ARE PHYSICIAN DISPENSING REFORMS SUSTAINABLE?

Transcription

1 ARE PHYSICIAN DISPENSING REFORMS SUSTAINABLE? Dongchun Wang Vennela Thumula Te-Chun Liu WC January 2015 WORKERS COMPENSATION RESEARCH INSTITUTE CAMBRIDGE, MASSACHUSETTS 1

2 COPYRIGHT 2015 BY THE WORKERS COMPENSATION RESEARCH INSTITUTE ALL RIGHTS RESERVED. NO PART OF THIS BOOK MAY BE COPIED OR REPRODUCED IN ANY FORM OR BY ANY MEANS WITHOUT WRITTEN PERMISSION OF THE WORKERS COMPENSATION RESEARCH INSTITUTE. ISBN PUBLICATIONS OF THE WORKERS COMPENSATION RESEARCH INSTITUTE DO NOT NECESSARILY REFLECT THE OPINIONS OR POLICIES OF THE INSTITUTE S RESEARCH SPONSORS. 2

3 ACKNOWLEDGMENTS Our gratitude goes to the two technical reviewers of the draft report, Dr. Gary Franklin and Mr. Frank Neuhauser. Their comments and suggestions not only helped us to improve the accuracy and clarity of the final report, but are also valuable for future studies. We also thank several other reviewers, Dr. Rochelle Henderson, Kristine Kennedy, Susan Piha, Maddy Bowling, and Joan Vincenz, for their comments, which helped improve the readability of the final report. Our colleague, Stephanie Deeley, contributed to the update of state policies and recent legislative and regulatory activities regarding fee schedules for workers compensation pharmaceuticals and related areas. We also benefited from timely updates of policy changes by Brian Allen and his colleagues at Helios and helpful comments provided in the past by Tron Emptage, Angela Jenkins, Kevin Tribout, and Matt Engels that helped us better understand pharmaceutical policies and practices. We are grateful for their generous contributions. Critical to the study was the indispensable assistance from Dr. Philip Borba and his team at Milliman, Inc. and Eric Harrison, Arlene Abueg, and other colleagues at WCRI. Their contributions, including pharmacy database construction, programming support, and quality assurance, made the study possible. We also thank Andrew Kenneally, the communications director at WCRI, for his fruitful effort in disseminating the research findings. Special thanks to Dr. Richard Victor, executive director of the Institute, for his invaluable input and guidance throughout the whole process. We wish to thank Sarah Solorzano and Callison Lawson for their superior administrative assistance that helped to improve the readability and accuracy of the report. Sarah Solorzano also managed the review and publication process. Of course, any errors or omissions that remain in the report are the responsibility of the authors. Dongchun Wang Vennela Thumula Te-Chun Liu Cambridge, Massachusetts January

4 EXECUTIVE SUMMARY Since 2007, 18 states enacted reforms to reduce the costs of physician-dispensed drugs. Most of these state reforms targeted the much higher prices paid to physicians for drugs they dispense, compared with filling the same prescriptions at a pharmacy. Often these price-focused reforms limited the reimbursements to an amount that was similar to the average wholesale price (AWP) assigned by the original manufacturer of the drug. New reimbursement rules like these help reduce prices paid for physician-dispensed drugs because prior to these reforms, physicians often dispense repackaged drugs at much higher prices. This is so because an intermediary (or repackager) often buys the original drug in bulk and repackages it in small quantities. The intermediary subsequently obtains a new National Drug Code (NDC) for the repackaged drug and assigns a new AWP to the repackaged drug, which is almost always much higher than the AWP assigned by the original manufacturer of that drug. Workers Compensation Research Institute (WCRI) studies have shown that the average prices paid for physician-dispensed drugs decreased after the reforms, often by large margins, as expected. However, many studies, including some by WCRI, have shown that when prices are reduced by regulation, the regulated parties sometimes find new ways to retain the high revenues they had prior to the reforms. This study identifies one way in which this seems to have arisen in two large states. The lessons should also be of interest in all other states where physician dispensing is permitted. The mechanism involves the creation of an opportunity to, once again, assign a much higher AWP to a physician-dispensed drug a practice targeted by the earlier reforms enacted in many states using language limiting reimbursement to a price based on the AWP of the original manufacturer. How can a new and higher AWP be set for physician-dispensed drugs? Consider a drug where the most common strengths are 5 milligrams and 10 milligrams. If a new strength, say 7.5 milligrams, comes to market, the original manufacturer of that new strength can assign a new AWP, and this AWP could be much higher than the 5-milligram and 10-milligram AWPs set by their original manufacturers. That is exactly what is seen in the data. For example, prior to the reforms in Illinois, there were two common strengths when cyclobenzaprine HCL was prescribed 5 and 10 milligrams. The average prices paid for cyclobenzaprine HCL of 5 and 10 milligrams ranged from $0.99 to $1.74 per pill. Prior to 2012, 7.5-milligram cyclobenzaprine HCL was rarely seen in the market. The 7.5-milligram products were introduced in 2012 and almost all were dispensed by physicians at an average price of $3.79 per pill in post-reform Illinois. The market share of physician-dispensed cyclobenzaprine HCL of 7.5 milligrams increased from 0 percent in the third quarter of 2012 to 21 percent in the first quarter of In California prior to 2012, 7.5-milligram cyclobenzaprine HCL was also rarely seen in the market. The average prices paid for 5- and 10-milligram cyclobenzaprine HCL, the two common strengths, ranged from $0.35 to $0.70 per pill. Since the introduction of the 7.5-milligram products in 2012, the market share of physician-dispensed cyclobenzaprine HCL of 7.5 milligrams increased from 0 percent in the fourth quarter of 2011 to 47 percent in the first quarter of 2013, when it became the strength of the drug most commonly dispensed by physicians. The average price paid for the new strength was $2.90 to $3.45 per pill. We found similar evidence in the data for a new strength of tramadol HCL (150 milligrams extended release), which appeared in By the end of 2012, the new strength of tramadol HCL extended release became one of the most common physician-dispensed drugs. The average price paid in California for the new 4

