The costs and payments for discarded medications from single-dose vials and how that cost and payment should be estimated have been a subject of much debate in recent years. In the 2019 Part B Discarded Drug Units Report, the Centers for Medicare & Medicaid Services (CMS) reported that in 2018 it had paid $725 million (or 2.1 percent of the total Medicare allowed amount for Part B drugs) for discarded medications administered in hospital outpatient clinics and physicians’ offices (CMS, 2019). CMS obtained that estimate by aggregating all Part B drug claims with the JW modifier associated with the claims. This calculation assumed that providers consistently use the mandatory JW modifier, which has not been verified empirically.1 Another analysis in 2016 estimated the money spent on discarded drugs in single-dose vials of cancer medications to be $1.8 billion (Bach et al., 2016).2 Furthermore, the analysis suggested that the amount of money spent on discarded medications by public and private health care payers swelled to $2.8 billion when physician and hospital markups were included in the estimate (Bach et al., 2016).
In addition to its formal Statement of Task, CMS charged this committee to use “available data sources to quantify the amount of waste associated with single-dose injectable drugs and biologics in billing units and/or proportion of available vial sizes and calculate the associated dollar amounts.” Both federal and private entities have attempted to
2 This study was conducted before the policy mandating the JW modifier.
assign a dollar value to these unused portions. To more fully understand the extent of health care costs—to the Medicare program, private health care payers, and patients—that arise from billing for discarded drugs and biologicals from single-dose vials, the committee commissioned two analyses of these costs. The researchers who conducted the analyses had a data-use agreement with CMS and IBM MarketScan that allowed them to conduct this work using the 2017 and 2018 Medicare and MarketScan databases.3 The focus of these analyses are not to assess the economic value of the discarded drug but rather to get a sense of the amount of drugs that are being discarded and to identify any inefficiencies in the current system of tracking the amount of discarded drugs yearly.
A committee subgroup was tasked with working directly with the researchers and guiding the analyses. The researchers made presentations to the full committee during closed deliberations and received guidance from the full committee, and the committee validated the findings from the analyses, which are summarized below. Further details from the commissioned analyses are provided in Appendix B.
The first commissioned analysis examined Medicare claims that used the JW modifier code,5 examining how much of total Medicare spending they represent; examined patterns in the claims using the JW modifier code and whether those claims accurately reflected the amount spent each year by Medicare on discarded drugs; and explored compliance with the JW modifier. The analysis was performed using a 20 percent random sample
3 The Medicare and MarketScan data are publicly available to any researcher, but a licensing fee and a data-use agreement with CMS are required to obtain the Medicare data, and a license fee and agreement are also required to obtain MarketScan data from IBM Watson Health. The issue of discarded drugs is not unique to Medicare and Medicare beneficiaries. Some private health insurance companies tend to follow CMS policy on discarded drugs as well, which is why the committee commissioned analyses that used both Medicare and MarketScan data.
4 This section draws on analyses commissioned by the Committee on Implications of Discarded Weight-Based Drugs on “The extent of health care costs—to the Medicare program, to private health care payers, and to patients—that arise from billing for discarded drugs and biologicals from single-dose vials” by Aditi P. Sen, Ph.D. (assistant professor, Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health) and Hui Zhao, Ph.D. (associate professor, Department of Health Services Research, Division of Cancer Prevention and Population Sciences, The University of Texas MD Anderson Center).
5 As described in Chapter 2, the JW modifier is the mechanism through which health care providers can bill and be reimbursed for the discarded portion of a single-dose vial in a way that formally documents the amount of drug that was discarded. The modifier can be used for any single-dose vial in which some drug is discarded, regardless of whether the drug is administered in a fixed dose or a dose based on weight.
of traditional fee-for-service Medicare outpatient and carrier claims from 2017 to 2018, with a focus on discarded drugs and biologicals. Claims for which Medicare was not the primary payer were excluded. The researchers specifically included the 77 drugs with the highest total allowed amounts6 in the 2018 Medicare Part B Discarded Drug Units Report (CMS, 2019). The discarded drugs and the associated spending were identified based on claim lines with the JW modifier code. According to that CMS report, the total Part B spending in 2018 on all drugs with discarded amounts was $34.8 billion; the report estimated that $725 million was spent on the discarded portions of these drugs. The 77 drugs in the committee’s commissioned analysis accounted for 83 percent of the total spending and 92 percent of the spending on the discarded portions. Of the 77 drugs, 34 are administered with weight-based dosing, 12 use body surface area (BSA)-based dosing, and 31 are given in fixed doses.7 (See Box 3-1 for a brief explanation of why certain fixed-dose drugs in single-dose vials could have portions discarded.) As discussed in previous chapters, in this report, weight-based drugs refers to drugs that are dosed either by weight or by BSA. Therefore, for the purposes of these analyses, weight- and BSA-based drugs are both described as weight based.
