The Role of Real-World Evidence, Patient-Reported Outcomes, and Economic Models in Oncology Formulary Decisions and the Role of Real-World Evidence in Oncology Product Registration

This publication was supported by Pfizer, Inc.

AJMC^® Clinical Brief summarizes key findings from studies regarding health plan use of real-world evidence, patient-reported outcomes, and economic models in oncology formulary decision-making and regulatory submissions to the FDA. These studies were originally presented at the Academy of Managed Care Pharmacy Nexus 2020 Virtual meeting, held from October 19 to 23, 2020. The results from these studies were later published as 4 separate articles.^1-4

REAL-WORLD EVIDENCE, PATIENT-REPORTED OUTCOMES, AND ECONOMIC MODELS IN ONCOLOGY FORMULARY DECISION-MAKING

Background

Randomized controlled trials (RCTs) are considered the gold standard for demonstrating safety and efficacy of a therapy. However, they may provide inadequate information for formulary decision-making, because the patient population in RCTs often does not reflect that in the real world.¹ This may be particularly relevant in oncology decision-making because of accelerated regulatory approvals and rare tumor types.¹ The implementation of the US FDA Safety and Innovation Act in 2012 allowed for accelerated approval of select drugs that treat serious or life-threatening disease if 1 or more clinically significant end point is considerably better than that of existing drugs.² The US FDA Safety and Innovation Act, enacted in 2012, facilitates the development and review of new drugs for serious or life-threatening diseases, allowing promising new drugs to receive accelerated approval based on surrogate markers of disease, while requiring that efficacy testing continues post-launch to confirm the initial results.³ Results from an analysis of oncology drugs approved from 2012 to 2017 showed that drugs with a breakthrough designation obtained their first FDA approval significantly sooner after their investigational new drug application (NDA) than drugs without breakthrough designation status (median, 5.2 years vs 7.1 years; P = .01 for difference).⁴ Although accelerated approvals make innovative therapies available to patients faster, payers are challenged with making decisions about coverage of breakthrough oncology drugs based on limited clinical data.³

Therefore, real-world evidence (RWE), defined by the Federal Food, Drug, and Cosmetic Act as “data regarding the usage, or the potential benefits or risks, of a drug derived from sources other than traditional clinical trials,” is often used to fill these knowledge gaps and guide oncology formulary decision-making.¹ Sources of real-world data (RWD) include electronic health records, medical claims and billing data, product and disease registries, and patient sources such as information stored on mobile devices.⁵

However, information on how RWE is used by payers to make oncology formulary decisions is unclear.³ Recent studies investigated how US payers use RWE to guide health care decision-making in oncology, which may help payers identify pertinent RWE and help manufacturers design relevant RWE studies.³

Payer Perceptions of the Use of Real-World Evidence in Oncology-Based Decision-Making

Methods

An interdisciplinary steering committee of payers, academic experts, and experts in health economics and outcomes research developed a survey to assess payer understanding about the use of RWE in decision-making for oncology products coverage.³ The content of the survey was based on literature reviews and current knowledge of RWE use in oncology decision-making. A pilot survey was sent to 5 payers, then reviewed and revised based on their feedback.³

The survey used in the study addressed 4 major areas on the use of RWE in decision-making for oncology coverage and was composed of 18 questions. Of these, 7 explored the types of decisions in oncology that can be guided by RWE and ways RWE is used by organizations in formulary decision-making, 5 covered internal and external sources of RWE, 2 addressed barriers to performing real-world analyses, and 4 questioned the experience and interest of payers in using RWE to guide outcomes-based contracting.³ The survey was distributed to a panel of 221 payers, which included Academy of Managed Care Pharmacy (AMCP)members who were pharmacy leaders involved in medication product evaluation or utilization management.³ The survey was also distributed to an additional 10 US payers who were not affiliated with this panel and agreed to participate in a live panel discussion in March 2020 about the results.³

Survey results were primarily presented as response frequencies, and dummy variables (yes/no) were used for responses on the multiple response questions.³ Pearson’s χ² test or Fisher’s exact tests (large [at least 1 million members] or small [less than 1 million members]), plan type, and scope of service (regional or national) were used to evaluate categorical responses according to plan size. Statistical significance was set at 2-tailed P values of .05 or less.³

