Search

39 Background: Public reports suggest increasing integration of in-house specialty pharmacies by independent and system-based oncology practices. Understanding the impact of such integration has important implications for state-level regulations limiting physicians’ ability to launch pharmacies in certain states and policies governing pharmacy network design. In this analysis, I document the growth of in-house pharmacies and estimate the impact of this growth on oral cancer drug use and quality. Methods: I used 2006-2017 data from Medicare Fee-for-Service and Part D claims, pharmacy names and characteristics from health plan pharmacy network lists and the National Plan and Provider Enumeration System, and a unique Health System and Practice Dataset, developed by National Bureau of Economic Research and Harvard University researchers, that tracks practice ownership relationships. I first linked physician organizations and pharmacies that they operate using a novel claims-based algorithm combining information on pharmacy characteristics, similarities in names of practices and pharmacies, and the share of total pharmacy spending attributable to a single practice or system. I then used an event study approach to compare oral oncolytic use and quality at practices launching a pharmacy between 2009 to 2014 compared with matched controls. Outcomes included total spending on oral oncolytics, use of oral drugs when an intravenous equivalent exists (capecitabine and 5-fluoruracil), timeliness of new prescription fills (# days from first fill to last office visit), medication adherence, and early discontinuation of lenalidomide, tyrosine kinase inhibitors, aromatase inhibitors, tamoxifen, enzalutamide, and abiraterone. Results: The number of independent or system-based oncology practices with an in-house pharmacy filling any oncology prescriptions increased from 135 in 2006 to 442 in 2017. In that time, the share of Medicare Part D spending on oral oncolytics filled at in-house pharmacies increased from 4% in 2006 to 27% in 2017. The launch of an in-house pharmacy was not associated with an overall increase in spending on oral oncolytics, but was associated with a 4.5 percentage point (p =.01) increase in the use of capecitabine relative to 5-fluorouracil, within 2 years of launch. In-house pharmacies were associated with a modest decrease in the time to fill an initial prescription (2.4 days, p

e18794 Background: Dysphonia is a rare and under-reported adverse event. It is associated with the use of several cancer drugs. Dysphonia has been included as possible side effect on packet insert (PI) for some medications, like regorafenib, lenvatinib and axitinib. However, dysphonia has not been reported on multiple other drugs. We aimed to conduct a more comprehensive study to generate signal for dysphonia as adverse effect of drugs used for oncologic indications. Methods: The United States Food and Drug Administration (FDA) Adverse Events Reporting System (FAERS) database, a pharmacovigilance database, was used to extract data. All reported cases of dysphonia in the database were filtered for an indication of cancer. Descriptive analysis was conducted using SPSS 26. Results: Total 41840 cases of dysphonia were extracted from FAERS database. Out of these, cancer as an indication for medication use was noted in 2996 cases. Dysphonia was reported as side effect of 30 medications. Some of the most common medications were regorafenib (n=463), cabozantinib (290), lenvatinib (178), axitinib (176) and palbociclib (126) as noted in Table. Dysphonia was not listed as adverse reaction on PI for cabozantinib, palbociclib, sorafenib and oxaliplatin. However, it has been reported for regorafenib, lenvatinib and axitinib. Median age at diagnosis was 66 years (interquartile range 59-73). Dysphonia was reported more commonly in males (51.5%) as compared to females (46.2%). Reactions were reported to the FDA more commonly by consumers (53.3% cases) as compared to healthcare professionals (44.5% cases). Conclusions: This study demonstrates signal of developing dysphonia in patients receiving certain cancer directed medications based on FAERS database. It is worth noting that some of the suspect medications identified in this study do not have dysphonia listed as known adverse effect on accompanying package insert. Further studies need to be conducted to confirm if causal relationship exists between use of these drugs and development of dysphonia. Detailed description of five drugs used for oncologic indications with highest reported cases of dysphonia as adverse effect.[Table: see text]

e18842 Background: Many cancer drugs used OL, eventually complete clinical development and are registered and reimbursed, while others are not. We aimed to describe OL indications' registration and reimbursement prospects in a universal healthcare system and identify groups of treatments that remain OL longer. Methods: We studied 612 OL oncology prescription requests for 115 unique indications and 29 drugs. All requests were approved by the Institutional Drug Committee in Rabin Medical Center, a tertiary center in Israel, between January 2016 and December 2018. OL indications were not registered by the Israeli Ministry of Health (MOH) and were not included in the Israeli National List of Health Services (NLHS) for reimbursement at the time of prescription. Additionally, 459 requests for 100 indications were not approved by the FDA at the time of prescription. We explored the subsequent regulatory and reimbursement milestones in the timeline following each prescription. We identified three temporal events: FDA indication approval, MOH indication registration, and NLHS indication inclusion of reimbursement. We then defined three time-to-event variables from the OL prescription to each of the three milestones. We applied Kaplan-Meier analysis and multivariable Cox regression to identify drug and indication properties, which are predictors for registration and reimbursement. Results: With a median follow-up of 34.5 months, 10.4% of indications were eventually registered in Israel, and 6% were included in the NLHS. 7% of non-FDA approved indications were subsequently approved. For MOH registration, positive predictors were immunotherapy, sufficient evidence (ESMO-MCBS A-B, 5-4), and a patient access program at the time of prescription (HR = 21.99; p< .0005, HR = 3.4; p= .02 and HR = 2.73; p< .0005, respectively). Negative predictors were metastatic setting and lung and gastrointestinal (GI) cancers (HR = 0.42; p= .002, HR = 0.17; p= .001, and HR = 0.24; p= .043, respectively). For NLHS inclusion, positive predictors were immunotherapy, sufficient evidence, and an access program (HR = 9.03, HR = 6.24, and HR = 2.56, respectively, all p< .0005). Negative predictors were lung and GI cancers and orphan disease designation (HR = 0.07; p= .002, HR = 0.14; p= .021, and HR = 0.13; p= .009, respectively). For FDA approval, positive predictors were immunotherapy, sufficient evidence, and a patient access program (HR = 12.11; p< .0005, HR = 4.34; p= .002, and HR = 2.25; p= .001, respectively). Negative predictors were metastatic setting, lung cancer, and orphan disease (HR = 0.43; p= .003, HR = 0.25; p= .006, and HR = 0.08; p= .018, respectively). Conclusions: Few OL indications are subsequently approved and reimbursed, usually due to clinical futility and insufficient evidence. However, we identified factors related to lack of initiative for further clinical development and registration.

e13572 Background: Improved understanding of the underlying biology of cancer has led to a paradigm shift in cancer drug development and has paved the way for many products to receive accelerated or regular approval based on non-randomized/single arm trials (SATs). Given the high unmet medical need of cancer patients, challenges with lengthy and confounded survival endpoints, and difficulty enrolling rare biomarker-defined subsets of disease, SATs have been used to evaluate a variety of cancer therapies. Unlike time to event endpoints, the objective and clinically relevant endpoint of response rate (RR) and duration of response is interpretable in SATs, as spontaneous tumor shrinkage is not expected. Methods: A search of FDA databases identified all drugs and biologics approved for malignant hematology and oncology indications from January 1, 2001, to December 31, 2020 based on SATs. Data sources included approval letters, U.S. prescribing information, and clinical review documents. The definition of response varied by setting and time period (e.g. RECIST, WHO, IWG, etc.). Results: Between January 1, 2001 and December 31, 2020, FDA granted 153 new indications based on SATs, including 102 accelerated approvals (AAs) and 51 regular approvals (RAs). Overall, 69 approvals (45%) were for new molecular entities and 84 (55%) were expanded indications. Response rate was the most common endpoint used in the trial providing substantial evidence of efficacy to support approval [120/153, (78%)]. The durability of response was also considered to support evidence of clinical benefit. Of the 102 AAs, 38 (37%) have fulfilled their post-marketing requirement (PMR) to verify clinical benefit, 59 (58%) are pending verification of benefit, and 5 (5%) have been withdrawn from the market. Of note, 88% (52/59) of AAs pending verification of benefit occurred in the last 5 years alone (22 AAs in 2020, 8 in 2019, 8 in 2018, 12 in 2017, and 2 in 2016). Between 2001-2020, 58 (38%) new indications were granted for kinase inhibitors, 34 (22%) for immune checkpoint inhibitors (CPIs), and 61 (40%) for drugs with other mechanisms of action including but not limited to antibody-drug conjugates, cytotoxic drugs, and non-CPI monoclonal antibodies. Conclusions: In the last two decades, SATs have been effectively used to study anti-cancer therapies in well-defined patient populations using durable RR as an objective and interpretable clinical endpoint. Although randomized clinical trials remain the gold standard in clinical research, SATs have allowed for rapid advancements in oncology drug development and will continue to serve an important role in bringing new therapies to patients with unmet need.

