The Rising Cost of Specialty Drugs Drove Spending Increases for People with Multiple Sclerosis How does the cost of prescription drugs affect the cost of care for individuals who rely on them? HCCI investigated how the cost of prescription drugs affects the total cost of care for people with multiple sclerosis (MS). MS is a central nervous system disease estimated to affect over 400,000 people in the U.S., costing an estimated $10 billion in health care spending each year. The rising prices of specialty drugs used to treat MS, called disease modifying therapies (DMTs), garnered media attention. The U.S. House of Representatives began investigating DMT prices. The role prescription drugs play in affecting the cost of care for people with diseases like MS will become more important as policy makers consider relaxing consumer protections created by the Affordable Care Act. What We Did. HCCI analyzed claims for people diagnosed with MS from 2009 to 2015. We limited our sample to individuals flagged with an MS diagnosis with 12 months of continuous insurance enrollment and prescription drug coverage. Our sample comprises individuals from the individual, employer sponsored health insurance, and Medicare Advantage markets. We aggregated total health care spending – the sum of negotiated payer spending and individuals’ out-of-pocket spending - and utilization for individuals within each year. Questions We Asked. 1. How did health care spending for people diagnosed with MS change from 2009 to 2015? 2. Which drugs drove changes in spending on Disease Modifying Therapies (DMTs)? 3. How did the use of DMTs change for people with MS? 4. Was increased spending on oral DMTs driven by changes in use or cost? 5. How did the cost of DMTs change from 2009 to 2015? What We Found. For people with MS, total health care spending per person nearly doubled from 2009 to 2015. Disease modifying therapies (DMTs) accounted for 82% of the increase in total spending. Overall DMT spending more than doubled from 2009 to 2015, primarily due to a rise in spending on newly introduced oral DMTs. Overall DMT use was slightly lower in 2015 than in 2009. Further, the average cost for each DMT studied rose by an average of between 9% and 23% per year from 2009 to 2015. While we do not provide causal evidence, our results suggest that the rising cost of DMTs were the primary driver of increased prescription drug spending - and total health care spending - for people with MS. On the other hand, overall DMT use did increase from 2012 to 2015 as the use of oral DMTs increased following their introduction. This increase in DMT use represents a potentially positive development to weigh against increased DMT spending from 2012 to 2015. April 2018 How did health care spending for people diagnosed with multiple 1 sclerosis (MS) change from 2009 to 2015? Health care spending per person nearly doubled from 2009 to 2015, driven by increased spending on disease modifying therapies (DMTs) We decomposed total health care spending – the sum of payer spending and individuals’ out-of-pocket spending – for people diagnosed with MS into spending on one of the following: medical services (inpatient, outpatient, professional services), DMTs (a specific class of specialty drugs used to treat MS), or all other prescription drugs (Figure 1, Table 1). Figure 1: Total Spending Per Person for People Diagnosed with MS, 2009 to 2015 Source: Authors' analysis of HCCI claims data.  By 2015, total spending per person for  DMTs accounted for 82% of the increase people with MS was $39,628: in total spending for people with MS from 2009 to 2015.  $11,820 in medical spending (excluding all prescription drugs)  By 2015, more than half (53%) of all health care spending for people with MS  $20,882 in spending on went towards injectable and oral DMTs. injectable, oral DMTs  The share of spending on all medical  $2,613 in spending on infused services (besides prescription drugs) DMTs dropped from 44% to 30% despite  $4,312 in spending on all other increased medical spending from 2009 prescription drugs to 2015. 