Results

Of the 231 survey recipients (221 AMCP plus 10 payer panelists), 106 completed the survey (response rate, 45.9%).³ Of the 99 respondents who provided demographic information, the largest proportions represented managed care organizations (MCOs; 47.5%) or pharmacy benefit managers (PBMs; 37.4%), with a similar proportion of regional (49.5%) and national (47.5%) plans represented.³ As for payer type, 64.6% of respondents represented commercial and government organizations, 19.2% represented government only, and 16.2% represented commercial only.³ The majority (89.9%) of respondents were pharmacists, with the most commonly reported roles being pharmacy administrators (55.6%) and community or clinical pharmacists (30.3%).³

Current State of Real-World Evidence Use

Of the respondents, 84.9% use RWE for making formulary decisions in oncology.³ According to these payers’ responses, RWE was most useful in comparing the effectiveness of interventions when no head-to-head trials were available and in providing supportive evidence for guidelines (eg, National Comprehensive Cancer Network [NCCN] guidelines).³ In the 10-payer panel discussion of the survey findings, the panelists concluded that interest in the use of RWE for decision-making continues to grow, particularly for comparing effectiveness between drugs, determining formulary placement, and guiding financial decisions.³ Along with RCT data and feedback from experts, RWE is also used by the Pharmacy and Therapeutics Committee to determine initial placement and coverage. Following product launch, RWE associated with a product can help to refine related decisions and therapy sequence.³

Sources of Real-World Evidence

A majority of respondents reported using RWE from claims data (79.2%), medical records data (68.9%), and data from prospective cohort studies (60.4%).³ PBMs and MCOs reported using similar types of data, although MCOs were more likely than PBMs to use prospective cohort studies (66.0% vs 56.8%) and patient registries (42.6% vs 27.0%).³ Nearly half of respondents stated that their plans perform their own real-world studies to guide decisions for oncology coverage (48.1% for large plans; 40.0% for small plans). Results are often used to garner data for budget-impact models or cost-effectiveness models to confirm RCT results within a payer’s own patient populations, and to support off-label use.³

The panelists identified high availability as the reason most payers use pharmacy and medical claims for RWE, although panelists expect the use of electronic health records to continue to increase.³ They noted that PBMs, with limited access to medical data, often need to rely on pharmacy claims data, which typically provide insight into only dose strength and therapy duration.³

Barriers to Use of Real-World Evidence

Limited resources and personnel was reported as the driving for plans that had difficulty conducting their own real-world studies; this was a greater concern for small plans than for large plans (91.1% vs 70.4%, respectively; P = .012).³ Other barriers commonly reported were insufficient analytic capabilities (55.6% of large plans and 57.8% of small plans) and a lack of appropriate in-house data (46.3% of large plans and 48.9% of small plans).³ Panelists noted that in-house production of RWE may be too costly for some payers. Some payers were concerned about whether data produced from in-house studies would be meaningful because of the small number of eligible patients for some of the innovative oncology drugs.³

Outcomes-Based Contracting

Interest in using RWE to guide outcomes-based contracting was reported by the majority of respondents—76.5% were moderately, very, or extremely interested; 17.6% were somewhat interested.³ However, only 17.2% of organizations stated that they had experience in outcomes-based contracting in oncology, with the most commonly reported barriers including challenges around the definition of outcomes to be measured and monitored over time (76.5%), complexity of outcomes-based agreements (74.5%), costs of implementation (48.0%), small size of the target population (48.0%), and lack of personnel (34.3%).³ Panelists cited uncertainty about how to identify and use data for contracts; they noted that outcomes and end points can be affected by factors including adherence, proxy measures, comorbidities, hospitalizations, and relapses.³ Furthermore, data access may be limited in some health plans, making measurement and tracking of the necessary data difficult.³

Conclusions

US payers perceive that RWE plays an important role in oncology decision-making, especially for innovative drugs indicated for niche populations where evidence from RCTs is often limited by accelerated approvals.³ However, the associated expense and limited resources make it difficult to conduct internal RWE studies, and many payers gain insights from RWE published in NCCN and other guidelines.³ Additionally, the timing of RWE availability is an issue for payers when making oncology formulary and contracting decisions because RWE that compares therapy effectiveness is often not available prior to product launch, which is when it would help payers assess the role of new oncology therapies for the patients who are most likely to respond.³ Additional types of RWE generated through clinical trials, such as cost and prevalence of a disease, approaches to treatment, and adverse events, may help guide optimal placement of a new product; however, the panelists stated that they rarely see this information. Finally, the increased use of RWE highlights the need to educate payers on the appropriate use of RWE for making oncology formulary decisions.³