e13597 Background: In recent years, efforts to improve the efficiency and speed of drug development and approval have driven a surge of Food and Drug Administration (FDA) approvals for cancer drugs. For many cancer therapies, the serious or life-threatening nature of the condition and unmet medical need confers eligibility for expedited programs. Many cancers are also rare diseases, and the increasing use of precision medicine principles to define cancer types further contributes to smaller trial sizes. With limited clinical evidence at the time of approval, cancer drugs may be subject to a greater burden of postmarketing requirements (PMRs). We analyzed PMRs for solid tumor therapies approved by the FDA over the past decade. Methods: The FDA’s novel drug approvals (2010-2019) were reviewed to identify drugs receiving primary approval for solid tumor indications. Approval letters were accessed via the Drugs@FDA database and analyzed for PMRs required under accelerated approval (AA), the Pediatric Research Equity Act (PREA) and the FDA Amendments Act of 2007 Section 505(o) (505(o)). Data are presented using descriptive statistics. Results: A total of 60 drugs received primary approval from the FDA for solid tumor indications between 2010 and 2019 (20 [33.3%] received AA, 33 [55.0%] received orphan designation, and 45 [75.0%] received Fast Track or Breakthrough Therapy designation). The proportion of drugs receiving AA doubled between the period 2010-2014 and 2015-2019 (Table). Of the 60 drugs approved, 52 (86.7%) received a total of 180 PMRs. All 20 drugs approved under AA received PMRs, with a total of 25 PMRs issued under AA. Data from new clinical trials were required for 22 (88.0%) of the 25 PMRs. No PMRs were issued under PREA. Additional safety data required under 505(o) comprised the largest proportion of PMRs; 155 total PMRs (86.1% of all PMRs) were issued for 45 (75.0%) of the drug approvals. Pharmacokinetic or other clinical safety data were required for 96 (61.9%) of the 155 PMRs. Conclusions: More than three-quarters of the cancer drugs approved for the treatment of solid tumors in the past 10 years were issued PMRs, with the majority requiring new safety data. The results of this study indicate that PMRs represent a critical mechanism by which FDA collects safety and efficacy for solid tumor therapies, and underscore the importance of PMR fulfillment. Post-marketing requirements (PMRs) for solid tumor drugs approved 2010-2019.[Table: see text]

e13599 Background: In the Netherlands, the health insurance package usually covers cancer drugs after EMA approval and positive recommendation of the Dutch Society of Medical Oncology (NVMO). However, the government can temporarily exclude newly licensed indications with high budget impact from coverage until the national HTA-body advises on the clinical effectiveness and cost effectiveness. Based on the recommendations of the HTA-body, the government negotiates a financial arrangement with manufacturers. Because submitting value dossiers, assessments and negotiations takes time there have been concerns regarding patients’ access to these novel treatments, especially for life threatening conditions such as cancer. Due to a lack of transparency, it is not known to what extent patients are given access to ‘free of charge programmes’ (FCPs). Methods: A list of oncology drugs and indications which were temporarily excluded from reimbursement due to negotiations with the government (July 2015-Dec. 2020). We contacted all the pharmaceutical companies that were engaged in the process above, to inquire about access to their products during the HTA evaluation and negotiations. Results: In total 17 oncology drugs (46 EMA registered indications) of 11 companies were subjected to financial agreements by the government. In 41,3% (19/46) of cases, FCPs were available via individual prescriber requests in a selected number of hospitals. In all other cases such arrangements were not needed because alternative treatment options were available (15,2%; 7/46) or because treatment was not recommended by the NVMO (13,0%; 6/46). For one indication FCP was available but the treatment eventually received a negative recommendation. In 30,4% (14/46) of the cases there was immediate reimbursement in place following a positive NVMO recommendation through umbrella agreements. Such agreements cover existing indications but also all the indications that will be licensed in the future. Umbrella agreements were especially set up for immune checkpoint inhibitors that showed benefit across multiple indications. As anticipated, some patient-access programmes had already started prior to EMA approval. In 5 cases an official compassionate use programme was set up through the Dutch Medicines Evaluation Board prior to authorisation. Conclusions: Our survey indicates that all manufacturers are committed to enabling patients’ and clinicians’ early access to their products pending payers’ approval in a high income country such as the Netherlands. However, the lack of transparency regarding the hospitals offering these programmes may result in inequity in access. Besides, there is no structural data collection. Recently, the NVMO has established a platform (DRUG Access Protocol) to harmonise all compassionate use programmes and FCPs, generating real-world data and ensuring equal access for all patients.

e18811 Background: An open-access, web-based platform was developed to provide standardized, evidence-based information, that is cancer regimen-specific and supportive care focused. Online content was provided in the English language only for more than 200 anti-neoplastic regimens. A survey was conducted to determine reasons adult patients and their caregivers seek internet-based cancer drug information, and to evaluate if there is an unmet need for improved oncology therapy education. Methods: Between February 2020 and January 2021, patients and caregivers completed an online anonymous, English language survey deployed on the website www.chemoexperts.com. Results: A total of 1,021 website users responded. The majority of users were from North America (75%). Female respondents comprised 65% of the population and users self-identified as being a patient (67%), or family member/caregiver (33%). The mean age of respondents was 60.7 years (+ 16.2). Roughly two-thirds (66.4%) had a college degree. The majority of respondents reported they were either currently receiving (59.6%), or planning to receive (27.4%) treatment, however patients of all education levels, and in all phases of treatment (before, during, and after) sought online drug information. Clinical drug information education was reportedly provided by a doctor (68.9%), nurse (40.9%), physician assistant or nurse practitioner (23.8%), or pharmacist (15%), while 23% did not receive any education from clinical staff. Modes of education received by participants included printed material (73.9%), teaching in person (52.3%), and/or other internet sites (27.5%). Reasons for visiting the website were reported as follows: seeking additional information (92%), did not know what questions to ask (12%), looking at treatment alternatives (10%), forgot to ask questions (5%), not enough information given (11%), not enough time to ask questions (5%), and afraid to ask questions (2%). Respondents could mark all reasons. Among content sections, the percentage of users reported the side effect section as being the most useful. However, section usefulness varied based upon age, phase of treatment, and baseline education level. Overall, 86% reported finding the information they were looking for on the website. Conclusions: The majority of respondents received education from a health care provider, but not all. It is possible that the current pandemic prevented some patients from receiving education when in-person teaching was not available. However, many patients and caregivers still use the internet to seek additional drug information. These data highlight a continued unmet need for patients using online sources when searching for cancer drug and supportive care information. Further work is required to determine whether supplemental online, education can improve outcomes while reducing adverse effects.

e18580 Background: Progression-free survival (PFS) is accepted as a surrogate endpoint for cancer drugs approval by the Food and Drug Association. In the expanding immunotherapy era, there is no convincing data for reliable surrogacy of PFS. Methods: We systematically searched PubMed and included all phase III trials of ICIs. Besides, we manually included phase III trials of ICIs presented in congresses. Trials including non-metastatic patients and hematologic malignancies were excluded. The primary endpoint was the relationship between treatment effects (i.e., hazard ratio (HR) of PFS and overall survival (OS)). The secondary endpoint was the relationship between the median PFS (mPFS) and median OS (mOS). We also performed subgroup analyses by ICIs subtypes (i.e., anti-programmed death-1 (PD-1) and anti-programmed death ligand-1 (PD-L1)), treatment lines (i.e., the first-line and subsequent treatments), ICI plus CT combination, and ICI monotherapy. To ensure parametric assumptions, we addressed logarithmic conversion to HRs by using log10. We used Pearson’s correlationand weighted linear regression for parametric variables and Spearman’s rank-order correlation for non-parametric variables. We assessed the relationship between variables by using correlation coefficient (r) and coefficient of determination (R2). Results: We included 57 phase III clinical trials with 39,525 patients. A total of 67 outcomes were compared. There was a good correlation between the logarithmic HR of PFS (log HRPFS) and OS (log HROS) (r=0.71, R2=0.50 pPFS and log HROS in anti-PD-L1, ICI plus CT, and the first-line treatment subgroups (r=0.61, R2=0.38 p=0.004 for anti-PD-L1; r=0.65, R2=0.43 p=0.008 for ICI plus CT; r=0.70, R2=0.49 pPFS and log HROS in the anti-PD-1, ICI monotherapy, and subsequent treatment line subgroups (r=0.81, R2=0.66 p2=0.51 p2=0.62 p

66 Background: This systematic review and meta-analysis compared incremental cost-effectiveness ratios (ICERs) of cancer drugs and their relationship with society’s willingness to pay (WTP) across different countries, scenarios and indications. Methods: We sought for cost-effectiveness studies in PubMed, Embase, Cochrane and LILACS published from December 2012 to December 2017. CAR-T cell therapies were excluded given short follow-up. We converted ICERs value and respective annualized 1time the Gross Domestic Product per capita (GDP) - used as a proxy for WTP threshold - into purchase-parity-power dollars (PPP) for better economic reality comparisons. Economics data came from the International Monetary Fund website. Characteristics studied in the distribution of ICERs values were compared using the Mann-Whitney or Kruskal-Wallis U-test with multiple comparison correction and simple linear regression. The chance of ICER being below the 1x GDP threshold was expressed as odds ratio and 95% confidence interval, calculated by logistic regression. Results: We retrieved 354 drug-versus-drug different assessments. PPPconversion increased median ICER of middle-income countries ($32k to $68k;P < 0.001). Median ICER was highest in USA ($154k), followed by other high-income countries ($76k; P < 0.001). In multiple regression, anti-VEGF (P < 0.001) and US-led (P = 0.03) studies had highest ICERs, while curative therapies (p = 0.02) had the lowest ones. 22% of studies were below the WTP bar, none of anti-VEGF, sipuleucel, anti-CD-30, castration-resistant, hepatocarcinoma, and renal cell ones. Immune checkpoint inhibition (OR 8.70; P = 0.045) and middle-income (OR 7.40; P < 0.001) studies were more likely above WTP, whereas curative therapies were more likely below WTP. Conclusions: Cancer therapies’ cost exceeding the WTPs is a worldwide pattern, with some factors related to ICER variation. These findings may foster better understanding and aid stakeholders deal with the global issue of high oncology care costs.

2060 Background: The FDA has issued hundreds of cancer drug indications, with many new drugs, expanded indications, and biosimilars approved in recent years. While the gold standard for regulatory approval is the randomized controlled trial (RCT), RCT design including selection of control arms can differ considerably. We sought to investigate trends and patterns in RCT trial design used to support FDA approvals in oncology. Methods: We reviewed the available FDA package inserts of oncology drugs (N=258) for RCTs cited to support initial and expanded indication approvals as of January 2020; biosimilars were excluded. RCTs were linked to the HemOnc ontology, which contains trial-level metadata including publication year, endpoints, and trial design. Log-linear regression was performed to evaluate trends in approvals over time by endpoint. Study drugs were categorized as cytotoxic therapy, targeted therapy, or immunotherapy. RCTs were categorized by four designs: escalation (adding a drug or increasing the drug dose in an established regimen), in-class comparison (comparing two drugs in the same therapeutic class), out-of-class switch (comparing drugs in distinct therapeutic classes), and de-escalation (removing a drug or reducing the drug dose in an established regimen). Results: We identified 556 registration trials, 372 (67%) of which were RCTs. Approvals have been increasing exponentially over time (R2 0.9, p2 0.77, p2 0.67, p

e19391 Background: Emerging oncology drugs such as targeted agents and immune checkpoint inhibitors (ICIs) have brought dramatic advances in cancer treatment, while rapid rise in the drug cost. In Korea, the policy of expansion of health insurance coverage has improved accessibility to high-priced oncology drugs through their registrations in reimbursement list. However, this made a considerable impact on the financial burden of national health insurance. The objective of this study was to assess the awareness and perceptions of medical oncologists about the value frameworks of cancer drugs in Korea. Methods: We collected the data from survey for 102 medical oncologists at cancer conference. They were asked about prioritization among multiple considerations when they prescribe cancer drugs, The second question was awareness of value frameworks such as American Society of Clinical Oncology (ASCO) value framework and European Society for Medical Oncology Magnitude of Clinical Benefit Scale (ESMO-MCBS). Additionally, potential application of value framework to the decision on reimbursement coverage was included. Results: The majority of respondents (90%) chose clinical efficacy as the most important factor when they selected oncology drug for cancer patients. Safety/tolerability and drug cost were followed as 5.9% and 1%, respectively. Meanwhile, the order of priority of considerations for ICIs treatment was as follows: clinical efficacy (74.5%), drug costs (13.7%), and safety/tolerability (9%). More than half of those questioned (53.9%) were already aware of ASCO value framework and ESMO-MCBS, while 26.5% did not recognize them at all. Over 90% of respondents agreed with the need for development of new value framework tool which can be complementary to current valuation system for oncology drugs. Oncologists responded positively to the necessity of the assessment tool as a criterion for reimbursement registration (84.3%) as well as post-reimbursement re-evaluation (89.2%). Conclusions: Our results suggest that value assessment tool of oncology drug be necessary for providing medical evidences for clinical decision and for determination to health insurance reimbursement in Korea.