Using the data from the Medicare claims, two types of analyses were performed. One was a general analysis of the use of the JW modifier for the 77 drugs, considered jointly and individually. The other looked at variations among physicians in their use of the JW modifier. The findings were then used to provide an estimate of the total nominal cost of discarded drugs—that is, the amount of dollars represented by the discarded drugs under the assumption that the discarded drugs could be valued at the same per-unit cost as unused drugs in single-use vials—and that number was compared with the total cost indicated by the JW modifier.
Of the claims involving the 77 drugs with the greatest total Medicare payouts for discarded amounts, a surprisingly small percentage actually
6 Total Medicare allowed amount for a given Part B drug includes the amount Medicare pays, the deductible and coinsurance amounts that the beneficiary is responsible for, and any amounts that a third party is responsible for.
7 The classification of the dosing regimen of the drugs for our analysis was abstracted by reviewing the package insert provided by FDA for each of the 77 drugs in the Medicare Part B Discarded Drug Units report included in the analysis. The insert includes prescription drug labeling resources, which lists detailed information on dosage and administration. If the dosage is based on patients’ weight or BSA, then we define the drug as weight based; otherwise, it is classified as fixed dose. The 31 drugs classified as fixed dose had the JW modifier associated with them.
included the JW modifier (i.e., differentiated the discarded amount from the administered amount when submitting their claims). In 2017, 16.2 percent of the claims had the JW modifier, and in 2018, 16.9 percent did. This finding suggests low compliance8 with the CMS mandate. The percentage of claims with the JW modifier was substantially higher among weight-based (26.9 percent in 2018) than fixed-dose (1.4 percent) drugs (see Table 3-1).9 JW modifier use was relatively concentrated, with 21 of the 77 drugs having over 25 percent of claims with the JW modifier. The drugs with the highest percentage of claims with the JW modifier fell into the following classes: oncology (30 drugs, with an average of 18 percent of claims having the JW modifier), rheumatology (12 drugs, 16 percent), and respiratory (6 drugs, 5 percent).
Looking only at the weight-based drugs in the Medicare sample, the researchers found that the use of the JW modifier was not prevalent. The
8 The level of compliance is determined by whether the JW modifier was used at all and not related to the specific amount of discarded drug recorded.
9 MarketScan analysis will be discussed later in this chapter.
TABLE 3-1 Percentage of Claims with the JW Modifier by Drug Dosage Category in Medicare and MarketScan Databases, 2017 and 2018
|Percentage of claims with the JW modifier, 2017
|Percentage of claims with the JW modifier, 2018
committee recognizes that in some cases, the JW modifier would not be reported because the entire contents of the vial were used in the treatment of a single patient, or there was no discarded drug because the vial may have been shared between patients in accordance with United States Pharmacopeia guidelines. However, approximately half of the drugs had the JW modifier coded in fewer than 10 percent of the claims; in 2018 the JW modifier was used in more than 10 percent of the claims for only 24 of the 46 drugs (see Figure 3-1, blue bars).
The analysis also examined the role of provider volume in terms of using the JW modifier. The rates of use were similar between health care providers with above and below the median number of claims for a particular drug and also between health care providers in the bottom 25th and top 25th percentiles of claims for that drug.13
The only consistent pattern identified across drugs was higher use of the JW modifier in physician office settings relative to hospital outpatient departments.14 Across the 40 drugs in the sample with at least 1 percent of claims with the JW modifier, a higher percentage of physician office settings used it for a given drug than hospital outpatient departments did. A number of factors may be driving this pattern, including billing practices in offices versus outpatient departments and differences in payment systems.
Given that use of the mandated JW modifier is a provider-level behavior, the committee explored whether noncompliance might be concentrated among a subset of providers and existing estimates of the amount of discarded drugs might thus be underestimates due to noncompliance. The commissioned analysis does demonstrate that a substantial proportion of physicians who prescribe drugs in the sample never use the JW modifier. In the 20 percent Medicare fee-for-service sample, a total of 163,646 health care providers had at least one claim for 1 of the 77 drugs. To look for a pattern of noncompliance, analyses were limited to health care providers with more than 10 claims, and among the 42,034 health care providers who had made at least 10 claims for at least one of the 77 drugs, 28,457 (67.7 percent) never used the JW modifier at all (see Figure 3-2).
