Concentration of spending and share of specialty drug spending in Medicare Part D over a 10-year period

Abstract

BACKGROUND:

In 2021, Medicare Part D gross prescription drug spending amounted to $216 billion, a number that has more than doubled over the last 10 years. Spending in Medicare Part D is concentrated on a small number of drugs, and spending on specialty drugs has increased in recent years. However, the extent to which concentration in Part D spending has changed over time and the drivers of this change have not been described.

OBJECTIVE:

To quantify the time trends in Medicare Part D spending and utilization, the concentration of spending, and the share of spending accounted for by specialty drugs from 2012 to 2021.

METHODS:

In this repeated cross-sectional study, we used data from the Centers for Medicare & Medicaid Services Part D Drug Spending Dashboard to investigate the time trends in total gross spending, prescriptions claims, and the average cost of a prescription claim for Part D drugs. We assessed the concentration based on the share of total gross spending and prescriptions by the drugs with the top 1%, 5%, and 10% of the highest spending and Lorenz curves and Gini coefficients. In addition, we stratified our analyses by specialty and nonspecialty drugs.

RESULTS:

Over the last 10 years, total gross drug spending in Medicare Part D increased by 103.5%, with a compounded annual growth rate of 8.2%. This change was driven by both increases in prescription claims and price increases of existing drugs to a similar degree. The concentration of spending intensified, with the top 1% of drugs accounting for an escalating share of total spending (from 31.4% to 41.1%). Over the 10-year study period, these top-spending drugs accounted for 5.6% of prescriptions but 34.6% of spending. Lorenz curves and increased Gini coefficients similarly showed that a smaller number of drugs accounted for increased spending over the study period. Specialty drug spending increased by 566.5%, with a compounded annual growth rate of 23.5%. The share of total spending on specialty drugs increased from 21.7% in 2012 to 71.1% in 2021. In 2021, specialty drugs accounted for 6.2% of prescriptions but 71.1% of total spending.

CONCLUSIONS:

Medicare Part D gross drug spending became increasingly more concentrated from 2012 to 2021, which was especially pronounced for specialty drugs. Increases in prices for specialty and other brand-name drugs will likely continue to drive gross spending upward. Although the Inflation Reduction Act provisions will likely reduce net spending on selected drugs, other policy changes may be warranted.

Plain language summary

Spending on the Medicare Part D outpatient prescription drug benefit doubled from 2012 to 2021. Over time, a small set of prescription drugs accounted for an increasingly higher share of the spending. Although the Inflation Reduction Act is likely to reduce Part D net spending, more policy changes may still be needed.

Implications for managed care pharmacy

Top-selling and specialty drugs account for a disproportionate share of spending in the Medicare Part D program. These findings reinforce the rationale of the Inflation Reduction Act provisions that allow Centers for Medicare & Medicaid Services to negotiate some drug prices, yet the extent to which provisions of the Inflation Reduction Act reduce Part D spending remains to be realized. Data on the concentration of drug spending can be informative in designing pharmacy benefits and assessing risk-pooling in insurance plans.

Prescription drug prices in the United States are higher than those in other countries and are a health care and financial concern for Americans.¹ Between 2016 and 2021, gross prescription drug spending in the United States increased by approximately 16%, exceeding $600 billion in 2021.² Over the last decades, there has also been a sharp increase in the concentration of prescription drug spending among patients. In 2001, half of all prescription drug spending was concentrated in 6.0% of the US population, but by 2018 this percentage had decreased to 2.3%.³

Medicare Part D, as a major payer of prescription drugs, is a significant component of the federal budget. In 2021, Medicare Part D plans covered more than 3,500 prescription drug products, with total gross expenditures of $216 billion.³ A high percentage of these expenditures is accounted for by a proportionally small number of drugs.^3,4 In 2021, the 10 top-selling Part D drugs with the highest total expenditures accounted for 0.3% of covered drugs but 22% of total gross Part D spending.³ This trend corresponds to the genericization of many prescription drugs and a strategic shift in the biopharmaceutical industry toward developing and marketing high-cost specialty drugs.^5,6

