Tomas Philipson, PhD

Tomas Philipson, PhD, says historical reform efforts in the United States have raised doubts as to whether or not the implementation of the Affordable Care Act will slow health spending growth.  “The problem with spending growth,” Dr Philipson says, “is that it's driven by medical innovation. These reforms are not designed to deal with medical innovation. Medical innovation is driven mainly by global innovation incentives.“

Dr Philipson also suggests that concepts of value-based insurance design must be considered differently when managing a chronic condition like diabetes. He reasons this is due to 3 factors. The first is that there is no risk for over overconsumption of diabetes prescriptions—if anything, under-usage is likely. Second, diabetes is a chronic disease, meaning most patients are not newly diagnosed and they therefore are merely continuing existing health treatment plans. Third, and most importantly, Dr Philipson argues that insurance design needs to be coupled with altruism.

“Health economists wrongly believe that low copays lead to moral hazard, which has the implication that Medicaid should raise copays,” Dr Philipson said, “But that's crazy because the reason we have low co-pays in Medicaid is because we basically have altruistic motives that want to stimulate consumption of people.”

Articles

Tomas Philipson, PhD, says historical reform efforts in the United States have raised doubts as to whether or not the implementation of the Affordable Care Act will slow health spending growth. Dr Philipson also suggests that concepts of value-based insurance design must be considered differently when managing a chronic condition like diabetes.

Session 2: Discussant

To develop a methodology for computing cost-effectiveness measures of a drug throughout its life cycle.

 We developed a set of models that measure the long-term cost-effectiveness of 2 oncology drugs, paclitaxel and docetaxel, throughout their life cycles.

 The study combined pricing history of the drugs, US Food and Drug Administration approval dates, drug utilization from Medicare claims, and clinical effectiveness information from phase III studies reported in the scientific literature. These data were used to estimate the incremental cost-effectiveness ratio (ICER) at the time of market entry and by year thereafter. The study population included patients with cancer who were treated with paclitaxel or docetaxel.

 The prices of paclitaxel and docetaxel dropped substantially due to patent expirations, while the number of users increased several fold because of subsequent empirical evidence and approval of new indications that resulted in greater efficacy. The ICER over a 10-year period was approximately 60% of the ICER at product launch for both drugs, and was further decreased when a longer-term perspective was taken.

We demonstrated that the ICER of a drug can decrease substantially over its life cycle. Thus, cost-effectiveness at drug launch might be a poor indicator of the longterm value of the drug. The results of this study are based on the analysis of 2 prominent oncology drugs, paclitaxel and docetaxel. The results may not be generalizable to other drug classes or other oncology drugs for which new indications are less common.

Rapid increases in drug-related expenditures over the last several decades have led payers to focus increased attention on cost containment and the trade-off between the benefits of a drug and the drug’s budget impact. An increasingly common method to estimate this trade-off is to perform cost-effectiveness analysis (CEA).

 CEA provides a summary of value that is considered by payers to help inform their reimbursement policy for a drug.

 The use of CEA by payers also influences drug expenditures and pricing,

 which can subsequently have a significant impact on the pace and arrival of drug innovation.

 One important limitation of the current approach to conducting CEAs is that they are typically performed and attract the most attention at the time of initial drug launch or for individual indications; therefore, such analyses often disregard the overall cost-effectiveness of a single agent in the longer term, which can change dramatically over the life cycle of a drug.

The change in cost-effectiveness over time is especially important in the context of oncology drugs. The initial indication for a new oncology therapy is typically treatment of patients with late-stage disease for whom all approved treatment options have failed. It is challenging to demonstrate a large clinical benefit in patients with refractory disease who may have received multiple lines of therapy and may also have experienced toxicity from prior treatments. Subsequent indications tend to be for earlier lines of therapy, when patients are more robust and a greater clinical response is more likely. These populations also tend to be larger and, therefore, reflect a more common experience of patients treated with the drug. Therefore, CEA analyses focusing on the initial indication might not accurately capture the value of a drug over its life cycle for 2 important reasons. First, the patient populations receiving the drug can change substantially as new empirical evidence of the drug’s effectiveness is published or new indications for the drug are approved. A second and important factor is that the costs of the drug for payers also change significantly over time due to loss of exclusivity and patent expirations, even though many oncology agents, such as paclitaxel, remain important elements of standard of care well after they have lost exclusivity.

The cost-effectiveness of a drug at initial launch, then, might be a poor approximation of the drug’s cost-effectiveness over its life cycle. However, summary measures of a drug’s cost-effectiveness that include the increasing efficacy in subsequent indications and as price declines are rarely presented. To our knowledge, only 1 other article, by Garrison and Veenstra,6 has included computations of the cost-effectiveness of a drug from a dynamic or life-cycle perspective. To do this, the authors simulated the change in cost-effectiveness of trastuzumab from the time of drug launch due to the introduction of a new indication.

In this article we add to the literature by estimating the cost-effectiveness of 2 oncology drugs, paclitaxel and docetaxel, over their life cycles. These 2 drugs were chosen because they have played an important role in improving cancer survival and because we are able to observe changes in prices and utilization of these drugs from product launch to the present. We extended prior research along several dimensions. First, we used data on drug utilization and prices to show how cost-effectiveness changes over a drug’s life cycle from product launch to post—patent expiration. Second, we documented changes in cost-effectiveness due to new clinical evidence of efficacy, US Food and Drug Administration (FDA) approval of new indications, the size of the patient population treated, and changes in drug prices. Finally, we used Medicare claims data to construct a nationally representative sample to depict the ways in which the utilization of these drugs evolved over time by cancer type and year.

We developed the incremental cost-effectiveness ratios (ICERs) of 2 oncology drugs from a life-cycle perspective. These measures describe the net value of products over their life cycles rather than just at the time of product launch. The underlying model framework that we propose builds on the model developed by Garrison and Veenstra.

To formally illustrate the difference between product life cycle and standard CEA measures, we first consider the ICER at product launch (t = 0), with the equation 

 is the incremental cost of the new drug at product launch and E

 is the measure of incremental effectiveness of the drug at product launch.

Next, we consider the ICER T time periods after launch:

 are the incremental cost and effectiveness of the drug T time periods after product launch, 

 is the population of users T periods after product launch. Given the definitions above, the ICER T time periods after product launch is simply the ratio of the present value of incremental costs associated with the drug until period T (numerator) and the present value of the incremental effectiveness associated with the drug until period T. The ICER from a product life-cycle perspective is simply ICER

, where T equals the end of the product life cycle.

Because drugs can last on the market for several decades, we calculate ICER

 for 10 and 20 years after product launch.

 only if the following 2 conditions are satisfied:

1) the incremental costs do not change over the life cycle IC

2) the incremental efficacy does not change over the life cycle E

If either of the above conditions are violated, then ICER

 would be a poor approximation of the true long-term cost-effectiveness of drugs.

As discussed earlier, the first condition is likely violated because drug prices can decrease dramatically over the life cycle, with prices falling to as low as 10% of the product launch price after patent expiration. Notice that the effects of reduction in prices after product launch will be magnified by the increased diffusion of the drug over time; in other words, lower prices mean that a larger population will use the drug. This would increase the weight attached to costs in later years, when prices are low, and would correspondingly decrease the weight attached to costs in earlier years, when prices are high.

