AMCP Partnership Forum: Managing Care in the Wave of Precision Medicine

Precision medicine, the customization of health care to an individual’s genetic profile while accounting for biomarkers and lifestyle, has increasingly been adopted by health care stakeholders to guide the development of treatment options, improve treatment decision making, provide more patient-centered care, and better inform coverage and reimbursement decisions. Despite these benefits, key challenges prevent its broader use and adoption. On December 7-8, 2017, the Academy of Managed Care Pharmacy convened a group of stakeholders to discuss these challenges and provide recommendations to facilitate broader adoption and use of precision medicine across health care settings. These stakeholders represented the pharmaceutical industry, clinicians, patient advocacy, private payers, device manufacturers, health analytics, information technology, academia, and government agencies. Throughout the 2-day forum, participants discussed evidence requirements for precision medicine, including consistent ways to measure the utility and validity of precision medicine tests and therapies, limitations of traditional clinical trial designs, and limitations of value assessment framework methods. They also highlighted the challenges with evidence collection and data silos in precision medicine. Interoperability within and across health systems is hindering clinical advancements. Current medical coding systems also cannot account for the heterogeneity of many diseases, preventing health systems from having a complete understanding of their patient population to inform resource allocation. Challenges faced by payers, such as evidence limitations, to inform coverage and reimbursement decisions in precision medicine, as well as legal and regulatory barriers that inhibit more widespread data sharing, were also identified. While a broad range of perspectives was shared throughout the forum, participants reached consensus across 2 overarching areas. First, there is a greater need for common definitions, thresholds, and standards to guide evidence generation in precision medicine. Second, current information silos are preventing the sharing of valuable data. Collaboration among stakeholders is needed to support better information sharing, awareness, and education of precision medicine for patients. The recommendations brought forward by this diverse group of experts provide a set of solutions to spur widespread use and application of precision medicine. Taken together, successful adoption and use of precision medicine will require input and collaboration from all sectors of health care, especially patients.

A dvances in research and technology have facilitated a growing adoption of precision medicine, which tailors health care to an individual's genetic profile while accounting for biomarkers, preferences, lifestyle, and environment. 1 Although personalized medicine is not a new concept, developments in molecular biology and analyses of large datasets are increasing its adoption across a diverse set of stakeholders to enhance patient-centered care. 2 Specifically, stakeholders are using precision medicine to guide development of novel diagnostics and therapeutics, facilitate regulatory approvals of indications for existing and new interventions, and inform therapeutic selection through gene and biomarker testing. In addition, precision medicine can improve the utility of data sources to inform patient-centered health care and appropriate drug coverage and reimbursement. Despite the growing use of precision medicine, barriers to its optimal use remain and highlight the need for greater stakeholder collaboration and engagement.

■■ Forum Aims
The Academy of Managed Care Pharmacy convened a forum of over 30 key health care stakeholders representing the pharmaceutical industry, clinicians, patient advocacy, private payers, device manufacturers, health analytics, information technology, academia, and government agencies on December 7-8, 2017, to discuss and provide recommendations on the following: 1. Identifying evidence needs for precision medicine 2. Overcoming challenges with data collection and interoperability 3. Innovating benefit design and reimbursement strategies for precision medicine 4. Overcoming current legal and regulatory barriers to precision medicine adoption