1 Which drugs drove changes in spending on Disease Modifying 2 Therapies (DMTs)? Spending on DMTs more than doubled from 2009 to 2015, primarily due to a rise in spending on newly introduced oral DMTs To better understand how spending on DMTs changed from 2009 to 2015, we decomposed total spending per person on DMTs by each specific DMT (Figure 2, Table 2). Figure 2: Total Spending Per Person on DMTs by People with MS, 2009 to 2015 Source: Authors' analysis of HCCI claims data.  Spending on Copaxone – the DMT with the  From 2009 to 2015, three oral DMTs highest spending per person – increased from were introduced: Gilenya (2010), $3,634 per person in 2009 to $6,163 in 2015. Aubagio (2012), and Tecfidera (2013).  Other injectable DMTs of note either saw  By 2015, Tecfidera had the second minimal increases in spending per person highest per person spending of any (Avonex, Rebif) or slight declines (Betaseron) DMT, $4,706. from 2009 to 2015.  From their introduction to 2015, oral DMTs accounted for: 74% of the increase in DMT spending: Tecfidera (41%), Gilenya (23%), Aubagio (10) 54% of the increase in overall spending: Tecfidera (30%), Gilenya (16%), Aubagio (7%) Methods note: In Figures 2 and 3, “All DMTs” refers to total per person spending on and use of injectable and oral DMTs. In particular, these figures omit spending on and use of infused DMTs. For further explanation, see the Methods section. 2 How did the use of Disease Modifying Therapies (DMTs) change for 3 people with MS? While overall DMT use was slightly lower in 2015 than in 2009, individuals shifted to using new oral DMTs To understand how DMT use changed from 2009 to 2015, we plotted the number of prescription filled days per person with MS for each DMT (Figure 3, Table 3). Figure 3: DMT Use Per Person by People with MS, 2009 to 2015 Source: Authors' analysis of HCCI claims data. While which DMTs people used changed, overall DMT use experienced a slight decline from 2009 to 2015. This suggests that the average cost of DMTs increased from 2009 to 2015.  By 2015, Tecfidera was the second most  Overall, use of injectable DMTs used DMT in our sample – 25 filled days per declined from 2009 to 2015: person – despite its introduction in 2013.  Avonex, 53% decrease  Due to increased use of oral DMTs, overall  Copaxone, 24% decrease DMT use increased from 95 filled days per  Betaseron, 58% decrease person in 2012 to 110 in 2015 (17%).  Rebif, 55% decrease  The overall increase in DMT use from 2012  Decreased use of injectable DMTs to 2015 represents a potential benefit to coincided with either increases or weigh against the overall increase in marginal decreases in spending. For a spending on DMTs over this time frame. further discussion, see the Appendix. 3 Was increased spending on oral Disease Modifying Therapies (DMTs) 4 driven by changes in use or cost? Increased spending on newer, oral DMTs was primarily driven by increased use despite increases in their cost Both spending on and use of oral DMTs (Aubagio, Gilenya, and Tecfidera) increased since their introduction to 2015. To understand changes in spending on each newly introduced oral DMT, we decomposed total spending per person into use (filled days per person) and average cost (total spending per filled day) (Figure 4). Figure 4: Comparing Changes in Per Person Spending, Use, and Avg. Cost by DMT Source: Authors' analysis of HCCI claims data. From the respective introductions of Aubagio, Gilenya and Tecfidera to 2015:  By 2015, oral DMTs accounted for 40% of all DMT filled days (Figure 3).  Each experienced larger changes in use than changes in the average cost, implying that increases in spending were primarily due to increased use.  