The second condition is also likely violated as the efficacy and effectiveness of the drug might change over its life cycle. As mentioned above, this is especially likely for oncology products, which are initially used as second- or third-line treatment for patients with late-stage cancers; after product launch many oncology drugs receive approval for additional indications, such as treatment of early-stage disease and/or other types of cancers. As was the case with changes in price, the effects of changes in effectiveness will be magnified by the increased diffusion of the drug over time; in other words, increased effectiveness translates into greater diffusion of the drug. This would increase the weight attached to effectiveness in periods when efficacy is increased. To compute the ICER from a life cycle perspective we need information on the measures described below.

We estimated incremental costs by indication by using cost-effectiveness data published in peer-reviewed journal articles whenever possible. For some indications approved in the early 1990s, when costeffectiveness data were often not available, we estimated the incremental cost of treatment using information on unit prices and treatment regimens from published data and/or FDA product labels, including dosage and average number of courses.

 To estimate the incremental costs for a given indication change over the life cycle of the drug due to changes in real drug prices, we obtained the wholesale acquisition cost (WAC) of paclitaxel and docetaxel in branded and generic forms (if applicable) from Analy$ourceOnline. These data were obtained from a national drug database, and are included with permission from and copyrighted by First DataBank, Inc. Information on the methodology of collecting price data is described at the First DataBank website: 

http://www.firstdatabank.com/Support/drug-pricing-policy.aspx

. When multiple prices were posted for a single year, the lowest WAC available for that year was used. All costs and prices shown in this article are inflation-adjusted to 2010 US dollars, assuming an annual inflation rate of 3%.

 For each drug we constructed nationally representative utilization rates by cancer type or indication from 1991 to 2007. The utilization data were derived from the surveillance, epidemiology, and end results (SEER) Medicare data and include the number of Medicare patients using each drug by major cancer categories, including breast, lung, prostate, and colorectal cancer. Because no utilization data are available for the time period after 2007, we assumed that utilization rates stayed at the 2007 levels for future years. The data were weighted to be nationally representative using information on national estimates of prevalence of different types of cancer by year. In particular, we assumed that the utilization rate of different drugs by cancer type is the same for Medicare and non-Medicare beneficiaries; therefore, national estimates of prevalence of different cancers multiplied by the utilization rates of different drugs by cancer type from SEER Medicare provide an estimate of the national utilization of drugs by cancer type.

 We also quantified life-cycle changes in efficacy. For each drug we identified the approval of new indications over the product life cycle from the FDA product labels, and the life-years gained for each new indication from peer-reviewed journal articles. We omitted certain indications (such as paclitaxel for second-line treatment of ovarian cancer and Kaposi’s sarcoma, and docetaxel for first-line treatment of non-small-cell lung cancer and gastric cancer), as no reliable data on cost-effectiveness were available or could be constructed for these indications. We combined these data with claims-based measures of the indication-specific utilization described above, in order to produce timevarying estimates of mean life-years gained. These estimates change with the introduction of new indications, as well as with shifts in the proportion of patients treated for newer or older indications over time.

Imputations in the Absence of Price and Utilization Data

In order to illustrate the long-run perspective, we calculated the ICERs of each drug for 20+ years since product launch. For paclitaxel we covered the years 1994 to 2016. For the 5 years beyond 2011, for which actual price data were not available, we assumed the real price of paclitaxel stayed constant at the 2011 level until 2016. For docetaxel, we covered the time period of 1998 to 2020. Because the docetaxel patent expired recently (in 2010), and we have not had a chance to observe price drops over a long period of time after the loss of exclusivity, as in the case of paclitaxel, we assumed its price would decrease at the same rate as that of paclitaxel after its loss of exclusivity, and imputed the annual prices of docetaxel between 2012 and 2020 using a linear prediction model. The data on utilization rates by indication were not available beyond the year 2007, and we assumed that the rates stayed at the same level of year 2007 for future years in our model.

The credibility of the results depends upon their robustness to alternate assumptions about the life cycle trajectory of each of the key model inputs described above. We therefore tested the robustness of our results with multiple 1-way sensitivity analyses, by varying the values of the incremental cost and life-years gained to plus or minus 25%.

New Indications and New Data on Cost-Effectiveness

Several new indications were approved for both paclitaxel and docetaxel within 10 years of product launch (

). In general, the effectiveness of the 2 drugs for new indications was greater than the effectiveness of their original indications at product launch. Two exceptions were a new indication for paclitaxel for the treatment of lung cancer and a new indication for docetaxel for the treatment of metastatic prostate cancer; however, both conditions are associated with a poor prognosis. Table 1 also shows that the empirical evidence on effectiveness in peer-reviewed journal articles usually predates its official FDA indication approvals by a few years.

Fifteen Years of Pricing History Since Product Launch

 shows the trends in real prices for paclitaxel and docetaxel in 2010 US dollars. The prices are rescaled so that the initial price in the year of product launch equals 100; this allows for easy comparisons of product prices over the life cycle relative to price at product launch. Figure 1 demonstrates that the price of paclitaxel decreased after product launch, with real prices being approximately 30% lower 6 years after product launch. There was also a sharp decline in prices after loss of exclusivity due to patent expirations; the price of paclitaxel 12 years after product launch was only 10% of its price at product launch. In contrast, the price of docetaxel increased gradually after product launch, rising to approximately 17% higher 6 years after product launch and 24% higher 12 years after product launch. Subsequently, the price of docetaxel declined by approximately 25% in a single year after it lost exclusivity in 2011 due to patent expirations.

Utilization Rates and New Evidence of Efficacy

 shows the time trends in the utilization of paclitaxel and docetaxel by cancer type, demonstrating that the utilization of both drugs changed in response to publication of clinical evidence of new indications and regulatory approval of new indications. Figure 2 also shows that for both paclitaxel and docetaxel, increased use for an indication began quickly after clinical data showed effectiveness, usually some time before the official approval of the new indication. In 2007 there were 8 times as many paclitaxel users as in 1994, when the first cost-effectiveness data on paclitaxel became available. Also in 2007 there were 9 times as many docetaxel users as in 1998, when docetaxel was first launched. Thus, the most common use of the products is generally for the approved indications for which efficacy is greatest. This use is much more widespread after loss of exclusivity than at the time of product launch.

The ICERs of paclitaxel and docetaxel over 22 years are shown in 

. The ICERs of both drugs decrease as one takes a longer-term perspective; for example, the ICER at year 0 or at product launch was $30,723 per life-year gained for paclitaxel and $37,537 per life-year gained for docetaxel. The ICER over a 10-year time period is only a fraction of the ICER at product launch. In particular, the ICER of paclitaxel over a 10-year time period was 61% of its ICER at product launch and, similarly, the ICER of docetaxel over a 10-year time period was 60% of its ICER at product launch. The ICERs of both drugs decreased even further when viewed with a 15- and 20-year perspective; specifically, the ICER of paclitaxel over a 20-year period was $8,563 per life-year gained (28% of the ICER at launch) and the ICER of docetaxel over a 20-year period was $15,468 per life-year gained (41% of the launch ICER). The 1-way sensitivity analyses produced a range of ICER values but do not change the main qualitative results of the study (

The ongoing debate over rising healthcare spending has led to additional scrutiny of the value of healthcare interventions, including pharmaceuticals. This movement has also resulted in keen interest in CEAs, which have been adopted as a standard for reimbursement approval in countries such as the United Kingdom. Although CEA is often conducted at product launch or for specific indications in isolation, and is an important component of dossiers submitted to public and private payers,

 the initial CEA or CEAs for new indications might not reflect the true long-term value of a particular drug. For this reason, the initial CEA needs to be placed into context against a summary measure of CEAs that reflects long-term value.

