By Russ Montgomery, Tom Valuck, Cynthia A. Bens, Andrea Ferris, Health Affairs | April 30, 2019
Personalized medicine, which targets care based on a patient’s genomics, genetics, medical status, personal circumstances, and care preferences, is transforming health care delivery. Meanwhile, value-based payment is also driving sweeping changes by tying payment to quality. However, the traditional approaches to quality measurement and improvement are based on standardization of care, which fundamentally conflicts with the goals of personalized medicine.
Using oncology as an example, in this blog post we outline three tensions that put quality measurement and personalized medicine into conflict and present recommendations for new approaches to quality measurement that better align with the goals of personalized medicine. Recognizing that quality measures are necessary but not sufficient, we also highlight other opportunities to further align quality and personalized medicine.
Now is the time to implement new approaches to measuring and improving the quality of personalized medicine. According to the Health Care Payment Learning and Action Network, more than one-third of health care payment is now tied to alternative payment models (APMs). In addition, most oncology drugs approved by the Food and Drug Administration (FDA) in 2018 included a biomarker as part of their indication, and the FDA recently announced plans to open a new office focused in part on implementing new review and approval processes for personalized medicines. As the regulatory approval process for personalized medicine is evolving, the processes used to ensure high-quality care must evolve, as well.
The tensions and recommendations presented in this blog post were identified through a scan and analysis of existing quality measures and a series of interviews with patient advocacy groups, practicing oncologists, researchers, and oncology professional societies. More information about the methods and recommendations is available in this issue brief.
Tensions Between Measurement And Personalized Medicine
In our measure scan, we found 26 measures focused on biomarker testing and targeted therapies in oncology. We analyzed these measures and identified three fundamental tensions between the focus and specifications of these measures and personalized medicine.
Highly Specific Clinical Criteria Leads To Low Measure Reliability
Existing measures of biomarker tests and targeted therapies focus on narrow groups of patients. With a decrease in both numerator and denominator for any given measure, the reliability of that measure also decreases. A given community oncologist will likely only see a handful of patients in such a narrow group in a given year and may have no viable personalized medicine treatment measures that meet minimum case threshold for certain quality reporting programs, such as the Merit-Based Incentive Payment System (MIPS). Even with the minimum threshold met, results will be highly variable and may be driven more by chance than performance.
At the same time, small patient numbers and related reliability issues mean that measures of rarer cancers and patient subtypes are unlikely to be developed in the first place. We found that existing measures focus almost exclusively on the most prevalent cancers, such as breast and colorectal.
Rapidly Evolving Science Leads To Out-Of-Date Measures
Due to rapidly evolving science, measures of personalized medicine have the potential to quickly become out of date. Measures are generally derived from clinical guideline recommendations, and processes are in place to ensure that measures keep up with evolving guidelines. However, guidelines themselves do not always keep up with the pace of new drug development. As biomedical scientific advances are made, maintaining up-to-date targeted treatment and biomarker measures will become a never-ending process that is difficult to achieve.
Lack Of Allowance For Patient Preference Can Lead To Worse Measure Performance
Engaging patients to understand their care preferences and applying that understanding to clinical decision making is a fundamental tenet of both value-based health care and personalized medicine. However, existing oncology measures that focus on patients receiving, or being spared, a targeted therapy based on mutation status do not make allowances for patients preferring another option. In this scenario, a clinician’s performance on the measure will suffer for respecting patients’ wishes, potentially resulting in adverse financial consequences.
Improving Oncology Quality Measurement
To address these tensions and establish a better foundation for quality in personalized medicine, we developed a set of recommendations aimed at stakeholders with an interest in oncology quality measurement and improvement. These recommendations were informed by our measure scan and analysis, and interviews with practicing oncologists, patient groups, researchers, and other stakeholders.