5 strength was $5.94 to $7.41 per pill, compared with $1.58 per pill for the same extended-release drug of other strengths. Because these new-strength drug products were almost all dispensed by physicians at much higher prices, we infer that the shift in strength was unlikely to be driven by new evidence about superior medical practices. Rather, it is likely that financial incentives drove some physicians to choose the strength for their patients. 5

6 1 INTRODUCTION As many states implemented changes to reimbursement rules to address additional costs of physician dispensing by capping the prices paid for physician-dispensed repackaged drugs, 1 the question arises as to whether the reforms are effective and sustainable in reducing prices paid for physician-dispensed prescriptions. Some evidence from post-reform California and Illinois suggests that an adaptive behavior in response to the new reimbursement rules presents a challenge for this type of physician dispensing reform. While Workers Compensation Research Institute (WCRI) studies have found substantial price reductions for many common physician-dispensed drugs across several post-reform states studied, 2 we identified an emerging issue of physicians dispensing new-strength drug products to injured workers at much higher prices compared with the same drugs of existing strengths. This was evident in post-reform California and Illinois and was the major force behind substantial increases in the average price paid for several common physician-dispensed drugs in the two states. These new-strength drug products, almost all dispensed by physicians, were paid at much higher prices, due to higher manufacturer average wholesale prices (AWPs) reported by the original manufacturer. This result may be explained by two reasons. First, these new-strength drug products were produced by the original manufacturer, not repackager, and therefore may not be targeted by the reforms in most states. Second, because these products had new strengths, no therapeutic equivalents could be found that would be associated with lower reimbursement rates. 3 Evidence of behavioral changes in response to price regulations has been provided by several studies. The changes in behavior of the regulated parties are often related to finding new ways to maintain the revenues they had prior to the reforms. A previously published WCRI study found an increased utilization of certain medical services and a shift to code more complex office visits in Tennessee after the introduction of the medical fee schedule in 2004 (Radeva et al., 2009). The authors of that report concluded that having substantial changes in prices paid may lead to changes in utilization of medical services. Evidence of physicians increasing utilization of medical services in response to a reimbursement change that has the effect of reducing prices paid has also been reported based on Medicare data (Jacobson and Newhouse, 2010; 1 As of January 2015, 18 states have made changes to the rules governing reimbursement to physicians for prescriptions that they write and dispense. Most of these reforms were aimed at reducing the additional costs of physician-dispensed prescriptions, without restricting physicians ability to dispense prescription drugs. Almost all of the reforms tie the reimbursement rates for physician-dispensed prescriptions to the AWP of the original drug used in the repackaging process. In several reform states, the maximum reimbursement amount is set to the lesser of the AWP of the original drug or the lowest price of the therapeutic equivalents. See Wang, Thumula, and Liu (2014a c) for more details. 2 See Wang, Liu, and Thumula (2014) and Wang, Thumula, and Liu (2014a c). 3 Reimbursement rates are tied to Medi-Cal rates in California. In Illinois, the reimbursement rates are tied to the AWPs of the original drugs. 6

7 Jacobson, Earle, and Newhouse 2011). The most relevant behavioral changes were observed in a study by the California Workers Compensation Institute. That study reported a rapid increase in use of higher-priced pharmaceuticals like compounds, co-packs, and medical foods after California adopted regulations to cap prices paid for repackaged drugs in 2007 (Ireland and Swedlow, 2010). According to the study, the prescription payments made for these products increased from 2.3 percent of total prescription drug costs in 2006 to 12 percent in In this report, we provide evidence that recent reforms are being challenged by use of higher-priced new drug products that have become commonly dispensed by physicians in post-reform California and Illinois. We describe a substantial increase in use and prices paid for three drug products. The three drug products are milligrams of hydrocodone-acetaminophen, 7.5 milligrams of cyclobenzaprine HCL, and 150 milligrams of tramadol HCL extended release. Hydrocodone-acetaminophen (brand names include Vicodin and Lortab ) and tramadol HCL (Ultram ) are the most common opioid drugs prescribed to injured workers for pain relief. 4 These two opioid drug products, approved by the U.S. Food and Drug Administration (FDA) in 2012, have new strengths that did not exist before. Cyclobenzaprine HCL (Flexeril ) is a muscle relaxant and often prescribed for treating musculoskeletal injuries and conditions. The 7.5-milligram strength was first introduced in 2006, and 7.5-milligram cyclobenzaprine HCL products have been produced by different manufacturers, but were rarely seen in the workers compensation market until We also discuss the impact of increased physician dispensing of these new drug products on price reduction in post-reform California and Illinois and the implications for other reforms states. Discussions of clinical benefits of these new strengths and overall cost implications are outside the scope of this report. The report is organized into five chapters. Chapter 2 describes the data and methods we used for this study. Chapter 3 provides the results that illustrate the emergence of two new strengths among physiciandispensed drugs in California. Chapter 4 presents the evidence of higher-priced new-strength drug products in post-reform Illinois. In the last chapter, we discuss some of the implications of the new strengths on price reductions. 4 The FDA recently reclassified hydrocodone-combined drug products (including hydrocodone-acetaminophen) from Schedule III to Schedule II and classified tramadol HCL as Schedule IV. 5 The 7.5-milligram cyclobenzaprine HCL that is often seen in the data is a drug product approved by the FDA in