There are several important reasons that standard CEA methods might not accurately capture long-term drug value, one of which is the impact of new indications and/or patient populations. This is particularly important for oncology products, since ethical considerations and the difficulty in obtaining regulatory approval typically dictate that an investigative compound be initially tested for second- or third-line therapy for patients with advanced disease, who are relatively small in number and have limited proven treatment options. If the new drug is effective in the initial clinical setting, new clinical trials will typically be initiated to determine safety and efficacy of the product as first-line treatment or adjuvant treatment (chemotherapy following curative surgery) for the same cancer or for other types of cancers. Our findings, which are supported by Garrison and Veenstra,

 show that these subsequent uses in less severely ill patients tend to be associated with better clinical benefit and more favorable cost-effectiveness. In addition, we have shown that patients treated for subsequently approved indications are more numerous, so that the experience of greater benefit and better cost-effectiveness is more characteristic of patients’ experience with the drug.

Another facet of standard CEA conducted at product launch is that it evaluates costs when they are at their maximum. Drug prices might decrease over their life cycles, especially after the entry of generics after patent expirations; therefore, prices at launch might be a poor indicator of costs of the drug over its life cycle.

Given that standard CEA might not capture the long-term value of drugs, innovative approaches to measuring the cost-effectiveness of new drugs over the entire product life cycle are needed in order to capture the overall value of new drugs over time. Comparing long-term value with value at product launch can inform the debate on appropriate use of CEA for reimbursement and coverage decisions.

In this article we presented our findings after developing and estimating the life cycle CEA for 2 important oncology agents. We demonstrated that the ICER of a drug can decrease substantially over its life cycle, thus creating value net of costs. The ICER of paclitaxel decreased by 39% at 10 years after product launch and by 72% at 20 years after product launch. Similarly, the ICER of docetaxel decreased by 38% at 10 years after product launch and by 59% at 20 years after product launch. The ICER from a life cycle perspective responded to the discovery of new indications when they were typically more cost-effective than indications at first product launch and were more typical of patients’ experience. The sharp decline in prices after patent expirations is also an important determinant of ICER, with its importance increasing over a longer-term horizon of 20 years.

It should be noted that the ICER calculated by a public plan may be different from that calculated by a private plan. For example, Medicaid prices are often lower than private payer prices due to mandated rebates. In addition, because there are various manufacturer rebates and copay assistance programs for patients with cancer, the ICER of a drug can even be different at the individual person level; however, such price differences should not change the general trend of a drug’s ICER continuing to decrease over time.

The findings of this study should be viewed in light of its limitations. First, the data on drug utilization are limited to Medicare beneficiaries and are thus not based on a nationally representative sample of individuals. To overcome this limitation, we weighted the data using supplemental information on national rates of cancer prevalence. Second, WAC does not represent the true acquisition costs of drugs, as it excludes rebates and discounts received by health plans. We are therefore likely overestimating drug costs and ICERs, especially before loss of exclusivity, when sizable rates are common. Third, our estimates do not capture improvements in the efficacy and cost-effectiveness of a drug over time as physicians learn how to better manage therapy. For example, in the case of high-risk leukemia and difficult-totreat neuroblastoma, event-free survival improved when the standard agents (methotrexate and melphalan, respectively) were administered differently or combined with different drugs.8,9 We were not able to measure such improvement in this article due to the difficulty of obtaining real-world data on adoption of best practices over time and relating the data to clinical outcomes. Finally, it is important to note that the results of this study might not be generalizable to other therapeutic classes or other oncology drugs. For example, highly complex biologics can remain competitive even after patent expirations because they are difficult to replicate, possibly resulting in less pricing reduction at the end of exclusivity. Nevertheless, Garrison and Veenstra6 showed that the ICER of trastuzumab, also a biologic, declined to less than half the original level even before loss of exclusivity because its second indication, to treat early-stage breast cancer, shows greater efficacy as well as more favorable cost-effectiveness and is applicable to a larger number of patients.

While evidence shows that various factors can cause the ICER of a drug to decrease in the long term, there remains great uncertainty at the time of a specific drug’s product launch about how its ICER will evolve over time. One direction for future research might be to create a predictive tool that estimates the predicted long-term dynamic ICER for a specific drug at product launch. Several sources of information might be useful in creating such a predictive tool. First, historical information on the use of other drugs in the same drug class or with the same mechanism of action might help shed light on how the use of a particular drug might evolve over time. Second, evidence from completed and ongoing clinical trials might suggest future new indications for the drug and might also be useful in predicting the risk of blackbox or safety warnings. Third, evidence from early postmarket studies might also help to predict how the use of a drug might evolve over its life cycle. All these sources of information can be used together to predict the probability of different events over a drug’s life cycle. Both positive events, such as new indications that expand the use of a drug, and negative events, such as black-box warnings and product withdrawals that restrict the use of the drug, should be considered. These data on the probability of different events should be combined with data on drug prices and potential patient populations to predict the long-term ICER of a drug. Finally, due to great uncertainty at product launch, sensitivity analyses should be performed to consider ICER under a range of future scenarios. Such data will eventually help decision makers make more informed decisions about drug coverage and reimbursement.

 Leonard D. Schaeffer Center for Health Policy and Economics and School of Medicine, University of Southern California, Los Angeles, CA (YL); Bristol-Myers Squibb, Plainsboro, NJ (JRP, SW); Leonard D. Schaeffer Center for Health Policy and Economics and School of Pharmacy, University of Southern California, Los Angeles, CA (NS); Harris Graduate School of Public Policy, University of Chicago, Chicago, IL (TP).

 This supplement was supported by Bristol-Myers Squibb.

Dr Penrod reports employment with Bristol-Myers Squibb. Dr Sood reports consultancy with Precision Health Economics. Ms Woodby reports consultancy with Bristol-Myers Squibb. Dr Philipson reports consultancy with Bristol-Myers Squibb and partnership with Precision Health Economics. Dr Penrod also reports stock ownership in Bristol-Myers Squibb. Dr Lu reports no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Concept and design (YL, JRP, TP, NS, SW); acquisition of data (TP, JRP, SW); analysis and interpretation of data (YL, JRP, TP, NS, SW); drafting of the manuscript (YL, JRP, TP, NS, SW); critical revision of the manuscript for important intellectual content (YL, JRP, TP, NS); statistical analysis (YL, NS); obtaining funding (JRP); and supervision (NS).

 Neeraj Sood, PhD, University of Southern California, 3335 S Figueroa St, Los Angeles, CA 90089. E-mail: 

		Sood N, de  Vries  H, Gutierrez I, Lakdawalla DN, Goldman DP. The  effect of regulation on  pharmaceutical revenues: experience in nineteen countries. 

 New York,  NY: Oxford University Press; 2004.

Dynamic Cost-Effectiveness of Oncology Drugs

 To quantify in the context of chronic myeloid leukemia (CML) the additional value patients receive when innovative treatments enable them to survive until the advent of even more effective future treatments (ie, the “option value”).

Observational study using data from the Surveillance, Epidemiology and End Results (SEER) cancer registry comprising all US patients with CML diagnosed between 2000 and 2008 (N = 9,760).

 We quantified the option value of recent breakthroughs in CML treatment by first conducting retrospective survival analyses on SEER data to assess the effectiveness of TKI treatments, and then forecasting survival from CML and other causes to measure expected future medical progress. We then developed an analytical framework to calculate option value of innovative CML therapies, and used an economic model to value these gains. We calculated the option value created both by future innovations in CML treatment and by medical progress in reducing background mortality.