Payer Perceptions on the Use of Economic Models in Oncology Decision-Making

Methods

An interdisciplinary steering committee of payers, academics, and industry health economics and outcomes research experts created a survey based on literature reviews and current knowledge of the use of economic models in oncology decision-making.⁶ The survey was composed of 13 questions about payer perceptions and use of economic models, including questions about payer interest in using different types of economic models (n = 1), experience of respondent organizations with using economic models in oncology decision-making (n = 8), barriers to using economic models (n = 2), and interest in partnering with manufacturers to develop or validate an economic model in oncology using plan data from the organization (n = 2).⁶

The survey was distributed to a panel of 221 payers, which included AMCP members who were pharmacy leaders involved in medication product evaluation or utilization management.⁶ The survey was also distributed to an additional 10 US payers who were not affiliated with this panel and agreed to participate in a live panel discussion about the results.⁶

Survey results were presented as frequencies of responses, coded as “yes” or “no” for multiple-response questions.⁶ Categorical responses were evaluated using Pearson’s χ² test or Fisher’s exact test according to plan size, plan type, and scope of service (regional or national). Statistical significance was set at 2-tailed P values of .05 or less.⁶

Results

Demographics

Of the 231 survey recipients (221 AMCP plus 10 payer panelists), 106 completed the survey (response rate, 45.9%).6 MCOs (47.5%) and PBMs (37.4%) were most commonly represented, with relatively similar proportions of regional (49.5%) and national (47.5%) plans.⁶ Commercial and government payers represented the majority (64.6%) of respondents.⁶ The majority (89.9%) of respondents were pharmacists, with pharmacy administrators (55.6%) and community or clinical pharmacists (30.3%) as the most commonly reported roles.⁶

Current Payer Perceptions and the Use of Economic Models

Compared with small plans (< 1 million lives covered), a higher proportion of large plans (≥ 1 million lives covered) reported reviewing economic models related to oncology products within the last 2 years (57.0% vs 35.6%; P = .043) and having individuals with expertise in evaluating economic models in oncology (55.6% vs 31.1%; P = .016).⁶ Reasons given for not reviewing economic models included lack of available models at the time of product review (44.1%), potential bias (38.2%), lack of personnel or training (33.3%), lack of transparency (31.4%), and excessively complicated models (28.4%).⁶

Payers ranked cost-effectiveness models (CEMs; 85.3% moderate/most interest) and budget-impact models (BIMs; 80.4% moderate/most interest) higher than cost calculators (34.3% moderate/most interest), and the majority of respondents reported sometimes, often, or always using CEMs (66.7%) and BIMs (62.7%) to inform oncology formulary decisions.⁶ Approximately two-thirds of payers (68.6%) stated that economic models can help inform decisions at least a moderate amount when considering therapies with a similar efficacy and safety profile.⁶

After reviewing a manufacturer-sponsored economic model, 43.1% of payers reported using internal data to perform additional analyses for oncology decision-making, 9.8% reported using the manufacturer-sponsored model without their own internal data to validate the model, and 31.4% reported not using a manufacturer-sponsored model to make coverage decisions.⁶

In the panel discussion, panelists confirmed their interest in economic models and stated that they valued data on standard of care, the anticipated role of the drug in therapy, estimated target population, costs of adverse events and hospitalizations in clinical trials, and total cost of care.⁶ Although they noted that data are often limited at the time of initial formulary decision-making (particularly for drugs with accelerated approval) and can change when the product enters the market, the panelists said that early economic data available from clinical trials are valuable, particularly if the data factor in total cost of care.⁶ Additionally, the panelists said that progression-free survival and overall survival are highly relevant end points to include in economic models, may be helpful for outcomes-based contracting, and should be included in follow-up RWE studies.⁶ The panelists added that real-world economic data are useful during the postlaunch stage to verify initial decisions and inform potential changes in coverage decisions or treatment pathways.⁶ However, the panelists noted concerns around prevalence in trial populations vs the plan’s population, lack of accurate and inclusive estimates for all components of total care costs, and treatment patterns in a trial that do not reflect those of physicians in the real world.⁶

Economic models can provide valuable supplemental information to inform formulary decision-making in oncology, but the models are often lacking at the time of product review.

