Developing a Value-Based Care Model for Neurology

Provider reimbursement today is primarily based on a fee-for-service (FFS) reimbursement model. The primary benefit of FFS is that it is simple to understand and operationalize. In FFS, the physician provides a service, bills for the service, and is paid for delivering the service. Although FFS is simple to implement, it has contributed to our current cost crisis in health care. Value-based care is a proposed solution to decrease costs and improve outcomes; however, the move from FFS to value-based care has been slower than expected. Among the primary reasons is the complexity of designing a reimbursement model that supports and incentivizes value-based care. It is unrealistic to expect providers to improve quality and lower cost if the payment system does not support their efforts. For neurology, the FFS model leaves proactive high-value services to prevent disease and complications undercompensated and creates an environment where reducing avoidable spending can decrease provider revenue. As we continue to struggle with high-cost health care producing suboptimal outcomes, we need to continue to push for value-based payment reform.

To date, there have been 2 primary types of value-based reimbursement models explored in the United States.¹ The first is a population health model where primary care physicians are held accountable for the total cost of care for their attributed population. The Medicare Shared Savings Program (MSSP)² is an example of a population health model, where providers are given a benchmark target for spending for the attributed population and can share in the savings or are responsible for paying back losses based on how they perform against the benchmark.

The second type of value-based care model is an episode-based payment model, typically triggered by an acute event and including the cost of care for the event and a defined time frame after. The Bundled Payments for Care Initiative (BPCI)³ and Advanced (BPCIA)⁴ model are good examples. The most widely cited episode is for a joint replacement, where all providers are held accountable to the cost for the surgery and 90-day postoperative care. Acute ischemic stroke is an area in neurology where this model is actively being used. Episodes have also been used to bundle care after an admission for an exacerbation of chronic conditions and could be considered for neurologic disorders that include flare-ups requiring hospitalization such as myasthenia gravis or multiple sclerosis.

Although both models have shown modest gains in decreasing costs and improving outcomes,^1,5 specialists have had limited input and engagement in the models for various reasons. In population health models, it is difficult for a specialist to understand how they can be successful when they have a relatively small contribution to the total cost of care, and their ability to improve outcomes may be limited. Attribution also becomes challenging due to the significant variability of direct management by a specialist vs a primary care provider.

By contrast, in episode-based payment models, specialists typically do understand how they can affect costs for conditions they treat. Unfortunately, the decision to participate in current episode-based models is typically made at the health system executive level, many times without the buy-in from the specialists who care for the patients, limiting engagement. In addition, current episode-based models are typically triggered by an acute inpatient admission, which is a high-cost event. Decreasing the cost of these acute episodes may not affect the total annual cost of care for the condition and limits the ability for the specialist to affect the cost and improve the outcomes.

We are still in need of a value-based reimbursement model that engages specialists early in the disease process to improve care and decrease costs for patients with chronic neurologic conditions. One solution is to develop a model around specialty-specific longitudinal care episodes.⁶ Engaging specialists early in the disease process can improve outcomes for patients.⁷ Specialists also understand their contribution to the care for these conditions and, in optimizing care delivery, could completely avoid many of the high-cost acute events that would trigger a typical episode.

There are many things to consider in developing a longitudinal care episode for neurologic conditions. First is defining the episode itself including the trigger, time frame, and any specific exclusions to care attributed to the episode. The trigger may be the beginning of the calendar year, the date of initial diagnosis, or the date the patient first saw the specialist. Unlike the current episode-based payment model, the model would not be triggered by an admission because the goal is to avoid the admission in the first place. For a longitudinal episode of care, it may be beneficial to look at the care for an entire year for the chronic condition. By evaluating the annual cost of care for a chronic condition, you will be able to account for improved care and outcomes that avoid unnecessary inpatient utilization. We will also need to define what is included in the episode. Some things to consider include medical and surgical costs, postacute care, care team/caregiver support, palliative care, and costs of medications. By developing a comprehensive approach to longitudinal care for neurologic conditions, all these items may decrease the total cost of care. In the United States, we typically silo these costs into different areas of spending, when we need to consider them holistically. For example, a very expensive medication may ultimately decrease the total cost of care by eliminating expensive acute care episodes.