8 2 DATA AND APPROACH The data used for this report came from payors that represented 46 and 51 percent of all medical claims, respectively, for California and Illinois. The detailed prescription transaction data were organized by calendar quarter so that for each quarter, all prescriptions filled for claims with dates of injury within 24 months of the observation quarter were included. 1 On average for each of the quarters reported, we included 219,572 prescriptions paid for 60,448 claims in California. The same figures were 43,034 prescriptions paid for 12,714 claims for Illinois. The detailed prescription data cover calendar quarters from the first quarter of 2010 though the first quarter of For the three drug products we focused on in this report that were dispensed by physicians, we have in our sample 5,169 prescriptions in the last quarter of 2012 and first quarter of 2013 for California and 709 prescriptions in the first quarter of 2013 for Illinois. 2 The prescription transactions included in this study were those for medications of prescription strengths and over-the-counter strengths (referred to as prescriptions throughout the report). These prescriptions could be filled or refilled by the injured worker at a pharmacy or physician s office and were paid for under workers compensation. Prescription medications that were dispensed at a hospital or administered parenterally by a medical provider (e.g., injections received at the physician s office) were excluded. Also excluded were nutrition supplements and medical supplies/equipment that were billed under National Drug Codes (NDCs). Physician-dispensed prescriptions were defined as those that were filled at the offices of independent practitioners, physician groups, or medical centers or clinics which may or may not have an on-site pharmacy. 3 Pharmacy-dispensed prescriptions were those dispensed at retail or mail-order pharmacies. 1 We constructed the quarterly data to only include prescriptions received by claims with dates of injury within 24 months of the observation quarter. By doing so, we hold the mix of claims with longer or shorter maturity (i.e., experience with prescriptions) constant across calendar quarters. We chose to use a 24-month cut off for constructing the quarterly data so that we could evaluate pre- and post-reform results for most of the reform states. Note that the frequency of physician dispensing and prices for common drugs may be different when imposing a longer claim maturity (e.g., 36 months). However, since we evaluate reforms by focusing on changes across the calendar quarters, the difference in the cut-off of claim maturity is unlikely to affect the findings of this study. 2 For California, the number of prescriptions used for the two quarters was 328 for milligrams of hydrocodoneacetaminophen, 2,469 for 7.5 milligrams of cyclobenzaprine HCL, and 2,372 for 150 milligrams of tramadol HCL extended release. The same figures in the latest quarter for Illinois were 310, 184, and 215, respectively. 3 We identified physician-dispensed prescriptions based on several critical data elements, including (1) the information provided by the payors that indicates if a physician or pharmacy was the provider of the medication; (2) the Medi-Span indicator that specifies repackaged drugs using the NDCs assigned for a repackaged drug by the U.S. Food and Drug Administration; and (3) the place of service. By our definition, prescriptions dispensed at and billed for by a medical center or clinic are considered physician-dispensed prescriptions. This is because, although the medical center may have an on-site pharmacy that functions like a retail pharmacy, prescriptions dispensed and billed for by the medical center as a financial entity are often reimbursed differently compared with retail pharmacies. Note that by our definition, prescriptions for repackaged drugs are considered physician-dispensed prescriptions. Based on the evidence we saw in our detailed data review and our understanding of the incentive mechanism in the business process, we believe that retail pharmacies rarely dispense repackaged drugs in practice. 8

9 We used the average price per pill for the medication dispensed by physicians or at pharmacies for the price comparison. The price per pill is computed by taking the amount paid for a prescription divided by the number of pills of the prescription. 4 When aggregated across prescriptions for the same medication, the price per pill takes into account the difference in the number of pills, making the price per pill more comparable between physician- and pharmacy-dispensed prescriptions. Note that the average price per pill paid for individual drugs reported is the price paid for generic versions of the drug. We do so for two reasons. First, a vast majority of physician-dispensed drugs are generics, but a considerable number of drugs dispensed at pharmacies are brand names, which may or may not be among those commonly dispensed by physicians. If the average price per pill is computed regardless of generic or brand-name status of the drug, the results may over-state pharmacy price substantially overall and for some drugs. Second, for common physician-dispensed drugs, the price comparison for generics is valid because most of the common physician-dispensed drugs are generics regardless of dispensing point. Although infrequent, outlier prices can be seen in the distribution of price per pill for a given drug, after controlling for drug name, dispensing point, generic or brand-name status of the drug, time period during which it was dispensed, and state of jurisdiction. These outlier values were likely due to key punch errors in the variables underlying the computation of price per pill (mainly amount paid and quantity). In our data, these were seen within the top 1 percent of the distribution for most of the common drugs and up to 3 percent for a few drugs. However, these outlier values may have a considerable impact on the accuracy of the average price per pill because of the large values. To address this issue, we established a set of thresholds in price per pill for individual drugs based on the percentile distribution controlling for the key variables described above, by taking 150 percent of the value at the 97th percentile. This set of thresholds was subsequently used to cap the price per pill for prescriptions of specific drugs. 5 We used the Medi-Span price history data to obtain the AWPs for unique NDCs, which were used in the analysis. It should also be noted that although the numbers reported here may be somewhat different from those previously published, the major findings and supporting data are consistent between this study and the previous WCRI studies on physician dispensing. The reasons for different numbers include (1) a larger sample of data, including the expansion to cover all medical claims (instead of claims with more than seven days of lost time); (2) different construction of the analysis data; (3) enhancements to the data and methodology (e.g., capping outlier prices); and (4) incorporation of our newly updated Medi-Span database. 4 A vast majority of the drugs that are commonly dispensed by physicians are in the form of tablets or capsules. 5 Because the underlying variables are amount paid and quantity, we implemented the capping so that if the amount paid for a prescription is greater than a converted value (i.e., the product of the price-per-pill threshold for capping and the quantity or number of pills for the prescription) for the drug in a given group, the amount paid is capped at the converted value. This way, the price capping will have an effect on other related prescription benchmark measures to maintain internal consistency. Note that the 1.5*97th percentile thresholds were directly applied to the three drugs included in this study. For a few other drugs reported in other WCRI prescription benchmarks studies, the value at the 97th percentile could still be unusually high, resulting in some residual outliers after the 1.5*97th percentile capping. In these cases, we checked the data to make sure that it was not caused by a data anomaly due to processing. Once we ruled out a systematic error, we adjusted the capping thresholds by using the maximum value of the capping thresholds for the same drug separately for generic and brand-name drug products. 9

