Performance Measurement is one of several fundamental components of alternative payment models (APMs), and specifically population-based payment (PBP) models. Clear performance goals addressing clinical quality, health outcomes, patient-care experience, and cost-of-care form the basis for APM agreements between providers and payers. Under these payment models, performance measurement plays a dual role of establishing accountability and encouraging improvement. It provides the set of standards by which providers, payers, purchasers, and patients can collectively assess whether health care and health targets are met.

The purpose of this paper is to help create a culture and environment conducive to designing and implementing measurement systems that are tailored to the demands of PBP models. It describes types of measures and categories of data needed to build a reporting infrastructure required for the long-term success and sustainability of PBP models. However, the paper does not include recommendations on specific measures that should be included in measure sets for PBP models. There are many other groups actively involved in developing consensus-based core measure sets, and the PBP Work Group did not think it was appropriate to duplicate efforts in this area.

The PBP Work Group believes measure sets for PBP models should be aligned to the greatest extent possible with common market segments. There is no principled reason why performance measurements should differ between public and private payers absent statutory or regulatory reasons. Nevertheless, the Work Group recognizes there are principled reasons to adopt different measure sets in market segments that target different populations, and as a result we do not expect complete alignment in measure sets across market segments.

One approach is to use statistical modeling to establish the “high-end” of what is possible to achieve a measure in “absolute” terms (e.g., a beta binomial distribution). Such techniques can also help clarify the minimal level of performance that the model will reward, again, in “absolute” terms. With this performance range in place, it would then be possible to system in which incremental improvements within the range are rewarded incrementally. So long as the minimum performance threshold is within reach of most providers, this type of system would motivate low performers to meet the performance threshold and subsequently clear the successive gates within the performance range. Similarly, so long as the high end of the performance range represents a standard of excellence, this type of system would motivate high performers to continue to improve until they meet the standard of excellence.

The paper recognizes the value of process measures and does not endorse eliminating them. Rather, the Work Group is proposing a new approach to using process measures to assess and reward performance. For the time being, as more outcomes-based measures are being developed, the Work Group believes it will necessary to combine process measures into aggregated Level 2 measures that are used to adjust payments in PBP models. For the purposes of adjusting PBP payments, the Work Group recommends replacing these aggregated process measures with Level 2 outcomes-based measures as they are developed and implemented. Nevertheless, the Work Group envisions that process measures will continue to be valuable for providers’ internal quality improvement activities, as higher performance on the process measures will likely translate into higher performance on the Level 2 outcomes-based measures that are used for payment.

The Work Group and LAN believe it is absolutely imperative to accelerate the transition to PBP models in light of current, unsustainable health-care spending in this country. Although existing performance measures are not sufficient to sustain long- term success of PBP models in achieving the goals of better care, better health, and lower costs, the Work Group strongly believes that existing measures are sufficient to initiate the transition to PBP. As discussed in the context of Recommendation 1, the Work Group believes it is appropriate to aggregate Level 3 measures together into Level 2 measures of health and quality. Until outcomes-based, “big-dot”  measures are developed, these aggregated Level 2 measures can be used to adjust payments in PBP models. However, we are acutely aware that these aggregated Level 2 measures are not ideal, and this is why the white paper provides very specific recommendations on 1) the types of measures that will be needed for PBP models; 2) the types of governance and coordination that will be needed to accelerate the development of measures for PBP models; and 3) approaches to refining and implementing measures that expedite their uptake and improve their reliability. The Work Group expects that adopting these recommendations will dramatically increase the availability of outcomes-based PBP measures in the short- to medium-term.

The PBP Work Group recognizes that a small number of Level 1 measures of health and well-being are in use for policy and research purposes. Information is beginning to appear about how summary measures of population health and well-being may be able to be used by health plans and accountable care organizations, but no reports on how they could be used for payment. (Reference Kottke et   ) Recommendation 1 of the White Paper notes that exploration of approaches to measurement and accountability models that might enable Level 1 quality and health measures should be developed and used. In the meantime, the White Paper highlights some important ways that the use of Level 2 measures of health and quality in PBP models brings social determinants of health into view, and causes providers to focus attention on factors outside of the four walls of clinical settings. This is a principal reason that the White Paper urges the importance of transitioning to more outcomes-oriented set of measures for the long-term success of PBP models.

As discussed in the context of Recommendation 1, TCOC is a truly “big dot,”  Level 1 measure for costs, because it evaluates ultimate outcomes (i.e., dollars spent) rather than the intermediary processes (e.g., emergency department utilization) that bring about these outcomes. TCOC is a vitally important measure in PBP models, forming the basis for accountability on cost (e.g., global budgets) and determination of shared savings or deficits. The use of TCOC for these purposes is addressed in detail in the Financial Benchmarking paper. Importantly, because of the use of TCOC in the risk-sharing arrangements that are central to PBP models, the Work Group holds strongly that “little dot” measures of cost, such as readmissions, ambulatory sensitive admissions, and non-urgent emergency room use, should not be included in PBP quality incentive measure sets. To include them in quality incentives would mean double paying providers for the same result—once through shared savings and again through quality payments. These little-dot measures are important to track and benchmark in PBP models, however, as they can be extremely valuable tools by which to support improvements on TCOC.

The Work Group acknowledges the ever-increasing importance of patient experience and patient outcomes in assessing quality and appropriateness of care. Likewise, because families are partners critical to the success of acute and especially chronic care, their perspective and experience must be included in the voices that have a say in what ultimately is important to measure. Therefore, both patients and family members are integral to the decision-making around selection, prioritization, and development of measures for PBP models.

The Work Group acknowledges that the incentives of PBP models differ significantly from those in traditional fee-for-service payment models. This is by design. While the fee-for-service payment model brought with it the risk of over-treatment due to incentives that reward volume and complexity of care, the introduction of budget constraints into payment, such as occurs with PBP models, brings the risk of under-treatment. A key role of quality and outcome measures in PBP models—and the reason they are given such extensive attention and focus in our work— is that the incentives tied to performance on these measures in PBP models are a key safeguard against under-use. Accordingly, the Work Group believes that the recommended approaches to performance measurement will help to ensure that PBP models do not lead to reductions in appropriate or necessary care. To the contrary, the Work Group believes that the proposed approaches to performance measurement and PBP model incentives will support significant improvements in quality, outcomes, and patient-care experiences.

The PBP Work Group recognizes that primary care is essential to the success of PBP models and that a lack of adequate primary care could conceivably skew the structure that supports and allows a PBP model to function properly. Consequently, the LAN has established a Primary Care Payment Model Work Group which will, among other considerations, address concerns around adequate primary care capacity and the implications of what that means for advancing PBP model implementation.