e16790 Background: Over the past 20 years, cancer drugs have contributed to increased life expectancy, reduced mortality, decreased hospitalization and decreased use of medical services. The economic value of these improvements is about as large as the value of the increase in the US gross domestic product during that time period. Recently, a health economic study presented at ASCO GI 2020 cited that every $1 (adjusted for inflation) spent on innovative PC treatments reduced non-drug expenditures by $9, thereby lowering the total cost of care for PC patients. Accordingly, the commercial opportunity of a new therapy should be measured by some combination of the clinical, economic and social value generated. We demonstrate the value of a novel PC drug from this perspective. Methods: Analysis of SEER survival and incidence data between 2008 and 2016 shows the introduction of new medicines for PC of all stages was associated with a cumulative increase of 26,456 life years, or 2.52 life years per patient. It was also associated with quality of life improvements, measured by a decline in hospitalizations rates and emergency room visits that can also lead to more days at work, at school and with family. Several studies have suggested the average value of an additional year of life, for the age of a typical patient diagnosed with PC, is at least $250,000. Using this figure, the value of 26,456 life years gained from 2008-2016 is $6.61 billion (26,456*$250,000) to patients, the healthcare system and society, as a result of advancing medical innovation for patients with PC. Results: The median annual list price of a life-enhancing cancer therapy is $150,000 per patient. Using the NCI treatment prevalence estimator (holding incidence constant), we estimate that between 2020-2025, there will be an additional 10,728 advanced PC patients requiring treatment who could benefit from innovative drugs. The total cost of these drugs for these patients would be $1.61 billion. However, the economic value of the life years saved would be $6.76 billion (10,728*2.52 life years*$250,000 = $6.76 billion). A review of cancer medicine payor coverage suggests a new PC therapy that produces such value would be able to obtain coverage from US payors given this value-based price. Conclusions: A value-based approach to estimating the opportunity for clinical and economic benefit reveals significant potential for new PC medicines.

7071 Background: The cost of many cancer drugs is very high, but it is unclear if these costs are associated with commensurate improvement in outcomes. We aimed to assess the association between the cost of cancer treatments and their clinical benefit, using the NCCN Evidence Blocks value assessment framework. Methods: The NCCN Evidence Blocks include 4 measures of clinical benefit: Efficacy, Safety, Quality of Evidence, and Consistency of Evidence. The NCCN assigns scores on each measure ranging from 1 (least favorable) to 5 (most favorable). We obtained the NCCN Evidence Blocks scores as of December 31, 2018 for all recommended cancer treatments for the 30 most prevalent cancers in the US. For each treatment, we calculated total treatment costs (including drugs, administration fees, and supportive care medications) using Medicare reimbursement rates. We categorized treatments as either “time-limited” or “time-unlimited” according to whether their costs are best reflected as per full treatment course (often, adjuvant/neoadjuvant treatments) (time-limited) or per month of therapy (often, treatments for advanced disease) (time-unlimited). We used generalized estimating equations, with clustering within treatment indications, to estimate the association between Evidence Blocks scores and treatment costs, modeling the expected change in cost associated with a one-unit increase in the score on an Evidence Blocks measure. Results: There were 541 time-unlimited and 845 time-limited treatments. Among time-unlimited treatments, monthly treatment cost ranged from $4 to $64,630. Monthly treatment cost was positively associated with Efficacy ($3,036, 95%CI: $1,782, $4,289) and Quality of Evidence ($1,509, 95%CI: $171, $2,847) but negatively associated with Safety (-$1,470, 95%CI: -$2,790, -$151) and Consistency of Evidence (-$2,003, 95%CI -$3,420, -$586). Among time-limited treatments, cost per course of therapy ranged from $0 to $775,559, and no measure was significantly associated with cost. Evidence Blocks scores accounted for little of the variation in treatment cost (linear model R-squared = 0.10 for time-unlimited, and < 0.01 for time-unlimited). Conclusions: The association between NCCN Evidence Blocks measures and treatment cost was inconsistent, and accounted for little of the cost variation among treatments for the same indication. The clinical benefit of cancer treatments does not appear to be a primary determinant of treatment cost, suggesting that current pricing models may be inadequate to incentivize the development and utilization of high-value treatments.

e14133 Background: Historically, pharmaceutical companies submitted phase III evidence for consideration of public reimbursement; however, phase II data is being more commonly used as primary evidence. Whether submissions with phase II data lead to similar rates of positive reimbursement recommendations as phase III data has not been comprehensively investigated. We compared frequency of reimbursement recommendations between phase II and phase III submissions for oncologic drugs and assessed for factors associated with a positive or conditional recommendation. Methods: We identified all submissions with phase II data from the CADTH pCODR’s expert review committee (pERC) recommendations from July 2011 to July 2019. We identified fourteen binary variables relating to clinical benefit, patient-based values, economic impact, and adoption feasibility. We used Fisher’s exact test to characterize associations between all variables and the final recommendation. We conducted multivariable analysis with logistic regression for three variables: feasibility of phase III study, hematologic indication, and unmet need. Results: We identified 139 submissions with a pERC final recommendation. Twenty-seven (19%) submissions were supported by phase II evidence, with 63% having a positive recommendation in comparison to 82% among submissions with phase III evidence. Clinical benefit (p < 0.001), gap in current treatment standards (p = 0.047), and patient alignment (p = 0.015) were associated with a positive recommendation, whereas the future feasibility of conducting a phase III study was associated with a negative recommendation (p = 0.040). No significant association was found between the recommendation and factors related to cost effectiveness or adoption feasibility. In multivariable analysis, only feasibility of a phase III study was significantly associated with a negative recommendation (p = 0.024, OR = 0.132). Conclusions: Oncologic submissions with phase II data were less likely to be recommended for public reimbursement than phase III studies. Positive or conditional recommendation was more likely if they demonstrated clinical benefit and aligned with patient values. pERC was less likely to recommend a submission with phase II if a phase III trial was either possible or already initiated.

2077 Background: A key clinical outcome for new cancer drugs is improvement in overall survival (OS), defined as time from the date of randomization to the death from any cause. However, many cancer drugs are approved by regulators based on changes to surrogate measures of OS, such as progression-free survival or overall response rate. When surrogate measures are not validated, they can provide misleading information about drug efficacy. We categorized pivotal trial endpoints for recently-approved cancer drugs in the US and Europe as showing improvements in OS vs non-OS surrogates, and evaluated the correlation with drug prices. Methods: We identified new drugs FDA-approved between 2009 and 2018 that were indicated to treat solid and hematologic tumors in adults and that had also been approved by the EMA and Swissmedic by December 2019. Launch prices were extracted and adjusted to average sales prices for monthly treatment costs in the US and compared to currency-adjusted ex-factory monthly treatment costs in Germany, Switzerland, and England. Pivotal clinical trial primary endpoints were collected from the drug labeling and FDA medical reviews for the US, and the EMA public assessment reports for Europe, and categorized as OS in any trial vs. not. Pearson’s correlation tests assessed the association between launch prices and OS vs non-OS endpoints in each country. Results: 54 drugs were approved by the FDA, EMA, and Swissmedic during the study period. In the US, 30 (56%) were approved based on OS by contrast to 35 (65%) in the EMA. The number of cancer drugs approved by the FDA based on OS decreased in the past years. By contrast, the number of approved cancer drugs by the EMA based on OS were stable. There was no association for the US (p = 0.05), Germany (p = 0.13) and England (p = 0.12), while Switzerland revealed an association (p = 0.03) between OS endpoint and price. Conclusions: Reductions in use of OS endpoints as the basis for cancer drug approval in the US is concerning. Drug pricing should be better aligned with the benefit that drugs provide to patients, as measured by clinical trial outcomes such as OS.

2006 Background: Cancer drug costs are rising in the US and Europe. While drug manufacturers set prices without restriction in the US, European countries have regulations that allow national authorities to directly negotiate drug prices at launch and over time. We analyzed and compared the launch prices and price developments of cancer drugs in the US, Germany, Switzerland and England. Methods: We identified new drugs indicated to treat solid tumors in adults that were FDA-approved between 2009 and 2019 and had also been approved by the EMA and Swissmedic by 31 December 2019. Launch prices and post-launch price changes as of 1 January 2020 were extracted and adjusted to average sales prices for monthly treatment costs in the US and compared to comparable currency-adjusted ex-factory monthly treatment costs in Germany, Switzerland, and England. A cross-sectional analysis was conducted to infer yearly trends in launch prices and post-launch price changes across the countries. Results: The study cohort included 42 drugs for solid tumors, of which 40 (95%) drugs were first approved in the US compared to Germany and England, and 41 (98%) to Switzerland. Average launch prices for monthly treatment costs per patient were $15,178 in the US vs $7,049 in Germany, $7,421 in Switzerland and $8,176 in England, i.e., 215% (interquartile range [IQR] 263%-187%), 205% (IQR 202%-185%) and 186% (IQR 166%-189%) higher in the US compared to Germany, Switzerland and England respectively. Post-launch prices of 36 (86%), 40 (95%), and 38 (90%) drugs decreased over time with total savings of monthly treatment costs for all drugs in the study cohort of $86,744, $44,936, and $1744 in Germany, Switzerland, and England respectively. By contrast, prices of 8 (19%) drugs decreased, while 34 (81%) increased post-launch in the US with total additional expenses of $128,192 for monthly treatment costs. Conclusions: Launch prices for cancer drugs are far higher in the US than in Germany, Switzerland, or England. These price disparities continue to increase substantially after market entry since cancer drug prices, in general, decrease over time in Europe and increase in the US. Spending on cancer drugs could be reduced in the US if it adopted the principles used to more effectively negotiate drug prices in Europe.