Discussion

Results from the survey and panel discussion showed that payers consider economic models to provide valuable information to add to traditional trial data when making oncology coverage decisions.⁶ The Institute for Clinical and Economic Review (ICER) provides information on the economic impact of formulary decisions that is generally available near the approval date for a product, which is often too late for consideration by payers.⁶ Therefore, access to economic data that contribute to the estimated total cost of care as early as possible enables payers to evaluate predicted costs with assumptions about patient population, rates of adverse events and hospitalizations, and medical costs.⁶ However, information needed for initial formulary review is often not available in the public domain or requires extensive searching by payers, and liaisons with manufacturers can improve understanding of early information on the product and preliminary economic data through guidance on preapproval information exchange.⁶

However, only about 50% of payers surveyed reported having experience reviewing economic models or performing their own economic studies in oncology, highlighting an opportunity to improve education on optimizing creation, validation, and use of the models for decision-making, particularly for small health plans.⁶ Possible strategies to address payer concerns about assumptions and transparency in economic models include involving payers in the development to ensure that the parameters chosen by manufacturers are relevant, providing references for the selected assumptions, includingsensitivity analyses, and explaining the origins of, rationale for, and changes in, assumptions over time.⁶

Limitations

Key limitations of the study included the underrepresentation of integrated delivery systems in the survey and decision-makers other than pharmacy directors of health plans.⁶ Additionally, the limited length of the survey did not allow for collection of additional information to explain answers, and respondents were not required to answer all questions, which may have introduced a bias.⁶ Furthermore, the study may have included multiple respondents from the same organization, and specific information on the respondents was unavailable because the results were deidentified.⁶

Conclusions

Economic models can provide valuable supplemental information to inform formulary decision-making in oncology, but the models are often lacking at the time of product review.⁶ Requests for preapproval RWE from manufacturers may help to bridge this gap, and payer education about the applications and limitations of economic models in oncology decision-making is beneficial to optimize the use of these models.⁶

Payer Perceptions on the Use of Patient-Reported Outcomes in Oncology Decision-Making

Methods

An interdisciplinary steering committee that included payers, experts in academic and industry health economics and outcomes research, and a measurement specialist created a survey to assess understanding on the value and use of patient-reported outcomes (PRO) evidence in oncology decision-making.⁷ The survey included 10 questions that addressed payer perceptions and use of PRO evidence: questions about the usefulness of PRO evidence from clinical trials and real-world studies (n = 2), the effect on formulary decisions of PRO evidence from clinical oncology trials or RWE (n = 4), the PRO version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE) (n = 2), the use of PRO evidence from clinical trials to provide context for safety data (n = 1), and the perceived role of PRO evidence for measuring value in innovative or value-based agreements (n = 1).⁷

The survey was distributed online through the AMCP Market Insights program to a panel of 221 payers, which included AMCP members who were pharmacy leaders involved in evaluation or utilization management of medication products, as well as 10 additional US payers who agreed to participate in a live panel discussion about the survey results.⁷

Results were presented primarily as response frequencies, with multiple-response questions coded as “yes” or “no.”7 Categorical responses were evaluated using Pearson’s χ² test or Fisher exact test according to plan size, plan type, and scope of service (regional vs national), and statistical significance was set at 2-tailed P values of .05 or less.⁷

Results

Of the 231 survey recipients (221 AMCP plus 10 payer panelists), 106 completed the survey (response rate, 45.9%).⁷ MCOs (47.5%) and PBMs (37.4%) were most commonly represented, and a similar proportion of regional (49.5%) and national (47.5%) plans were represented.⁷ The majority (64.6%) of respondents’ organizations were private payer plans covering commercial and government lives, and all health plan sizes were well represented.⁷ Most respondents (89.9%) were pharmacists, with pharmacy administrators (55.6%) and pharmacists/clinical pharmacists (30.3%) being the most commonly reported roles.⁷

