October 29, 2016

Centers for Medicare & Medicaid Services
U.S. Department of Health & Human Services

To Whom It May Concern:

The National Committee for Quality Assurance (NCQA) would first like to thank you for the opportunity to comment on your Request for Information on State Innovation Model Concepts. We believe working directly with states to expand payment and delivery reform at the state level is a highly effective approach that can leverage public payer leadership to accelerate multi-payer alignment and health care transformation on a wide scale.

NCQA generally agrees with all objectives laid out in the RFI. We particularly support implementing models that could qualify as Advanced Alternative Payment Models (APMs). APMs not only have greater ability to improve quality because of their sophisticated structure, they are subject to financial risk and therefore also have strong incentives to do so without increasing costs.

Multi-payer State Based Strategies for APMs

Access to data has and will remain an immediate obstacle to any multi-payer alignment efforts. For example, self-insured payers are reluctant to share data on their enrollees. Partially or fully self-insured plans cover as many as 92 million lives according to Kaiser Family Foundation – an effective multi-payer strategy must address the issue with these plans. Behavioral health providers are often similarly reluctant to share patient data. Although it may not be feasible to compel these stakeholders to share certain data, federal authorities can develop models that offer incentives to do so.

Another challenge is that many payers have already invested significant dollars in quality improvement and payment reform. Before implementing a new standardized care intervention, states must first talk to each other and the payers in their markets to reach consensus about an APM that would both be of interest and also be accepted by Medicare. The Health Care Payment Learning & Action Network is a great example of information sharing and is a model for future collaboration across states.

Securing multi-payer participation is contingent upon incorporating existing efforts. For example, because of Star Ratings, Medicare Advantage plans have long-standing incentive structures and risk agreements in place that drive substantial portions of their revenues. Many Medicaid programs have also invested heavily in bundled or global payment initiatives. Any national alignment effort would need to acknowledge these as well as commercial payer models for those plans and clinicians that participate in both public and private programs. This would mitigate the risk of conflicting payment models that generate cross-purpose incentives and also ensure that MCOs remain financially viable. Similarly, models must be flexible enough to ensure that model participants can meet the needs of their specific populations. For example, commercial insurers within Tennessee’s public employee benefits program were unable to participate in the state’s SIM initiative because that particular model was so heavily focused on Medicaid.

Addressing these challenges and others will be critical to the success of multi-payer delivery system reforms. However, there are several other factors to consider as well:

  • Patient-centered design: Models must put patient safety, outcomes and experience at the forefront of delivery reform.
  • Core set of quality measures: There should be consistent quality measures across payers, specified to the unit of accountability, applicable to both primary and specialty care, that can be used to facilitate apples-to-apples comparisons. Consistency also reduces the reporting burden on clinicians.
  • Technical assistance and support for practice management infrastructure: Financial and technical support for care management platforms and other workable health IT solutions will be critical to both the success of a model and the participating clinicians. An agreed-upon health information exchange is essential, with additional support provided to practices looking to begin electronic clinical data reporting. Models should require audited oversight to ensure the integrity of the data and the accuracy of reporting prior to attaching payment to performance. Tools that analyze data on claims and gaps in care and also generate actionable feedback to clinicians would be similarly helpful. Especially on cost measures, clinicians need more robust performance reports to identify areas for improvement.
  • Risk adjustment: NCQA does not support the risk adjustment of clinical quality measures for sociodemographic factors. We do however support adjusting payments to clinicians to account for the greater resources needed to care for complex patient panels. Adjusting the measures themselves would merely hide without addressing gaps in quality – especially among more disadvantaged and complex Medicaid populations.
  • Transparent benchmarking: Transparent, agreed upon benchmarking methodologies will be helpful for all clinicians. These are particularly critical however for small and rural practices that may have less control over the total cost of care. Consensus on methodologies that promote accountability for lowering the cost of care will also encourage the kind of collaboration among clinicians that’s necessary for improving patient outcomes.

Medicaid programs face their own unique challenges in addition to those discussed above. Constant shifts in eligibility and enrollment impact the ability to uniquely identify patients across the continuum of care. This in turn makes it difficult to manage the care provided to Medicaid beneficiaries and fairly attribute each member to a clinician. Outdated state systems add to these challenges – limitations in data collection result in some states only measuring what is easy to measure because their systems are currently unable to collect anything else.

Social factors that influence health tend to have a greater impact on this population as well, resulting in greater incidence of chronic comorbidity and greater challenges in adherence to medications and care plans. These factors are especially pronounced in rural areas in states with limited access where fair comparisons of patient outcomes may be difficult to achieve.

To get an accurate evaluation of readiness and continued transformation toward patient-centered care, we recommend using a consistent practice assessment tool. Standardized models such as NCQA Patient-Centered Medical Home (PCMH) and Patient-Centered Specialty Practice (PCSP) could provide the desired consistency.

For example, the State of Vermont Blueprint for Health used NCQA PCMH as a multi-payer standardized measure of practices’ capability to provide advanced primary care. This provided the necessary foundation for those practices to participate more broadly in reform efforts through State Innovation Model testing.

The basic tools and technical assistance listed above are the foundational elements that CMS can support for states launching multi-payer reforms. To reiterate, there must be an emphasis on health IT. New support for building health information exchanges and analytics engines will facilitate regional exchange and processing of clinical quality data; this should be the linchpin of any sophisticated delivery reform effort. Usability testing of these solutions is also necessary to ensure they can be effectively integrated into clinical workflows. CMS could further bolster this effort by requiring public reporting of scoring to drive faster innovation and improvement in usability and interoperability.