For a recently diagnosed CML patient, the option value of innovative therapies from future medical innovation amounts to 0.76 life-years. This option value is worth $63,000, equivalent to 9% of the average survival gains from existing treatments. Future innovations in CML treatment jointly account for 96% of this benefit.

 The option value of innovative treatments has significance in the context of CML and, more broadly, in disease areas with rapid innovation. Incorporating option value into traditional valuations of medical innovations is both a feasible and a necessary practice in health technology assessment.

Traditional health-technology assessment, such as cost-effectiveness analysis, measures the value of a health innovation by comparing benefits (eg, survival gains or improved quality of life) to costs, assuming no further improvements in medical technology aside from the innovation in question.

 Such methods overlook the fact that life-extending innovations can allow patients to live until the next breakthrough, which may offer additional survival opportunities. This possibility of benefiting from future improvements in medical technology is the “option value” of a health innovation. As long as medical technology continues to advance, there will be option value in survival-improving innovations. Failing to account for this value ignores a potentially important source of benefit to patients, understating the effectiveness of new therapies, especially in areas of rapid innovation.

The existence of option value has been demonstrated in the context of end-of-life care from the perspective of economic theory

 option value has not been incorporated into existing assessments of pharmaceutical innovation. Because the literature lacks both a guide to incorporating option value into health-technology assessment and a method for calculating the option value of current technologies, real-world demonstrations are needed to illustrate the feasibility and necessity of incorporating option value into practice. In the context of healthcare management, option value assessments are key to achieving a better understanding of the overall effectiveness of innovative treatments that may ultimately lead to significant improvements in patient care.

The latest breakthroughs in chronic myelogenous leukemia (CML) treatment serve as a prominent example of the relevance of the option-value concept. (See the 

 for a list of agents used for CML treatment since 1959 and the dates of their approval by the US Food and Drug Administration.) Historically, survival has been 3 to 5 years from the time of CML diagnosis.

 However, the introduction of tyrosine kinase inhibitors (TKIs), an innovative class of targeted therapies, has led to substantial survival gains in the treatment of CML, with overall survival of patients in clinical trials increasing to 89% at 5 years.

 Treatment with the first-generation TKI imatinib rapidly became the standard of care after the drug was introduced as first-line treatment in 2001.

 Imatinib remained the only TKI approved for CML treatment for several years until the introduction of 2 second-generation agents. Dasatinib was approved for second-line use in June 2006, representing a breakthrough for CML patients failing imatinib; long-term follow-up of these patients showed that the vast majority (78%) were alive at 5 years.

 In 2007, nilotinib was also approved for second-line use.

 While dasatinib and nilotinib were initially approved for second-line use, these agents demonstrated superiority to imatinib in the first-line setting and were approved for this indication in 2010.

In this study we quantified the extent to which the firstgeneration TKI served as a “bridge” to subsequent CML treatments, and how treatment with first- and second-generation TKIs is allowing patients to survive into periods of future improvements in CML treatment and background mortality. For this purpose we developed an analytical framework to enable incorporation of option value into healthtechnology assessment. We used a Cox proportional hazards model to estimate historical real-world CML survival trends, and employed the Lee-Carter method to forecast future survival from any cause of death in the Surveillance, Epidemiology and End Results (SEER) cancer registry. Based on these estimates we quantified the option value of innovative TKI therapies and used an economic model

As discussed previously, option value is the additional survival benefit patients receive when innovative treatments enable them to experience future medical breakthroughs; stated more simply, it is the benefit derived from the opportunity to live longer. Measuring the option value of a treatment is feasible to the extent that there are available data on treatment effectiveness and future (post-treatment) survival. As an illustration of this point, assume that half of the patients suffering from a terminal disease experience a 1-year life extension due to innovative Therapy A, while the remaining half do not survive; hence, the life expectancy gain from Therapy A is 0.5 life-years. Next, suppose that, after 1 year, surviving patients gain access to a new medical innovation (Therapy B) that allows them to live an additional 4 months (ie, 0.33 life-years = 4/12); thus, the estimated option value of Therapy A is 2 months (or 0.16 life-years, computed as the product of 0.5 life-years from Therapy A effectiveness and 0.33 life-years from future access to Therapy B). A more detailed explanation of this concept is presented in the eAppendix, along with the analytical framework to assess the value of a health innovation and incorporate its option value.

In the context of CML, initial assessments of benefits comparing TKI therapy with the previous standard of care underestimate the value of TKI therapy by ignoring its option value. In particular, the option value of first-line treatment includes: 1) reductions in background mortality through future medical innovations, and 2) survival improvements that enable patients who develop treatment resistance to survive with greater probability into a period during which second-generation agents become available. These benefits have not been previously estimated.

SEER is the gold standard registry for cancer in the United States. SEER registries identify incident cases of malignancies based on uniform reporting, according to the 

International Classification of Disease for Oncology, 3rd Edition (ICD-03)

. As the only national cancer database that follows patients over time to trace survival, SEER contains information on 21,558 individuals diagnosed with CML between 1973 and 2008. The data set includes type(s) of cancer (up to 10 tumors), age at diagnosis, year of diagnosis, gender, race, ethnicity, marital status, county of residence, and time and cause of death.

Patients with CML were identified in the SEER registry based on 

 codes 9863 (chronic myeloid/myelogenous leukemia), 9875 (chronic myelogenous leukemia, BCR/ ABL-positive), and 9876 (atypical chronic myeloid leukemia, BCR/ABL-negative). Because imatinib was introduced in 2001, we restricted the sample to CML patients diagnosed in 2000 or later. Patients diagnosed in 2000 served as the control (pre-treatment) group in order to measure the cost of delaying treatment by 1 year. 

 presents summary statistics for the restricted sample; specifically, for patients diagnosed between 2001 and 2005 (the time period when firstgeneration TKI treatment became available) and for patients diagnosed after 2006 (the time period when second-generation treatments became available). As a sensitivity analysis, we considered control groups diagnosed with CML up to 5 years prior to 2000. The results are quantitatively similar and shown in the eAppendix. The 2000 to 2008 sample consists primarily of white (82%) and male (58%) patients. More than half (52%) of the patients are married, and the average age at diagnosis was 62 years. Approximately 5 of every 6 patients have only 1 primary cancer (CML), and the average mortality from CML and other causes in the period of analysis is 26% and 11%, respectively. Aside from mortality, patient characteristics are similar across control and treatment groups. Mortality from CML was 25% lower among patients diagnosed with CML between 2001 and 2005, relative to patients diagnosed before the introduction of firstgeneration treatment. Other-cause mortality, or mortality due to causes different from CML, is similar across the firstgeneration treatment and control groups. Additionally, both CML and other-cause mortality are substantially smaller among patients diagnosed between 2006 and 2008 relative to patients diagnosed before 2005. This is driven largely by a shorter follow-up period for the second-generation treatment group, resulting in greater censoring of CML and other-cause mortality.

We combined 2 distinct statistical analyses to compute the option value of TKIs. First, we employed a Cox proportional hazards model19 to calculate past improvements in CML survival and survival from diseases different from CML between 2000 and 2008. Second, we used the Lee-Carter method to forecast future improvements in CML and othercause mortality several decades into the future.