Patient–Drug Pairs Level of Analysis
Not only is there significant variation among physicians in the use of the JW modifier, but many individual physicians are themselves not consistent about it, even when filing claims for the same patient on the same drug. It is reasonable to assume that if a certain amount of a particular drug for a particular patient is marked with the JW modifier on one occasion, the modifier is likely appropriate in all instances of that patient receiving that drug within a relatively short period, since the dosage is based on weight or body size. Yet, an analysis of the five drugs with the greatest Medicare spending on claims with the JW modifier (Table 3-2, left panel)15 found that the modifier was used inconsistently in a significant number of cases where a patient was given the same drug on multiple
15 MarketScan data on the right panel will be discussed later in this chapter.
TABLE 3-2 Health Care Provider Use of the JW Modifier for a Patient–Drug Combination in Medicare and MarketScan Data, 2018
|Medicare JW Modifier Compliance (%)
|MarketScan JW Modifier Compliance (%)
NOTES: Top five drugs in terms of spending on claims with the JW modifier. Patient–drug combinations with at least two observations. Trastuzumab use refers to reference biologic only.
occasions. For example, in 23 percent of patient–drug combinations with bortezomib, the JW modifier was used in some but not all cases, while the percentage of inconsistent usage was 44 percent for romiplostim.
Similarly, the analysis found substantial variation in how likely health care providers were to use the JW modifier from one drug to another. This part of the analysis was restricted to the 25 drugs in the sample with at least 10 percent claims with the JW modifier and was also restricted to providers with at least 11 claims for the drug. The number of providers included in the analysis varied widely according to the particular drug, from 9 providers with at least 11 claims for ipilimumab to close to 2,000 providers with at least 11 claims for onabotulinumtoxinA. In this sample, the percentage of providers who never used the JW modifier varied across drugs, from 8 percent of those administering brentuximab vedotin to 70 percent of those administering infliximab.16
The large percentages of health care providers with more than 11 claims who used the modifier in some cases but not others present clear evidence that providers are using the JW modifier inconsistently from one drug to another. Equally important, over two-thirds of the providers with 11 or more claims for at least one of the 77 drugs never used the JW modifier at all.
Estimated Costs of Discarded Drugs
Given the evidence that many physicians are not using the JW modifier at all and that those who do apply it are doing so inconsistently from
one drug to the next and even from one treatment to the next for the same patient, it seems clear that a significant percentage of discarded drugs are never reported via the JW modifier and so the estimates of the total cost of discarded drugs based on the aggregation of Medicare claims with the JW modifier are likely to be significantly low. To get a sense of just how low, the analysis examined the top five drugs in terms of spending associated with claims with the JW modifier and estimated the amount of unreported spending on discarded drugs.
First, the researchers calculated a patient drug–specific ratio of JW-associated spending to total spending for claims with the JW modifier and then applied that ratio to calculate additional potential spending for observed encounters where the patient received the drug but there was no claim with the JW modifier.17 The total spending on discarded drug as a percentage of total spending for the drug was 9–27 percent without accounting for the amount of unreported spending. After including the estimated amount of unreported spending using the approach described above, the total spending on discarded drug was 11–28 percent of total spending. Even after including the estimated unreported spending, these figures are likely an underestimation because it is not possible to project potential spending for patient–drug combinations where the JW modifier was never used, which accounts for a sizable percentage of patient–drug combinations.
Another approach to estimate costs of discarded drugs is to include only health care providers who used the JW modifier. Excluding health care providers who never used the JW modifier, the average percentage of claims with the JW modifier was 59 percent and the average spending on these claims was 11 percent of total spending among the 25 drugs with at least 10 percent of claims with the modifier in 2018. These levels compare to 50 percent and 10 percent, respectively, when all providers (including those who never used the modifier) are included in the analysis.18 In some drugs (e.g., omalizumab), the difference in the percentage spending on JW claims could be as large as 50 percent (i.e., 10.3 percent versus 6.3 percent).
The committee recognizes the limitations of this analysis. The researchers who performed the analysis did not have access to durable medical equipment claims, for example, and these would likely capture additional amounts of discarded drugs (e.g., for products delivered via inhalers). Also, data from Medicare Advantage, which now covers approximately 36 percent of Medicare beneficiaries, were not included (Freed et al., 2020). Finally, although the Medicare claims data suggest that beneficiaries are likely to be responsible for coinsurance on discarded
products, it was not possible to precisely capture beneficiary out-of-pocket spending due to the unavailability of data on supplemental private or public coverage (e.g., Medigap), which commonly covers cost-sharing for some patients.
The second set of analyses examined the use of the JW modifier among privately insured patients and estimated the cost associated with discarded drugs for both patients with private health care and their payers.