Meaningful evaluations will be important for measuring the efficacy and extent of new delivery reform models. However, such evaluations are contingent upon several factors that must be built into a model itself to ensure accuracy and fairness. These factors include risk stratification models, appropriate risk adjustment to payment, consistent measures for each population as well as consistent methodologies for data collection, calculation and reporting. Each process should also be audited, either through certification or other third party methods to ensure data integrity.

We encourage CMS to emphasize patient-reported outcomes measures as a way to improve the fairness of evaluations. However, we do have concerns about using statewide survey data, such as CAHPS, to measure outcomes in a model. Although it may be expedient, it could potentially disconnect measurement from accountability because the results may not be sufficiently timely or actionable for individual facilities or clinicians. We believe you should incorporate ways to measure outcomes that provide actionable results that can in turn be used for care processes, quality improvement, and accountability.

We also recommend special consideration for safety net facilities and practices that may need more support to accommodate their unique circumstances.

Total Population Health

Achieving accountability for total population health will require all of the tools and resources listed above and more. Again, this includes special focus on technical and financial support for robust care management platforms as well as data extraction processes. Care management fees provide a good foundation but CMS should provide resources specific to total population health. NCQA PCMH and PCSP standards provide the necessary infrastructure for practices looking to assume this kind of accountability and incentivizing NCQA recognition could help ensure that practices are actually prepared to do so. The Medicare Access and CHIP Reauthorization Act (MACRA) recognizes our programs as meeting national standards for Clinical Practice Improvement. This is an ideal opportunity to create alignment across programs by encouraging use of a single standard like NCQA PCMH to meet a specific goal such as total population health accountability.

Again, general health IT adoption and use to improve care must also remain a top tier priority. However, rather than tie financial incentives to adoption, it may be useful to incentivize progress on the specific features of health IT. For example, usability and interoperability are not evolving at an adequate pace.  MACRA includes provisions that encourage progress in these areas but incentives at the state level could provide an additional policy lever to catalyze further development.

For example, you could incentivize development of platforms that enable hospitals to send real-time admission alerts to care coordinators. You could further incentivize clinicians themselves to actually use those systems to coordinate follow-up care.

We encourage CMS to offer support for all-payer claims databases and alignment of data structures for the purpose of supporting attribution models as well as quality and cost measure calculations. It will be important, however, to standardize data access rules as these rules currently differ across states. We also believe it’s important to address the challenges in getting data from health plans to any kind of integrated database. Third party certification should be used to verify the integrity of the source data before it’s integrated into any such database.

For population health measures, we again encourage use of patient-reported outcomes. Vital Signs and Patient-Reported Outcomes Measurement Information System are great resources for patient-centered measures that monitor physical, mental, and social health. These will offer the kind of whole-person assessment that will be critical for accurately measuring outcomes.

Another area of opportunity is behavioral health where privacy rules limit data sharing.  We encourage CMS to work more closely with SAMHSA to outline the types of data that can be shared across clinicians to encourage more care coordination. Specifically, we encourage you to provide further clarification on the 42 CFR Part 2 regulations that dictate substance abuse confidentiality. Behavioral health is a critical aspect of patient outcomes and lack of clarity around data sharing rules remains a major obstacle for care coordination between primary and behavioral health care providers.

Assessing the Impact of Multi-State Care Interventions

It is our experience that states are willing and excited about participating in new delivery and payment initiatives. It will be important, however, for states and the payers in their markets to come to consensus on a model that is both appealing and rigorous enough to qualify as an Advanced APM under MACRA. Vermont is an example of a state that leveraged Medicaid authority to test the impact of statewide multi-payer support for patient-centered medical homes. The Vermont PCMHs have lowered annual health care expenditures by as much as $450 per patient.

Another area of opportunity is in addressing disparities, as noted in the recently updated Medicaid Managed Care Rule. Effective models to address disparities should include extensive, mandatory data collection and use imputation strategies where data availability lags. Stratified reporting of that data will help identify specific drivers of regional disparities and the corresponding payment model can provide adjustments to account for those complexities. Plans and clinicians will need assistance with implementation so the model should also provide technical assistance.

NCQA developed a set of standards for plans and community based organizations delivering Managed Long-Term Services and Supports (MLTSS) and these could be used for models specific to addressing disparities. This accreditation program addresses the unique needs of individuals receiving LTSS in the home and community, including the non-medical supports necessary to provide well-coordinated, comprehensive care. Requiring this accreditation for programs such as the Financial Alignment Demonstrations could offer the alignment and standardization that is critical to a model’s success.

States can leverage External Quality Review Organizations to focus on quality improvement specific to mitigating health disparities. Such an approach should implement person-centered care processes and evaluate beneficiary experience through patient engagement in all relevant languages. Collaboration with community health organizations can help engage people who may be difficult to reach. The NCQA Multicultural Health Distinction is an accreditation option for organizations to distinguish themselves as meeting the Office of Minority Health Culturally and Linguistically Appropriate Services Standards.

Another challenge to consider is that Medicaid beneficiaries often cycle in and out of eligibility. Tracking and coordinating care for this population across the care continuum is therefore difficult. Direct facilitation of all-payer claims databases and health information exchanges could mitigate some of those tracking issues.

Finally, we would like to praise your efforts at implementing sophisticated multi-payer alignment models such as Comprehensive Primary Care Plus. Continuing to do so will encourage greater collaboration between payers and help clinicians prepare to participate in Advanced APMs. Future patient-centered models should test interventions such as behavioral health integration and also focus on fostering community linkages that can help address the social determinants of health. 

Thank you for the opportunity to comment on this Request for Information. We look forward to working with CMS and the Innovation Center as you develop new priorities and new models for health care payment and delivery. Please contact Kristine Thurston Toppe at 202-955-1744 or toppe@ncqa.org if you have any questions.


Margaret O’Kane