Current Payer Perception and Use of PRO Evidence

The majority of respondents stated that PRO evidence is useful for formulary decision-making, with similar responses between small (< 1 million covered lives) and large health plans (≥ 1 million covered lives; 60.0% vs 57.4%; P = .794). Respondents also stated that PRO evidence from real-world studies is useful (small plans, 57.8% vs large plans, 51.9%; P = .555).⁷ Nearly half of payers stated that a lack of PRO evidence from oncology clinical trials would negatively affect formulary review at least somewhat (48.2% of large plans vs 42.2% of small plans; P = .538).⁷ Respondents considered PRO evidence to be useful for providing additional safety-related context (78.2%), oncology formulary decision-making when selecting between therapies (34.3%), and measured value for innovative or value-based agreements (51.5%).⁷

Panelists in the 10-member panel discussion agreed with the survey findings and added that PRO may help with development of treatment pathways to address health-related quality of life (HRQOL)in oncology, shared decision support, care management designs, and dialogues between provider and patient.⁷ They added that PRO evidence may address patient functionality during progression-free survival and development of post-progression clinical pathways that identify adverse effects affecting concepts measured in PROs.⁷

The panelists said that PRO evidence from real-world studies was more generalizable and realistic than data from clinical trials, but noted that RWD are often unavailable at the time of initial formulary decision-making, particularly with accelerated drug approvals.⁷ They added that PRO evidence from real-world, postlaunch studies is valuable but may be difficult for payers to find and incorporate into future decision reviews because they may have insufficient resources to thoroughly review literature post launch.⁷ PRO evidence in the package insert is also more likely to be easy to access and accepted by payers than PRO evidence from other sources.⁷

Total cost of care was also identified as a concern, and the panelists stated that improved knowledge of PRO evidence may affect treatment decisions and models of care delivery used by network providers, which ultimately influence the total cost of care.⁷ Although interest in using PRO evidence for outcomes-based or value-based contracts was expressed by many panelists, they preferred more objective measures and could not identify specific examples in which PRO data were used for outcomes contracts.⁷

Discussion

Although payers may have a bias against PRO evidence based on a perception of its subjectivity, the effect of PRO on total cost of care and utilization of health care resources is increasingly recognized by the FDA and value frameworks from multiple organizations, including ICER and the American Society of Clinical Oncology.⁷ PRO evidence related to HRQOL may be particularly relevant in symptomatic late-stage or metastatic oncology; therefore, real-world studies that include HRQOL-related PROs may be of interest to payers and providers.⁷ However, better methods are needed for collecting and reporting PRO evidence that is meaningful and actionable for payer decision-making, as providers are less likely to record patient-reported measures (eg, pain or fatigue) in the patient’s medical chart than objective measures that can be tracked over time (eg, blood cell count).⁷ Technologic advances for gathering and measuring PRO data (eg, from member digital apps and online portals) have been developed, and further developments should focus on increasing the usefulness and actionability of gathering and reporting PRO data.⁷ The PRO-CTCAE measurement system, designed to evaluate patient-reported toxicity symptoms for participants in oncology clinical trials, is 1 example of a tool for tracking PROs, but 81.1% of payers in the current study stated that their level of awareness of PRO-CTCAE was low or very low, and 67.7% stated that they did not have sufficient information to answer questions about its usefulness.⁷ Therefore, education of payers about the value of PRO-CTCAE and other PRO measurement systems is needed.⁷

Although the panelists were unfamiliar with contracts that were based on PRO measures, PRO evidence also has the potential to be used more fully in value-based contracting with manufacturers.⁷ CMS and some commercial payers have been investigating ways to include PRO evidence in payment and accountability programs (although these are still not widely used in oncology), and widespread utility of PRO evidence in outcomes- or value-based contracts requires continued development and enhancement of standardized and relevant PRO measures.⁷

Limitations

Although MCOs, PBMs, and large payers were well represented, integrated delivery systems accounted for only 10.1% of respondents, and decision-makers other than health plan pharmacy directors (eg, medical directors) were not surveyed and may have different perspectives.⁷ Additionally, perspectives from these US-based payers may not be generalizable to those in other countries, and the opinions of participants in the panel discussions may not represent those of all respondents.⁷

Conclusions

PRO evidence is a complementary addition to traditional efficacy and safety data from clinical trials used by payers when making formulary decisions, particularly in oncology where HRQOL carries relatively high importance and traditional evidence is limited because of the large number of accelerated approvals.⁷ PRO evidence has expanded from that limitation of patient-physician interactions to a valuable information source that payers can use to assess patient experience and HRQOL in their plan members undergoing oncology treatment.⁷