10 3 HIGH-PRICED NEW STRENGTHS IN CALIFORNIA In 2007, California made changes to the pharmacy fee schedule that capped the maximum reimbursement amount for physician-dispensed drugs at the same level as for the same drugs dispensed at pharmacies. A previous WCRI study reported that California s rule changes resulted in substantial price reductions, and during two years post-reform (in ), the prices per pill paid for most physician-dispensed common drugs were close to or slightly above the prices paid to pharmacies for the same drugs. 1 The price differentials between physician- and pharmacy-dispensed prescriptions for most common physician-dispensed drugs remained small in post-reform California until the first quarter of Figure 3.1 shows a rapid increase in the average price per pill paid for physician-dispensed cyclobenzaprine HCL (a muscle relaxant with common brand names including Flexeril ), from $0.43 per pill in the fourth quarter of 2011 to $2.11 per pill in the first quarter of The large increase in the price for physician-dispensed cyclobenzaprine HCL was driven by frequent dispensing of 7.5-milligram cyclobenzaprine HCL with higher prices. Before generic cyclobenzaprine HCL of 7.5 milligrams was introduced, there were two prescription strengths for the drug 5 milligrams and 10 milligrams. First introduced in 2006, 7.5-milligram cyclobenzaprine HCL products have been produced by different manufacturers, but rarely seen in the workers compensation market until Figure 3.2 shows a large increase in physician dispensing of 7.5- milligram cyclobenzaprine HCL starting in the first quarter of 2012, which was coupled with large decreases in the frequency of physician dispensing of the same drug of the two existing strengths. By the first quarter of 2013, prescriptions for 7.5-milligram cyclobenzaprine HCL increased to 47 percent of all physician-dispensed cyclobenzaprine HCL prescriptions. When 7.5-milligram cyclobenzaprine HCL prescriptions were dispensed by physicians in California, the price per pill paid was much higher than that for the existing strengths, as shown in Figure Since California s pharmacy fee schedule is tied to Medi-Cal, the state Medicaid fee schedule, the maximum reimbursement amount for a drug, regardless of dispensing point, is substantially lower than the AWP of the drug. See Wang (2012) for a more detailed discussion of the post-reform experience for California. 10

11 Figure 3.1 Average Price per Pill Paid for Physician- and Pharmacy-Dispensed Cyclobenzaprine HCL in California Average Price per Pill Paid for Physician- and Pharmacy-Dispensed Rx $2.50 $2.00 $1.50 $1.00 $0.50 $ Q1 2010Q2 2010Q3 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 2012Q2 2012Q3 2012Q4 2013Q1 Physician-Dispensed Rx Pharmacy-Dispensed Rx Notes: The underlying data include prescriptions filled from 2010Q1 to 2013Q1 for all medical claims that had injuries occurring within two years prior to a given quarter. 2013Q1 refers to the first quarter of Similar notation is used for other quarters. California made changes to the pharmacy fee schedule in March 2007 that set the maximum reimbursement for physician-dispensed prescriptions at the same level as for the same prescription dispensed at a pharmacy. The data reported for all calendar quarters are post-reform for California. Key: Rx: prescriptions. Figure 3.2 Distribution of Physician-Dispensed Prescriptions for Cyclobenzaprine HCL in California, by Strength 80% % of Prescriptions for Cyclobenzaprine HCL, by Strength 70% 60% 50% 40% 30% 20% 10% 0% 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 2012Q2 2012Q3 2012Q4 2013Q1 7.5 Milligrams Price per Pill Paid: $2.90 $ Milligrams Price per Pill Paid: $0.35 $ Milligrams Price per Pill Paid: $0.49 $0.70 Notes: The underlying data include prescriptions filled from 2011Q1 to 2013Q1 for all medical claims that had injuries occurring within two years prior to a given quarter. 2013Q1 refers to the first quarter of Similar notation is used for other quarters. California made changes to the pharmacy fee schedule in March 2007 that set the maximum reimbursement for physician-dispensed prescriptions to the same level as for the same prescription dispensed at a pharmacy. The data reported for all calendar quarters are post-reform for California. 11