We estimated the direct survival gains from TKIs by first applying the Cox model to the SEER sample of patients diagnosed with CML between 2000 and 2008. We included year of diagnosis as the key explanatory variable to generate year-specific survival curves. To mitigate potential bias from unobservable factors, we controlled for a rich set of patient characteristics that affect mortality independently of year of diagnosis, including age at the time of CML diagnosis, month of CML diagnosis, number of other cancers at diagnosis, and other demographic variables. The eAppendix describes the Cox proportional hazards model and shows the regression estimates. Next, we developed an algebraic method to identify survival associated with first- or secondline TKI treatment from the year-specific survival curves observed between 2000 and 2008. This algebraic method is explained in the eAppendix. Year-specific survival curves were weighted by population survival probabilities in the United States from the Human Mortality Database (HMD) to obtain nationally representative survival data of the CML population. HMD life tables are based on population estimates from the US Census Bureau, and birth and death reporting from the Census Bureau and the National Center for Health Statistics. We accounted for treatment resistance by incorporating estimates of the probability of resistance to first-line treatment from the existing literature.20 Given the sample’s limited time span, survival past 7 years from diagnosis was assumed to decline at the rate of the general population based on the HMD. A sensitivity test to evaluate the implications of this assumption is presented in the Discussion section.

The second part of the statistical analysis forecasted improvements in CML and other-cause survival from 2000 to 2055 by applying the Lee-Carter method to the year-specific (2000-2008) SEER survival curves. This method involved modeling age-specific mortality rates in terms of a long-term trend in overall mortality and an age-specific response to the overall trend.

 Furthermore, sensitivity analyses tested the robustness of our results to alternative survival forecasts. The eAppendix describes the Lee-Carter method in detail and presents sensitivity analyses where the option value of innovative CML treatments is recalculated using the lower and upper bounds of the survival forecasts’ 95% confidence intervals.

We calculated the option value of innovative TKI treatments relative to the survival profile of a CML patient with the average age in the sample (62 years) diagnosed in 2000. Finally, we used an economic model based on previous work by Becker et al

 to express option value in monetary terms. This economic model is described in the eAppendix.

The option value of innovative TKI therapies, accounting for medical progress in CML and background mortality, is shown in 

. These gains were quantified in the context of 3 possible scenarios. The first scenario (column 1) represents the survival gain and option value derived from first-line therapy for patients who have not developed resistance to treatment, relative to patients in the control group. This option value arises from future reductions in all-cause mortality, future advances in first-line treatment of CML, and future availability of second-line therapies if first-line treatment fails. The second scenario (column 2) represents the survival gain and option value of second-line therapy for patients in whom first-line treatment fails, as compared with patients in the control group. This option value is generated by future reductions in all-cause mortality, advances in second-line treatment of CML, and future availability of third-line therapy if second-line treatment fails. Calculating the weighted average of columns 1 and 2 reveals the third scenario (column 3) as the expected survival gain and option value for a recently diagnosed hypothetical CML patient with access to second-line therapies if first-line treatment fails.

Per the results in Table 2, there is almost negligible option value among CML patients from medical progress against background mortality (ie, 0.02-0.04 life-years) even though TKIs provide non-negligible survival benefits against other diseases (ie, 0.82-1.99 life-years). The option value of firstline treatment with imatinib as a bridge to further innovation in CML treatment alone (and not other diseases) accounts for 96% of the total option value.

First-line treatment provides an estimated total life expectancy gain of 9.54 life-years in CML patients who do not experience treatment failure, as compared with the control group (Table 2, column 1). In addition, future medical progress in CML treatment and background mortality provides patients who have not developed resistance to first-line treatment an option value of 0.78 life-years, or 8.2% of the total life-expectancy gain. Conversely, second-line treatments provide an estimated total life-expectancy gain of 6.82 life-years in CML patients who have experienced first-line treatment failure (Table 2, column 2). For these patients, the option value of second-line treatment from future medical progress against CML and background mortality is 0.71 lifeyears, or 10.4% of the total life-expectancy gain. On average, the option value of innovative TKI treatments for a recently diagnosed CML patient is 0.76 life-years, or 8.7% of the 8.73 expected life-year gain from first- and second-line treatments (Table 2, column 3).

Expressing all life-year gains as a private monetary valuation,

 the option value of first-line treatment is worth approximately $66,000. For patients who develop resistance to first-line treatment, the option value of second-line treatment is worth nearly $57,000. Newly diagnosed patients have an average (or expected) option value of 0.76 life-years, worth approximately $63,000.

For this study we developed an analytical framework for incorporating option value into the standard calculation of the value of a health innovation, and applied this framework to the case of innovative CML therapies. While economic theory suggests the presence of option value in the context of medical treatment, this source of value has been ignored in the literature on health technology assessment. This study is among the first applications of option value to the assessment of pharmaceutical innovation, and the first to measure option value prospectively.

Our findings show that, in the context of CML, the option value of innovative TKI therapies is equivalent to 9% of their average survival gain. Thus, ignoring the option value of these treatments underestimates their effectiveness by approximately 9%. In monetary terms, an option value of $63,000 is a sizable benefit, especially when compared with the annual costs of TKI therapies, which range from $30,000 to $70,000.23 Although a thorough cost-effectiveness analysis of these therapies would require an assessment of the costs incurred in the additional years of life extension, our findings suggest that the option value may fall within the range of costs for the first year of TKI treatment.

An intrinsic limitation of our study is that option value estimates depend on forecasts of future improvements in CML and other-cause survival. Although forecasts are inherently uncertain and not definitive, ours were based on highquality and representative data on historical survival gains. A sensitivity analysis addressed this issue by recalculating option value at the lower and upper bounds of the survival forecasts’ 95% confidence intervals; this sensitivity analysis is shown in the eAppendix. Proceeding from the least to the most optimistic forecast scenario, the average option value of a newly diagnosed patient increased from 0.36 life-years (with a net worth of $34,000) to 1.16 life-years (with a net worth of $92,000).

Given the limited time-span of our sample (2000-2008), we assumed survival past 7 years from diagnosis declined at the same rate as in the general population, based on the HMD. Although this assumption forces survival curves in the treatment and control groups to decay at the same rate, it is conservative to the extent that the control group’s survival curve may actually decay at a faster rate than those of the treatment groups. Nevertheless, we performed an even more conservative sensitivity test by limiting the analysis to 7 years post-diagnosis. Only 30% of survival gains and 20% of the estimated option value accrued in the first 7 years after diagnosis. In this more conservative scenario the option value as a percentage of the survival gains from TKI treatments was reduced from 9% to 7%, given the limited time frame for the benefits from future medical breakthroughs to accrue.

Because option value captures the benefits from surviving long enough to take advantage of the next medical breakthrough, it is determined exclusively by the effectiveness of present and future treatments. In this regard, the greater option value of first-line treatment relative to second-line TKIs can be attributed largely to the substantial survival benefits of second-generation therapies. Several therapeutic approaches for a cure for CML are under evaluation,

 suggesting that actual innovation in second- and later-line CML treatment may develop faster than forecasted. In such a scenario, future research may provide a larger benefit from first- and second-line treatment than estimated here.

Given the available survival data for CML patients, this study, by necessity, focused on imatinib as the only first-line TKI treatment; however, both dasatinib and nilotinib have been found to offer better cytogenetic response than imatinib in recently completed trials,

 Although our study did not directly quantify the option value of using second-generation TKIs in firstline treatment, based on the survival improvements relative to first-generation treatment that have been recently documented in the literature,

 we could expect larger option value from the newer TKIs if given as first-line therapy.

