Implementing Episode-Based Payment for Perinatal Care in the Medicaid and Commercial Markets

The Arkansas Health Care Payment Improvement Initiative (AHCPII) presents an opportunity to explore the impact of bundled payment in a multi-payer setting. Between 2012 and 2021, Arkansas ran the AHCPII – a statewide bundled payment program with mandatory provider participation. The AHCPII was implemented as a partnership of payers in the state, including Arkansas Medicaid and Blue Cross Blue Shield (AR BCBS). Under the AHCPII, the payers set spending targets for care episodes managed by Principal Accountable Providers (PAPs), who bore responsibility for the quality and cost of care. The AHCPII followed a "FFS with reconciliation" model, meaning providers in the program received fee-for-service payments throughout the episode but the PAP was responsible for a portion of any episode spending above the target. PAPs could also share in gains if total episode spending was below the target, as long as they hit certain quality measures. Arkansas’ program initially covered five episodes: perinatal care, attention-deficit hyperactivity disorder (ADHD), hip and knee replacement, congestive heart failure, and upper respiratory infection. 

Our team studied the effects of bundled payment on perinatal care in the commercial market under the AHCPII. The partnership between Medicaid and commercial insurers in Arkansas covered roughly 70% of births in the state, limiting providers’ ability to opt in or out of different payment systems. We compared outcomes in Arkansas between 2010 and 2014 to outcomes in four neighboring states that did not implement bundled payment programs—Alabama, Kentucky, Louisiana, and Oklahoma. Our findings include:

Average commercial spending per perinatal episode declined by 3.8% ($396 per episode) in the first year that the bundled payment program was fully implemented (2014), compared to neighboring states.

Reductions in hospital facility-based spending accounted for over 80% of the overall episode savings. Specifically, intrapartum facility spending—which covers hospital services provided during labor and delivery—decreased by 6.6% ($332 per episode). The “FFS with reconciliation” design encouraged physicians to reduce spending on “indirect” services, which they influenced but were not directly compensated for with FFS payments (e.g., hospital services). There were weaker incentives to reduce spending on direct physician services, for which physicians continued to receive fee-for-service reimbursement. 

Lower spending on facility-based care was due to lower prices for those services, likely due to a shift in physician referrals that favored less expensive facilities. However, we were unable to directly test this mechanism.   

Other than changes in hospital choice, our research found few changes in care patterns in response to the new payment model. Most utilization and quality measures – including C-section rates and how long patients were in the hospital – remained stable. A notable exception, however, was large increases in chlamydia screening rates, which were tied to gainsharing under the AHCPII. Baseline screening rates were below recommended clinical guidelines prior to bundled payment implementation, giving physicians a direct financial incentive to increase testing in an area linked to improved perinatal outcomes.

The findings from Arkansas lend support to two reforms. First, commercial insurers should consider expanding episode-based bundled payment programs for perinatal care. The initiative in Arkansas shows the effectiveness of bundled payment for defined, hospital-centered episodes like perinatal care where there is a natural “accountable” provider and episode volume is predictable. 

Second, policymakers should pursue multi-payer initiatives with commercial insurers and Medicaid for perinatal care. Multi-payer agreements are crucial to expanding bundled payment to conditions that have low Medicare shares or where the market shares of individual payers are limited. Without such agreements, bundled payment by individual payers may be insufficient to move physician behavior, or cause physicians to favor other payers that continue to use FFS reimbursement. Even in cases where individual payers have sufficient market share to sidestep these concerns, they may be reluctant to incur the costs and uncertainty of bundled payment implementation, knowing that the benefits of lower spending may be shared with competitors through spillover effects. The federal government could incentivize state Medicaid agencies to adopt bundled payment approaches through experimental models at the Center for Medicare and Medicaid Innovation or by approving waivers during the regular plan approval process. The AHCPII was supported by a three-year State Innovation Model Testing grant from the Centers for Medicare and Medicaid Services (CMS).

For future bundled payment models, payers and policymakers should take note of key design choices:

Level of market competition. Because the savings under the AHCPII depended on physician agency over referrals, the model is most likely to generate savings in markets with hospital competition and meaningful price variation among them (Interrante et al. 2026). Other states with more labor and delivery units per capita could see even higher savings than Arkansas due to greater local competition. 

Level of reliance on FFS reimbursement. Arkansas used a FFS with reconciliation approach, which incentivized accountable physicians to reduce unnecessary inputs and lower costs. However, the strongest incentives were to reduce services delivered by other providers, where physicians did not receive direct FFS payment. This contrasts with prospective bundled payment, which offers one upfront payment for all providers involved in the episode to share.

Mandatory vs. voluntary participation. Arkansas’ program required doctors to participate, which may have contributed to its positive outcomes. Mandatory participation addresses provider selection, since voluntary programs attract self-selected early adopters, whose results may not reflect what the average provider would achieve (Einav et al. 2022). Although smaller, choice-based programs can also produce savings, those initiatives may not be replicable at scale.