Still, the average cost of each DMT increased by an average of almost 10% per year.  Changes in use for each DMT were primarily due to changes in the number of people using each DMT, rather than the amount of each DMT people were using.  Similar plots for older, injectable DMTs can be found in the Appendix. 4 How did the cost of Disease Modifying Therapies (DMTs) change from 5 2009 to 2015? The average cost of DMTs more than doubled from 2009 to 2015, rising similarly for each DMT From 2009 to 2015, spending on DMTs increased (Figure 2) and overall use remained similar (Figure 3), implying that the average cost of DMTs increased. However, there was also a shift towards using to newly introduced oral DMTs. To explore whether the average cost of DMTs rose due to a change in which DMTs people were taking, we compared the total spending per filled day for each DMT, highlighting oral DMTs (Figure 5, Table 4). Here, total spending refers to the sum of negotiated payer spending plus individuals’ out-of-pocket spending. Figure 5: Spending Per Filled Day for Oral DMTs vs. Injectable DMTs (gray), Quarterly, 2009 to 2015 Source: Authors' analysis of HCCI claims data.  Spending per filled day for the average DMT more than doubled from 2009 to 2015. The cost of each DMT followed similar increases, averaging between 9% and 23% per year.  By 2015, a filled day of Gilenya cost $197. This translates to $5,503 for a month’s supply. Table 5 reports the cost of a month’s supply for each DMT.  The similar increases in the cost of each DMT suggest that the average cost of DMTs increased primarily because each DMT became more expensive from 2009 to 2015, rather than changes in which DMTs people with MS used. Methods Note: The cost per filled day data reported above are quarterly averages of total spending per day on the most common NDC code for each DMT. For this reason, the average cost per filled day plotted in Figure 5 are slightly different than Figure 4. For more information, see the Methods section. For similar plots for the older, injectable DMTs, see the Appendix. 5 Data and methods Sample Construction: Using HCCI claims data from 2009 to 2015, we flagged individuals as diagnosed with multiple sclerosis (MS) if they had a claim with either an ICD-9 diagnosis code for MS (340) or a MS-DRG code (058, 059, 060). Once flagged, individuals remained flagged for the duration of our sample. We omitted claims prior to the initial diagnosis. We limited our sample to person, year observations of individuals flagged with an MS diagnosis with 12 months of continuous insurance enrollment and prescription drug coverage. Our sample is best thought of as a repeated cross section ranging from 42,279 people in 2009 to 58,608 in 2015. Our sample comprises individuals from the individual, employer-sponsored health insurance, and Medicare Advantage markets. Around 75% of the sample are female. In our sample, 10% of person, year observations are under the age of 35, 73% of people are between the ages of 35-64, and 17% are over the age of 65. We aggregate total spending – the sum of payer spending and individuals’ out-of-pocket spending within each year. We decompose total spending into spending on all medical services (inpatient admissions, outpatient services, and professional services), and prescription drug spending. We further decompose prescription drug spending into spending on disease modifying therapies (DMTs) and all other prescription drugs. To calculate spending on infused DMTs, we aggregate all medical spending on days on which an individual receives an infusion as spending on infused DMTs. We further reclassify them as spending on disease modifying therapies. We decomposed spending on DMTs by the type of each DMT. Throughout the brief we report per person numbers which