Real-World Evidence in Support of Oncology Product Registration: a Systematic Review of New Drug Application and Biologics License Application Approvals From 2015-2020

Methods

A systematic review of NDA and biologics license application (BLA) approvals was performed to identify oncology products from the Center for Drug Evaluation and Research (CDER) calendar year approvals, CDER Drug approvals by month, and the Center for Biologics Evaluation and Research biological approvals by year; the corresponding FDA review documents were assessed for RWE.⁸ Supplemental NDA and BLA approvals were also identified in a similar manner, with CDER efficacy supplement approvals and hematology/oncology approvals and safety notifications used as additional sources.⁸ Automated text extraction was used to extract text into a consolidated database and to search for keywords related to RWE. After drug approvals with RWE were identified, pivotal information related to the trial, application, and RWE were extracted from the FDA review documents.⁸

Results

Eleven of the 133 approvals for oncology NDAs and BLAs included RWE supporting efficacy.⁸ The median time from investigational new drug submission to approval was 5.5 years (range, 2.8-8.4).⁸ The NDAs or BLAs with RWE were for avelumab, axicabtagene ciloleucel, entrectinib, erdafitinib, polatuzumab vedotin-piiq, selinexor, avapritinib, capmatinib, tafasitamab, and tazemetostat.⁸ Of 573 oncology supplemental NDAs and BLAs identified, 249 were for new oncology indications, and 2 (0.8%) of these included RWE in support of efficacy (for blinatumomab and palbociclib).⁸

All 13 products with RWE for efficacy were FDA approved after enactment of the 21st Century Cures Act, with 11 applications that received orphan drug designation, 12 that received priority review, 9 that have been designated as breakthrough therapies, 6 that received fast-track designation, and 3 that received all of these designations.⁸ The 4 products that received full approvals had designations of orphan disease, breakthrough therapy, priority review, no advisory committee meeting called by the FDA, and high level of primary efficacy.⁸

RWE was used for contextualization in 8 applications, in comparison with clinical trial results in 3 applications, and both purposes in 2 applications.⁸ Chart review from clinical sites was the most common source of RWD, and overall response rate (ORR) was the most common end point.⁸ RWE for 3 therapies used a real-world population that matched the eligibility criteria to the trial, and of the 11 new approvals, RWE sample sizes were more than 50% larger than in the pivotal trial in 4 (36%), similar in 3 (27%), and more than 50% smaller in 4 (36%) cases.⁸

FDA-identified sources of bias related to the use of historical controls and RWD included selection bias, residual or unobserved confounding, different methods of outcome assessment and frequency of measures compared with the pivotal trials, lack of comparability between trial participants and external controls, incorrect classification of outcomes, and inadequate statistical adjustment of differences between comparator groups.⁸

Discussion

RWE may help accelerate the drug development process, but it has inherent limitations, which prompted the FDA to make several recommendations for improving its quality. These recommendations include early engagement to ensure appropriateness of the data source and decide the purpose of this RWE study, development of a prespecified study protocol to improve transparency, and selection of the RWE population that is comparable with that of the pivotal trial.8 Additionally, the FDA encouraged the use of concordance analysis to address the discrepancy between using real-world tumor assessment and Response Evaluation Criteria in Solid Tumors [RECIST] criteria to evaluate ORR, descriptions of ways to minimize confounding (eg, appropriate index date and reduction in missing values), and quantitative bias analysis to describe effects of unmeasured confounding.⁸

The FDA tended to have fewer critiques on RWE quality when the pivotal trials had a high ORR and the RWE had a low ORR; however, the magnitude of primary efficacy in the pivotal trial was associated with receiving full approval (vs accelerated approval), with ORRs of more than 70% and median duration of response (DOR) of at least 9 months observed in these trials.⁸ Accelerated approvals had a wide range of ORRs (13% to ≤ 70%) and median DOR (4 months to not reached).⁸ However, an ORR of greater than 70% should not be interpreted as the benchmark for full approval.⁸