■■ Evidence Needs for Precision Medicine Evidence Requirements
When evaluating the benefits and risks of a new precision medicine diagnostic, stakeholders uniformly rely on measures of analytic validity, clinical validity, and clinical utility. These diagnostics must demonstrate that they measure and accurately detect what they intend through analytical and clinical validity. A test also should have clinical utility or benefits for patients through information about a diagnosis, treatment, or management. 3 While these are conceptually standard metrics, implementing and analyzing them in a consistent way can be difficult. Clinical trials and registries often measure outcomes differently, making it challenging to assess validity and utility uniformly.
Notwithstanding the advances in genomics and development of new tests, evidence is often lacking on clinical validity and clinical utility. This missing and evolving evidence 2. Evidence pertaining to precision medicine should be integrated into clinical practice guidelines to inform clinical decision making and coverage policies. 3. There is a need for consistent definitions of the health episode, lifetime budget impact, and target cost or reimbursement in advance of clinical or economic valuations of precision medicine. 4. A multistakeholder effort is needed to identify and standardize best practices for gathering RWE on precision medicine treatments. 5. Interoperability of electronic health records (EHRs) and linkages between these and other data sources will facilitate multisourced data collection, such as RWE, for precision medicine.
Novel Trial Designs 1. While novel trial designs are viewed as a viable alternative to traditional randomized controlled trials (RCTs) in precision medicine, a taxonomy should be developed that appropriately matches trial design with precision medicine evidence needs, disease, and treatment type. 2. To develop rigorous evidence for precision medicine, there is a need for enhanced approaches to enrolling patients into clinical trials, such as educational and promotional efforts with patient and advocacy organizations.
Value Assessment Frameworks 1. Value assessment frameworks should incorporate patientcentered measures that explicitly address methods to account for emerging evidence over time that reflect the potential increased value and patient benefits of precision medicine that are not evident at the time of regulatory approval or formulary decision making. 2. Frameworks should recognize the value in increased treatment efficiency (e.g., ability to avoid unnecessary tests) and place greater emphasis on RWE to determine the benefits and risks of precision medicine interventions in clinical practice. 3. When applicable, frameworks should also evaluate a therapy's associated diagnostic or test to demonstrate the comprehensive value of the treatment pathway.

■■ Overcoming Challenges with Data Collection Precision Medicine Data Collection, Analysis, and Application
A broad range of data and evidence sources are used to inform precision medicine decision making. Among these are databases such as PharmGKB, consortiums such as the Clinical Pharmacogenetics Implementation Consortium (CPIC), EHRs, insurance claims, and operations data (e.g., employee and supply chain data). In addition, data directly collected from patients, such as wearable technologies and genetic swab testing, are increasingly being used. Analytical advances such as machine learning, deep learning, and natural language processing are enabling the discovery of scientific relationships, can present significant challenges, especially for payers when determining whether to cover certain tests or therapies. Rather than delay coverage, payers are increasingly using coverage with evidence development, which allows patients to access tests or therapies while payers continue to collect and monitor real-world evidence (RWE) associated with these treatments. 4

Trial Designs
Traditional clinical trials require sample sizes and durations sufficient to detect clinically and statistically significant treatment effects. Based on the genetic characteristics of a disease, target populations for precision medicine therapies may be more narrowly defined. As a result, clinical trial designs predicated on large sample sizes may no longer be feasible when testing a therapy targeted to a subpopulation for which few indicated patients are available. Furthermore, the clinical utility of a test may not accrue until after a traditional clinical trial period is over. As such, traditional clinical trials may underestimate the value of a precision medicine test or diagnostic, posing risks for payers when determining whether to cover a test without sufficient evidence. "Basket" studies, "umbrella" studies, "N of 1" designs, and various adaptive and novel trial designs are emerging as viable options to overcome traditional clinical trial design limitations. 5,6 These designs enable more efficient allocation of limited patient populations and testing of multiple mutations, and help discern differences in health outcomes attributable to genomics and other individual or small group heterogeneity. To date, these designs have often been used in oncology; however, they are increasingly applied to other disease areas. 7

Value Assessment Frameworks
In recent years, various organizations have developed value assessment frameworks to support health care decision making and resource utilization. These frameworks largely rely on clinical evidence (e.g., overall survival) and economic information (e.g., wholesale acquisition cost) to assess value with measures such as quality-adjusted life-years (QALYs). With the approval and use of targeted therapies, however, reliance on traditional value measures may not be appropriate in current frameworks. These measures may not adequately convey the full value of a precision medicine therapy, which often accrues over time. Furthermore, QALYs do not account for unique aspects of precision medicine tests and diagnostics, such as the spillover of information to family members. 3

Recommendations
In light of these challenges, forum participants identified several recommendations for best practices in evidence generation and collection and trial designs in precision medicine.
Standardization, Collection, and Dissemination of Evidence 1. Participants recommended working groups to define and agree on basic sets of standards for tests used in precision medicine.
cause and effect or associations between phenotypes and potential outcomes, identification of risk factors, care patterns, behavioral associations, and health outcomes of precision medicine evidence. The generation of big data related to precision medicine poses challenges to efficient management and application of findings. A lack of system interoperability makes valuable information unavailable or inaccessible to clinicians and other health care providers. For example, data generated by a pathology lab may not be readily shared with or interpretable by other departments to inform real-time clinical decisions. Moreover, a lack of standard data formats and current business interests may limit meaningful sharing, storing, and interpretation of data, slowing decision support.
The current medical coding system also presents challenges for precision medicine data collection. Through innovations in research and analytics, an increasing number of diseases are understood to be very heterogeneous. For example, within cancer tumor types (such as breast), certain mutations or expressions may be absent or present, which can significantly affect treatment response or outomes. 8 The current medical coding system, however, was not designed to account for these distinctions within diseases. Therefore, when heterogeneity is present, clinicians and health systems cannot have a comprehensive and complete understanding of their patient population, track health care quality, or make informed decisions about resource allocation.