To get a sense of how often the JW modifier is used in claims on private insurance policies, the researchers used IBM MarketScan research databases to identify it in both the Commercial Claims and Encounters and Medicare Supplemental databases from 2017 to 2018. Using the same 77 drugs selected in the above Medicare analysis, they examined the patterns of JW code use in the private insurance market, identifying claims for the 77 drugs from MarketScan data by using the Healthcare Common Procedure Coding System codes for those drugs.
Out of more than 4 million claims in 2017 and 2018, the overall percentage with the JW modifier was very low—just 3.62 percent. Fifteen of the 77 drugs had zero JW claims. Among the 62 drugs with at least one JW claim, less than 1 percent to 22.2 percent, for azacitidine (J9025), had the JW modifier in the claims. As Table 3-1 shows, the use of the JW modifier in weight-based drugs in the MarketScan database was substantially lower than the pattern found in the Medicare analysis, as not all private insurance claims require health care providers to use it. When limiting the analysis to only weight-based drugs, findings from the analysis of MarketScan data showed that the use of the JW modifier was even more concentrated than the pattern observed in Medicare data. As Figure 3-1 shows, only 15 of the 46 weight-based drugs had more than 10 percent of the claims submitted using the JW modifier.
The JW modifier was used more often for weight-based than fixed-dose drugs, and oncology, neuromuscular, and rheumatology drugs were the most likely to have it. The JW modifier was more prevalent among patients with Medicare supplemental insurance than among those with private, employment-based insurance, those who received treatment in 2018 versus 2017, and those who resided in the South.
Patient–Drug Pairs Level Analysis
Although the MarketScan data did not allow for the sorts of physician-to-physician comparisons that were described above from Medicare data, it was possible to carry out an analysis with patient–drug pairs to gain insight into how consistently the JW modifier is used in claims on private insurance. The researchers categorized modifier compliance into three levels: none, partial, and full. The analysis used the same top five drugs as in the analysis of Medicare fee-for-service claims data; Table 3-2 shows the pattern of compliance based on patient–drug pairs in the MarketScan data. For patients with Medicare, the range of full compliance for these five was 36–60 percent. The corresponding JW compliance in MarketScan data was much lower, 3.9–22.8 percent.
Out of the 62 drugs with at least one JW claim, 25 drugs had more than 5 percent of JW claims at the patient–drug pair level; azacidine had the highest JW compliance rate, about 42 percent combined partial and full compliance.19
Estimated Costs of Discarded Drugs
Working from the MarketScan data, the researchers calculated the costs of discarded drugs as reflected in the JW modifier filings. They observed that 18 drugs had discarded amounts with a total cost greater than $1 million. The discarded amount for these 18 drugs accounted for about 89 percent of the $80.6 million total cost for the discarded amounts of all 77 drugs in MarketScan data.
The $80.6 million figure was calculated using claims with the JW modifier. However, private payers grossly underused the JW modifier.20 The researchers imputed the potential additional spending on discarded drugs using the same approach described in the Medicare analysis by applying a patient drug–specific ratio to the partially compliant group. The researchers found that using the approach employed by CMS, discarded drugs accounted for 0.81 percent of total insurance payment on the above 18 drugs. However, this percentage was based on the underused JW modifier. When the potential additional spending imputed from the partially compliant group was included, the percentage of total insurance payment on discarded drugs increased to 1.6 percent.21 This analysis again demonstrated the underestimation of costs associated with discarded drugs when the calculation relied on the JW modifier.
The analyses of both the Medicare and MarketScan claims data demonstrated the inadequacy of relying solely on the JW modifier to estimate costs associated with discarded drugs because the current billing practice largely does not adhere to the CMS discarded drug policy. To better understand the financial impact of discarded drugs, the researchers conducted an additional analysis that used patients’ actual weight and height to determine the discarded percentage of their drugs.
In the MarketScan data, a subset of patients participated in the Health Risk Assessment (HRA) survey. The HRA data include self-reported weight, and height information. From the 77 Part B drugs, 45 weight-based drugs were found in the claims of patients who responded to the HRA survey. The researchers abstracted the recommended dosage and vial size for a drug to treat a specific disease based on IBM Micromedex and collected the claims of these 45 drugs from the 2017 MarketScan data. Then they linked the HRA data with the drug claims to select patients with known weight/height as a study cohort of 4,688 patients.22 They computed the discarded drug units by subtracting patients’ treatment drug units from the full vial sizes and calculated the percentage of the discarded drug units by dividing the discarded units over the full vial sizes. Then, they multiplied the percentage of discarded drug units with the total net pay or out-of-pocket cost for the discarded drug amount for payers and for patients.