1. Continue To Develop Measures That Promote Use Of Appropriate Biomarker Testing, Including Broader Measures That Apply Across Cancers
Access to biomarker testing is foundational to personalized medicine. Measures of biomarker testing increase rates of appropriate testing, and the results can then be used to inform care plans, increase participation in appropriate clinical trials, and improve access to appropriate targeted therapies. However, significant measure gaps remain for important biomarker tests tied to treatment. For example, LUNGevity Foundation has prioritized increasing access to comprehensive biomarker testing as a way to promote precision medicine in lung cancer treatment. They identified a lack of biomarker testing measures as a significant barrier to increased access and are evaluating options for the development of a measure of comprehensive lung cancer biomarker testing that would address this gap.
While quality measures are viable for biomarker tests for cancers with relatively higher prevalence, such as lung cancer, small patient numbers and related reliability issues make biomarker measures less feasible for less prevalent cancers. As additional targeted therapies are developed and patient subgroups become narrower, there is a need for measures of comprehensive biomarker testing that can be used across multiple cancers. For example, there is no existing measure of next generation sequencing technology, which has applications to multiple cancers.
2. Incorporate Allowances For Patient Preferences Into All Biomarker Testing And Tailored Treatment Measures
Allowances for patients’ preferences should be included in all measures of biomarker testing and targeted therapy so clinicians are not penalized for following patients’ wishes. For instance, measures focused on whether a patient received—or was spared—a targeted therapy could, where appropriate, be defined to focus on whether the clinician recommended the therapy instead of whether the therapy was ultimately received. Or, a patient could be removed from the denominator if they refuse the test or treatment. Existing measures addressing the use of adjuvant hormonal therapy and combination chemotherapy for breast cancer take this approach.
3. Develop And Use Measures Related To Patient Goal Setting And Concordance With Patient Goals
During our interviews, patient advocates and practicing oncologists expressed interest in the development and use of patient goal attainment measures. Patients and providers are increasingly working together to set treatment goals and “co-create” a care plan. This approach is required in the Oncology Care Model, an APM implemented by the Centers for Medicare and Medicaid Services (CMS). However, in that model, there is no formal review to assess whether care is ultimately concordant with the plan. A care plan concordance measure would provide a patient-centered approach to assessing quality of treatment and could be used across cancers, which would reduce reliability issues.
4. Prioritize The Development And Use Of Patient-Centered Outcomes Measures, Including Measures Of Quality Of Life And Patient Experience That Apply Across Cancers
There is a dearth of cancer-specific measures that assess quality of life and patient experience. In fact, there are no cancer-specific quality-of-life measures in use in CMS programs. Quality-of-life measures should be used to assess the side effects of chemotherapy and other treatments, and patient experience measures should be used to assess the quality of cancer care received. These measures can be applied across different cancer types, reducing issues of small numbers and measure reliability, and are not subject to becoming out of date as new therapies are developed.
5. Promote Quality Using Tools And Initiatives Beyond Measurement
There will always be limitations to quality measurement. Measures should be considered one tool in a toolbox that includes other ways to ensure high-quality, personalized cancer care. To provide high-quality personalized care, oncology practices must transform and implement new clinical care models. This can be encouraged through facility accreditation programs, practice recognition programs such as the Patient-Centered Specialty Practice, and APMs. In addition, patients must be educated and empowered to take a more active role in decision making. Patient engagement activities such as shared decision making and goal setting should be encouraged through MIPS Improvement Activities.
While these programs and activities provide structure and incentives, transformation is difficult, and practices need assistance addressing transformation challenges. To meet this need, the Personalized Medicine Coalition (PMC) is engaged in a project to identify barriers that impact the speed at which personalized medicine is integrated into care. Based on this learning, the PMC will recommend specific practice improvements to address these challenges.
Conclusion
There are fundamental tensions between quality measurement and personalized medicine, and a new approach is needed for measures to be patient-centered, reliable, adaptive to changing evidence, and meaningful and fair to providers. However, measurement alone is not sufficient for ensuring that personalized medicine is high quality and fully meets the needs of patients. Clinical practice transformation and patient education and empowerment are essential.