Implementation of a Precision Oncology Program as an Exemplar of a Learning Health Care System in the VA

Million Veteran Program

The Million Veteran Program (MVP) is a VA ORD initiative that asks veterans to share their medical data, lifestyle, and genetic data with researchers to allow for the discovery of correlations between their genetic profile and their health, disease and response totreatments. Currently more than 430,000 veterans have agreed to participate and have donated data and blood samples, and researchers are performing the first projects to use this resource.

Although the knowledge gained from these studies will be indirectly relevant to veterans in general the MVP presents an opportunity to present specific findings to individual participants that will directly affect their care. While reuse of the MVP resource for precision medicine is under consideration, there are important cultural and technical barriers that must be addressed. Like POP, integration of the MVP research program with clinical care should be carried out with consideration of a community of stakeholders and not driven exclusively by a research agenda.

Challenges in Moving Forward

Central to the implementation of a learning mechanism in health care systems is the recognition by administrators of the importance of the activity and appreciation of the business argument favoring the investment. This runs counter to the current notion of separate silos for health care and medical research whereby health care systems are liberated from the cost of investigation but then suffer from a dearth of knowledge relevant to their operation.

Additionally, research enterprises are not structured for such activities. Academic investigators are incentivized to create knowledge and generate publications and they understand best the currency of grant funding. Their world is not geared to reinvent or engineer solutions for health care systems. In light of these considerations, a decentralized approach that creates institutions for local learning needs to be developed and “owned” by individual and groups of medical centers with engagement of administration, patient, scientific, and community stakeholders. The Patient-Centered Outcome Research Institute (PCORI) and the consortia it has funded, PCOR-Net, have adopted this approach.³

Importantly, a new set of ethical and regulatory standards that distinguish it from traditional research must accompany progress in the creation of a learning health care system (LHS). Sharing of patient data to benefit fellow patients must come to be expected and without the formalized sharing agreements that are required in traditional research activities. Although the digitization of medical records makes most of what this article discusses possible, execution requires access to information technology resources and a talented staff.

More than a decade ago, the decision was made to dis-integrate the Office of Information Technology from VHA. This was executed with no provision to support the small army of VA clinician-informaticists who had done much in support of patient care, including the creation of the initial iteration of the VA EHR. Although the VA includes small pockets of this clinical informatics culture throughout its organization, the community has been largely silenced and taken refuge at academic affiliates. Access to VA information systems and funding opportunities for development and implementation of tools essential for learning will draw this intellectual capital back to the VA and allow for the VA to lead in this critical arena.

The VA Precision Oncology Program

Precision medicine is a medical model that incorporates the results of genetic diagnostic testing to customize or tailor medical decision making and treatment for the individual patient. Characteristics of the VA health care system that create a favored environment for introducing precision medicine include the single-payer model, where implementation decision and authority are centralized, a standardized EHR that enables informatics requirements, and a clinician and patient culture that supports innovation. To date, the benefits of precision medicine are most robust in cancer care. Under the leadership of Michael Mayo-Smith, MD, the VA New England Healthcare System has completed a regional pilot project in precision oncology that demonstrated feasibility of incorporating a precision medicine program in the clinical care environment.

For the majority of patients with lung cancer, DNA sequencing of tumor tissue identifies driver mutations—alterations believed responsible for tumor growth and behavior. The abundance of both driver and passenger mutations (those alterations whose significance is unknown) identified within an individual cancer specimen and the diversity of alterations found across the spectrum of all patients with cancer virtually assures the unique genetic profile (hence behavior) of any given patient’s tumor. The new generation of antineoplastic agents are targeted therapies that disrupt the downstream effects of these alterations and result in improved anticancer effects and reduced toxicity compared with conventional chemotherapy. The POP approach to cancer treatment determines the mutation profile of malignancies and identifies targeted therapies with the highest likelihood of treatment success. Although many driver mutation-targeted therapy combinations have been FDA approved, many more are in development and are available only as investigational agents.

Work Accomplished

Developed over the past 2 years in VISN 1, POP is a demonstration project that standardizes the processes necessary to deliver precision oncology care for veterans with lung cancer. With approval of the cancer care specialist, targeted sequencing of cancer genes (multiple biomarker panels) is performed on formalinfixed, paraffin-embedded tissue from newly diagnosed lung cancers as part of routine POP cancer care. Samples are shipped within 48 hours of diagnosis to Personal Genome Diagnostics (CancerSelect-88 targeted genome panel: PGD, Baltimore, MD) or Personalis (ACE Extended Cancer Panel: Menlo Park, CA). Following the sequencing of the targeted gene regions for mutations, a formal report of identified genomic aberrations is collated, annotated, and transmitted for inclusion in patient medical records. Both PGD and Personalis use N-of-One (Lexington, MA) to curate the medical literature and provide mutation annotations. The VA Computerized Patient Record System shares mutation results with the treating clinician, and a consultation service, offered through Specialty Care Access Network-Extension for Community technology, is available to help clinicians incorporate the test results into a treatment plan for the patient.

The POP is highly interdisciplinary: design and implementation required buy-in and coordinated efforts from the clinical medicine, laboratory medicine, pathology, pharmacy, radiology, and research services as well as from contracting, human resources, information technology, and procurement. With more than 150 specimens processed, procedures for tissue selection, processing, shipment, and tracking have been refined, and the informatics challenges met.

A Learning Health Care System Approach

Although the standard of care in oncology is evolving to include sequencing for all solid tumors and hematologic malignancies, the lack of correlated mutation status, patient outcomes data available for analysis, and difficulties in identifying subjects eligible for clinical trials of novel therapeutics combine to slow progress. The former problem arises from the effort required to aggregate EHR data from disparate systems as well as technical and cultural barriers to data sharing. The latter problem stems from the relative rarity of patients (and the difficulty identifying them) with a given mutation that determines eligibility for a clinical trial of a particular targeted therapy.