Over the past decade, specialty drugs have accounted for a growing share of new drugs introduced to the market, and they have been introduced at substantially higher prices than nonspecialty drugs. Total US spending on specialty drugs grew from $211 billion in 2016 to $301 billion in 2021²; and the share of drug spending on specialty drugs increased from 32% in 2012 to 51% in 2022, driven by growth in immunology and oncology products.⁷ In Medicare Part D, the Congressional Budget Office (CBO) found that net spending on specialty drugs increased from $8.7 billion in 2010 to $32.8 billion in 2015, accompanied by a growing share of total net drug spending from 13% to 31%.⁸ The CBO anticipates a continued rise in the proportion of Medicare Part D spending allocated to specialty drugs in the coming years.⁸

Medicare Part D spending is concentrated on relatively few drugs. However, it is unclear to what extent the concentration of spending (1) has changed over 10 years and (2) is driven by utilization (ie, greater number of prescription claims for a smaller set of drugs), by the cost per prescription of these drugs, or by both. Also, up-to-date estimates of the impact that specialty drugs have had on the total spending and increased spending concentration in Medicare Part D are needed. Data on the concentration of drug spending are an important instrument for managing drug benefit plans and understanding appropriate risk-pooling in insurance plans.³ Further, analyzing the concentration of prescription drug spending can provide insights into potential budget-impact consequences of programs like the Medicare Drug Negotiation Program.

In this study, we examined Medicare Part D spending over a 10-year period (2012-2021) to address the following objectives: (1) to quantify time trends in spending and utilization of Medicare Part D drugs, (2) to describe the concentration of Part D spending, and (3) to calculate time trends in spending, utilization, and concentration of spending among specialty and nonspecialty drugs.

Methods

STUDY DESIGN AND DATA SOURCE

We conducted a descriptive study using the Centers for Medicare & Medicaid Services (CMS) Part D Drug Spending Dashboard data (repeated cross-sectional analyses for 2012 to 2021) to obtain information on drug utilization and spending.⁹ The dashboard is an interactive, web-based tool that summarizes all prescription transactions covered by Medicare, linked by National Drug Codes, and aggregated across all strengths and dosage forms.¹⁰ The spending data represent the gross drug cost, including ingredient, dispensing fees, and sales tax, and amounts paid by Medicare, the Part D plan sponsor, and the beneficiary, but the data do not reflect any manufacturer rebates or discounts. Because we extracted data over a 10-year period, we used the Consumer Price Index for Medical Care to adjust all cost data to 2021 US dollars.¹¹ We relied on publicly available data that did not include data on human participants, and this study was deemed nonhuman research exempt from formal approval by the University of Florida Institutional Review Board.

10-YEAR TIME TREND AND CONCENTRATION ANALYSES

In the time trend analysis of Medicare Part D prescription drug utilization and spending, we calculated 3 primary measures for each drug and year in the study period: total prescription drug spending, total number of prescription claims, and the average cost per prescription claim. Total prescription drug spending and the total number of prescription claims were calculated by annually adding up the total amount paid for each prescription drug and the total number of prescription claims. The average cost per prescription claim was calculated by dividing the annual spending for each drug by its annual prescription claims. Then, we calculated the overall average cost per prescription claim across the entire sample of products every year. Data on different formulations of a single drug product were aggregated. For each of the 3 primary measures, we calculated the annual value, percentage change from 2012 to 2021, and the compounded annual growth rate (CAGR). We also assessed whether the change in total prescription drug spending was primarily driven by changes in utilization, measured as the number of prescription claims, or driven by changes in the drug price index, measured as the average cost per prescription claim while holding the utilization and price fixed, respectively.

We used 2 approaches to describe the concentration of Medicare Part D prescription drug spending. In the first approach, we sorted drugs in descending order of total spending per drug per year. Then, we grouped the top-selling drugs that accounted for each year’s top 1%, 5%, and 10% of total spending. Finally, we calculated the share of total spending; the share of total prescription claims accounted for by the top-selling 1%, 5%, and 10% of drugs for each year; and the averaged share across 10 years as well as the 10-year change.

