Editor’s note: This post is the first in a periodic Health Affairs Blog series, which will run over the next year, looking at payment and delivery reforms in Arkansas and Oregon. The posts will be based on evaluations of these reforms performed with the support of the Robert Wood Johnson Foundation. The authors of this post are part of the team evaluating the Arkansas model.

In 1932, Supreme Court Justice Louis Brandeis famously described states as laboratories of democracy. The idea that states can try “novel social and economic experiments without risk to the rest of the country” is certainly relevant today as the United States reforms its health care system. Many states are developing novel strategies.  Arkansas is one, adopting an innovative model of delivering and financing health care.

The Arkansas transformation model, known as the Arkansas Payment Improvement Initiative (APII), aims to create payment incentives that motivate and reward patient centered management of outcomes and reductions in total cost of care. The payment model relies on several types of reforms, including payments based on “episodes of care” and incentives surrounding Patient Centered Medical Homes for health needs and Health Homes for populations with additional, more complex support needs. In this post, we focus on the episodic payment component. Arkansas’s system is a multi-payer model, with participation by Medicaid, the self-insured State Employee and Public School Employee health insurance plans and the two largest insurers in the state, Arkansas Blue Cross and Blue Shield (ABCBS) and QualChoice of Arkansas (QCA), as well as Walmart, the state’s largest private self-insured employer purchaser.

To build the episodic payment system, the APII’s leadership in consultation with stakeholders targeted select care conditions to find areas of practice variation and inefficiency in common procedures or illnesses. All associated events were grouped together and labeled “episodes of care.”  A growing list of episodes has been developed with input from consumers, providers and other stakeholders collected through expert interviews, online surveys and public workgroup meetings.  Episode selection and definition reflects an extensive public stakeholder and provider engagement process, review of evidence-based guidelines, examination of claims data and consultation with both state staff and national experts.

The initial episodes, launched at the end of 2012, were Pregnancy, Attention Deficit/Hyperactivity Disorder (ADHD), Hip and Knee Replacement, Congestive Heart Failure (CHF) and Upper Respiratory Infections (URI).  Four more, Colonoscopy, Cholecystectomy (Gallbladder Removal), Tonsillectomy and Oppositional Defiance Disorder (ODD), were designed and implemented in 2013.  An additional six are planned for 2014, including Coronary Artery Bypass Grafting (CABG), Percutaneous Coronary Intervention (PCI), Asthma, Chronic Obstructive Pulmonary Disease (COPD), ADHD/ODD Comorbidity and a group of neonatal conditions.

For each episode, a work group was established to analyze Arkansas-specific data, create quality metrics and exclusion criteria, determine risk adjustment and define outliers. The work groups relied on published clinical literature as well as collaboration from Arkansas providers, payers and consumer stakeholders.

Assigning Accountability

Arkansas’s approach relies on identifying a Principal Accountable Provider (PAP).  In most cases, the PAP is determined based on triggering events (e.g. childbirth or cholecystectomy).  For a chronic illness, such as ADHD, the PAP is assigned retrospectively based on provision of the majority of services over the duration of the episode. One PAP is assigned for each episode. Typically, the PAP is the main decision-maker for most of the care delivered within the episode, has the most ability to influence and improve care for the patient and has the greatest opportunity to receive financial incentives when care is effectively and efficiently managed.

The PAP is accountable for all care delivered within an episode over a predetermined time period. Participation is not voluntary; if a provider bills for a triggering service, they are included in the episode profiling process. Each PAP must meet a minimum caseload per episode (which varies by episode) to qualify for the opportunity to receive gain or risk sharing.

Financial Incentives

All episodes generated from a PAP’s caseload during a performance period contribute to their performance profile.  At the conclusion of a performance period (typically 12 months), an average cost for all episode-related care is generated for each PAP with specified exclusions and adjustments applied for comorbidities.

Initially, payers participating in the initiative proposed a single up-front payment for each episode. This met with significant provider resistance due to concerns about financial disruption and a lack of system integration. Following discussions among payers, providers and health system leaders, Arkansas moved to a model based on setting a standard spending target for each episode, providing fee-for-service payments throughout the episode and reconciliation based on the target and actual spending after the episode has ended. This approach maintains the existing claims submission process but adds financial incentives for quality and cost savings available to a PAP for an episode. This reflects the challenge that, in a state lacking tightly integrated delivery systems, with both provider shortages and a low-density, largely rural population, fully integrated global capitated systems of care are years away (if not impossible).