The FDA also noted that selection of appropriate RWD requires consideration of selection bias and confounding, which are inherent with chart reviews. Certain chart reviews lack generalizability to a wider population and have incomplete data from cancer-specific electronic health record databases.⁸ However, RWE may highlight an unmet therapeutic need addressed by the investigational therapy by showing the disease burden and poor prognosis with real-world standard of care in the target population, and the strength of trial data is a key factor for determining need for RWE in regulatory decision-making.⁸ Additionally, some limitations of RWE related to study design and analysis may have been overcome by communicating with the FDA early in the application process, and the FDA encourages applicants to maintain transparency about study design and analysis before initiating the study and register observational studies on ClinicalTrials.gov.⁸ Common weaknesses in RWE study design include differences in follow-up duration, inclusion/exclusion criteria, and end point definitions between the pivotal trial and the RWE study, as well as the inability to compare outcome measures with those of the pivotal trial and definitions of the index date that led to immortal time bias.⁸

In situations when RWE and the pivotal trial were directly compared, residual confounding and other assumptions related to propensity score weighted analysis were key statistical analysis concerns identified by the FDA.⁸ To address these concerns, the FDA recommended incorporation of statistical methods to evaluate the relative effect of these biases, such as an array approach and quantitative bias analyses for unmeasured cofounding, and multiple imputation and tipping point analysis for missing values.⁸ Despite the inadequacies of RWE that were noted, the FDA encouraged inclusion of RWE in applications and stated that observational studies with external comparators may provide supportive data for NDA and BLA submissions.⁸

Although this study was limited to FDA-approved oncology applications that had publicly available information, included a limited number of examples, and was unable to fully characterize accelerated approvals, it highlighted that RWE complemented efficacy data from single-arm trials in successful approvals of oncology products.⁸

CLOSING REMARKS

RWE is valuable for health plans in oncology formulary decision-making, particularly for innovative oncology products or for those that serve small, niche populations that have high unmet needs because traditional evidence is often limited with these accelerated approvals.³ Education of payers on optimal use of RWE, including PRO and economic models, is beneficial for informing oncology decisions.³ Although the passage of the 21st Century Cures Act in 2016 increased interest in using RWE to support approval of new indications for current therapies or to provide data for postapproval requirements, current use of RWE remains limited.⁸ Increasing the use of RWE in regulatory submissions helps increase the amount of information available about a therapy before launch of a product and provides additional data for oncology formulary decision-making by health plans.⁶

REFERENCES

1. Oderda G, Brixner D, Biskupiak J, et al. Payer perceptions on the use of real-world evidence in oncology decision-making: results from an online survey and payer panel. Presented at: Academy of Managed Care Pharmacy Nexus 2020; October 19-23, 2020, virtual. Accessed September 16, 2021.
2. Food and Drug Administration Safety and Innovation Act. Pub L. No. 112-144, 126 Stat. 993 (2012) https://www.congress.gov/112/plaws/publ144/PLAW-112publ144.pdf
3. Brixner D, Biskupiak J, Oderda G, et al. Payer perceptions of the use of real-world evidence in oncology-based decision making. J Manag Care Spec Pharm. 2021;27(8):1096-1105. doi:10.18553/jmcp.2021.27.8.1096
4. Hwang TJ, Franklin JM, Chen CT, et al. Efficacy, safety, and regulatory approval of Food and Drug Administration-designated breakthrough and nonbreakthrough cancer medicines. J Clin Oncol. 2018;36(18):1805-1812. doi:10.1200/JCO.2017.77.1592
5. Framework for FDA’s real-world evidence program. United States Food and Drug Administration. December 2018. Accessed September 16, 2021. https://www.fda.gov/media/120060/download
6. Biskupiak J, Oderda G, Brixner D, Burgoyne D, Arondekar B, Niyazov A. Payer perceptions on the use of economic models in oncology decision making. J Manag Care Spec Pharm. 2021;27(11):1560-1567. doi:10.18553/jmcp.2021.27.11.1560
7. Oderda G, Brixner D, Biskupiak J, et al. Payer perceptions on the use of patient-reported outcomes in oncology decision making. J Manag Care Spec Pharm. 2021;28(2):188-195.. doi:10.18553/jmcp.2021.21223
8. Arondekar B, Duh MS, Bhak RH, et al. Real-world evidence in support of oncology product registration: a systematic review of new drug application and biologics license application approvals from 2015-2020. Clin Cancer Res. Published online October 19, 2021. doi:10.1158/1078-0432.CCR-21-2639

Download PDF