Second is the development of the payment model. This involves defining attribution to a provider and aligning on retrospective, prospective, or a combination payment model. One of the benefits of value-based care is that it frees providers from the confines of FFS, allowing them to provide care that has tremendous value but is not adequately compensated in FFS. Providers typically need some up-front funding to implement types of care that are not otherwise adequately compensated such as home-based services and caregiver support. In a retrospective bundle, providers are held accountable to a target cost for the episode, FFS is billed, and at the end of the episode, the costs are reconciled, and the provider receives a share of any savings below the target.⁸ This type of model does not provide any up-front funding for care that is not directly compensated, so a prospective payment model or a combination (prospective care management fee with a retrospective reconciliation) may make the most sense.

Finally, we need to consider the clinical care for the episode. This includes developing a model of care, quality metrics, and measurable outcomes. Specialists who care for this population are likely the best suited to develop the care and outcome criteria for the model. It is important in the development of the model to consider health equity, patient-reported outcomes, the patient and caregiver voice, and quality metrics that are timely and predict the outcome you want to achieve.

In evaluating conditions that would lend themselves to a longitudinal episode of care, it is important to first evaluate what conditions have the most opportunity to improve outcomes and decrease unnecessary spending.

The AAN and several other specialty societies, previously worked to design episode-based alternative payment models. The AAN developed models for headache and epilepsy care and presented them to the CMS Innovation Center (CMMI) in 2017 and 2018. These conditions were selected based on disease prevalence, ability for the neurologist to affect outcomes and cost, and the proportion of spend on neurology services. The models included episodes for both acute and chronic care within these 2 conditions. Neither model was implemented, and CMMI feedback included the need to incorporate higher levels of risk and the need to consider whole person care.

CMS recently released the new “GUIDE” model demonstration project for dementia care, which provides infrastructure and monthly funding to providers who care for patients with dementia. This model does not include a total cost of care target but has the goal of incorporating savings into other CMMI models.^9,10 The release of this model emphasizes the need to develop neurologic-specific longitudinal bundles.

Some of the lessons learned from feedback of prior models include advocating for model simplicity, assuring the ability to operationalize the model clinically and from a payer perspective, assessing applicability of models across multiple domains of neurologic care, and considering whether a condition is best suited to an acute episode or a longitudinal episode based on the chronicity of the condition. Finally, it would be beneficial to pilot any proposed model with a payer partner to show the ability to operationalize the model and validate the proposed improved outcomes and lower cost of care.

As we look to reign in health care expenditures, it is critical to engage specialists in this work. Episodic care models focused on acute conditions still have a place in value-based care; however, longitudinal episodes of care for chronic conditions in neurology is an area of untapped potential. By actively engaging to align payment models with outcomes, neurologists can improve the care that they provide and the outcomes for patients and their families.

Appendix. Authors

Name	Location	Contribution
Martha M. Walsh, MD, MHSA, FACOG	Blue Cross Blue Shield of Michigan, Detroit, MI	Drafting/revision of the article for content, including medical writing for content
Daniel J. Ackerman, MD, FAAN	St. Luke's University Health Network, Fountain Hill, PA	Drafting/revision of the article for content, including medical writing for content
Robert M. Kropp, MD, MBA, CPHI, FAAN	Robert M Kropp, MD Consultants, St. Petersburg, FL	Drafting/revision of the article for content, including medical writing for content
Meghan Eigenbrod, MPH	American Academy of Neurology, Minneapolis, MN	Drafting/revision of the article for content, including medical writing for content

Study Funding

The authors report no targeted funding.

Disclosure

The authors report no relevant disclosures. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.