Quality measurement. Sponsors of bundled payment programs should prioritize investment in quality measurement alongside cost tracking to ensure that spending reductions are not coming at the expense of patient well-being. Current evidence on the quality of care under episode-based payment remains limited. Most bundled payment programs—including the AHCPII—use quality measures focused on care processes rather than actual patient health outcomes. Additional attention should be paid to how meaningful health measures can be tied to gainsharing under bundled payment programs, without imposing undue risk on providers. 

Financial sustainability of the gainsharing structure. The AHCPII ran from 2012 to 2021, and its wind-down is instructive. Arkansas Medicaid discontinued its bundled payment program after gainsharing payments to providers grew larger than the program could sustain (Arkansas Department of Human Services [ADHS] 2020). Even so, state officials highlighted successful elements of the program, namely stabilizing episode costs and quality—separate from challenges with the design of gainsharing (ADHS 2020). AR BCBS, which phased out its own episode program at the end of 2018 as part of a shift toward other value-based payment programs, also faced high gainsharing payments (ACHI 2019). For perinatal episodes in the AHCPII's first full year of implementation, AR BCBS paid roughly $462,000 in gainsharing dividends to 92 accountable providers while collecting about $10,000 in risk-sharing penalties (Carroll et al. 2018). Gainsharing structure has also been a challenge in Medicare’s bundled payment programs, where bonus payments have eroded program savings (Ryan et al. 2026). Setting gainsharing thresholds and bonuses is particularly challenging for commercial bundled payment programs, where providers may more easily opt out of programs that are not financially advantageous. Nevertheless, payers should consider periodically re-basing spending thresholds so that the shared savings payouts do not outstrip realized savings. 

Considerations from similar state programs. Similar to the AHCPII, Ohio and Tennessee launched episode-based payment programs in 2014 that included perinatal care (MACPAC 2019; Dickson et al. 2022). Participation is mandatory for Medicaid plans and is voluntary for commercial payers, with commercial payers in both states opting to implement episode-based payment models (Dickson et al. 2022). Evidence on the spending effects of the Ohio and Tennessee programs is descriptive, rather than causal, and limited to the Medicaid setting. TennCare reported a 9.2% decrease in perinatal spending relative to projections in 2018 (MACPAC 2021). In the Ohio Medicaid program, perinatal episode costs increased modestly in the initial years of bundled payment (Moody 2018), but there is no evidence on how these costs compare to projections or counterfactual spending – i.e., what spending would have been without bundled payment. Both states reported no declines in care quality from their programs; a federal evaluation compared Tennessee and Ohio with nearby states, finding that Tennessee’s model was associated with fewer cesarean sections and Ohio’s model was associated with improved screening for group B streptococcus (RTI International 2021).

AHCPII reduced average commercial spending per perinatal episode by 3.8% ($396 per episode) in the first full year of implementation, compared to neighboring states. Based on research from MarketScan claims data, we estimate that commercial spending on perinatal care totaled about $37 billion in 2023, with labor and delivery hospital care comprising the majority of those costs.  That amount reflects an average commercial episode cost of approximately $20,000 and roughly 1.8 million births (51% of all births) (Truven Health Analytics 2013; Valencia et al. 2022; Duffy et al. 2024; Osterman et al. 2025; Winger et al. 2025). Medicaid spending on childbirth care totaled roughly $15 billion in that same year, assuming an average cost of $10,000 per birth across 1.5 million births (42% of all births) (Valencia et al. 2022; Osterman et al. 2025).

Given the 3.8% savings from Arkansas’ episode-based bundled payment approach for perinatal care, broad application of such a model among commercial payers could save $1.4 billion in private health care spending (in 2023 dollars). That figure represents 0.03% of total national health spending and 0.1% of private health care spending in 2023. Additional savings could accrue to participating state Medicaid programs, and there could be gains from improved quality of care beyond direct financial savings.

Separate from episode-based bundled payment reforms, FFS reimbursement for perinatal care is shifting. Effective January 2027, the American Medical Association and the American College of Obstetricians and Gynecologists will restructure the CPT codes that govern physician reimbursement for perinatal care. Instead of using global obstetrical fees to cover professional services delivered throughout the perinatal period, there will be disaggregated codes for antepartum care, labor management, delivery care, and postpartum services. Because this reform changes the underlying CPT code set, it reaches all payers. Implementation of this reform became more complicated when CMS proposed to adopt the disaggregated codes in 2027 but also solicited comments on creating an additional set of G-codes that would reproduce the current global fee structure. While the G-codes would be designed for the Medicare Physician Fee Schedule, CMS’ decision could have broader effects if other insurers follow Medicare’s lead. 

Because of the increase in distinct billable services, there is concern that the new maternity code structure will incentivize an inefficient increase in care provision (Andrews 2026). Under the new coding system, therefore, bundling services through episode-based bundled payment could have potentially larger effects. Notably, episode-based payment is broader than the global fee, which covers only professional services and does not include hospital services or diagnostic services such as ultrasounds and lab work. Arkansas layered its episode-based bundled payment program on the existing global fee—and the savings there came almost entirely from facility prices, which the CPT coding changes leave untouched.