In the second approach, we used Lorenz curves and Gini coefficients to describe the degree of concentration of Medicare Part D gross spending distribution. The Lorenz curves graph the cumulative proportion of drug spending against the entire set of drugs ranked by spending and has been frequently used to analyze inequality in health and income over patient populations.^12-14 This approach describes the concentration of drug spending over time in a dimensionless metric unaffected by population size increases. We used the Lorenz command in R (version 4.3.2, R Foundation) to graph each year’s Lorenz curve and estimate the corresponding Gini coefficient. The Gini coefficient is a measure bounded by 0 and 1, where a Gini coefficient of 0 would imply that all drugs account equally for total Part D spending, whereas a Gini coefficient of 1 would imply that 1 single drug accounts for all the Part D spending. A higher Gini coefficient is associated with a higher concentration of spending. Changes in the Gini coefficient over the 10-year period illustrate changes in the concentration of Medicare Part D spending.

SPECIALTY DRUG STRATIFIED ANALYSES

To assess the impact of specialty drugs, we stratified both the 10-year time trend analysis and the change in concentration of spending analysis by specialty and nonspecialty drugs. Because there is no universally accepted definition of specialty drugs, we relied on the CMS specialty cost per month tier ($600 from 2012 to 2016 and $670 from 2017 to 2021), which we proxied based on the average cost per claim, as has been done in previous analyses of the Spending Dashboard.^4,15,16 We calculated time trends in total spending, total prescription claims, and average cost per prescription claim for both specialty drugs and nonspecialty drugs as described above. In addition, we summarized the share of total spending and total prescription claims accounted for by specialty and nonspecialty drugs. Finally, we produced Lorenz curves and Gini coefficients separately for specialty and nonspecialty drugs to identify their relative contributions to the concentration of Part D spending.

REBATE ANALYSES

Because the CMS data source reported gross drug spending rather than net drug spending, we repeated all analyses with average rebate assumptions based on a CBO analysis as 12% rebate for specialty drugs, 47% rebate for nonspecialty brand-name drugs, and no rebate for generic drugs.^5,17

Results

10-YEAR TIME TRENDS IN PART D SPENDING AND UTILIZATION

Over the 10-year period, total gross Part D spending and the average cost per prescription claim increased at a higher rate (103.5% and 245.6%, respectively) than the number of total prescription claims (33.3%) (Table 1). Total gross Part D spending increased from $106.0 billion in 2012 to $215.7 billion in 2021 (103.5% increase, 8.2% CAGR). The increased gross spending was driven by both increases in prescription claims and price increases of existing drugs to a similar degree, whereas the introduction of new covered drugs had a smaller impact on increased gross spending (Table 2). The number of total prescription claims increased from 1,125.4 million to 1,500.4 million (33.3% increase, 3.2% CAGR). The average cost per prescription claim increased from $962.94 to $3,327.63 (245.6% increase, 14.8% CAGR).

TABLE 1.

Medicare Part D Prescription Drug Gross Spending and Utilization (Inflation-Adjusted 2021 US Dollars)

Year	Total gross spending (billions)	Total prescription claims (millions)	Average cost per prescription claim (95% CI)
2012	$106.0	1,125.4	$962.94 ($794.76-$1,131.12)
2013	$124.1	1,308.7	$1,248.09 ($1,042.42-$1,453.77)
2014	$146.0	1,400.3	$1,370.86 ($1,168.77-$1,572.95)
2015	$162.2	1,446.7	$1,676.25 ($1,450.29-$1,902.21)
2016	$162.6	1,444.7	$1,742.16 ($1,513.21-$1,971.11)
2017	$165.8	1,394.9	$2,241.67 ($1,923.41-$2,559.92)
2018	$180.0	1,455.8	$2,600.91 ($2,241.06-$2,960.76)
2019	$194.5	1,493.1	$2,983.55 ($2,540.41-$3,426.70)
2020	$202.4	1,496.6	$3,170.71 ($2,741.27-$3,600.15)
2021	$215.7	1,500.4	$3,327.63 ($2,906.91-$3,748.35)
Change 2012-2021, %	103.5	33.3	245.6
CAGR, %	8.2	3.2	14.8

CAGR=compounded annual growth rate (CAGR calculation: [value in 2021/value in 2012] ^(1/9) −1).

TABLE 2.