are the sum of total spending on or use of each DMT divided by the number of individuals in the sample in each year. Importantly, this measure is distinct from spending per person taking each drug. All dollars values reported are nominal. DMT Classification: We categorized drugs as DMTs in accordance with Hartung et al. (2015), and following publications from the National Multiple Sclerosis Society. Throughout the brief we report spending and use per person for the most common DMTs over the duration of our sample. We group all remaining injectable DMTs taken by people in our sample as “other injectable DMTs” (Extavia, Glatopa, and Plegridy). We categorize infused DMTs (Tysabri, Lemtrada, and Novantrone) separately from oral and injectable DMTs because they are administered differently than prescription drugs; their use is measured in the number of days with an administered infusion rather than prescription filled days. To facilitate the comparison of spending and use per person we omit spending and use of infused DMTs from our analysis on pages 2 and 3. As seen in Figure 1, though, infused DMTs are not a leading driver of increased health care costs for people with MS in our sample from 2009 to 2015. 6 Data and methods (continued) Calculating Average Cost for Each DMT: To measure average cost of each DMT, we computed total spending per filled day. Total spending refers to the sum of negotiated payer spending plus individuals’ out-of-pocket spending per filled day. To calculate the spending per filled day reported in Figure 5, we used subset of prescription drug claims by people with MS in our sample. To address potential outliers, we limited the subset of our sample to claims for the most common amount of filled days for each NDC code for each DMT. To account for the possibility that drugs change formulations over time, we used the most frequently used NDC code in our sample for each DMT. For each claim we divided total spending by the number of prescription filled days. We subsequently took the quarterly average for each DMT to compute the average cost per filled day for each DMT used in Figure 5, Table 4. The average cost per month’s supply in Table 5 is computed as the product of the average cost per filled day by the most common amount of filled days for each NDC code for each DMT. Due to this procedure, the spending per filled day plotted in Figure 5 and reported in Table 4 are slightly different than the yearly average spending per filled day used to compute percent changes from 2012 to 2015 in Figure 4. In Figure 4, average cost was computed as total spending by people with MS on each DMT divided by the total number of filled days used by people with MS on each DMT in each year; this included data from all NDC codes and for all prescriptions filled for each DMT. 7 Limitations This study has several limitations that affect the interpretation of the findings presented. This brief presents per person spending and use trends for adults flagged as diagnosed with multiple sclerosis (MS) and those not flagged as being diagnosed with MS. Because this study was based on claims data, HCCI could not identify individuals with MS who did not file medical claims with their health insurer or had undiagnosed MS. Therefore, individuals identified in this dataset as having MS are by construction more likely to have any health care spending than the average adult with MS. It is important to note that over 60% of people in our sample appear in multiple years. Consequently, it is possible that the trends reported in this issue brief may reflect increases in the cost of care as individuals’ course of MS progresses as well as changes in the cost of care for all individuals with MS. While we see evidence that changes in health care spending in our sample were not driven by changes in the demographic make up of our sample, we