7011 Background: Regulatory approval of oncology drugs are often based on data presented in the primary publication of clinical trials (CT). However, clinically relevant data, such as long-term overall survival (OS) and quality of life (QOL), are often reported in subsequent publications. Therefore, this study aimed to evaluate the ASCO-VF NHB at the time of drug approval and over time as further evidence is published. Methods: All FDA approved oncology drug indications from 01/06-12/16 were reviewed to identify CTs that were scorable using the ASCO-VF version 2. Subsequent publications of included CTs relevant for scoring were identified from Web of Science with a follow-up time of 3 years from approval. Using ASCO-defined threshold scores of ≤40 for low benefit and ≥45 for substantial benefit, changes in classification of benefit were assessed at 3-years post-FDA approval. Results: We identified 57 FDA approved indications (40.4% OS, 59.6% progression-free survival (PFS) as primary endpoints) with scorable ASCO-VF CTs. Among those 57 indications, 36.8% at the time of FDA approval demonstrated substantial benefit, 10.5% demonstrated intermediate benefit, and 52.6% demonstrated low benefit. We then identified 96 subsequent publications relevant to scoring within 3-years of FDA approval, consisting of primary endpoint updates (29.2%; 14.6% OS, 12.5% PFS), secondary endpoint updates (44.8%; 16.7% OS, 7.3% PFS), new reporting of secondary endpoint (4.2% OS), safety updates (28.1%), and QOL reporting (43.8%). Upon reassessment of the NHB in subsequent publications, there was an overall change from initial classification of benefit in 36.8% of trials (17.5% became substantial, 8.8% became low, and 10.5% became intermediate). Changes in scores were mainly the result of an updated hazard ratio (35.1%), change in scoring endpoints from PFS to OS as per ASCO-VF endpoint hierarchy (8.8%), toxicity updates (57.9%), new tail of the curve bonus (12.3%), palliation bonus (14.0%), or QOL bonus (22.8%). Overall, at reassessment at 3 years post-FDA approval, 42.1% were substantial, 10.5% were intermediate, and 47.3% were low benefit. Conclusions: Only a modest proportion of FDA approved drugs have demonstrated substantial NHB at time of approval. As further evidence was published, a substantial proportion of indications have a change in classification of NHB, resulting in a small increase in the overall proportion of indications being deemed to have substantial benefit at 3 years post-approval.

7083 Background: Studies assessing wastage of discarded weight (WT)-based cancer drugs (can-RX) derived their estimates from Medicare Average Sales Prices + 6% mark-up. With higher mark-ups, financial impact of discarded can-RX can be much higher in private insurance market. In addition, no study has examined the financial burden for patients(pts) even though they are subject to out-of-pocket payment (OOP) of the discarded RX. This study fills in these important knowledge gaps. Methods: We obtained the list of WT-based can-RX from the 2017 Centers for Medicare and Medicaid Services Part B Discarded Drug Units Report and gathered information on package size and recommended dose from IBM Micromedex. We identified pts who received WT-based can-RX from 2017 MarketScan data and linked to MarketScan Health Risk Assessment to obtain WT and height for a subset of pts. For each claim, we derived the recommended dose based on pt’s RX and his/her WT or body surface area (calculated from WT and height) and estimated the % of discarded dose. We quantified OOP as the sum of deductible, copayment, and coinsurance, and classified pts’ insurance by their enrollment to high-deductible (HD) plans. For claim-level analysis, we applied beta regression (for fractional outcomes) to determine factors associated with % discarded while accounting for within pts correlations. Covariates were age group (grp) ( < 50, 50-59, > = 60), WT grp ( < 150, 150-199, > = 200lb), HD plan (yes/no), region, cancer type, and can-RX. For pt-level analysis, we calculated the amount payers and pts spent on all and discarded can-RX in 2017 and compared OOP b/t pts in HD and non-HD plans. Results: Of 27,883 claims of WT-based can-RX, the median discarded % was 0.10 (mean 0.19, SD = 0.21). In addition to can-RX and cancer type, significantly higher discarded % was found in the lowest WT grp (0.03, P < 0.001, vs. WT > = 200lb) and significantly lower % was found in the middle age grp (-0.01, P = 0.02 vs. age < 50). Of 1963 pts, 778 (40%) had HD plans. Payments by payers (NETPAY) and pts (OOP) for total and discarded can-RX and the comparison b/t HD and non-HD plans are shown in Table. Conclusions: Payers incurred substantial financial burden from discarded can-RX. While OOP costs were modest for most privately insured pts, 5.3% of pts (~8% in HD plans) paid > $400 for discarded RX. [Table: see text]

641 Background: Having standardized and high quality reporting of patient reported outcomes (PRO), especially in clinical trials that establish standards of care in oncology is important for patient centered care. This study assessed the status of reporting and quality of analysis of PROs in FDA approved drugs for genitourinary malignancies. Methods: We conducted a systematic review of the FDA archives to identify urological cancer drugs approved between 2007-2018. We retrieved the clinical trials that led to these drug approvals from ClinicalTrials.gov and PubMed. We systematically screened for PROs and reviewed their analytic tools and interpretation methods reported in their published manuscripts and study protocols. A clinical trial was considered to include PROs if they were reported in either the primary or a subsequent manuscript. Results: We identified 22 clinical trials leading to FDA approval of urological cancer drugs between 2007-2018. Only 63% of trials had published PROs. PROs were reported in the primary clinical trial manuscripts for two drugs (9%), and in a secondary PRO manuscript for 12 drugs (54%). The median time between the primary and secondary papers was 12 months (IQR:7.5-26 months). Among the 14 published PRO papers, the hypothesis was broad in 79%, and not reported in 21%. PROs were never included as a primary endpoint of a study. Instead, PROs were reported as secondary endpoints in 5 (36%) and as exploratory endpoints in 7 (50%) studies, while two papers (14%) did not mention PRO reporting in their endpoints. The most common PRO instruments were EQ-5D (64%) and FACT-P (50%). In 92% of PRO papers, statistical analyses were conducted to account for missing data. Control for type I error was needed but not done in 57% of the trials. Conclusions: Delays in publication of PROs occur regularly in trials leading to drug approval in GU malignancies. Our study highlights the need to enhance standardization of the analysis and interpretation of PROs to maximize the value of this data for drug development and approval.

809 Background: Proportionate representation of women in health research is an area for improvement. This study aims to assess the representation of women in gastrointestinal (GI) cancer clinical trials leading to FDA cancer drug approvals over the past 10 years. Methods: FDA cancer drug approvals between 07/2008-06/2018 were identified and trial reports supporting approvals sourced. The ratio of female to male (F:M) enrollment was compared with F:M cancer incidence in the U.S., and U.S. cancer prevalence and mortality. Results: Although F:M enrollment for all 229 trials leading to FDA cancer drug approvals in this period was similar to overall F:M cancer incidence in the U.S. (0.89 vs 0.86; Odds Ratio for female enrollment (OR) 1.05, 95% Confidence Interval (CI) 1.03-1.06, P

185 Background: 5-Fluorouracil (5FU) is one of the most commonly used cancer drugs. FdUMP is a metabolite from 5FU which inhibits thymidylate synthase (TS). TS inhibition causes a rise in the intracellular pool of the natural TS substrate dUMP leading to increased global levels of deoxyuridine (dUR). In order to increase TS inhibition, 5FU is combined with folates, usually leucovorin (LV). However, LV needs enzymatic conversion to become active. The LV conversion capacity differs between patients, with the consequence that many patients will have a reduced effect of 5FU-based treatment. Arfolitixorin is a biologically active folate, 5,10-methylenetetrahydrofolate, which doesn´t need enzymatic conversion; hence it could potentially be more efficacious. Methods: ISO-CC-005 is a phase I/II safety and tolerability study in mCRC patients receiving 5FU/arfolitixorin alone or in combination with irinotecan or oxaliplatin ± bevacizumab. Efficacy was evaluated as Overall Response Rate (ORR) after 4 cycles of chemotherapy. A LC-MS/MS method was developed for quantification of plasma dUr and 5FU. Blood samples were collected before and 24 h after 5FU administration at the first (C1) and fourth (C4) treatment cycle (n = 33). The C4/C1 ratio for dUr and 5FU, respectively, were calculated.The clinical response, which is an ordered categorical data, was modelled using nominal logistic regression using the ratio C4/C1 and the absolute values C1 and C4 as explanatory variables. Results: A positive correlation between dUr and 5FU levels was seen (r = 0.89, p = 0.0011) at C4. The C4/C1 ratio for dUr and 5FU in plasma correlated positively (r = 0.95, p = 0.0001). A positive correlation was also found between arfolitixorin dose and plasma levels of dUr (r = 0.44, p = 0.016) at C1, but not at C4. In the model test, the C4/C1 dUr ratio, as well as the individual dUr levels at C1 and C4, had a highly significant effect on ORR (p = 0.0075). Conclusions: The results demonstrate that the arfolitixorin dose has an impact on plasma levels of dUr and also on clinical response. dUr could potentially be an early surrogate marker for TS inhibition in mCRC patients treated with 5FU/arfolitixorin. Clinical trial information: NCT02244632.

