Innovative payment models for complex populations

July 20, 2016 / By Richard Fuller, MS, Norbert Goldfield, MD

In developing our classification tools we are continually called upon to make judgement calls on the relative clinical complexity of individuals and the interaction between comorbid chronic conditions on quality outcomes such as preventable complications and/or the resources required to treat these individuals. Working with health professionals to create clinical models that emphasize the interaction between diseases  is a major differentiator of the categorical clinical model as compared to models that rely more heavily on statistical matching of what is already observed.1

The separation of the clinical description to form a normative categorical model of what “should be” compared to the positive model of “what is,” obtained from current patterns of spending and clinical outcomes, is essential when trying to foster meaningful payment reform – that which redirects patterns of care and resource allocation. This is essential because it is the core of payment reform and it is instantly recognizable when considering two of the most complex patient populations: Those with serious mental health/substance abuse disabilities (MHSA) and those potentially requiring long- or short- term residential (nursing home) care.

In a just-published Open Forum article in the journal Psychiatric Services, we highlight the challenges facing care for those with severe MHSA disabilities in the United States and provide detailed recommendations to help evolve the care system from one of crisis management to one of comprehensive care.2 There are three key components for addressing the issues:

New variables that increase the validity of risk-adjustment and can assist us in tracking outcomes used for this complex population need to be routinely collected. Essentially, we need to incorporate other elements into the patient description that are not routinely available, such as filled prescription history, educational attainment, homelessness, employment (paid or unpaid) history and history of incarceration.

We also need to reverse current patterns of care that equate inpatient admissions with suicide prevention rather than identifying the circumstances under which respite and/or inpatient care is appropriate as part of the overall care of the individual. Financial incentives to care for those with severe MHSA disabilities, combined with improved risk-adjustment, are necessary but not sufficient. We must provide a specific set of financial incentives focused on improved outcomes. In addition, payers need to encourage the exchange of information on best practices in combination with increasingly clinically-valid information on outcomes performance.

This is a tangible need and requires a normative view rather than one simply based upon making the current system more cost efficient. That is, we need to differentiate across patients and rationalize spending while promoting more appropriate outcomes (e.g. ability to stay employed) rather than capitate spending upon what we observe today.  Identifying other societal metrics that are important to “whole person” care is essential, as noted by the New York Medicaid Commissioner Jason Helgerson during a recent United Hospital Fund conference.3

Similar conditions exist when considering the complex populations increasingly falling under Managed Long-Term Services and Supports (MLTSS). A recent primer on rate setting for MLTSS, published by Mathematica, outlines in detail the issues and policy options facing Medicaid programs.  A key conclusion drawn by the authors is that strategies are necessary for diverting nursing home residents to community care and for better matching resources to patients needs while avoiding adverse risk selection.  They note that this requires “needs based risk-adjustment.” For those familiar with the Clinical Risk Grouping (CRG) classification system, it is no surprise that we have moved a significant way down this path with the introduction of the “mental and functional status grouper”5 embedded within the CRG 2.0 software. The separation of the clinical and financial elements allows payment system design that can reform the existing pattern of care.  It does this by recognizing the need to blend existing program costs (e.g. acute and custodial care) with new program costs (such as social support costs) as well as with new and better ways to identify the complexity and need of those whose care is being managed. Moreover, the creativity for matching payment to need should also be linked to the collection of additional data elements that are useful for both improved risk-adjustment and more precise specification of desired outcomes. Examples of needed elements, particularly for risk-adjustment, include the extent and quality of available social support and marital (significant other) status.

It is our hope that as we learn more about those distinctions that are meaningful for classification, we can continue to push for the routine collection of variables that capture this information and determine which metrics can measure meaningful outcomes for these vulnerable populations. As previously noted, financial incentives, combined with improved risk-adjustment, are necessary but not enough. Payers need to encourage the exchange of information on best practices in combination with increasingly clinically-valid information on outcomes performance.  Our next blog will continue this conversation with a focus on the just-released Institute of Medicine report on Accounting for Social Risk Factors in Medicare Payment, within the context of President Obama’s recently-published article in JAMA assessing our progress to date on health reform.

Richard Fuller, MS, is an economist with 3M Clinical and Economic Research.

Norbert Goldfield, MD, is medical director for 3M Clinical and Economic Research.


  1. Fuller RL, Averill RF, Muldoon JH, Hughes JS. Comparison of the Properties of Regression and Categorical Risk-Adjustment Models. J Ambul Care Manage. 39(2):157-165. doi:10.1097/JAC.0000000000000135.
  2. Goldfield NI, Fuller RL, Vertrees JC, McCullough EC. How Encouraging Provider Collaboration and Financial Incentives Can Improve Outcomes for Persons With Severe Psychiatric Disorders. Psychiatr Serv. 2016;
  3. Goldberg, Dan Helgerson lays out a vision for Medicaid that looks beyond health care.
  4. Dominiak M, Libersky J. Medicaid Rate-Setting for Managed Long-Term Services and Supports: Basic Practices for Integrated Medicare-Medicaid Programs.; 2016.
  5. Fuller RL, Hughes JS, Goldfield NI. Adjusting population risk for functional health status. Popul Health Manag.