The researchers then limited the analyses to drugs with more than 10 claims for each drug (38 drugs). The analysis included the mean, minimum, and maximum discarded percentage for each drug.23 The mean discarded percentage ranged from less than 1 percent to close to 40 percent. Variations in patients’ weight or BSA resulted in a noticeable range between minimum and maximum percentages for some drugs. The within-drug variation in the discarded percentage highlights the challenges of adhering to the JW modifier mandate, as providers will need to make the calculation for each patient based on weight at the time of treatment, which can fluctuate from one chemotherapy administration to another. It also raises a question whether payers will have sufficient real-time biometric data to verify the information submitted in claims with the JW modifier.
The researchers analyzed payment variables from claims data to estimate the financial burden of discarded drugs for private payers and patients. On average, private insurance paid $36,652 per patient on weight-based drugs, and of this amount, $2,992 (8.2 percent) was spent on discarded drugs. The mean out-of-pocket cost was $1,043, and of this,
$78 (7.5 percent) was spent on discarded drugs. Although the average out-of-pocket cost for discarded drugs was moderate, a small subset of patients disproportionately bore the financial burden, especially those enrolled in high-deductible plans. For these patients, the top 10 percent paid more than $380 for discarded drugs, whereas for patients not in high-deductible plans, the top 10 percent paid more than $260. A recent survey on the economic well-being of U.S. households reported that 4 in 10 adults could not cover an unexpected expense of $400 (Board of Governors of the Federal Reserve System, 2018). Using $400 as the benchmark value, this analysis found that approximately 10 percent of patients with high-deductible plans paid more than that on discarded drugs versus 4 percent for those who did not enroll in high-deductible plans.
The committee recognizes two major limitations of these analyses and other estimates of the payments for discarded drugs produced by other researchers. First, the information from the JW modifier is insufficient to provide accurate estimates of the amounts of drugs that are discarded. Any estimate that depends on data from the JW modifier will be an underestimate because of the noncompliance with the CMS discarded drug policy, and the resulting data may not be representative. Second, the estimates are only nominal values that are based on the assumption that every ounce in a vial has an equal value and that the value of an ounce administered to treat a patient is equal to the value of an ounce of drug that is discarded. The calculations do not reflect the reality of the excess costs to patients and the system because of the factors that go into determining the price of drugs and how they are paid for. Chapter 5 explicitly challenges the assumption underlying the calculations and addresses the economics of discarded drugs using the committee’s analytical construct.
- Not all health care providers use the JW modifier, and among those health care providers that use the modifier, the JW modifier is not reported consistently, even though it has been required by the Centers for Medicare & Medicaid Services since January 2017.
- Current estimates on the payments for discarded drugs produce nominal values that do not accurately capture the economic value of discarded drugs because the estimates are based on the assumption that every ounce in a vial has an equal value and that the value of an ounce administered to treat a
patient is equal to the value of an ounce of drug that is discarded, which in reality is not the case.
- Currently, attempting to estimate a single value for the cost of discarded drugs from single-dose vials to the Medicare program using the JW modifier would be incomplete and underestimated because the modifier is underused by health care providers and the resulting data may not be representative. Also, information on discarded drugs from Medicare Advantage plans and other commercial insurance plans are currently not available.
- The use of the JW modifier as a basis for data collection and measurement of the amount of discarded drugs from single-dose vials has no current practical application to addressing the issue of discarded drugs.
RECOMMENDATION 3-1: While the current Medicare Part B drug reimbursement system is in place, the Centers for Medicare & Medicaid Services should discontinue the use of the JW modifier.
Bach, P. B., R. M. Conti, R. J. Muller, G. C. Schnorr, and L. B. Saltz. 2016. Overspending driven by oversized single dose vials of cancer drugs. BMJ 352(352):i788.
Board of Governors of the Federal Reserve System. 2018. Report on the economic well-being of U.S. households in 2017. https://www.federalreserve.gov/publications/files/2017report-economic-well-being-us-households-201805.pdf (accessed October 1, 2020).
CMS (Centers for Medicare & Medicaid Services). 2019. Medicare Part B discarded drug units report. https://www.cms.gov/research-statistics-data-systems/cms-drug-spending/medicare-part-b-discarded-drug-units-report (accessed August 7, 2020).
Freed, M., A. Damico, and T. Neuman. 2020. A dozen facts about Medicare Advantage in 2020. Kaiser Family Foundation. https://www.kff.org/medicare/issue-brief/a-dozen-facts-about-medicare-advantage-in-2020 (accessed October 13, 2020).