112 Background: Rising costs of cancer therapies calls for strategizing towards sustainable care delivery models from a hospital planning, payer as well as policy-making perspective. This topic becomes increasingly significant as there is exponential growth of novel, high-cost immunotherapy drugs making it imperative for players to adopt and practice value-based oncology. Objective: This study aims to evaluate increasing trends in use of new and transformative cancer therapies, and associated drug costs in a comprehensive cancer center in Massachusetts. Methods: Study period was fiscal year (FY) 2015-2018. Utilization was defined by the number of patients receiving infusion and number of visits made by them to the center during the FY. Cost was defined as expense to hospital. Data source was hospital billing database. ASP (Average Sales Price) of drug was obtained from CMS website and was used as an indicator for cost per unit of the drug. Results: Top 10 drugs were identified based on total cost incurred in the study period and contributed to almost a third of center’s total cost. Over the three years, number of visits for these drugs grew up to 700% and treated patients grew up to 350%. Use of chemotherapy in isolation decreased from 35% of treated patients in FY15 to 26% in FY18, whereas its use in combination with immunotherapy increased from 22% in FY15 to 28% in FY18. Average drug cost to hospital per patient for the study period ranged as high as $120,000 (excluding non-drug treatment costs). Conclusions: While clinical value of the new cancer therapies is unquestionably significant, there is a dire need for policy-makers, providers and payers alike to pay continued attention towards its high cost implications as observed in this study and, continue striving towards establishing more sustainable pricing policies through alternative payment models.

6513 Background: The Breakthrough Therapy program was established in July 2012 to expedite drug development and approval by the FDA. We compared the characteristics of clinical trials leading to FDA approval as well as the magnitude of clinical benefit and value framework scores of breakthrough-designated and non-breakthrough-designated cancer drugs. Methods: We searched the Drugs@FDA website for cancer drug approvals from July 2012 and December 2017. For each indication, we applied the value frameworks and used thresholds of high clinical benefit developed by American Society of Clinical Oncology Value Framework version 2 (ASCO VF v2; scores ≥45), the ASCO Cancer Research Committee (OS gains ≥2.5 months PFS gains ≥3 months), the European Society for Medical Oncology-Magnitude of Clinical Benefit Scale version 1.1 (ESMO-MCBS v1.1; grade of A or B for trials of curative intent and 4 or 5 for those of non-curative intent), and the National Comprehensive Cancer Network (NCCN) Evidence Blocks (scores of 4 and 5). Trial characteristics and value framework scores were compared using Chi squared or Mann Whitney U tests. Results: We identified 106 pivotal trials supporting the approval of 52 individual drugs for 96 indications. Of these indications, 38 (40%) received breakthrough designation. Compared with trials for non-breakthrough drugs (n = 62), trials for breakthrough drugs (n = 44) had smaller sample size (median 373 vs 612, P= .03), were less often randomized (57% vs 86%; P= .001) and more likely to be open-label (84% vs 53%, P= .001). Trials for breakthrough drugs were more likely to demonstrate high clinical benefit using ASCO VF (68% vs 31%, P= .002) and NCCN Evidence Blocks (86% vs 56%, P= .002). A similar proportion of trials supporting breakthrough and non-breakthrough drugs demonstrated high clinical benefit using the ASCO Cancer Research Committee (82% vs 68%, P= .25) and ESMO-MCBS (35% vs 33%; P= .87) frameworks. Conclusions: In patients with advanced solid tumors, cancer drugs approved under breakthrough therapy designation were more likely to demonstrate high clinical benefit as defined by the ASCO VF and NCCN value frameworks. A similar proportion of approved breakthrough and non-breakthrough therapy drugs met the high benefit thresholds using the ASCO Cancer Research Committee and ESMO-MCBS frameworks.

e18135 Background: African Americans comprise 12% of the total United States population. Lifetime probability of developing cancer in African Americans is 37.55%. The 5-year relative survival rate is 68% in Caucasians and 61% in African Americans for all cancers combined. Concerns exist that African Americans are underrepresented in trials of cancer medications. We sought to examine African Americans participation for pivotal cancer medications trials submitted to the U.S. Food and Drug Administration (FDA) supporting marketing applications. Methods: On the basis of publicly available FDA reviews, the authors assessed enrollment of African Americans in trials supporting 75 drug approvals from 2014 to 2018. Prevalence-corrected estimates for the participation of African Americans were calculated as the percentage of African Americans among trial participants divided by the percentage of African Americans in the disease population (participation to prevalence ratio [PPR]), with a range between 0.8 and 1.2 reflecting similar representation of African Americans in the trial and disease population. Results: A total of 61763 patients enrolled in clinical trials, which resulted in subsequent FDA approval. The proportion of African Americans enrolled in the clinical trials that led to cancer medication approval in the 5-year period was 7.44%. The calculated PPR for participation of African Americans in clinical trials that led to drug approval for all cancers combined was 0.31. The results were consistent across major cancer subtypes including breast cancer (PPR of 0.29), prostate cancer (PPR of 0.18), and lung cancer (PPR of 0.15). Skin cancer, colorectal cancers and hematological malignancies results were also consistent with low PPR. Conclusions: African Americans were under represented in trials of drugs for all cancer types combined. Efforts should be made to increase the proportion of participation of African Americans in clinical trials.

6638 Background: Given rising cancer drug costs, Medicare recently proposed to tie some US drug prices to average prices paid by comparable countries. To understand the potential scope of this policy, we assessed differences in cancer drug prices in the US and selected European countries. We also evaluated the correlation between drug prices and their clinical benefit, as measured by two value frameworks: the American Society of Clinical Oncology Value Framework v2 (ASCO VF) and the European Society for Medical Oncology Magnitude of Clinical Benefit Scale v1.1 (ESMO-MCBS). Methods: We identified all new drugs for adult solid and hematologic cancers, approved by the FDA from 2009-2017 and that have also been approved by the EMA by December 31, 2018. US average sales prices (and if not available, wholesale acquisition costs) were extracted as of February 1, 2019, and compared to comparable currency-adjusted ex-factory drug costs in England, France, Germany, and Switzerland. ASCO VF and ESMO-MCBS scores were assessed for pivotal trials supporting solid tumor drugs; in case of multiple trials, we focused on the highest score. Consistent with the developers of the rating scales, “high benefit” was defined as scores of A-B (neo/adjuvant setting) and 4-5 (palliative setting) on the ESMO-MCBS scale and scores≥45 on the ASCO VF. Linear regression models and non-parametric Kruskal-Wallis test and were used to assess the association between drug prices and benefit scores. Results: The study cohort included 63 drugs approved by the FDA and the EMA during the study period. 46 (73%) were approved for solid tumors, and 17 (27%) were approved for hematologic malignancies. Overall, median cancer drug prices in included European countries were 52% (interquartile range: 37-72%) lower than US prices. There was no statistically significant association between monthly treatment cost and ASCO-VF or ESMO-MCBS scores in any country. There was also no association between price differential between US and median European drug prices and either ASCO-VF (P = 0.599) or ESMO-MBCS (P = 0.321) scores. Conclusions: Cancer drug prices in the US were significantly higher than in the compared European countries. Drug prices of cancer drugs were not associated with clinical benefit in the US or in European countries.

6631 Background: Rising pharmaceutical costs threaten affordability and access to cancer care. Methods: The 2013 and 2015 Medicare Part D claims databases were queried for CareFirst’s 2017 list of oral cancer drugs. Drugs with on-label indications for conditions other than cancer were excluded. Cost in 2013 were adjusted for inflation to 2015 dollars according to the Bureau of Labor Statistics consumer price index. Results: The total cost of oral cancer drugs paid by Medicare nearly doubled from 2013 to 2015, rising from over $4.8 billion to over $9.1 billion. The number of prescriptions increased by 21.6% from 640,193 to 778,357, suggesting that increased spending is driven more by rising drug prices than by increased utilization. Cost per patient increased for 81% (46) of the 57 drugs prescribed in both years and decreased for the remaining 19% (11). While only seven drugs cost more than $50,000 per patient in 2013, more than fifteen drugs cost more than $50,000 per patient in 2015. The overall average cost per beneficiary increased by 56% from $7,521.64 to $11,734.68, and seven drugs more than doubled in per patient cost. Cost per patient increased by 4.5 times for imbruvica (ibrutinib), 3.5 times for targretin (bexarotene), 3.2 times for soltamox (tamoxifen citrate), 3.1 times for nilandron (nilutamide), 2.6 times for gilotrif (afatinib dimaleate), 2.6 times for cyclophosphamide, and 2.5 times for iclusig (ponatinib hydrochloride). Conclusions: Medicare costs for oral cancer drug nearly doubled from 2013 to 2015, largely due to increased per patient costs. [Table: see text]

e18129 Background: Cancer drug stockouts occur at high frequencies globally, however their clinical effects are understudied in sub-Saharan Africa (SSA). We aim to describe prescription patterns and cost of systemic chemotherapy in cancer patients in Botswana during periods of stockout. Methods: Using a retrospective cohort study of the ten most common solid tumor malignancies treated with systemic chemotherapy at Princess Marina Hospital (PMH), Gaborone in 2016, we conducted a subset analysis of suboptimal events, defined as a cycle with ≥ 7 days delay or therapy switch from initiated guideline regimen, that occurred during drug stockout vs non-stockout periods. We estimated financial cost of therapy per cycle using Management Sciences for Health International Price Indicator Guide. Chi-squared and Wilcoxon rank sum were used for comparisons. Results: 167/378 patients contributed to 320 suboptimal events. 63% (201/320) of events occurred during a drug stockout, of which 43%, 43% and 14% were delays, switches, or both, respectively. There were significantly more delays (56% vs 44%, p < .0001) and switches (75% vs 26%, p < .0001) during stockout periods vs no stockout. The majority of switches during drug stockouts occurred in breast cancer patients receiving curative therapy: 48% (20/42) were “paclitaxel + trastuzumab” ($4673) to “paclitaxel alone” ($35) in HER2 positive patients resulting in a 99% cost decrease; and 29% (12/42) were paclitaxel ($35) to docetaxel ($108) resulting in a 209% cost increase per cycle switched. Colon cancer patients receiving palliative-intent therapy were the second most frequent patients with therapy switches during stockout periods: 42% (8/19) were “capecitabine + oxaliplatin” ($259) to “capecitabine alone” ($105) resulting in a 59% cost decrease. Conclusions: Breast cancer patients form the majority of patients treated with systemic chemotherapy at PMH and experienced the most delays and switches in therapy during drug stockout periods. Changes in drug prescription patterns during stockout periods may be associated with switches leading to inferior but less costly regimens, and in some cases costly regimens with higher toxicity. Interventions that minimize cancer drug stockouts are imperative and further studies to understand impact of stockout on survival are needed in SSA.