12 For physician-dispensed tramadol HCL (Ultram ), an opioid drug classified by the Drug Enforcement Administration (DEA) as a Schedule IV controlled substance, 2 the average price per pill paid to physicians also increased dramatically in California after a period of time during which the prices paid to physicians for the drug were similar to pharmacy prices for the same drug (Figure 3.3). Figure 3.3 Average Price per Pill Paid for Physician- and Pharmacy-Dispensed Tramadol HCL in California $3.50 Average Price per Pill Paid for Tramadol HCL, by Dispensing Point $3.00 $2.50 $2.00 $1.50 $1.00 $0.50 $ Q1 2010Q2 2010Q3 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 2012Q2 2012Q3 2012Q4 2013Q1 Physician-Dispensed Rx Pharmacy-Dispensed Rx Notes: The underlying data include prescriptions filled from 2010Q1 to 2013Q1 for all medical claims that had injuries occurring within two years prior to a given quarter. 2013Q1 refers to the first quarter of Similar notation is used for other quarters. California made changes to the pharmacy fee schedule in March 2007 that set the maximum reimbursement for physician-dispensed prescriptions to the same level as for the same prescription dispensed at a pharmacy. The data reported for all calendar quarters are post-reform for California. Key: Rx: prescriptions. Behind the large price increase for physician-dispensed tramadol HCL was a new drug product, 150 milligrams extended release. 3 Table 3.1 shows that in the fourth quarter of 2012 and the first quarter of 2013, about half of the physician-dispensed tramadol HCL prescriptions shifted from 50 milligrams regular release to 150 milligrams extended release. Note that 150-milligram extended release prescriptions were rarely dispensed at pharmacies. Over the same period, prescriptions for other strengths of tramadol HCL extended release (e.g., 100, 200, and 300 milligrams) filled at pharmacies remained at 3 4 percent of all pharmacy- 2 There are five schedules of controlled substances classified by the DEA under federal law. They are based on a drug s medical usefulness and abuse potential. Other examples of scheduled opioids include oxycodone HCL (OxyContin ), oxycodone-acetaminophen (Percocet ), and hydrocodone-acetaminophen (Vicodin ), which are all classified as Schedule II. 3 This new drug product was introduced to the market in 2012 and almost all prescriptions for this drug product were dispensed by physicians. 12

13 dispensed prescriptions for tramadol HCL (Table 3.1). 4 Table 3.1 Distributions of and Prices Paid for Physician- and Pharmacy-Dispensed Tramadol HCL of Different Strengths in California Average Price per Pill Paid for the % of Rx for Tramadol HCL by Drug Product Drug Products Drug Product 2012Q1 2012Q2 2012Q3 2012Q4 2013Q1 2012Q1 2012Q2 2012Q3 2012Q4 2013Q1 Physician-dispensed Rx Extended release 150 milligrams s/c $6.09 $6.79 $5.94 $7.41 0% 2% 31% 51% 49% Regular release 50 milligrams $0.34 $0.30 $0.28 $0.29 $ % 98% 69% 49% 51% Pharmacy-dispensed Rx Extended release 150 milligrams n/d n/d n/d n/d n/d n/d n/d n/d n/d n/d Extended release 100, 200, or 300 milligrams $1.58 $1.58 $1.58 $1.57 $1.58 4% 3% 3% 3% 4% Regular release 50 milligrams $0.27 $0.27 $0.27 $0.26 $ % 97% 97% 97% 96% Notes: The underlying data include prescriptions filled from 2012Q1 to 2013Q1 for all medical claims that had injuries occurring within two years prior to a given quarter. 2013Q1 refers to the first quarter of Similar notation is used for other quarters. California made changes to the pharmacy fee schedule in March 2007 that set the maximum reimbursement for physiciandispensed prescriptions to the same level as for the same prescription dispensed at a pharmacy. The data reported for all calendar quarters are post-reform for California. Key: n/d: not seen in the data; Rx: prescriptions; s/c: small cell size (fewer than 30 prescriptions). As mentioned earlier, we focused on three new drug products in this study. The third drug product was milligrams of hydrocodone-acetaminophen (Vicodin ), which had a substantial increase among physician-dispensed prescriptions in post-reform Illinois (discussed in the next section). In California, we observed an increase in use of this new-strength drug, but the increase was moderate compared with Illinois. We do not know exactly what may explain this, but observations we made suggest several reasons for the moderate use of milligrams of hydrocodone-acetaminophen in California, including how frequent hydrocodone-acetaminophen was dispensed by physicians prior to the introduction of the new strength, how often higher strength drug products were used, as well as how the workers compensation medical care delivery system is organized and the distribution mechanism for drugs. 4 One benefit of having an extended-release drug may be for the patient to take fewer pills in a day. If this is the reason for the shift, one would expect to see similar doses per claim before and after the shift. We observed that the average dose per claim for tramadol HCL remained similar before and after physicians shifted to dispense the 150-milligram product in California, but the same figure increased in post-reform Illinois. 13