Risk sharing payments – which may either be additional payments (gain sharing) to providers or required recoupment (penalties) from providers – are based on a retrospective review of spending for qualifying clinical episodes throughout a predetermined performance period (generally 12 months).  To set payment targets, each work group reviewed, discussed and scrutinized Arkansas-specific data to form an appropriate understanding of variation in spending across providers and benchmark against Arkansas-specific experiences. Within each episode, the work group determined thresholds of “acceptable,” “commendable” and “gain sharing limit” expenditures (separately, for each payer to adhere to anti-trust regulations). These thresholds are communicated to providers prior to the start of an episode performance period.  Episode PAPs with average costs above the acceptable threshold are assessed penalties.  Episode PAPs with average costs between acceptable and commendable do not receive gain sharing or a penalty. Episode PAPs with average costs below the commendable threshold are eligible for gain sharing only if they perform well enough on quality measures.

Even though the episode model is based on payments for specific acute events or conditions, reducing heterogeneity relative to a global payment model, case-mix differences could still be an issue.  For this reason, each episode also includes exclusion criteria and payments are risk adjusted based upon modeling of Arkansas data during episode development.  Together, these approaches reconciled provider concerns about fairness with payer needs for accountability.  At the close of an episode performance period, these adjustments are applied retrospectively and a PAP’s average risk adjusted episode cost is calculated. We have found physicians to be receptive to this framework, in which an improved practice pattern will often result in gain sharing.

The Pregnancy Example

Arkansas Medicaid covers 60 percent of all births in the state. In this high volume episode, the PAP is the delivering physician (or group) for a woman under care for at least the last eight weeks of a pregnancy. An episode extends from 40 weeks prior to delivery to 60 days post delivery.  All related services provided by any provider within the episode attribute to the PAP.  The PAP receives quarterly performance reports with quality metrics, clinical information and cost reports for their episodes.  If the PAP has more than 5 patients who meet the criteria for an episode (the minimum caseload for pregnancy), the PAP will face downside risk at the end of the performance period and, if quality requirements are met, they will be eligible for gain sharing.  Thus, the PAP has aligned incentives from multiple payers to improve prenatal care, avoid unnecessary tests, minimize elective C-sections and optimize post-partum care.

Figure 1













Figure 1 provides an example of Arkansas Medicaid’s thresholds for the perinatal episode.

Clear variations in spending for the perinatal episode were evident before the model was introduced, as displayed in Figure 1, with the highest cost perinatal provider on the left and lowest cost for the same set of services on the right.

This example depicts a perinatal PAP’s average episode cost across all of their completed Medicaid perinatal episodes during the 12 month performance period as $3,000. The Arkansas Medicaid program currently shares 50 percent of the gain or risk share opportunity with the PAP. So, for this particular PAP, $197 per episode, or 50 percent of the difference between $3,394 and $3,000, will be shared. Say the PAP oversaw 200 completed perinatal episodes during the performance period, then, in this scenario, ($394*50 percent)*200 episodes would result in a $39,400 incentive payment to the provider in addition to previously paid claims.

Quality Assurance

A series of quality metrics have been developed through a stakeholder process that identifies and defines each episode. In some cases, thresholds for metrics must be met to qualify for gain sharing.  In other cases, metrics must only be tracked and are not tied to gain sharing.  Table 1 (below) lists the current metrics included in the perinatal episode.

Table 1


Provider Data Support

Three months prior to the start of a performance period, each PAP receives an episode summary report of the previous year’s experience – consistent across all payers. These reports give potential PAPs the opportunity to review their historical quality, cost and utilization data relative to their peer group in the context of the gain and risk sharing incentive thresholds for the active episode. Arkansas PAPs access their quarterly reports through an online portal system. At one stop, a provider can download and view their reports from both public and private payers. These reports allow providers to determine whether they need to focus attention on certain quality metrics and to determine whether utilization statistics are comparable to peer providers. This permits providers to determine if a select patient’s episode summary warrants further exploration to understand what is driving their cost or utilization metrics outside of the predetermined acceptable ranges.

Payer Flexibility and Alignment

With the exception of payment and threshold rates (which vary a bit by payer), most methodology surrounding episode design and implementation is aligned across participating payers. This alignment leads to consistent expectations among providers and motivates them to thoughtfully consider how to reach the best patient outcomes for each acute care episode.  Without multi-payer alignment on key elements of the model, providers would feel overburdened by a changing system individually implemented by each payer, thereby lessening the opportunity for successful delivery system transformation. With alignment and associated financial incentives, PAPs have new and consistent opportunities and reinforcements to engage with fellow providers, modify practice patterns, coordinate care, identify and avoid adverse events and deliver higher quality, more efficiently achieved outcomes.