While the primary analysis used SEER data over the period 2000 to 2008, 2 sensitivity analyses reported in the eAppendix were used to determine whether CML-specific survival improved at an uncharacteristic rate over this period. Including patients diagnosed with CML as early as 1995 in the Cox proportional hazards regression provided quantitatively similar results and, subsequently, estimating option value at the lower and upper bounds of the survival forecasts’ 95% confidence intervals also yielded similar results.

We also assessed the sensitivity of our option value estimates by varying the probability of first-line treatment failure over a wider spectrum than previously suggested.20 Estimates were not overly sensitive to this key parameter, as the average option value for a recently diagnosed CML patient from future innovations in CML treatment ranged from 0.62 lifeyears (with a failure probability of 0.05) to 0.82 life-years (with a failure probability of 0.50) (see eAppendix).

While our analysis is tailored to the specific context of innovative TKI treatments for CML, our methodology also can be applied to other diseases. The feasibility of this method will naturally be favored in those areas for which comprehensive survival data are available (as is the case with SEER data in oncology). Although it is still possible to measure option value in disease areas for which survival data are scant but sufficient to identify treatment efficacy and posttreatment survival, such analyses will inevitably be subject to greater uncertainty.

Recent innovations in the treatment of CML have increased patient survival not only by improving the management of the disease, but also by allowing patients to live long enough to experience more effective future treatments. For a recently diagnosed CML patient, the option value of innovative therapies in terms of future medical innovations in CML and other diseases amounts to 0.76 life-years. This option value is equivalent to 9% of the average survival gains from existing treatments and is worth approximately $63,000. Future innovations in CML treatment jointly account for 96% of this benefit.

Our study demonstrates to researchers and payers that, by ignoring option value, the traditional methods of health technology assessment undervalue life-extending healthcare innovations, particularly those in areas of medicine in which rapid innovation is producing substantial survival gains. We provided methods for incorporating option value into standard health-technology assessment, and identified the scenarios for which this methodology could be applied in the future. It is worth noting that ongoing medical progress against a particular disease guarantees that the option value of available treatments will always be positive.

We have provided just 1 example of how option value could be applied in health technology assessment. While this study does not make any inference about the generalizability of the CML case to other therapies, the results underscore the need to account for option value in the assessment of health innovation, as failure to do so will likely result in the underestimation of overall value. Our framework is, however, applicable to various types of analyses, from survival benefits and cost-effectiveness analyses involving life-years or quality-adjusted life-years, to measures of an innovation’s economic value. As medical progress continues to improve survival in a broad spectrum of disease areas, investigators performing future analyses may find the methodology developed in this study useful for calculating the option value of such innovations. Ultimately, a better understanding of the overall effectiveness of innovative treatments foreshadows significant improvements in the management of patient care.

 Bristol-Myers Squibb, Plainsboro, NJ (JRP); Harris Graduate School of Public Policy, University of Chicago, Chicago, IL (TP); Precision Health Economics, Los Angeles, CA (XLQ, YS, JTS); Leonard D. Schaeffer Center for Health Policy and Economics, and Sol Price School of Public Policy, University of Southern California, Los Angeles, CA (JR).

 Dr Penrod reports employment with and stock ownership in Bristol-Myers Squibb. Dr. Philipson reports consultancy with Bristol-Myers Squibb and partnership in Precision Health Economics. Dr Qiu, Dr Romley, Dr Sanchez, and Dr Snider report employment with Precision Health Economics.

 Concept and design (JRP, TP, JR, YS, JTS, XLQ); acquisition of data (TP, JR, YS, JTS, XLQ); analysis and interpretation of data (JRP, TP, JR, YS, JTS, XLQ); drafting of the manuscript (JRP, TP, JR, YS, JTS); critical revision of the manuscript for important intellectual content (JRP, TP, JR, YS, JTS, XLQ); statistical analysis (XLQ, YS); and obtaining funding (JRP).

 Yuri Sanchez, PhD, Precision Health Economics, 11100 Santa Monica Blvd, Ste 500, Los Angeles, CA 90025. E-mail: 

		Weinstein MC, Stason WB. Foundations of cost-effectiveness analysis for health and  medical practices. 

		Philipson TJ,  Becker G, Goldman DP, Murphy KM. Terminal care and  the  value of life  near its  end. NBER  Working Paper. Published 2010.

		Thornton Snider J, Romley JA, Vogt  WB, Philipson TJ.  The option value of innovation. Forum for Health Economics & Policy. Manuscript 1306. Published 2012.

		Besa E, Woermann U. Chronic myelogenous leukemia. 

http://emedicine.medscape.com/article/199425-overview#aw2aab6-b2b3aa

. Published 2011. Accessed November 10,  2011.

		Druker  BJ, Guilhot F, O’Brien SG, et al.  Five-year follow-up of patients receiving imatinib for chronic myeloid leukemia. 

		Baccarani M, Saglio G, Goldman J, et al.  Evolving concepts in the  management of chronic myeloid leukemia: recommendations from an expert panel on behalf of the  European LeukemiaNet. 

		Baccarani M, Cortes J, Pane F, et al.  Chronic myeloid leukemia: an  update of concepts and  management recommendations of European LeukemiaNet. 

		Kantarjian H, Shah NP, Hochhaus A, et al.  Dasatinib versus imatinib in newly diagnosed chronic-phase chronic myeloid leukemia. 

		de  Lavallade H, Apperley JF, Khorashad JS,  et al.  Imatinib for newly diagnosed patients with  chronic myeloid leukemia: incidence of sustained responses in an  intention-to-treat analysis. 

		Druker  BJ, Guilhot F, O’Brien SG, et al.  Five-year follow-up of patients receiving imatinib for  chronic myeloid leukemia. 

		O’Brien SG, Guilhot F, Larson RA, et  al.  Imatinib compared with  interferon and  low-dose cytarabine for  newly diagnosed chronic-phase chronic myeloid leukemia. 

		Shah NP, Kantarjian HM, Kim DW, et al.  Intermittent target inhibition with  dasatinib 100  mg  once daily preserves efficacy and improves tolerability in imatinib-resistant and -intolerant chronic-phase chronic myeloid leukemia. 

		Talpaz M, Shah NP, Kantarjian H, et al.  Dasatinib in imatinib- resistant Philadelphia chromosome-positive leukemias. 

		Shah NP, Cortes JE, Schiffer CA, et al.  Five-year follow-up of patients with  imatinib-resistant or -intolerant chronic-phase chronic myeloid leukemia (CML-CP) receiving dasatinib. 

		Eiring AM, Khorashad JS,  Morley K, Deininger MW. Advances in the  treatment of chronic myeloid leukemia. 

		Kantarjian H, Giles F, Wunderle L, et al.  Nilotinib in imatinib-resistant CML and  Philadelphia chromosome-positive ALL. 

		Saglio G, Kim DW, Issaragrisil S, et al.  Nilotinib versus  imatinib  for newly diagnosed chronic myeloid leukemia. 

		Becker GS, Philipson TJ,  Soares RR. The  quantity and quality of life  and  the  evolution of world inequality. 

		Cox  DR. Regression models and  life-tables. 

. Series B (Methodological). 1972;34(2):187-220.

		Quintas-Cardama A, Kantarjian HM, Cortes JE. Mechanisms of primary and  secondary resistance to imatinib in chronic myeloid leukemia. 

		Lee RD, Carter L. Modeling and  forecasting U.S. mortality. 

The Option Value of Innovative Treatments in the Context of Chronic Myeloid Leukemia

Although clinical trial data have quantified patient survival gains associated with tyrosine kinase inhibitors in chronic myeloid leukemia, the overall value of these benefits is unknown.