14 4 EARLY EVIDENCE OF HIGH-PRICED NEW STRENGTHS IN ILLINOIS We saw the same issue in the first four months after Illinois reforms. 1 Since Illinois rule changes went into effect in the fourth quarter of 2012, one would expect to see a substantial price reduction for most common physician-dispensed drugs in the first quarter of We saw large price reductions for most of the top drugs dispensed by physicians. However, the average price per pill paid to physicians increased by 22 percent for hydrocodone-acetaminophen (one of the most commonly-prescribed opioid pain medications with brand names including Vicodin ), 45 percent for cyclobenzaprine HCL (a muscle relaxant with brand names including Flexeril ), and 152 percent for tramadol HCL (a commonly-prescribed opioid pain medication with brand names including Ultram ), between the third quarter of 2012 pre-reform and the first quarter of 2013 post-reform (Table 4.1). Note that over the same period, the pharmacy prices for the same drugs remained stable. The large increases in prices paid for these physician-dispensed drugs were mainly because of the same drugs with new strengths that were introduced into the market and frequently dispensed by some physiciandispensers at much higher manufacturer prices, compared with the AWPs of the manufacturer drugs with existing strengths. Table 4.2 shows prescription distributions by strength for each of these three common physician-dispensed drugs. The average price per pill paid for prescriptions of the drug with a given strength is also provided in the table, for the third quarter of 2012 (pre-reform) and the first quarter of 2013 (postreform). The three new drug products seen in the data are milligrams of hydrocodone-acetaminophen, approved by the FDA in 2012, 7.5 milligrams of cyclobenzaprine HCL, approved by the FDA in 2011, and 150 milligrams of tramadol HCL extended release, approved by the FDA in Table 4.2 shows that the frequency of physicians prescribing these new-strength drug products increased dramatically after Illinois reforms. The number of physician-dispensed new-strength hydrocodone-acetaminophen prescriptions increased from almost none in the third quarter of 2012 (pre-reform) to 25 percent of prescriptions for the drug (across all different strengths) in the first quarter of 2013 (post-reform). The same figure increased to 21 percent for cyclobenzaprine HCL and to 29 percent for tramadol HCL. When dispensed at physicians offices, these three drug products were also paid at higher prices compared with what was paid to physicians for the same drugs with other existing strengths. In the first 1 In our recent physician dispensing studies, we examined the impact of physician dispensing reforms for the states for which we have post-reform data: Arizona, California, Connecticut, Georgia, Illinois, Michigan, Oklahoma, South Carolina, and Tennessee. Note that we did not see similar issues in the other reform states studied, but we caution the reader that the issue could emerge in any of the states. 14

15 quarter of 2013, for example, the average price paid for physician-dispensed hydrocodone-acetaminophen of milligrams was $3.04 per pill, double to quadruple the prices paid to physicians for the same drug of other existing strengths (Table 4.2). Because these new-strength drug products were produced by a manufacturer, not repackager, and the new strengths did not have therapeutic equivalents in the market that had lower AWPs, it was not surprising to see that these drug products were paid for at much higher prices, compared with the same drug with other existing strengths. 2 Like cyclobenzaprine HCL and tramadol HCL in California, these drug products with new strengths were rarely dispensed at pharmacies. Table 4.1 Average Price per Pill Paid for Physician- and Pharmacy-Dispensed Drugs, before and after Illinois' Reform a Rx Drugs Commonly Dispensed by Physicians (70% of all MDRx in pre-reform Illinois) 2012Q1 (pre-reform) Average Price per Pill Paid for the Drugs 2012Q2 (pre-reform) 2012Q3 (pre-reform) 2012Q4 (transition) 2013Q1 (post-reform) Percentage Change 2012Q3 2013Q1 Hydrocodone-acetaminophen (Vicodin ) Physician-dispensed Rx $1.21 $1.26 $1.31 $1.19 $ % Pharmacy-dispensed Rx $0.53 $0.53 $0.53 $0.52 $0.52-2% Cyclobenzaprine HCL (Flexeril ) Physician-dispensed Rx $1.58 $1.64 $1.73 $1.71 $ % Pharmacy-dispensed Rx $1.02 $1.00 $0.99 $0.98 $0.98-1% Tramadol HCL (Ultram ) Physician-dispensed Rx $1.42 $1.48 $1.43 $1.53 $ % Pharmacy-dispensed Rx $0.76 $0.76 $0.76 $0.76 $0.75-1% Notes: The underlying data include prescriptions filled from 2012Q1 to 2013Q1 for all medical claims that had injuries occurring within two years prior to a given quarter. 2013Q1 refers to the first calendar quarter of Similar notation is used for other quarters. a Effective November 20, 2012, Illinois changed the reimbursement rules to set the prices for physician-dispensed prescription drugs to the average wholesale price of the original drug used in the repackaging process and explicitly require that dispensing physicians provide the National Drug Code of the underlying drug. Data for 2012Q1 2012Q3 are pre-reform for Illinois. Data for 2012Q4 reflect a mix of pre- and post-reform experience. Calendar quarter 2013Q1 is the first full post-reform quarter for Illinois, which is the latest quarter of the study period. Key: MDRx: physician-dispensed Rx; Rx: prescriptions. 2 We saw that the prices paid for these drug products were close to the AWPs of original manufacturers. The AWPs for the same drugs of other existing strengths were much lower. 15