Engaging Patients in Data Collection
While there have been advances in evidence collection and interpretation, patient engagement remains an essential component in demonstrating the value of precision medicine tests and treatments. Fundamentally, however, there is a lack of patient literacy around genomics and how individual data can be shared and used. In addition, patient-reported outcomes (PROs), which represent direct measures of a patient's perceived health, largely are not collected and/or used in a manner that is optimal for precision medicine. In clinical trials, for example, it is challenging to achieve statistical power because the sample sizes for precision medicine tests and therapeutics often are small. 9 However, they are an important complement to the clinical and biometric components of precision medicine and can further demonstrate the value of new treatments by providing a direct measurement of the treatment's effects (e.g., side effects) that patients experience.
Analytic advances and other data collection methods have facilitated the generation of significant precision medicine data; however, using these data to make clinical advancements is not feasible without greater linkages of clinical, pathology, treatment, and cost data within a health system (e.g., a single hospital) and across systems (e.g., competing systems). Currently siloed data systems are limiting determinations of clinical utility, or value, at the individual and population levels.

Recommendations
In light of these limitations, forum participants developed recommendations for ways to make precision medicine evidence more readily available, usable, and actionable for all stakeholders.
Data Collection for Novel Trial Designs 1. Participants noted the need for novel forms of primary data collection and validated tools and methodologies to advance precision medicine trials beyond traditional RCTs. Coupling these tools with natural language processing can better generate more structured and comparable trial data. 2. Wraparound services, such as data analytics and machine learning, can reduce workflow for clinicians and other staff, improve data availability and utilization, and help identify the most appropriate patients for novel trials. 3. There should be greater integration of biometrics (e.g., wearables) into novel trials to better link trial data with RWE. Biometric data have become more readily available and measured, thereby providing a continuous source of data (e.g., sleep and blood pressure) to measure treatment effects. 11 Data Collection for Safety and Pharmacovigilance 1. Current EHRs are not fully interoperable with each other, creating challenges in tracking and identifying the occurrence of adverse events (AEs) across care settings. Participants recommended the need to develop minimum standards to facilitate EHR interoperability and integration.

Voluntary platforms and initiatives that continuously collect patient data, such as the American Society of Clinical
Oncology's CancerLinQ, should be more systematically integrated across disease areas to provide more data on AEs, which could prospectively identify patients at risk for these AEs. 3. Participants also highlighted the need for pharmacists to have access to patients' genomic data to support drug utilization reviews and drug utilization evaluations, which potentially can reduce the risk of AEs. Greater training and education of pharmacists about pharmacogenomics is needed to support this patient care. 12 Data Collection for Clinical Decision Making 1. Currently, payers and clinicians may not have sufficient incentives to collect precision medicine data to inform decision making. Generating RWE and communication of the ultimate benefits of precision medicine related to improved outcomes and lower health care costs could improve data collection, supporting therapy decisions and patient engagement.

Current coding systems, including the International
Classification of Diseases and Current Procedural Terminology systems, should be refined to account for the emerging specificity of genetically defined patient subgroups and genomic testing. These modifications should be done in such a way as to enable future modifications, reduce uncertainty, and minimize time and resources. 3. A neutral third party is needed to establish guidelines for data sharing to overcome inherent incentives and interests for businesses to hold data internally.

■■ Innovating Benefit Design and Reimbursement Strategies for Precision Medicine
Payers face challenges in evaluating and determining coverage of precision medicine tests and therapies. Evidence is often limited because of small participant sizes and short time horizons of most clinical trials. Also, patients frequently change insurers and do not carry their data with them, which may misalign payer and patient incentives with respect to episodes of care and time horizons for health outcomes. Payers also may be unsure about how to assess increasing volumes of data generated by precision medicine tests and diagnostics for coverage and reimbursement determinations. Participants provided recommendations to overcome these challenges in benefit design and reimbursement models, including approaches to integrating precision medicine in value-based contracting (VBC).