do not test for this possibility directly. The findings in this study are descriptive and not causal. In particular, we did not account for the direction of the relationship between a MS diagnosis and spending. Further, while our results provide evidence, for example, that the cost of disease modifying therapies rose over our sample time period, we do not assign causality to changes in the cost of such drugs and changes in spending on them or their use. HCCI considers its work a starting point for analysis and research on health care spending for people diagnosed with MS. Throughout the issue brief, we focus on total health care spending – the sum of payer spending and individuals’ out-of-pocket spending. Total health care spending on prescription drugs cannot account for any rebates received by payers for prescription drugs. We also cannot account for any out-of-pocket assistance received by patients which may affect their true out-of-pocket burden. Consequently, the changes in total spending we report here may overstate the true changes in the cost of care payers face in covering individuals with MS, and individuals may face to cover the cost of their own care. Authors Contact Bill Johnson Health Care Cost Institute, Inc. 1100 G Street NW, Suite 600 Washington, DC 20005 202-803-5200 Copyright 2018 Health Care Cost Institute, Inc. Unless explicitly noted, the content of this report is licensed under a Creative Commons Attribution Non-Commercial No Derivatives 4.0 International License. 8 A1 Appendix: How did prices of DMTs change from 2009 to 2015? Changes in spending on older, injectable DMTs was primarily due to changes in their cost To understand changes in spending on each DMT we decomposed total spending per person into use (filled days per person) and average cost (total spending per filled day). Figure A1: Comparing Changes in Spending, Use, Average Cost of Injectable DMTs, 2009 to 2015 Source: Authors' analysis of HCCI claims data.  Use of Avonex, Betaseron, Copaxone, and Rebif declined from 2012 to 2015 (Figure A1, Table 3).  Despite the decline in their use, injectable DMTs still accounted for 60% of all DMT filled days in 2015 (Figure 3).  The average cost of each injectable DMT increased from 2012 to 2015. Spending per filled day increased an average of around 10% per year for each injectable DMT.  Due to the increase in their average cost, changes in spending on each older, injectable DMT did not match their decreased use from 2012 to 2015. 9 A2 Appendix: How did the average cost of older injectable DMTs change from 2009 to 2015? The average cost of injectable DMT rose similarly from 2009 to 2015 Figure A2: Comparing Spending per Filled Day of Older, Injectable DMTs to Newer, Oral DMTs (gray), 2009 to 2015 Source: Authors' analysis of HCCI claims data. Methods Note: Spending per filled day reported above are quarterly averages of total spending per filled day on the most common NDC code for each DMT. Values that aggregate fewer than 10 observations are censored. For more information, see the Methods section. 10 Table 1: Total Spending Per Person Diagnosed with MS by High Level Service Category, 2009 to 2015 Service Category 2009 2010 2011 2012 2013 2014 2015 Total $23,890 $25,184 $27,286 $29,564 $32,087 $35,050 $39,628 Medical Total $10,601 $10,907 $11,153 $11,391 $11,590 $11,562 $11,820 Inpatient $2,821 $3,149 $3,208 $3,325 $3,352 $3,421 $3,202 Outpatient $3,059 $3,209 $3,368 $3,506 $3,599 $3,694 $3,934 Professional $4,721 $4,550 $4,577 $4,560 $4,638 $4,448 $4,684 Services DMT Total $10,638 $11,309 $12,862 $14,685 $16,823 $19,507 $23,496 Injectable $9,420 $9,963 $10,731 $11,664 $11,766 $11,380 $12,432 Oral $0 $15 $583 $1,133 $3,014 $5,949 $8,450 Infused $1,219 $1,331 $1,548 $1,888 $2,043 $2,178 $2,613 All Other RX $2,651 $2,968 $3,270 $3,488 $3,675 $3,980 $4,312 Source: HCCI, 2018 11 Table 