Causes of Increases in Medicare Part D Gross Spending

Percentage increase in spending owing to price/utilization/new drugs	2012-2013	2013-2014	2014-2015	2015-2016	2016-2017	2017-2018	2018-2019	2019-2020	2020-2021	Average
Increase owing to increase in claims for existing drugsa	12.12	9.99	5.10	−1.67	−4.64	3.37	4.50	3.78	4.04	4.07
Increase owing to increase in prices for existing drugsb	5.47	6.67	5.09	2.93	3.16	4.28	2.46	0.23	2.20	3.61
Increase owing to introduction of new drugsc	1.37	1.93	2.25	0.00	8.55	0.87	1.29	0.40	0.28	1.88
Total percentage increase in spending	17.11	17.63	11.11	0.19	2.01	8.56	8.05	4.08	6.55	8.37

^a This holds prices constant from 1 year to another and gives the increase in spending only because of the change in claims.

^b This holds prescription claims constant from 1 year to another and gives the increase in spending because of changes in prices.

^c This is the increase in spending because of the inclusion of new Part D drugs from 1 year to another. The average price (spend per claim) is substantially higher for drugs that are newly introduced in Part D compared with already existing drugs.

CONCENTRATION OF PART D SPENDING

The share of total gross spending was highly concentrated among the top 1% of drugs over the 10-year period. Over the 10 years, the top 1% of drugs accounted for only 5.6% of all prescriptions but 34.6% of total spending (Supplementary Table 1 ^{(515.3KB, pdf)} , available in online article). The top 1% of drugs included mostly diabetes, cardiovascular disease, and cancer drugs, and 27 of the 1% top-spending drugs are specialty drugs (Supplementary Table 1 ^{(515.3KB, pdf)} ). The share of gross spending was also concentrated in the top 5% and 10% of drugs over the 10-year period, but the relative difference between the share of gross spending and prescriptions was less pronounced for the 10% of drugs (Supplementary Table 2 ^{(515.3KB, pdf)} ).

When evaluating the change in concentration of gross spending over time, the share of total gross spending accounted for by the top-selling 1% of drugs increased from 31.4% in 2012 to 41.1% in 2021, whereas the corresponding share of total prescription volume decreased from 8.5% in 2012 to 5.1% in 2021 (Figure 1). The changes in shares of total gross spending and prescriptions for the top-selling 5% and 10% of drugs followed a similar pattern, although the magnitudes of the changes were not as pronounced (Figure 1 and Supplementary Table 2 ^{(515.3KB, pdf)} ).

FIGURE 1.

Medicare Part D Average Share of Total Gross Spending and Prescription Claims Among the Drugs With the 1%, 5%, and 10% Highest Spending (2012-2021)

Figure 2 shows the Lorenz curves and the associated Gini coefficients of Part D gross spending for each year from 2012 to 2021. The Gini coefficients, taking the distribution of all drugs into account, were all higher than 0.86 and increasing year by year, confirming a higher concentration of gross spending for each year with a smaller number of drugs accounting for more and more total spending.

FIGURE 2.

Lorenz Curves of Prescription Drug Gross Spending Distribution in Medicare Part D, 2012-2021

10-YEAR TIME TRENDS OF SPECIALTY DRUGS AND NONSPECIALTY DRUGS

Between 2012 and 2021, Table 3 shows that there was a substantial increase in specialty drug spending as measured by total gross spending (566.5% increase, 23.5% CAGR) and average cost per prescription claim (60.5% increase, 5.4% annual growth rate). In contrast, total gross spending decreased for nonspecialty drugs (−24.9%). The average cost per prescription claim did increase for nonspecialty drugs but by a much lower magnitude than for specialty drugs (10.4% vs 60.5%, respectively).

TABLE 3.

Medicare Part D Prescription Drug Gross Spending and Utilization for Specialty and Nonspecialty Drugs (Inflation-Adjusted 2021 US Dollars)

