweighted and weighted by PDP enrollment¹⁴ in January of each year. We used the Cochrane-Armitage test¹⁵ (for trends in proportions; two-tailed) and Jonckheere-Terpstra test¹⁶ (for trends in the distribution of quantity limits and OOP costs; one-tailed) to evaluate changes in these measures over time.

RESULTS

With the exception of long-acting oxycodone, the opioid-dose combinations considered were covered by 100% or close to 100% of PDPs throughout the years (Table 1). Long-acting oxycodone 20 mg was covered by about one-third of all PDPs in 2015-2017. This proportion increased to 50%-60% in 2018-2021 during which time Xtampza ER—the abuse-deterrent formulation of long-acting oxycodone—was increasingly added to formularies.

TABLE 1.

Medicare PDP Coverage, Coverage Restrictions, Cost-Sharing Tiers, and OOP Costs for Opioids Commonly Used for Cancer-Related Pain, 2015-2021

Proportion of Medicare PDPs adopting a prior authorization requirement increased dramatically for two of the three long-acting opioids considered (Fig 1): For long-acting morphine 30 mg, it increased from 0% in 2015 to 9.8% in 2018 and to 48.8% in 2021 (P < .001); for fentanyl patches 25 μg/h, it increased from 7.7% in 2015 to 15.6% in 2018 and to 52.0% in 2021 (P < .001). For long-acting oxycodone 20 mg, of the 30%-60% of PDPs covering the drug, close to one quarter adopted a prior authorization requirement in 2015 and about one in five adopted the requirement in 2021.

FIG 1.

Medicare PDP adoption of a prior authorization requirement for six opioids. PDP, Prescription Drug Plan.

A very high proportion of PDPs (> 90%) adopted a quantity limit for all six opioids throughout the years (Table 1). The median quantity limit for hydrocodone-acetaminophen 10 mg/325 mg was reduced from 360 tablets in 2015-2018 to 180-240 tablets (with consistent reductions at the 25th and 75th percentiles) in 2019-2021 (P < .001). For all other opioid-dose combinations (except long-acting oxycodone 20 mg), despite stable median quantity limits, the distribution of quantity limits across PDPs shifted overtime to being more restrictive over the years (P < .001).

We found that an increasing proportion of PDPs placed the opioid-dose combinations on tier 3 or 4 (P < .001 for all six opioids; Fig 2A; tier 5 is typically used for specialty drugs; no PDP in our data adopted a tier 5 or above for the six opioids considered). For example, for hydrocodone-acetaminophen 10 mg/325 mg, short-acting morphine 15 mg, and long-acting morphine 30 mg, this proportion increased from below 50% in 2015 to over 70% in 2021. For fentanyl patches 25 μg/h, it increased from 58.5% to 76.6%. Of note, for all opioid-dose combinations except long-acting oxycodone, of the PDPs that put a given opioid-dose combination on tier 3 or above, 100% or close to 100% of the PDPs had a generic opioid on the lowest tier (ie, tier 3 or above, Appendix Table A1, online only). For long-acting oxycodone 20 mg, this proportion was 0% in 2015, about one-third in 2016-2018, and ranged from 11% to 13% in 2019-2021 because there was no generic medication in 2015, and since 2019, many PDPs covered the abuse-deterrent Xtampza ER (a branded drug) only.

FIG 2.

Medicare PDP (A) adoption of tier 3 or above and (B) estimated OOP cost for six opioids. OOP, out-of-pocket; PDP, Prescription Drug Plan; USD, US dollars.

For four opioid-dose combinations, median OOP costs across PDPs for a 30-day supply increased dramatically over the past 7 years (Fig 2B). The median OOP costs more than quadrupled for short-acting morphine 15 mg from $10 US dollars (USD) in 2015 to $42 USD in 2021, more than tripled for hydrocodone-acetaminophen 10 mg/325 mg (from $12 to $40 USD) and fentanyl patches 25 μg/h (from $13 to $37 USD), and more than doubled for long-acting morphine 30 mg (from $17 to $42 USD). Long-acting oxycodone 20 mg, dominated by OxyContin in 2015-2016 but increasingly represented by the abuse-deterrent agent Xtampza ER in 2017-2021, had the highest OOP costs for a 30-day supply (over $45 USD in most years) among all opioids considered, albeit limited changes in OOP costs over the 7 years. Mean OOP costs across PDPs for the six opioids changed in a similar pattern as median OOP costs (Table 1).