To estimate the total value of survival gains associated with first- and second-line TKI therapy in chronic myeloid leukemia (CML) and the fraction of tyrosine kinase inhibitor (TKI)- related survival-gain value retained by patients and drug companies.

 This retrospective study identified CML patient data from the Surveillance, Epidemiology and End Results registry, dasatinib clinical trials, and insurance claims data sets.

 Multivariate Cox proportional hazard models were used to estimate improvements in CML survival associated with the introduction of first-line imatinib therapy. Survival gains associated with second-line dasatinib treatment were identified via retrospective analyses and published clinical outcomes. An economic model was developed to calculate the social value of survival gains derived from first- and second-line TKI treatment. TKI costs were used to estimate the fraction of survival gain value retained by patients and drug companies.

The introduction of TKIs in 2001 was associated with a hazard ratio of 0.833 (P <.01). Cost analyses indicate that the TKI drug class in CML therapy has created more than $143 billion in social value. Approximately 90% of this value is retained by patients and society, while approximately 10% is recouped by drug companies.

These estimates indicate that the introduction of TKI drugs to treat CML has generated significant social value as a result of survival gains, the vast majority of which has accrued to patients.

Cancer chemotherapy has long relied on generalized and nonspecific treatments that often involve substantial patient toxicity. One of the major chemotherapeutic advances over the past decade has been the discovery and deployment of targeted therapies directed at cancer-specific molecules and signaling pathways. In some cases this treatment specificity has proved very effective, appreciably improving survival rates for certain cancers.

Considered broadly, there have been substantial increases in cancer survival rates over the past several decades. The 5-year relative survival rate for all cancers diagnosed between 1996 and 2003 was 66%, up from 50% in the mid-1970s.

 Improvements in survival have been particularly dramatic for breast cancer, colon cancer, and non-Hodgkin lymphoma—conditions for which there have been important advances in therapy, screening, or both.

 It has, however, also been argued that some new cancer therapies only marginally extend life, and at a high cost.

 Hence, the cost-effectiveness of expensive new cancer therapies has been called into question, with recommendations made for more comprehensive economic evaluations, as well as innovative approaches that consider the unique economic impact of cancer treatments.

 The results of the recommended comprehensive economic analyses have important implications for the assessment of managed care, as they promote a better understanding of the total value of medical technologies.

The study described in this article examines the value of survival gains in chronic myeloid leukemia (CML) achieved by treatment with a new class of targeted therapy, tyrosine kinase inhibitors (TKIs), in first- and second-line therapy. In one multicenter, international, open-label, phase 3 randomized study, investigators compared the TKI imatinib with interferon as first-line treatment for CML, and demonstrated an 18-month survival rate of 97.2% for patients treated with imatinib.

 These results were durable; at 60-month follow-up, the overall survival (OS) rate for imatinib-treated patients was 89% (95% confidence interval, 86-92),

 and 7-year event-free survival and OS rates were 81% and 86%, respectively.

 Two newer TKIs, dasatinib and nilotinib, have also been approved for first-line CML treatment.

Studies of second-line TKI therapy in CML have also demonstrated efficacy. Patients receiving the newer TKIs (dasatinib and nilotinib) after imatinib failure experience significant cytogenetic response,

 and can expect substantial survival benefits. Indeed, in a randomized, international, multicenter, open-label phase 3 trial of dasatinib as second-line treatment following imatinib failure or intolerance, the overall 60-month survival rate was 78%.

Based on these clinical trial data, survival gains for CML patients treated with TKIs appear to be substantial. However, survival gains experienced in the community have not been studied, and the overall value of these survival gains—both in aggregate and relative to the cost of treatment&mdash;is unknown. In addition, the value of survival gains subsequently returned to patients and society has not been adequately evaluated. Prior research by Sun and colleagues

 revealed that, between 1988 and 2000, a substantial majority (~80%) of US cancerrelated survival gains were due to treatment improvements (as opposed to cancer detection strategies). Subsequently, Lakdawalla and colleagues

 evaluated, from a societal perspective, the economic gains due to cancer treatment associated with these improved overall survival rates; the study used an economic model developed by Becker et al,

 which was designed to assess patients’ willingness to pay for improved survival rates achieved by cancer treatment. Predicated on this model, the estimated average increase in life expectancy of 4 years between 1988 and 2000 was associated with 23 million life-years and $1.9 trillion gained in social value. The model also demonstrated that drug companies acquired 5% to 19% of this total value as revenue.

In this article we apply a similar approach to estimate the social value of TKIs in the treatment of CML. Observational data and clinical trial results were used to assess survival gains stemming from the introduction of both first- and second-line TKI therapies. Subsequently, an economic framework was employed to calculate the social value of improvements in survival gains due to TKI treatment. Lastly, to estimate the fraction of TKI treatment value retained by patients and society, as well as the proportion recouped by drug companies over time, the cost of TKI treatment was estimated from lifetime individual and population-level perspectives.

Multivariate Cox proportional hazard models were used to estimate survival improvements in CML associated with the 2001 introduction of first-line imatinib therapy. CML patient survival data were obtained from the Surveillance, Epidemiology, and End Results (SEER) registry. The SEER registry has tracked patients diagnosed with CML from 1973 to 2006 (the year of the most recent registry update). The key advantages of the SEER registry are that it is the only national cancer data base that: 1) follows patients over time in order to track survival, and 2) is large enough to contain sufficient sample sizes for survival analysis.

The sample for the survival analysis was restricted to CML patients diagnosed in 2000 or later. This ensured that the study captured the survival impact of the 2001 introduction of imatinib on the incident patient cohort, whose response to TKI therapy best represents the first-line survival effect of TKIs on present and future CML patients. Using multivariate hazard analysis, a rich set of demographic controls was applied, including age and squared age at diagnosis, gender, marital status, and separate indicator variables for white, black, Hispanic, Asian/Pacific Islander, and other races. To account for secular trends in CML survival unrelated to the introduction of imatinib, flexible controls for year and month of diagnosis were also included.

The key variable of interest in the proportional hazard model was the post-TKI indicator variable, which took the value of 1 in all years following 2000 and corresponded to the 2001 introduction of imatinib for first-line CML treatment. We report the estimated hazard ratio (HR) associated with the post-TKI variable, as well as the implied survival rate ratio.

The timing of second-line TKI approvals (2006 US Food and Drug Administration [FDA] approval for dasatinib; 2007 for nilotinib) precluded retrospective survival analysis of these drugs using the SEER registry. We estimated the community survival effect of new TKIs in second-line therapy by interpolation. First, we calculated the clinical survival effect of second-line relative to first-line treatment, which we obtained by dividing the 60-month OS rate in the second-line dasatinib clinical trial by the 60-month OS rate in the imatinib first-line clinical trial. We then applied this relative effectiveness rate of second-line therapy to the community-based survival effects estimated for imatinib in the retrospective SEER-based survival analysis. In performing this interpolation we assumed that the proportional difference in clinical and community survival rates for imatinib was the same as the proportional difference in clinical and community survival rates for dasatinib. Among newer TKIs we focused on dasatinib for this component of the analysis given the availability of recent 60-month survival results,

 which are not yet available for nilotinib.