2: Total Spending Per Person Diagnosed with MS on Disease Modifying Therapies (DMTs), 2009 to 2015 (Excluding Infused DMTs) DMT: 2009 2010 2011 2012 2013 2014 2015 DMT Total $9,420 $9,978 $11,314 $12,797 $14,780 $17,330 $20,882 (Excluding Infused) DMT Injectable $9,420 $9,963 $10,731 $11,664 $11,766 $11,380 $12,432 Avonex $2,719 $2,671 $2,726 $3,077 $3,138 $2,984 $2,811 Betaseron $1,221 $1,213 $1,200 $1,194 $1,094 $973 $1,154 Copaxone $3,634 $4,207 $4,824 $5,299 $5,342 $5,399 $6,163 Rebif $1,845 $1,855 $1,963 $2,076 $2,181 $2,006 $1,968 Other DMT $1 $19 $18 $18 $11 $19 $336 DMT Oral $15 $583 $1,133 $3,014 $5,949 $8,450 Aubagio $10 $288 $646 $1,159 Gilenya $15 $583 $1,122 $1,496 $1,880 $2,585 Tecfidera $1,229 $3,424 $4,707 Source: HCCI, 2018 12 Table 3: Prescription Filled Days Per Person Diagnosed with MS of Disease Modifying Therapies (DMTs), 2009 to 2015 DMT: 2009 2010 2011 2012 2013 2014 2015 DMT Total 112 102 99 95 95 100 110 (Excluding Infused) DMT Injectable 112 102 95 87 76 66 66 Avonex 33 28 25 24 21 18 15 Betaseron 14 12 11 9 7 6 6 Copaxone 43 42 41 38 34 31 33 Rebif 22 20 18 16 14 11 10 Other DMT 0 0 0 0 0 0 2 DMT Oral 4 7 19 35 44 Aubagio 0 2 4 6 Gilenya 0 4 7 9 10 13 Tecfidera 8 20 25 Source: HCCI, 2018 13 Table 4: Average Cost (Total Spending Per Filled Day) of the most common NDC code for each DMT by Quarter, 2009 to 2015 Year Quarter Aubagio Avonex Betaseron Copaxone Gilenya Rebif Tecfidera 2009 1 $80 $80 $78 $80 2009 2 $85 $85 $85 $83 2009 3 $85 $91 $86 $86 2009 4 $85 $90 $86 $88 2010 1 $93 $90 $94 $90 2010 2 $96 $96 $98 $91 2010 3 $99 $100 $103 $95 2010 4 $101 $105 $103 $136 $95 2011 1 $106 $105 $117 $135 $101 2011 2 $109 $113 $118 $133 $103 2011 3 $112 $113 $118 $133 $112 2011 4 $113 $119 $118 $140 $113 2012 1 $122 $128 $135 $145 $121 2012 2 $129 $130 $136 $146 $129 2012 3 $132 $136 $136 $160 $132 2012 4 $130 $140 $140 $143 $161 $135 2013 1 $127 $143 $150 $157 $168 $149 2013 2 $128 $144 $152 $157 $169 $155 $153 2013 3 $139 $153 $158 $156 $169 $161 $152 2013 4 $149 $159 $158 $156 $169 $165 $153 2014 1 $161 $165 $162 $170 $177 $175 $166 2014 2 $168 $165 $171 $171 $178 $175 $167 2014 3 $172 $167 $177 $177 $178 $176 $168 2014 4 $178 $174 $188 $183 $187 $175 2015 1 $185 $173 $196 $205 $187 $191 $177 2015 2 $185 $181 $194 $205 $196 $196 $185 2015 3 $188 $181 $206 $192 $200 $190 2015 4 $197 $192 $205 $197 $209 $197 Source: HCCI, 2018 14 Table 5: Average Cost (Total Spending) Per Month’s Supply of the most common NDC code for each DMT by Quarter, 2009 to 2015 Year Quarter Aubagio Avonex Betaseron Copaxone Gilenya Rebif Tecfidera 2009 1 $2,233 $2,243 $2,326 $2,233 2009 2 $2,384 $2,385 $2,540 $2,327 2009 3 $2,369 $2,539 $2,570 $2,394 2009 4 $2,390 $2,532 $2,572 $2,451 2010 1 $2,594 $2,534 $2,805 $2,526 2010 2 $2,678 $2,690 $2,935 $2,553 2010 3 $2,786 $2,797 $3,092 $2,652 2010 4 $2,820 $2,944 $3,091 $3,804 $2,648 2011 1 $2,961 $2,937 $3,520 $3,775 $2,829 2011 2 $3,041 $3,158 $3,541 $3,735 $2,897 2011 3 $3,140 $3,161 $3,541 $3,720 $3,136 2011 4 $3,174 $3,330 $3,553 $3,926 $3,162 2012 1 $3,426 $3,583 $4,045 $4,049 $3,385 2012 2 $3,614 $3,640 $4,082 $4,080 $3,603 2012 3 $3,688 $3,810 $4,084 $4,491 $3,682 2012 4 $3,642 $3,913 $3,913 $4,285 $4,501 $3,790 2013 1 $3,557 $4,003 $4,188 $4,699 $4,710 $4,163 2013 2 $3,594 $4,020 $4,249 $4,723 $4,731 $4,350 $4,595 2013 3 $3,887 $4,291 $4,416 $4,689 $4,734 $4,506 $4,573 2013 4 $4,160 $4,464 $4,422 $4,692 $4,734 $4,634 $4,583 2014 1 $4,494 $4,632 $4,546 $5,109 $4,945 $4,914 $4,976 2014 2 $4,698 $4,631 $4,802 $5,131 $4,986 $4,913 $5,016 2014 3 $4,829 $4,675 $4,970 $5,313 $4,987 $4,916 $5,025 2014 4 $4,987 $4,872 $5,629 $5,114 $5,246 $5,249 2015 1 $5,172 $4,856 $5,480 $6,144 $5,231 $5,346 $5,313 2015 2 $5,173 $5,058 $5,440 $6,140 $5,496 $5,487 $5,540 2015 3 $5,264 $5,080 $6,169 $5,386 $5,596 $5,692 2015 4 $5,515 $5,386 $6,151 $5,503 $5,851 $5,898 Source: HCCI, 2018 15