In contrast, for ondansetron HCL 4 mg, the proportion of PDPs adopting tier 3 or 4 increased slightly from 13.0% in 2015 to 16.2% in 2021; median OOP costs for a month's supply of ondansetron HCL 4 mg remained around $8-$10 USD over the years (Table 1). All other features of the formulary design remained stable or changed to be less restrictive over the years (eg, a decrease in % of PDPs adopting prior authorization and quantity limit).

The results weighted by PDP enrollment (Appendix Table A2, online only) presented similar patterns, as seen in unweighted results.

DISCUSSION

We observed multifold increases in Medicare PDP adoption of prior authorization requirements, high cost-sharing tiers, and median OOP costs over 2015-2021 for several opioids commonly used for cancer-related pain. Such changes seem to have targeted opioids, but not other drugs used to alleviate side effects of cancer treatment such as ondansetron. Coverage restrictions and OOP costs apply to all patients enrolled in a PDP, regardless of clinical indication or risk-benefit considerations associated with drug therapies. The extraordinary increases in coverage restrictions and OOP costs seen likely have limited access to opioids for cancer-related pain.

The rapid increase in PDPs' adoption of prior authorization requirements for long-acting opioids could be related to a larger prescription opioid safety initiative implemented by the Centers for Medicare and Medicaid Services in 2019 as part of the Comprehensive Addiction and Recovery Act of 2016.¹⁷ A previous study found that prior authorization for long-acting opioids in a Medicaid program was associated with a 70% reduction in initiation of long-acting opioids by opioid-naive patients and a 33% reduction in the likelihood of having a new prescription irrespective of prior opioid prescriptions.¹⁸ Two additional studies found strict prior authorization policies targeting high-dose opioids or long-acting oxycodone substantially reduced prescriptions of targeted opioids by Medicaid patients.^19,20 The increases seen in this study thus pose concerns for inadequate pain control for patients with cancer needing sustained pain relief. In addition to limiting new prescriptions, increasing use of prior authorizations could also interfere with ongoing chronic opioid therapy by requiring frequent renewals of these authorizations, adding to provider burden and potentially leading to delayed access and opioid withdrawal, although our data do not contain information on specific rules of prior authorization.

The emergence of many novel and expensive anticancer interventions has drawn attention to the growing financial burden faced by patients and the concerning reduction in treatment initiation and continuation associated with high OOP costs.²¹ Although pain medications are not typically considered significant cost centers for cancer treatment, our findings suggest substantially higher financial burdens associated with opioid therapies in recent years. Our data showed that generic opioids were increasingly placed on tier 3 or higher, suggesting increasing designation as nonpreferred generics and therefore higher cost-sharing. Our estimates of OOP costs suggest that, in 2021, patients with cancer requiring round-the-clock opioid therapies as well as short-acting opioids for breakthrough pain could expect to pay $100 USD out of pocket monthly just for opioid medications. Previous studies found that patients with a history of cancer were more likely to report changing their prescription drug use (eg, not filling or skipping) for financial reasons than individuals without a history of cancer,²² reflecting heightened susceptibility of patients with cancer to cost-related nonadherence. The dramatic increases in OOP costs for opioid therapies seen in Medicare PDPs, although largely not directed at patients with cancer-related pain, were likely to have disproportionately affected access by patients with cancer.