V ariable	Total gross spending (billions)		Total prescription claims (millions)		Average cost per prescription claim (95% CI)
Year	Specialty drugs (%)	Nonspecialty drugs (%)	Specialty drugs (%)	Nonspecialty drugs (%)	Specialty drugs	Nonspecialty rugs
2012	$23.0 (21.7)	$83.0 (78.3)	12.7 (1.1)	1,112.7 (98.9)	$5,329.15 ($4,371.09-$6,287.21)	$172.48 ($164.13-$180.83)
2013	$29.6 (23.9)	$94.5 (76.1)	14.4 (1.1)	1,294.3 (98.9)	$6,180.19 ($5,156.74-$7,203.64)	$186.50 ($177.95-$195.06)
2014	$44.3 (30.4)	$101.7 (69.6)	17.5 (1.2)	1,382.9 (98.8)	$5,712.97 ($4,872.06-$6,553.88)	$190.18 ($181.84-$198.52)
2015	$62.1 (38.3)	$100.2 (61.7)	23.7 (1.6)	1,422.9 (98.4)	$6,072.69 ($5,271.18-$6,874.21)	$194.55 ($186.21-$202.89)
2016	$63.3 (38.9)	$99.3 (61.1)	22.6 (1.6)	1,422.1 (98.4)	$5,988.57 ($5,221.54-$6,755.60)	$192.34 ($183.83-$200.86)
2017	$69.2 (41.7)	$96.6 (58.3)	23.0 (1.6)	1,371.9 (98.4)	$7,246.19 ($6,241.23-$8,251.16)	$199.68 ($190.22-$209.13)
2018	$83.5 (46.4)	$96.5 (53.6)	33.2 (2.3)	1,422.6 (97.7)	$7,835.41 ($6,776.15-$8,894.68)	$204.03 ($194.58-$213.47)
2019	$104.2 (53.6)	$90.3 (46.4)	49.8 (3.3)	1,443.3 (96.7)	$8,456.43 ($7,214.61-$9,698.24)	$201.41 ($192.32-$210.50)
2020	$120.9 (59.7)	$81.6 (40.3)	61.7 (4.1)	1,434.9 (95.9)	$8,398.67 ($7,282.33-$9,515.02)	$195.10 ($186.38-$203.82)
2021	$153.4 (71.1)	$62.3 (28.9)	92.9 (6.2)	1,407.5 (93.8)	$8,553.59 ($7,499.29-$9,607.89)	$190.40 ($181.87-$198.93)
2012-2021, %	566.5	−24.9	631.7	26.5	60.5	10.4
CAGR, %	23.5	−3.1	24.7	2.6	5.4	1.1

CAGR=compounded annual growth rate (CAGR calculation: [value in 2021/value in 2012] ^(1/9) −1).

Over the 10-year period, the percentage of total gross Part D spending attributed to specialty vs nonspecialty products has nearly reversed (Table 3 and Supplementary Figure 1 ^{(515.3KB, pdf)} ). The share of total gross spending on specialty and nonspecialty drugs was 21.7% and 78.3%, respectively, in 2012 compared with 71.7% and 28.9% in 2021. Despite the growing share of gross spending on specialty drugs, the percentage of specialty drug prescriptions rarely exceeded 6% during the 10-year period (Table 3).

The Lorenz curves and Gini coefficients for specialty and nonspecialty drugs show very similar and highly concentrated spending in 2021 (Supplementary Figure 2 ^{(515.3KB, pdf)} ). The Gini coefficients also indicate that the increased concentration of spending over time has primarily been driven by specialty drugs, as the concentration of nonspecialty drugs spending has been fairly constant, whereas the Gini coefficient for specialty drugs increased from 0.818 to 0.891.

REBATE ANALYSES

The results with assumed rebate are provided in the supplementary materials (Supplementary Tables 3-5 ^{(515.3KB, pdf)} and Supplementary Figure 3 ^{(515.3KB, pdf)} ). These results followed a similar pattern as the main analyses.

Discussion

Medicare Part D gross spending more than doubled from 2012 to 2021, and our analysis shows that the growth was both increases in prescription claims and price increases of existing drugs to a similar degree. Spending was disproportionately concentrated, with the top 1% of drugs by spending accounting for only 5.6% of prescriptions but 34.6% of spending over 10 years. This high concentration of spending has become even more pronounced over time and may continue in this direction given the focus on high-cost specialty drug development.

The concentration of Part D spending on an increasingly smaller number of high-cost products reinforces the rationale of the Inflation Reduction Act (IRA) provisions that allow CMS to negotiate some drug prices.¹⁷ The Medicare Drug Price Negotiation Program uses a phased approach to create a list of single-source brand-name drugs or biologics without generic or biosimilar competitors for which the Secretary of the Department of Health and Human Services will negotiate prices with drug companies. In August of 2023, CMS named the first 10 drugs subject to negotiation,¹⁸ and additional drugs will be added in the coming years. Not surprisingly, the first 10 drugs subject to negotiation were among the top 1% of drugs identified in our analyses (Supplementary Table 1 ^{(515.3KB, pdf)} ). Negotiations have been projected to generate savings ranging from $1.8 billion^19,20 to $3.7 billion,²¹ and our findings suggest that the concentration of Part D spending may decrease in response.