Purpose To evaluate to what extent results of completed trials of cancer drugs conducted in the United States are publicly available at ClinicalTrials.gov, as required by the Food and Drug Administration Amendments Act (FDAAA), or are published in journals. Methods We searched ClinicalTrials.gov for cancer trials governed by the FDAAA: phase II to IV trials assessing drugs in the United States with a primary completion date between December 26, 2007, and May 31, 2010. For each trial, we also searched PubMed to identify the publication of results. We assessed the cumulative percentages of posted or published results over time by using the Kaplan-Meier method. Results We identified 646 trials, including 209 randomized controlled trials (RCTs). At 12 months after completion of the trials, the cumulative percentages of trials with results posted at ClinicalTrials.gov, published in journals, and available either at ClinicalTrials.gov or in journals were 9% (95% CI, 7% to 11%), 12% (95% CI, 10% to 15%), and 20% (95% CI, 17% to 23%), respectively, and for RCTs, the percentages were 12% (95% CI, 8% to 16%), 5% (95% CI, 2% to 8%), and 17% (95% CI, 12% to 22%), respectively. At 36 months, these percentages were 31% (95% CI, 28% to 35%), 35% (95% CI, 31% to 39%), and 55% (95% CI, 51% to 59%), respectively, and for RCTs, they were 38% (95% CI, 31% to 45%), 32% (95% CI, 25% to 39%), and 56% (95% CI, 48% to 62%), respectively. Public availability of phase III trials was 15% (95% CI, 7% to 23%) at 12 months, 39% (95% CI, 27% to 49%) at 24 months, and 64% (95% CI, 50% to 73%) at 36 months. Conclusion Despite the FDAAA, results for nearly half the trials of cancer drugs in the United States were not publicly available 3 years after completion of the trials.

41 Background: Unlike most other countries, Canada has a dedicated HTA process to review cancer drugs. pCODR, a program of CADTH, conducts thorough and objective evaluations of clinical and economic evidence; as well as considering clinician and patient perspectives, and using this information to make recommendations to the participating jurisdictions to guide their drug funding decisions. Previously, Canadian provinces had separate regional drug review processes to inform their local funding decisions. The pCODR process reduces duplication of effort by individual funders and ensures that reviews are done in a timely and consistent manner. Methods: In a retrospective review, we identified all anticancer drugs reviewed by pCODR from July 2011 to March 2018. Results: As of March 31, 2018, pCODR has issued 103 notifications to implement a final recommendation. Of note, 96% of the submissions include patient group input and 88% of the submissions include clinician input. The median time to complete a review is 146 business days. The pCODR Expert Review Committee has issued 21 negative recommendations, while the remainder were either positive recommendations (n = 10), or conditional recommendations (n = 72). The “condition” that must be addressed most frequently in the conditional recommendations is the cost-effectiveness of the drug. Over 75% of the 82 positive and conditional recommendations have received uptake from one or more participating jurisdictions. The concordance rates are as follows: Conclusions: With the implementation of the pCODR process, there is greater harmonization in cancer drug funding decisions and supports equitable access across Canada.[Table: see text]

100 Background: In Canada, the Canadian Agency for Drugs and Technologies in Health’s pERC makes reimbursement recommendations for cancer drugs based on pCODR reviews of best available evidence, which in some circumstances, is from Phase II NRTs. To date, the majority of pERC recommendations based on NRT evidence have been for blood cancers. Objective: To examine aspects of NRT evidence that may influence pERC reimbursement recommendations for blood cancers. Methods: Final pERC Recommendations on blood cancer reviews supported by NRTs were included (July 2011 to June 2018). Factors that influenced Final Recommendations, such as clinical benefit, alignment with patient values and cost-effectiveness, were extracted. Results: As of June 2018, 10 conditional and 6 negative decisions were made in 13 Final Recommendations. Among conditional reimbursement recommendations, substantial need for treatment options and poor prognosis with available therapies were commonly noted. Assessment of the feasibility of randomized controlled trials (RCT) varied. The magnitude of benefit was impressive or a substantial benefit was seen in subgroups with greater need. Some recommendations noted benefit above historical outcomes or consistent evidence with other indications or trials. Most recommendations reported an improvement in quality of life (QoL) and a manageable toxicity profile. Limitations included short trial follow-up. Factors affecting cost-effectiveness and alignment with patient values varied. Among negative recommendations, there was less certainty about burden of illness and need. Uncertainty about magnitude of clinical benefit was attributed to lack of direct or indirect comparison to available options, lack of long term data or comparison to historical evidence, limited QoL data, and variability in toxicity. All cases were not cost-effective and partially aligned with patient values. Conclusions: pERC may accept evidence from NRTs to make reimbursement recommendations for blood cancers when there is a high burden of illness, unmet need, reasonable demonstration of efficacy and manageable toxicities. Feasibility of RCT was not a consistent factor.

42 Background: In Canada, through the pan-Canadian Oncology Drug Review (pCODR) process, CADTH conducts evaluations of clinical, economic, and patient evidence on cancer drugs to provide public reimbursement recommendations. The pCODR Expert Review Committee (pERC) makes these recommendations based on a drug’s overall clinical benefit, alignment with patient values, cost-effectiveness, and feasibility of adoption into the health system. Methods: In February 2016, pCODR launched a pilot process that allowed eligible clinicians (individually or as groups) not directly involved in the reviews to participate in the pCODR process, to provide: (1) input at the outset of a review; and, (2) feedback on the Initial Recommendation made by the pERC. Eligible clinicians are those who: (1) are actively practising physicians; (2) are members of a provincial cancer agency or similar body or a national cancer organization; and, (3) submit a declaration of conflict of interest. Results: As of March 31, 2018, 177 oncologists have registered to participate in the pCODR process. Of 43 submissions, 38 (88%) included clinician input. Fifteen submissions received individual input, and 33 received group input, the latter involving three to 13 clinicians or groups. Between April 2016 and March 2018, clinician input by tumour type was as follows: lung (n = 8), leukemia (n = 5), lymphoma (n = 5), gastrointestinal (n = 5), myeloma (n = 4), breast (n = 3), melanoma (n = 2), gynecological (n = 2), endocrine (n = 2), sarcoma (n = 1), and genitourinary (n = 1). Clinician input has answered several key questions, including current treatments, eligible patient populations, relevance to clinical practice, sequencing and priority of treatments, and companion diagnostic testing. Conclusions: Clinician engagement has provided value-added information on local issues from a practice perspective and insights into areas of unmet need. Continuous process improvement is important, however, and the pCODR program completed consultations in April 2018 to enhance clinician participation, proposing to: (1) customize the template that clinicians complete; and, (2) broaden the eligibility of clinicians to oncology pharmacists and oncology nurses.

6607Background: Physicians who accept payments from a pharmaceutical company are more likely to prescribe that company’s cancer drug[s]. Whether long-term physician-industry relationships are more ...

6516Background: United States Food and Drug Administration (FDA) primarily takes into account the therapeutic index before approval of a new cancer drug whereas in countries with public funding of ...

e23569Background: The vision of a precision medicine-guided approach to novel cancer drug development is challenged by high intra-tumor heterogeneity, interpatient diversity, and complex interactio...

6588Background: Data suggest that newly approved cancer drugs have worse safety and tolerability profiles than older drugs used as control groups in trials. However, less is known about the toxicit...

6515Background: Canada has an established publicly-funded healthcare system but a complex drug approval and funding process. After proof of efficacy (POE; key publication/presentation), before beco...

Duringthelastfewdecades,arapiddevelopmenthastakenplace in the exploration and development of cancer drugs, including cytostatics,endocrineagents(firstgenerationoftargeteddrugs),andmore recently, a new generation of targeted drugs. This development was facilitated by a marked increase in knowledge on how to best run well-designed clinical studies within academic collaborative groups, mainlyintheUnitedStatesandEuropeandfrequentlywiththepharmaceutical industry as sponsors for the studies. Theaimswereandstillaretoregisterandestablishnewdrugsand indications for broad groups of patients based on conventional morphologic classification systems of the primary cancers. The inclusion criteria are frequently aimed at defining a clinically homogeneous patient population, although the interpatient molecular signatures may be markedly different between patients included, not investigated, and not detected. More recent studies have had large sample sizestobeabletodetectsmalldifferences.Anotherwaytoexpressthis is that a positive effect in few patients will be recorded as a statistically

Purpose To analyze the age-related enrollment of cancer patients onto registration trials of new drugs or new indications approved by the US Food and Drug Administration from 1995 to 2002. Patients and Methods This study involved retrospective analyses of demographic data of cancer patients enrolled onto registration trials. The data on 28,766 cancer patients from 55 registration trials were analyzed according to age distributions of ≥ 65, ≥ 70, and ≥ 75 years. The rates of enrollment in each age group for each cancer were compared with the corresponding rates in the US cancer population. The age distributions of the US cancer population were derived from the Surveillance, Epidemiology, and End Results Program of the National Cancer Institute for the period 1995 to 1999 based on the 2000 US Census. Results The proportions of the overall patient populations aged ≥ 65, ≥ 70, and ≥ 75 years were 36%, 20%, and 9% compared with 60%, 46%, and 31%, respectively, in the US cancer population. Statistically significant under-representation of the elderly (P < .001) was noted in registration trials for all cancer treatment except for breast cancer hormonal therapies. Patients aged ≥ 70 years accounted for most of the under-representation. Conclusion Elderly were under-represented in the registration trials of new cancer therapies. Various strategies may be needed to evaluate cancer therapies for the elderly in prospective clinical trials and to improve cancer care in the elderly population.