The survival effects of imatinib were calculated as the average across all CML patients, regardless of whether they received imatinib; this was interpreted as the “intent-totreat” (ITT) effect. To estimate imatinib’s “treatment-onthe- treated” (TOT) effect, the ITT effect was scaled based on the percentage of CML patients who received treatment. The fraction of CML patients who received TKI treatment following its introduction was estimated using Ingenix Touchstone data,

 a database of private-sector healthcare claims, including prescription drug claims and inpatient emergency ambulatory medical claims from approximately 35 Fortune 500 firms. Individuals with CML were identified based on the presence of 2 or more International Classification of Diseases, 9th Edition, Clinical Modification (ICD-9-CM) diagnoses (primary, secondary, or tertiary) with medical claims codes of 205.1x, or 1 ICD-9-CM code and TKI therapy use after diagnosis. Based on this data set, between 2001 and 2002, 40% of CML patients received treatment with imatinib; therefore, the TOT effect was calculated as the ITT survival effect divided by 0.4.

The economic model developed by Becker et al

 was employed to calculate the social value of survival gains due to first- and second-line TKI therapy. Clinical evidence suggests that, relative to TKI therapy, prevailing treatments are associated with relatively rapid mortality in CML, not slight increments in mortality risk. This economic model calculates patients’ willingness to pay (ie, value to patients) for longevity gains, accounting for discrete (rather than marginal) increases in survival probabilities. The technical details regarding the application of this methodological approach are presented in 

In keeping with this methodology, the benefit derived by CML patients was calculated based upon survival gains pertaining to TKI treatment and the average income level among CML patients. In particular, patients’ lifetime value was estimated using income data from the 2005 and 2006 Medical Expenditure Panels Study (MEPS). Due to the small number of CML patients in the MEPS, all MEPS cancer patients were used to estimate CML patient income. The estimated value was then aggregated across all CML cohorts to obtain the total social value of survival gains due to TKI treatment.

The Ingenix Touchstone Data database provided estimates of annual imatinib treatment costs, as well as information on drugs utilized, total drug costs, and patient costsharing information. Because claims-based costs for dasatinib were not available, these costs were estimated via imputation by applying the proportional difference between the manufacturers’ prices of imatinib and dasatinib to the claims-based cost estimates.

As with the lifetime value of TKI treatment, lifetime treatment cost was determined by calculating the present value of annual imatinib treatment costs given the life expectancy implied by the post-TKI survival curve. The total value of TKI therapy was estimated by summing lifetime costs for a CML patient across all CML patients in a cohort, and then summing present discounted lifetime costs across all present and future cohorts. This took into account the decreased costs associated with the patent expiration of imatinib (2015) and dasatinib (2020). It was assumed that current lifetime and total cohort costs prevailed until patent expiration, after which costs were assumed to be zero. In addition, MEPS and SEER data were used to estimate the size of each CML cohort (4,500), of which 30% were expected to be resistant or intolerant to imatinib based on prior evidence in the literature.

 shows the patient characteristics of all CML patients in the SEER registry and the SEER CML sample. Incident cohorts of CML patients diagnosed in 2000 or later were included in survival-gain estimations associated with the introduction of first-line imatinib treatment. Restricting the data to SEER CML registrants for whom key demographic controls were reported generated a final analytical sample of 5,523 patients. Those in the CML subsample were demographically similar to patients from the full SEER CML sample, with the exception of 2 variables; namely, survival time (which was expected given the greater probability of survival among CML patients diagnosed at later dates) and Hispanic race (which likely resulted from a broader definition of “Hispanic” in more recently diagnosed patients).

 shows the results from the multivariate Cox proportional hazard analysis of survival impact following the 2001 introduction of first-line TKI therapy. The independent variable of interest is the post-TKI variable, which takes the value of 1 in 2001 and thereafter. The coefficient of this variable can be interpreted as the HR corresponding to the introduction of imatinib. Table 2 also depicts the most flexible form of the secular effects of time trend and time since diagnosis; year-of-diagnosis indicators and quadratic trend for month-of-diagnosis were included as controls for the secular trend in survival. As shown, the post-TKI variable was associated with an HR of 0.833 (P <.01). The implied TOT survival ratio of 1.307 can be interpreted as imatinib causing an immediate ~30% increase in the probability of CML survival at 90 months. This effect can be seen in the 

, which shows the estimated pre-TKI and post-TKI survival curves for the treated patients. The survival ratio is equivalent to the vertical distance between curves at 90 months. Alternatively, the HR can be characterized by the median survival length (the life expectancy of a 50th percentile CML patient; ie, the horizontal distance between the curves). The HR of 0.833 corresponds to an estimated increase in median survival from 90 to 210 months.

As previously discussed, the effect of imatinib was estimated from the sustained shift in the survival curve that occurred in 2001, the year of the drug’s introduction. As a sensitivity analysis, we evaluated whether shifts were observable in the survival curve at other years, as a large, sustained shift in survival occurring in any other year might imply that the estimated survival effect at 2001 was driven by factors unrelated to imatinib. To evaluate this, the survival analysis was repeated using a post-year variable for each year between 1998 and 2005. Reported P values for each estimated HR of the post-year variable are available in 

. The only statistically significant coefficient corresponded to the post-2001 (ie, post-TKI) variable.

OS rates reported in first-line imatinib studies13,26 were averaged, with response rates weighted by study sample size. This was then compared with OS rates in second-line dasatinib therapy.20 As shown in 

, the relative effectiveness of dasatinib in second-line therapy, as compared with imatinib in first-line therapy, was 91%. Under the assumption that the differences in clinical versus community survival rates would be the same for dasatinib and imatinib, we applied this 91% to the survival rate ratio of imatinib estimated in the SEER-based proportional hazard analysis.

 shows the cost of treatment, the value of survival gained through treatment, and the value of gains for patients treated with imatinib as a first-line response to CML. Estimated annual and lifetime costs are reported in the first line of 

. The annual value of TKI-related survival gains (column 1) is equivalent to the increase in annual income necessary to make a CML patient indifferent to the pre- and post-treatment survival curves, which is also equivalent to the patient’s willingness to pay for treatment. Column 2 shows the present lifetime value of survival gains per individual treated with imatinib, column 3 illustrates the lifetime value of TKI treatment across all newly diagnosed US CML patients over a 1-year period, and column 4 shows the lifetime value of TKI treatment for all present and future US CM L patient cohorts. Note that the fraction of total value recouped by drug manufacturers is lower in column 4 than in columns 1 through 3. This reflects the assumption that the cost to future CML cohorts of treatment after patent expiration is negligible relative to prices under patent. Columns 1 to 4 in

 show corresponding values for dasatinib in second-line therapy.

These results indicate that patients place an annual value of $110,600 on first-line TKI treatment relative to annual costs (under patent) of $30,500. This implies that, for all patients in present and future CML cohorts, the present social value of first-line TKI therapy is $88 billion. The present value of these costs, accounting for future patent expirations, was estimated to be $8.24 billion. Hence, approximately 91% of the social value of TKIs in first-line therapy is retained by patients, while 9% is recouped by drug companies. Likewise, dasatinib used as second-line CML therapy creates $55 billion in social value, of which patients retain roughly 90%, while 10% is recouped by drug companies.

This study evaluated the social value of CML survival gains achieved by treatment with TKIs. This analysis attributes the discrete 2001 HR drop in CML deaths observable in SEER data to the effect of imatinib treatment; this finding was robust to flexible controls for secular time trends in the survival analysis.

Tomas Philipson, PhD

Articles

Session 2: Discussant

Dynamic Cost-Effectiveness of Oncology Drugs

The Option Value of Innovative Treatments in the Context of Chronic Myeloid Leukemia

Value of Survival Gains in Chronic Myeloid Leukemia