Not all drugs in the top-spending 1% are eligible for the Medicare Drug Price Negotiation Program. For example, Revlimid and Humira were not eligible because of the presence of generic/biosimilar competition.²² Specialty drugs are another category of potential negotiated products that could also reduce total spending. In our analysis, specialty drug spending was a major contributor to the increased concentration of spending. Total spending on specialty drugs increased 5.7-fold over the 10-year study period, and 27 of the 1% top-spending drugs are specialty drugs. In 2021, specialty drugs accounted for a disproportionate share of spending, with only 6% of prescriptions accounting for 71% of total spending. Experts have predicted that specialty drugs, such as Jakafi, Ofev, and Xtandi, may be added to the Medicare Drug Price Negotiation Program in future years.²² Notably, drugs eligible for negotiation in 2026 and beyond may be drawn from the pool of drugs paid for by Medicare Part B, which will affect the potential savings for the Part D program.

Challenges and unanswered questions about the Medicare Drug Price Negotiation Program and how it will affect spending and concentration of spending remain, including methods for selecting therapeutic alternatives, optimizing patient preferences, and ensuring that benefits are achieved across sociodemographic subgroups.^23-27 Indeed, the limited scope of the IRA prompted the Biden administration to issue an executive order directing the Center for Medicare and Medicaid Innovation to evaluate other models that would lower drug costs for beneficiaries.²⁸ The Center for Medicare and Medicaid Innovation proposed 3 models, only 1 of which is applicable to Part D plans.²⁹ The proposed Medicare High-Value Drug List would promote the use of high-value generic drugs by limiting patient cost sharing to $2 per fill. This program has the potential to improve adherence and stabilize patient out-of-pocket costs,³⁰ but it would likely do little to offset the costs of the top-spending 1% and specialty drugs that are driving Part D spending, many of which do not have generic alternatives.

Our study aligns with and expands on industry reports by providing results specific to the Part D program. Prior trend reports aggregate results across all payer types,^2,7 are nearly 10 years old,⁸ do not stratify by specialty,³¹ or do not include a time trend analysis over 10 years.^3,4 Although none of these industry reports are peer-reviewed, our findings are generally consistent with the reported trends.

LIMITATIONS

We relied on the CMS Part D Gross Drug Spending data for our study, which do not include information about drug-specific rebates or discounts. Thus, our values are inflated over actual net prices. We attempted to account for this by applying specialty and nonspecialty drug-specific average rebates in our sensitivity analysis, but we acknowledge that this approach has limitations because rebates vary by individual drug and by plan. Furthermore, we defined utilization as the total number of prescription claims rather than the total dosage units. Consequently, we used the average cost per prescription claim as the price index. This approach can potentially impact our findings, because the cost per prescription claim does not account for the different days supply associated with each prescription. Finally, given that there is no universally recognized definition of specialty drugs, we defined specialty drugs as those with an average cost per claim at or above the dollar-per-month thresholds in the CMS guidelines. Because monthly costs are not directly available in the CMS dashboard, this approach may limit the precision of our specialty drug classification and affect the comparability of our results with studies that use a different definition of specialty drugs.

Conclusions

Over 10 years, total gross drug spending in Medicare Part D more than doubled, driven by both increased claims and increased prices. The concentration of Medicare Part D drug spending has continuously increased, with the top-spending 1% of drugs accounting for only 5.1% of prescriptions but more than 41.1% of total spending in 2021. Similarly, specialty drugs accounted for an increasingly disproportionate share of spending over the study period, with only 6.2% of prescriptions accounting for 71.7% of total spending by 2021. Specialty drugs have contributed to the increased concentration of Medicare Part D spending. Price increases for specialty and other brand-name drugs will likely continue to drive spending upward. Although the IRA provisions will likely reduce net spending on selected drugs, other policy changes may be warranted.