16 Table 4.2 Distribution of and Price per Pill Paid for Three Common Drugs, by Strength for Physician- and Pharmacy-Dispensed Prescriptions in Illinois a Hydrocodone-acetaminophen (Vicodin ) 2012Q3 (pre-reform) 2012Q4 (transition) 2013Q1 (post-reform) 2012Q3 (pre-reform) 2013Q1 (post-reform) Physician-dispensed Rx b,c milligrams n/d <1% 25% n/d $ milligrams 38% 39% 32% $1.29 $ milligrams 15% 14% 10% $1.28 $ milligrams 11% 11% 7% $1.20 $ milligrams 31% 33% 25% $1.43 $0.94 Other strengths 4% 3% 2% n/a n/a Pharmacy-dispensed Rx b milligrams n/d n/d n/d n/d n/d milligrams 33% 34% 38% $0.54 $ milligrams 15% 13% 10% $0.40 $ milligrams 13% 12% 13% $0.57 $ milligrams 34% 35% 34% $0.61 $0.57 Other strengths 5% 6% 5% n/a n/a Cyclobenzaprine HCL (Flexeril ) Physician-dispensed Rx 5 milligrams 6% 4% 6% $1.74 $ milligrams b n/d 2% 21% n/d $ milligrams 94% 94% 73% $1.70 $1.29 Pharmacy-dispensed Rx 5 milligrams 15% 15% 15% $1.43 $ milligrams b n/d n/d n/d n/d n/d 10 milligrams 85% 85% 85% $0.99 $1.00 Tramadol HCL (Ultram ) % of Rx for the Drug by Strength Average Price per Pill Paid Physician-dispensed Rx Regular release 50 milligrams 100% 97% 70% $1.49 $1.01 Extended release 150 milligrams b <1% 3% 29% s/c $10.43 Other extended release (100, 200, 300 milligrams) <1% <1% 1% s/c s/c Pharmacy-dispensed Rx Regular release 50 milligrams 97% 96% 95% $0.73 $0.73 Extended release 150 milligrams b n/d n/d n/d n/d n/d Other extended release (100, 200, 300 milligrams) 3% 4% 5% $3.38 $2.00 Notes: The data included are the prescriptions filled from 2012Q3 to 2013Q1, for medical claims that had injuries occurring within two years prior to a given quarter. 2013Q1 refers to the first calendar quarter of Similar notation is used for other quarters. a Effective November 20, 2012, Illinois changed the reimbursement rules to set the prices for physician-dispensed prescription drugs to the average wholesale price of the original drug used in the repackaging process and explicitly require that dispensing physicians provide the National Drug Code of the underlying drug. Data for 2012Q3 are pre-reform for Illinois. Data for 2012Q4 reflect a mix of pre- and post-reform experience. Calendar quarter 2013Q1 is the first full post-reform quarter for Illinois, which is the latest quarter of the study period. b The three new-strength drug products are milligrams of hydrocodone-acetaminophen, approved by the FDA in 2012; 7.5 milligrams of cyclobenzaprine HCL, approved by the FDA in 2011; and 150 milligrams of tramadol HCL extended release, approved by the FDA in c The two numbers in the strength denote the strengths for the two main ingredients in the drug. For example, milligrams stands for 2.5 milligrams of hydrocodone and 325 milligrams of acetaminophen. Key: FDA: U.S. Food and Drug Administration; n/a: not applicable; n/d: not seen in the data; Rx: prescriptions; s/c: small cell size (fewer than 30 prescriptions). 16

17 5 IMPLICATIONS AND CONCLUSION As of January 2015, 18 states have made changes to reimbursement rules for drugs that physicians write and dispense at their offices. 1 These reforms were expected to lead to substantial price reductions by capping the prices paid for physician-dispensed repackaged drugs. However, the results from this study suggest a behavioral change after the reforms on the supply side of drug products and on the part of some physiciandispensers who wrote and dispensed three new-strength products and were paid higher prices. While it is debatable whether there are clinical benefits of prescribing these new-strength drug products, the findings raise a question about whether the reforms could lead to effective and sustainable price reductions for physician-dispensed drugs. We observed an emerging trend of drug products with new strengths in post-reform California and Illinois. These new drug products had higher AWPs compared with the AWPs for the same drugs of existing strengths, and they were almost all dispensed by physician-dispensers. Based on the patterns we observed, we infer that the shift to the new strengths was unlikely to be driven by new evidence about superior medical practices. Rather, it is likely that financial incentives drove some physicians to choose the new strengths for their patients. Because these new drug products were provided by an original manufacturer, at least labeled as such, these drugs may not be considered as repackaged drugs for the purpose of determining reimbursement. If these drug products have new strengths, they may not have therapeutic equivalents among the same drugs of other existing strengths. As a result, many prescriptions for these new drug products were reimbursed at the 1 Among the 18 states, 13 states made changes prior to July 2013, as described in Wang, Liu, and Thumula (2013b). Since July 2013, five new states (Delaware, Hawaii, Idaho, North Carolina, and Pennsylvania) have made similar changes. Effective July 1, 2013, Idaho s medical fee schedule requires that reimbursement to physicians for repackaged medicine shall be the AWP for the medicine prior to repackaging, identified by the original manufacturer NDC, without a dispensing fee. Effective September 11, 2013, Delaware made changes to reimburse for physician-dispensed prescription drugs at the AWP of the underlying drug as identified by the original NDC, with no dispensing fee. Effective July 1, 2014, Hawaii passed legislation that caps prices paid for all forms of prescription drugs (including repackaged and relabeled drugs) at 140 percent of the original manufacturer s AWP. In August 2014, North Carolina established a pharmacy fee schedule that limits reimbursement for prescription medications and professional pharmacy services to 95 percent of the AWP of the drug products, requiring that bills for prescription medication include the original manufacturer s NDC. If the original manufacturer s NDC is not submitted in the bill, the reimbursement will be set at 100 percent of the AWP of the least expensive therapeutically-equivalent medication. North Carolina s pharmacy fee schedule also requires that no outpatient provider other than a licensed pharmacist be reimbursed for more than an initial five-day supply of a Schedule II or Schedule III medication. In October 2014, Pennsylvania passed a legislation (House Bill 1846) that caps reimbursement for physician-dispensed repackaged drugs at 110 percent of the AWP of the original drug product used in the repackaging process. Dispensing physicians are required to submit the original NDCs for the repackaged drugs they dispense. If the original NDC is not provided, the payor has the option to select the NDC of the least expensive clinicallyequivalent product. Moreover, provisions in the bill also limit reimbursement for Schedule II drugs to an initial 7-day supply, if dispensed by any non-pharmacy outpatient providers, with an additional 15-day supply if the injured worker requires a medical procedure including surgery. The dispensing of other non-schedule II medications is limited to a 30- day supply by any non-pharmacy outpatient providers, and physicians will not be reimbursed for over-the-counter drugs. 17