Purpose: To describe the clinical findings, occurrence rates, causality evidence, and dissemination media for serious cancer drug–associated adverse drug reactions (ADRs) reported in the postmarketing setting. Methods: ADRs were termed serious if they resulted in death or severe organ failure. ADR information for oncology drugs from package insert (PI) revisions, so-called Dear Doctor letters, and journal articles was evaluated to identify serious ADRs reported from 2000 to 2002. Timing and content of information disseminated was assessed. Results: Twenty-five serious ADRs associated with 22 oncology drugs were identified after approval. Approximately half of these serious ADRs are associated with drugs approved before 1995. ADRs were described in articles in medical journals (17 ADRs), PI revisions (18 ADRs), and Dear Doctor letters (12 ADRs). PI revisions occurred less than 1 year after peer-reviewed publication for four ADRs. These revisions often differed for similar ADRs that occurred with drugs of the same class. Five of the seven ADRs lacking PI changes occurred with off-label use, for which PI change is not recommended by US Food and Drug Administration (FDA) policy. No cancer drug was withdrawn from the market during the observation period. Conclusion: Our findings demonstrate that serious ADRs may be discovered as long as 36 years after a drug receives FDA approval. This suggests a need for continued vigilance and efficient strategies for dissemination of information about ADRs associated with cancer drugs.

e14028 Background: Traditionally, only grade 3 and 4 adverse events (AEs) occurring during cycle 1 are used for dose escalation decisions in phase I oncology trials. With the shift in cancer drug development away from cytotoxic chemotherapy towards molecularly targeted agents (MTAs), assessment of lower grade AEs and those in later cycles is considered to be of increasing relevance. This study evaluated differences in AE incidence between trial types (combination cytotoxic vs. single MTA) and cycle number (1 vs. 2-6). In line with previous work suggesting ethnic and age related differences in toxicities, we evaluated AE incidence across a range of patient demographics. Methods: Retrospective analysis of toxicity data in patients enrolled onto phase I clinical trials at a single UK centre between 2006 and 2016. All AEs in the first six cycles of treatment deemed at least ‘possibly related’ were recorded. Results: A total of 912 AEs (664, grade 1; 194, grade 2; 54, grade 3 and 4) were identified in 127 patients across 15 different trials (420 treatment cycles). Mean AE totals for patients on cytotoxic combinations and MTAs respectively were 4.51 and 2.94 in cycle 1, 3.47 and 3.04 in cycles 2-6. Patients on cytotoxic combinations experienced a significantly higher mean number of AEs in six cycles compared to those on MTAs (7.98 vs. 5.98, p = 0.0135). Of the patients experiencing grade 3-4 AEs, 57% (cytotoxic combinations) and 75% (MTAs) occurred for the first time after cycle 1. No significant difference in incidence of AEs was found between any subgroups of patient demographic variables. Conclusions: Incidence of AEs is less in MTAs than for cytotoxic combinations across cycles 1-6. The significant incidence of first grade 3-4 AEs occurring from cycle 2 onwards supports incorporation of AE assessment beyond cycle 1 into recommendations for phase II doses in MTAs.

6607 Background: Intravenous drugs administered through body-surface area (BSA) or weight-based dosing may cause wastage due to large and/or limited fixed vial sizes, and vial sharing restrictions. Drug wastage leads to incremental costs without incremental value to patients. Bach et al. (2016) estimated 10% of revenue ($1.8 billion) from cancer drugs would result from wastage in 2016. The pan-Canadian Oncology Drug Review (pCODR) committee provides recommendations on which drugs to publicly reimburse by reviewing clinical and economic evidence. There is considerable potential that drug wastage could impact the economic evaluations. We sought to determine the impact of modeling cancer drug wastage on the results of economic evaluations. Methods: Economic evaluations submitted by drug manufacturers and reviewed by the pCODR Economic Guidance Panel (EGP) were assessed for frequency of wastage reporting and modeling. Cost-effectiveness analyses and budget impact analyses were conducted for scenarios in which “no wastage” and “wastage” of drugs occurred. Sensitivity analyses were performed to determine the effects of BSA and weight variation. Results: 12 drugs for use in 17 indications were analyzed. Wastage was reported in 71% and incorporated in 53% of manufacturer’s models, resulting in a mean incremental cost-effectiveness ratio (ICER) increase of 6.1% (range: 1.3% to 14.6%). EGP reported and incorporated wastage for 59% of models, resulting in a mean ICER increase of 15.0% (2.6% to 48.2%). When maximum wastage (i.e. the entire unused portion of each vial is discarded) was incorporated in our independent analysis, the mean ICER increased by 24.0% (0.0% to 97.2%) and the mean 3-year total incremental costs increased by 26.0% (0.0% to 83.1%). Over a 3-year period, wastage can increase the total incremental drug budget cost by CAD $102 million nationally. Changing the mean BSA or body weight caused 45% of the drugs to use a different vial size (if available) and/or quantity, resulting in further increased drug costs. Conclusions: Wastage can have an under-recognized and significant impact on economic evaluations of intravenous chemotherapy drugs. Guidelines are needed to promote uniform and optimal modeling of drug wastage in economic evaluations.

10009 Background: Older adults are a growing segment of our oncology population, with an expected increase in cancer incidence of 67% from 2010 to 2030 in people age 65 and older. However, older adults have been proportionally underrepresented in clinical trials. We sought to analyze the age-related enrollment of cancer patients onto trials supporting registration of new drugs or new indications approved by the US Food and Drug Administration from 2005 to 2015. Methods: This study involved retrospective analyses of demographic data of cancer patients enrolled onto trials supporting registration from 2005-2015. The data on 224,766 cancer patients supporting 105 drug applications were analyzed according to age distributions of

2543 Background: Trials using randomized designs have been conducted for decades to demonstrate efficacy of novel anti-cancer drugs (NACD). Recently, several NACD have shown high antitumor activity in early phase studies, prompting suggestions that NRCT could expedite drug development. We sought to determine what changes have occurred in numbers of NACD RCT vs NRCT conducted from 2004-2016. Methods: We reviewed a database of NACD clinical trials conducted by INC (excluding phase I and I/II trials) and classified them by RCT vs NRCT, grouped by year (≤ 2010 or > 2010). We queried Citeline Trialtrove database for industry sponsored, phase 2 trials (P2T) initiated from 2006-2016 and examined numbers of RCT vs NRCT by year.A more detailed analysis of non-small cell lung cancer (NSCLC) clinical trials based on drug type category - Immunooncology (IO) vs. all other mechanims of action (NIO) was performed. Results: 190 INC-conducted trials were reviewed. 58 trials (31%) were performed ≤ 2010 and 132 trials (69%) > 2010. Over this period, NRCT (n = 107, 56%) outnumbered RCT (n = 83, 44%). Whereas RCT outnumbered NRCT from 2004-2010 (74% vs 52%), after 2010, NRCT outnumbered RCT (58% vs 42%). Citeline Trialtrove search revealed 4776 industry sponsored P2T initiated from 2006-2016. The total number of P2T started annually was highest in 2007 (n = 621), decreasing to a low of 375 in 2016. The proportion of phase 2 RCT demonstrated an increase from 27% (n = 166) in 2006 to a peak plateau of 37-39% from 2011-2014, followed by a drop to 33% in 2015 and 29% in 2016. Among IO studies, RCT declined in 2015-6 vs. previous years, and a decreased for all NACD in 2016 vs. previous years also was noted. For studies in NSCLC, declines in RCT were evident from 2015-6 vs. previous years ( 45% in 2007-14 vs. 25% in 2015-6). Conclusions: Our data indicate a trend toward fewer trials of NACD using randomized designs and more studies using non-randomized designs, with overall fewer P2T initiated in the past year. This change reflects shifts in NACD development pathways, related to a better understanding of cancer biology, drive to develop personalized treatment and a more flexible regulatory drug approval process.

6526 Background: Concerns about the increase of cancer drug prices emphasized the need to base pricing decisions on clinical added value (CAV). In France, the CAV assessment is done by the National Authority for Health (HAS) on a 5-point scale based on clinical data. The magnitude of the CAV is partly used by the economic committee for healthcare products to determine drugs prices: a major to moderate CAV leads to the highest prices. Recognizing the importance of presenting clear statements, we analyzed which criteria are taking into account for CAV appraisal. Methods: A retrospective and descriptive study analysis of all new hematology/oncology cancer indications assessed by HAS between 2010 and 2015 has been conducted. For each appraisal, information regarding the level of evidence and clinical effect has been collected. Results: 78 new cancer indications (58 drugs) have been assessed: 20 had a major to moderate CAV (25%), 32 a minor CAV (41%) and 26 no CAV (33%). The percentage of major to moderate CAV was higher when a comparative pivotal trial was conducted (25%, 16/62 vs. 20%, 4/16). Among the 62 comparative trials, 58 had a superiority design (94%) supporting a major to moderate CAV in 27% (16/58) of the cases. Overall survival (OS) was measured in 91% of the pivotal trials (n = 71) and in 24% (n = 19) as a primary outcome. A higher proportion of major to moderate CAV were observed in trials using OS as a primary endpoint (31%, 6/19) than progression free survival (PFS) (23%, 9/39) or response rate (25%, 5/20).When a major to moderate CAV was obtained, a median absolute gain of 4 and 8 months for OS and PFS respectively were observed. All non-inferiority or non-conclusive designs led to no or minor CAV whatever the quantity of clinical effect. Conclusions: CAV is assessed in a two-step process where clinical effect is taken into account when level of evidence is appropriate. Consequently, there is not a unique profile of outcomes supporting a major to moderate CAV but a combination of several key elements such as study design, primary endpoint, and statistical significance. HAS appraisals remain multi-factorial, and medical need add a nuance to HAS expectations in terms of level of evidence.