Recommendations
Benefit Design and Reimbursement Models 1. Flexible condition-specific bundles, which cover a patient's entire health episode, could reduce some risk to payers by holding providers accountable for care, while enabling the use of precision medicine. 2. There is a need for provider and patient incentives to promote portability of patient records, which could reduce the need for unnecessary retesting and support future treatment decisions. 3. When clinically appropriate, the window from prescribing to dispensing should be shortened through automated prior authorization and smart exception processes, such as linking timely genetic test results with formulary information, so that patients could avoid an appeals or exception process. 4. Better distinctions between germline (mutations that arise in germ cells and can be passed to offspring) and somatic (mutations that arise in a single cell) testing are needed to support evidence-based payment policies pertaining to precision medicine testing.
Value-Based and Outcomes-Based Contracting 1. VBC for precision medicine requires consideration of how risk is allocated across stakeholders. This includes careful selection of outcomes used for payment determinations.
2. Participants highlighted the need to focus on long-term outcomes and benefits, which are where many benefits of precision medicine may accrue. Payers should be supported or incentivized to cover precision medicine tests and therapies, including instances in which the benefits of these interventions accrue years later when some beneficiaries have switched payers. 3. Although most VBC to date has focused on drugs, this approach should also be extended to diagnostic technologies that guide therapy (e.g., imaging). 4. VBC should incorporate surrogate endpoints that have been validated as predictive of health care outcomes in the short and long term.

■■ Overcoming Legal and Regulatory Barriers
Current regulations and laws, such as the Health Insurance Portability and Accountability Act (HIPAA), the Genetic Information Nondiscrimination Act (GINA), and the Protecting Access to Medicare Act (PAMA) are meant to protect patients and their personal health information. However, they can pose barriers in precision medicine. For example, they can inhibit data sharing among clinicians and other health care providers and may not allow for precision medicine tests to be reimbursed in a value-based way, such as the Medicare Payment Advisory Commission's national limitation amounts. 9 In addition, there often is disconnect between the medical and pharmaceutical sides of health plans, which can be bridged by pharmacists.
Precision medicine also introduces potential ethical issues related to the reporting of findings and patient consent. For example, genetic testing can detect elevated risks (e.g., Alzheimer's disease) or mutations that are not related to the primary test indication and may not clinically manifest themselves. Patients and clinicians must determine the circumstances in which secondary findings are communicated. 10 Precision medicine is also increasingly marketed directly to consumers, but the data conveyed may not be evidence-based.

Recommendations
1. GINA should be expanded to include protections for life and disability insurance coverage, coverage decisions, and the military. This may incentivize patients to more readily participate in research and share their genetic data and information. 2. The scope of practice, and subsequent compensation, for pharmacists should be expanded to include ordering and interpretation of genetic tests, which supports team-based patient care, better drug use, and overall management of patient outcomes. 3. There is a need for regulations to ensure that patients' precision medicine genetic data and information are communicated in ways that are interpretable and meaningful to patients with different health literacy.

4.
Best practices on patient consent should be established so that patients, regardless of health literacy, can understand how their genetic information would be shared.

■■ Conclusions
In recent years, advances in research, analytics, and data management have expanded the use and capabilities of precision medicine by a number of health care stakeholders. Precision medicine diagnostics and therapies are enabling more individualized care, thereby improving patient health and allowing for better allocation of resources. A number of key operational and scientific barriers remain, however, that prevent more widespread use. Forum stakeholders discussed operational and scientific challenges, including the need for better identification of patient subgroups, incompatible datasets, patient identifiers and EHRs, lack of well-defined endpoints, underuse of novel trial designs and RWE, lack of patient education around genomics, and the need for regulations that promote patient safety while facilitating data sharing. In light of these challenges, stakeholders developed recommendations to promote stronger evidence generation, data collection, benefit and reimbursement design, and laws and regulations (Table 1).
While participants shared stakeholder-specific perspectives and recommendations, there were 2 overarching areas of consensus. First, there is a greater need for common definitions, thresholds, and standards to guide evidence generation in precision medicine. Second, current information silos are preventing the sharing of valuable precision medicine data. Collaboration among stakeholders is needed to support better information sharing, awareness, and education about precision medicine for patients. Fundamentally, however, advances in precision medicine will only be possible with multistakeholder initiatives that are centered on improving patient health, such as the recommendations provided here by forum participants.