Our findings suggest immediate strategies pertaining to PDP formulary designs and pharmacy dispensing systems to ensure adequate access for patients prescribed opioids for cancer-related pain. First, prior authorization should be exempted based on indication of cancer-related pain, determined by either specialty of the prescriber (eg, hematology/oncology, palliative care) or an explicit indication by the prescriber in the script. This could be operationalized by automated overriding during the pharmacy adjudication process. In addition, prior authorization should be subject to less frequent renewal requirements for cancer-related than for non–cancer-related pain. Second, plans should develop or modify policies to lower OOP costs of opioids for patients with cancer-related pain. Although technically possible, it is not a common practice for PDPs to assign a different/lower tier to a medication prescribed for certain conditions, which may require strong advocacy to bring to bear. Our findings do suggest that plans could assign generic opioids commonly prescribed for cancer-related pain to lower, preferred generic tiers to better ensure access to opioid therapies for cancer-related pain.

Our analysis has several limitations. Tiering and our estimated OOP costs apply to the postdeductible, precoverage gap phase of Medicare Part D coverage, without low-income subsidy. Whether tiering matters and the actual OOP costs faced by patients may depend on the phase of coverage they are in and by low-income subsidy status. In addition, tiering is not standardized over time or across PDPs (eg, a tier 3 may mean preferred brands for one PDP while nonpreferred generics for another). However, the proportion of PDPs adopting a 4+ tier formulary was close to 100% (ranging from 94% to 97%) over the years (data not shown), suggesting that nonspecialty tier designation (preferred generics, nonpreferred generics, preferred brands, nonpreferred brands) was stable. Although the estimated OOP cost may be a more informative measure of cost-sharing than the measure of tier 3 or above, we retained the measure of tier 3 or above because policy implications are directly tied with tiering. We selected six opioids that are commonly used for cancer-related chronic and acute pain conditions. Our results may not reflect the totality of coverage policies for all opioids used for cancer-related pain. The six opioids selected, however, represent essential therapies for cancer pain³ and provide the stability to evaluate formulary changes over time. Although we characterized changes in adoption of prior authorization requirements over time, specifics of those policies (eg, how frequently prior authorization needs to be renewed) are not included in the data. For opioid-dose combinations with multiple NDCs covered by a PDP, because we assigned the least restrictive parameter to a PDP, our estimates reflected the lower bound of coverage restrictions/OOP costs faced by patients.

Management of pain among patients with cancer is clinically nuanced, calling for careful assessment of patient needs for effective pain control and risks associated with opioids and other controlled substances. This is especially true among patients with risk factors for nonmedical use of opioids, including existing mental health and/or substance use disorders.²³ Although the focus of this study is on barriers to appropriately prescribed opioids for cancer-related pain, it does not undermine the importance of clinical management that balances and mitigates the risks associated with opioid use along with access to resources such as multimodal pain therapies and mental health counseling.²⁴

Medicare PDP coverage restrictions and OOP costs for opioids increased rapidly in recent years. The extent to which such increases may have inadvertently restricted access to effective pain control by patients with cancer-related pain needs to be closely monitored. Meanwhile, incremental changes should be implemented in formulary designs and pharmacy dispensing systems to effectively exempt cancer-related pain from restrictive coverage rules and high cost-sharing requirements. Future studies should examine the extent of actual implementation of exemption practices if and when such data are available.

APPENDIX

TABLE A1.

% of PDPs That Put an Opioid-Dose Combination on Tier 3 or Above (if covered) and, of PDPs Assigning Tier 3 or Above, % That Covered a Generic on the Lowest Tier

TABLE A2.

Medicare PDP Coverage, Coverage Restrictions, Cost-Sharing Tiers, and OOP Costs for Opioids Commonly Used for Cancer-Related Pain, 2015-2021, Weighted by PDP Enrollment in January of Each Year

AUTHOR CONTRIBUTIONS

Conception and design: Yuhua Bao, Lisa R. Witkin, Judith A. Paice

Provision of study materials or patients: Yuhua Bao, Daniel M. Hartung

Collection and assembly of data: Yuhua Bao, Hao Zhang, Daniel M. Hartung

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

Medicare Part D Coverage Restrictions and Patient Cost-Sharing for Opioids Commonly Used for Cancer Pain, 2015-2021

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. 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/op/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).