e14036 Background: Proper assessment of harms of cancer treatment is essential to make risk-benefit trade off decisions in oncology clinic. CONSORT guidelines require the authors of RCTs to provide exact data for adverse events (AEs) grade 3 or higher (severe AEs), serious AEs (SAEs) and fatal AEs (FAEs) and avoid using general terms (such as well-tolerated) that downplay the harms. We aimed to assess the incidence and risks of severe AEs, SAEs and FAEs in such RCTs of cancer drugs that downplayed the harms. Methods: We extracted all the phase 2 and 3 RCTs of cancer drugs from the five major journals (NEJM, Lancet, Lancet Oncology, JAMA, JCO) from 2016 January to 2016 December. The abstracts/full texts of the papers were studied to assess if the harms of the experimental arm were downplayed. Use of any of the following terms while describing toxicities or AEs or safety was assumed to downplay the harms: acceptable, feasible, tolerable (or well-tolerated) , favorable, safe, manageable. Data on severe AEs, SAEs and FAEs for both the experimental and control cohorts were extracted from these RCTs and pooled using random-effects model to determine the overall incidence and risk. Results: Of 122 phase 2/3 RCTs of cancer drugs identified, 53 (43.4%) reports used terms downplaying the toxicities of the experimental arm and were included in this analysis. The overall incidence of severe AEs, SAEs and FAEs with these cancer drugs were 50.6% (95% CI: 41.5%-59.7%), 21.9% (95% CI: 16.3%-28.7%) and 1.6% (95% CI: 1.2%-2.2%) respectively. Compared with control, the risk of severe AEs and SAEs were significantly increased with the use of these cancer drugs: severe AEs (RR 1.15, 95% CI: 1.04-1.27, p = 0.005) and SAEs ( RR 1.49, 95% CI: 1.26-1.77, p < 0.001) . However, the risk of FAEs was similar across the experimental and control arms ( RR 0.89, 95% CI: 0.72-1.11, p = 0.306) Conclusions: Our analysis shows that considerable proportions of patients suffer severe AEs, SAEs and FAEs and significantly increased risk of severe AEs and SAEs in the trials that downplay the toxicities of cancer drugs. Terms that falsely downplay the toxicities of cancer drugs should be avoided in the reporting of RCTs to encourage accurate risk-benefit assessment.

e18130 Background: Understanding the clinical pharmacology properties of new drugs is essential, enabling correct prescribing with respect to dose, interactions and organ impairment. In the era of targeted therapy and resurgence in oncology R&D there is a need to consider how and when to define clinical pharmacology, since many drugs are given accelerated reviews, limiting the development time. This assessment was undertaken to understand how FDA review the clinical pharmacology components of NDAs. Methods: The Clinical Pharmacology and Biopharmaceutics Review documents of small molecule approvals in oncology by FDA from 2011 -2016 were analysed. Information regarding clinical pharmacology studies, in addition to post marketing requirements (PMR) was analysed. Results: Of 42 small molecules identified, 35 had clinical pharmacology review information. There were 16 accelerated approvals and in only 3 approvals were there no PMRs issued. A particular issue concerns the dose, 14 NDAs had identifiable issues in review, resulting in a 10 PMRs which include conducting a further pivotal study with additional doses. The majority of submissions (33) included clinical assessment of drug effect on the QTc interval. 7 PMRs were issued, 4 requesting completion of ongoing studies. A number of PMRs (24) were issued to assess organ impairment and DDI assessment (12). In all orally administered drugs a food interaction study was conducted with 1 PMR issued. In 32 submissions an excretion balance study was conducted, only 1 of the remaining 3 drugs was issued a PMR. Conclusions: Whilst a number of PMRs for clinical pharmacology studies were issued it is encouraging that these did not limit approval, ensuring patient access. For accelerated approvals, there will be insufficient time to complete many studies prior to filing. However, there were a significant number of PMRs regarding dose, reflecting FDA concerns that there is inadequate characterisation of the dose. Clinical pharmacology characterisation of NMEs is still required, but could be conducted through PMRs to enable early patient access. A major focus for sponsors should be a clearer understanding of the dose response relationship.

6597 Background: As cancer drug prices rise and insurance plans shift toward greater cost sharing, studies link patients’ high out-of-pocket (OOP) costs to non-adherence and early discontinuation of treatment. Meanwhile, few oncologists routinely discuss OOP costs with their patients. Using qualitative methods, we explored barriers and facilitators of cost transparency (i.e., disclosure of financial risks of cancer treatment). Methods: We performed semi-structured interviews with cancer patients (n = 22) and providers (n = 19) at an academic medical center and three affiliated community practices between August, 2015 and May, 2016. Two analysts coded the transcribed interview data using textual thematic methods, and hypotheses were generated employing grounded theory method. Results: We grouped themes that emerged into three major domains: 1) barriers, 2) facilitators, and 3) patient preferences. Patients and providers both expressed a strong aversion to making tradeoffs between financial and physical health outcomes. While patients feared being “profiled” based on their ability to pay, providers feared that cost transparency might threaten the doctor-patient relationship by exposing personal or institutional financial conflicts of interest. Pragmatic barriers included time constraints and difficulty in providing accurate cost estimates. Important facilitators were strong doctor-patient relationships and availability of support staff with financial expertise. We detected substantial heterogeneity in patient preferences. While some patients wanted to discuss costs with their provider, others feared “distracting” providers from their primary roles as health advocates. Conclusions: With implementation of OOP cost transparency, oncology practices will need to consider patient/provider aversion to financial/health tradeoffs, patients’ sensitivity to socioeconomic “profiling,” provider- and practice-level financial incentives, time constraints, accuracy of cost estimates, and variability in patient preferences. Meanwhile, strong provider-patient relationships and availability of support staff will facilitate OOP cost transparency.

6623 Background: The French National Authority for Health (HAS) is responsible for health technology assessment (HTA), providing opinion on drugs for reimbursement and pricing purposes. For all indications with a positive opinion for reimbursement, HAS assesses the clinical added value (CAV) on a 5-point scale for pricing negotiations based on clinical data. A major to moderate CAV leads to the highest prices, a minor CAV leads to a higher price than the comparator, and no CAV leads to lower price than the cheapest comparator. Countries increasingly face the policy challenge of harnessing the benefits of cancer drugs while managing healthcare budgets. In this context, we aimed to analyze cancer drugs assessment by HAS. Methods: a retrospective and descriptive analysis comparing all new hematology/oncology cancer indications versus all others new indications assessed by HAS between 2010 and 2015 has been conducted. Results: 87 cancer indications (60 drugs) have been evaluated, representing 17% (87/510) of all new drugs indications assessed by HAS. Almost all cancer indications (92%) obtained a favorable opinion for reimbursement. Seven (8%) had an unfavorable opinion versus 20% in other therapeutic areas. However, 5 of these 7 indications were related to a drug included on the list of reimbursed drugs for another indication and consequently drugs are available. Overall, only 2 drugs were not reimbursed: nintedanib in non small cell lung cancer and tegafur/gimeracil/oteracil in gastric cancer (no impact on survival and lack of results transposability). Of the 80 indications with a favorable opinion, 20 had a major to moderate CAV (25%), 32 a minor CAV (40%) and 28 have no CAV (35%). Major to moderate CAV are mostly composed of hematology drugs (12, 60%). The proportion of major to moderate CAV assessment is higher in oncology than in other therapeutic areas (66% vs. 30%). Conclusions: France benefits from a universal healthcare system offering wide coverage and large access to drugs. Almost all of the new hematology/oncology cancer drugs assessed had a favorable opinion and are fully reimbursed (100%) by health insurance. A high proportion (66%) of CAV is recognized in oncology. Nevertheless, the over DRG list can inadvertently limit access to these new cancer drugs.

6549 Background: Payers are moving from fee-for-service (FFS) reimbursement models towards value-based Alternative Payment Models (APMs). In Medicare, this trend has emerged via MACRA legislation and the Oncology Care Model (OCM). Reimbursement changes carry financial implications for payers, providers, and drug manufacturers. Methods: Financial impact of APMs was modeled as a function of Volume Factors and Price Factors in sales of a cancer drug product. Volume Factors included utilization patterns and market share. Price Factors included launch price, price increases, and both mandated and negotiated discounts. The model described three APM categories: Buy & Bill Plus (similar to OCM), Episode-Based Payment, and Third-Party Buy & Bill. The model included assumed timelines of APM implementation and the share of commercial insurer and Medicare markets implementing. Assumptions were generated in a multi-stakeholder workshop applying a modified Delphi method and Nominal Group Technique to arrive at consensus. The model included input ability for existing product forecast data including expected patient numbers, market share, payer mix, channel mix, price, 340B discount parameters, and additional concessions. Combinations of inputs and assumptions were run through the model as scenarios to generate results. Results: Across all scenarios, Episode-Based Payment resulted in revenues averaging 34% below baseline forecast at the ten-year mark. Third-Party Buy and Bill resulted in revenues averaging 100% of baseline at the ten-year mark. Buy and Bill Plus resulted in revenues averaging 66% below baseline at the ten-year mark. Market share as a function of clinical differentiation appeared to be a significant factor influencing revenue impact, with highly differentiated products outperforming others in all APMs. Conclusions: While much attention is paid to the impact of reimbursement reform on oncology provider economics, with subsequent impacts to patient access, less attention has been focused on impact to drug manufacturers. APM implementation could significantly impact manufacturer revenues, which in turn may impact both access to such therapies and research investment towards new anti-cancer therapies.

10029 Background: Elderly patients are commonly believed to derive less benefit from cancer drugs, even if they fulfil clinical trial eligibility. We aim to examine if novel oncology drugs provide differential treatment outcomes for elderly and young patients on clinical trials. Methods: A systematic review of randomized control trials (RCTs) cited for clinical efficacy evidence in novel oncology drug approvals by the Food and Drug Administration, European Medicines Agency, and Health Canada between 2006 and 2015 was conducted. Studies reporting age-based subgroup analyses for overall or progression free survival (OS/PFS), were considered. Independent reviewers extracted survival hazard ratios (HRs) and confidence intervals (CIs) for age-based subgroups. Meta-analyses based on an inverse variance random effects model were performed to examine patient subgroups < 65 and ≥ 65 years separately, and pooled HRs were compared to examine if differences in relative survival benefits existed between patient subgroups. Sensitivity analyses were conducted specific to cancer type, primary endpoint, and the type of systemic treatment. Results: Eighty-five RCTs, including 55,512 patients, reported age-based survival outcomes and were included. One study reported age-based toxicity and no studies age-based quality of life results. Pooled HRs [95% CIs] for patients < 65 and ≥ 65 years were 0.60 [0.56-0.65] and 0.66 [0.61-0.72], respectively with no difference between the two subgroups ( P= 0.08). All sensitivity analyses revealed similar results. Conclusions: Our results suggest that elderly and young patients derive similar relative survival benefits from novel oncology drugs. In settings where there is no other direct high-level evidence of elderly patients deriving less benefit than younger patients, it is reasonable to consider offering novel oncology drugs to elderly patients who fulfil trial eligibility. There is, however, a need to report age-based toxicity and quality of life results to support patient discussions regarding the balance of treatment benefit and harm, to encourage informed individualized decision-making.