18 AWPs of the drug products reported by the manufacturer, which were much higher than those for other existing strengths of the same drug. In these cases, the specific language of the reforms that targeted prices for repackaged drugs in some states may not be as effective at reducing the prices paid for physician-dispensed drugs. We should note that the use of these new-strength drug products did not appear to be as prevalent in the other reform states we studied. 2 However, this emerging issue, if not addressed, could be a challenge to the physician dispensing reforms in other states. We hope that the findings of this study are helpful for policymakers and stakeholders who are interested in finding the most effective solutions to the issues related to physician dispensing. The data for Illinois were early, reflecting four months of post-reform experience in the state. We will continue to monitor the impact of the reforms and identify emerging issues that are important for public policies regarding physician dispensing. 2 Other reform states studied: Arizona, Connecticut, Georgia, Michigan, Oklahoma, South Carolina, and Tennessee. 18

19 GLOSSARY Average wholesale price (AWP): The AWPs of individual drug products are the prices reported by manufacturers and labelers. They do not necessarily represent the actual prices charged or paid in sales transactions. In workers compensation systems, the AWP is often used as a price benchmark for pharmacy reimbursements of prescription drugs. Medi-Span : A publisher that offers a series of comprehensive drug databases, tools, and applications utilized by health care professionals. Medi-Span is part of Wolters Kluwer Health, Inc. National Drug Code (NDC): A unique 11-digit code assigned by the U.S. Food and Drug Administration to each medication in the United States that is intended for use on humans. The number identifies the specific drug product, its strength and dosage, package size, manufacturer, and repackaging firm. Original drug: The drug product produced by a manufacturer that is used to make repackaged drugs by a repackager. The AWPs of original drugs are established based on the reported prices by the manufacturers, while the AWPs of repackaged drugs are based on prices reported by repackagers. Physician-dispensers: Physicians who dispense drugs at their offices directly to patients. Physician dispensing: Refers to the practice of physicians dispensing prescription drugs at their offices directly to patients. Pharmacy fee schedule: A schedule of maximum reimbursement levels for drugs dispensed at pharmacies and/or physician offices, as part of workers compensation laws. Repackaged drug: A drug product supplied by a repackager. In the past decade, repackaging firms (also known as repackagers) often purchased large quantities of a given medication and repackaged the pills into single-prescription-sized packages (e.g., 30 pills). Once registered with the U.S. Food and Drug Administration, a repackager obtains a new NDC for the drug they repackage and assigns an AWP for the repackaged drug. The AWP of a repackaged drug is almost always higher than the AWP of the original manufacturer drug used in the repackaging process. 19

20 REFERENCES Alpert, A., M. Duggan, and J. Hellerstein Perverse reverse price competition: Average wholesale prices and Medicaid pharmaceutical spending. Working Paper Cambridge, MA: National Bureau of Economic Research. Eggleston, K Prescribing institutions: Explaining the evolution of physician dispensing. Asia Health Policy Program Working Paper #24. Stanford University. Iizuka, T Experts agency problems: Evidence from the prescription drug market in Japan. RAND Journal of Economics 38 (3): Ireland, J., and A. Swedlow The cost and utilization of compound drugs, convenience packs, and medical foods in California workers compensation. Oakland, CA: California Workers Compensation Institute. Jacobson, M., and J. Newhouse Expect the unexpected? Physicians responses to payment changes. Expert Voices by NIHCM Foundation. November Jacobson, M., C. Earle, and J. Newhouse Geographic variation in physicians responses to a reimbursement change. The New England Journal of Medicine 365 (22): 2,049 2,052. Lipton, B., C. Laws, and L. Li Workers compensation prescription drug study: 2011 update. Boca Raton, FL: NCCI Holdings, Inc. Neuhauser, F., A. Swedlow, and B. Wynn Impact of physician-dispensing of repackaged drugs on California workers compensation, employers cost, and workers access to quality care: A study for the Commission on Health and Safety and Workers Compensation. Oakland, CA: Commission on Health and Safety and Workers Compensation. Radeva, E., C. Telles, R. Yang, and R. Tanabe Monitoring the impact of Tennessee fee schedules: CompScope benchmarks, 9th edition. Cambridge, MA: Workers Compensation Research Institute. Swedlow, A., B. Gardner, and J. Ireland Difference in outcomes for injured workers receiving physiciandispensed repackaged drugs in the California workers compensation system. Oakland, CA: California Workers Compensation Institute. Swedlow, A., and J. Ireland Analysis of post-reform outcomes: Changes in pharmaceutical utilization and reimbursement in the California workers compensation system. Oakland, CA: California Workers Compensation Institute. Wang, D Physician dispensing in workers compensation. Cambridge, MA: Workers Compensation Research Institute. Wang, D., and T. Liu. 2011a. Prescription benchmarks, 2nd edition: Trends and interstate comparisons. Cambridge, MA: Workers Compensation Research Institute b. Prescription benchmarks for Florida, 2nd edition. Cambridge, MA: Workers Compensation Research Institute. Wang, D., T. Liu, and V. Thumula Impact of physician dispensing reform in Georgia, 2nd edition. Cambridge, MA: Workers Compensation Research Institute a. Impact of reform on physician dispensing and prescription prices in Georgia. Cambridge, MA: Workers Compensation Research Institute b. Prevalence and costs of physician-dispensed drugs. Cambridge, MA: Workers Compensation 20