Much of the focus in health policy circles and the media has been on the price transparency provisions of the recently released executive order on improving price and quality transparency. However, the executive order also calls for the establishment of a health quality roadmap that outlines a process for alignment of measures across all federal programs and care settings, adoption of common measures, and elimination of “low-value or counterproductive measures.”
The health quality roadmap will need to address several key questions including:
- What is the purpose of measurement?
- What do “high-value and productive” measures look like?
- What is a feasible pathway to evolving measurement?
In this post, we address these and other important questions and discuss the importance of developing patient-centered outcomes measures. We provide an overview of our approach, which is described in more detail in a just-released white paper and was developed through multistakeholder convenings and a review of existing measures and measurement frameworks. We argue that a confluence of factors makes this an ideal time to move forward with these next generation of measures.
Purpose Of Measurement
There have been many uses for measures, but the primary application for measures in any quality roadmap should be accountability. We use the term accountability to encompass two specific use cases: assessing and rewarding provider performance, and enabling informed provider selection by consumers.
With the growth in alternative payment models and the movement toward value, there has been an increasing recognition of the need for measures that are better suited for these models, which include accountable care organizations and episode-based payments. They address care across the continuum comprehensively and therefore require parsimonious and meaningful performance assessments, rather than assessments based on individual patient encounters such as receipt of a hemoglobin A1c test.
Similarly, these process measures have not been useful to consumers as they seek information to inform their choice of providers. In fact, the success of any transparency initiative aimed at patients and consumers will be dictated by our ability to provide useful price and quality performance information to individuals. We fully recognize the importance of more granular measures for use by providers in quality improvement but posit that clinicians and health systems should have the flexibility to choose such measures based upon their particular circumstances and quality improvement goals.
Principles Of Measures For Accountability
Irrespective of whether they are culled from existing measures or are newly developed, measures used for accountability purposes should be outcome-oriented, patient-centered, sourced from different types of data, and longitudinal and site-agnostic:
Measures should be focused on outcomes or be primarily outcomes-oriented and include measures sourced from different types of data: administrative, clinical, and patient-reported. They should also incorporate the concept of “appropriateness,” for example, avoiding unnecessary imaging studies for low back pain or ensuring appropriateness of pharmacologic prescribing, which involves ensuring that patients receive the “right” treatments.
As transformation efforts place patients and their preferences at the forefront of their care, measures used in accountability programs will need to mirror this shift.
Measures should be patient-centered and incorporate new approaches to assessing patient health status and patient experience. Such measures would include assessment of clinical outcomes, patient-reported outcome measures, as well as new approaches to evaluation of patient experience. For example, many patients use Yelp-style reviews or social media to access provider performance information; we need to find valid and reliable ways of tapping into these avenues or creating rigorous Yelp-like platforms for the purposes of assessing health system performance. A variety of data sources will be needed for these different types of patient-centered outcomes measures. Measures that assess complications post-surgery could be based on claims or a combination of claims and clinical data. Functional status measures would rely on patient-reported data collected using validated instruments such as the HOOS, JR. (hip disability and osteoarthritis outcome score)–KOOS, JR. (knee injury and osteoarthritis outcome score).
Longitudinal And Site Agnostic
Measures should be longitudinal and (as appropriate) agnostic of site of care. In other words, they should assess performance of the delivery system over time and across settings.
Developing And Implementing Patient-Centered Outcomes Measures
To progress toward wide-scale implementation of patient-centered outcomes measures will require identifying priority areas or conditions as test cases for this approach. These test cases should involve conditions that are high cost, high prevalence, or both; exhibit variations in care; and offer opportunities for ensuring appropriateness and adherence to clinical guidelines and further catalyzing the adoption of patient-centered outcomes measures.
The following are potential priority areas for subsequent measures work: depression, maternity care, patients with multiple chronic conditions, and osteoarthritis. Across these areas, availability of existing measures that meet the principles outlined above varies significantly, highlighting the need for holistic, new measure development. There are, however, a set of needed activities that are common across all these conditions. These include:
- Defining the population that will be the focus of measurement;
- Identifying the outcomes measure concepts relevant to the particular population and that meet the principles outlined earlier;
- Selecting from existing measures and funding new measure development;
- Establishing a network of payers and providers who can test implementation of these measures on a broader scale; and
- Addressing cross-cutting issues such as risk adjustment and data infrastructure needs in a consistent and efficient manner.
Patient-Centered Outcomes Measures: The Time Is Now
A confluence of enabling factors has set the stage for the immediate development and adoption of patient-centered outcomes measures. The executive order requires the development of a quality roadmap within 180 days of issue. At the same time, there is an opportunity to build on ongoing efforts to define populations for whom providers will be held accountable. An example is work at the Duke-Margolis Center for Health Policy and the Duke Department of Orthopedic Surgery to define patient populations and triggers for degenerative joint disease episodes; these efforts will use diagnosis and procedure codes initially, with a view to eventually using patient-reported functional status and shared decision-making aids.
There is also an opportunity to build on existing Center for Medicare and Medicaid Innovation efforts to collect patient-reported outcomes under the comprehensive joint replacement model. And as required by the Medicare Access and CHIP Reauthorization Act, the Centers for Medicare and Medicaid Services has awarded cooperative grants to several organizations to develop measures that tie to depression and osteoarthritis.
In some of the priority areas listed above, such as depression, outcome measures sourced from patient data that assess performance are readily available for widespread implementation. In fact, a measure of depression remission at 12 months is currently included in the Medicare Shared Savings Program measure set as well as the core set of measures released by the Core Quality Measures Collaborative. It will be important to learn from the experience of Medicare and other regional collaboratives in implementing this measure and scaling such efforts nationally.
Finally, one of the significant barriers in moving toward patient-centered outcomes measures that are sourced from different data sources has been the inability to easily exchange health care data. The 21st Century Cures Act has several provisions that address this important issue, and the recently released proposed rules to advance interoperability and data liquidity demonstrate that this is a top priority for the Trump administration. In addition, advances in apps and other technology to collect, transfer, and analyze data are gaining greater traction. Examples include the recently announced app PRISM that can help patients report health outcomes data.
We cannot overemphasize the urgency of implementing patient-centered outcomes measures, and the executive order further underscores that urgency. The concepts described in this post need additional vetting; we look forward to working with relevant stakeholders to develop a roadmap of patient-centered outcomes measures that extends beyond the federal alignment called for in the executive order to alignment between public and private participants.
The development and implementation of this roadmap will require sustained multistakeholder leadership and involvement from payers, providers, purchasers, and patients who can champion the ongoing development, adoption, and evolution of patient-centered outcomes measures. We envision a roadmap that will address in greater detail the list of activities across all priority areas discussed earlier, as well as identify feasible pathways for broad-scale testing and implementation of these measures. We look forward to collaborating with a broad range of parties to accomplish these goals.
Mark B. McClellan, MD, PhD, is an independent board member for Alignment Health Care, Cigna, and Johnson & Johnson. He co-chairs the Accountable Care Learning Collaborative and the Guiding Committee for the Health Care Payment Learning and Action Network, and serves as an adviser for Cota and MITRE. He has received research grants from the American Heart Association, the Commonwealth Fund, and the Gary and Mary West Foundation and WestHealth Policy Center.