How Value-Based Care With Provider Enablement Improves Maternal and Infant Outcomes in Medicaid

ABSTRACT

Objectives: Medicaid holds significant responsibility for improving maternal and infant health in the US. Utilizing value-based care (VBC) that offers additional support to providers is one strategy by which the Medicaid system can improve these outcomes. In this analysis, we examined a Medicaid managed care plan’s incentive-only VBC program, which is supported by a provider enablement team to assist care providers in meeting program goals.

Study Design: Cross-sectional analysis of deliveries occurring between July 2020 and June 2022 from Elevance Health–affiliated Medicaid managed care plans operating in 16 states.

Methods: This study primarily relied on medical claims data to compare maternal, infant, and cost outcomes in Medicaid members with a care provider participating in a supported VBC program vs those with a care provider not participating in supported VBC. A propensity-balanced multivariable regression model was used to estimate the impact of participation vs nonparticipation in supported VBC on delivery, cost, and quality outcomes.

Results: Members with a care provider supported in the VBC program had significantly lower neonatal intensive care unit (NICU) lengths of stay, preterm births, and low birth weights; significantly better timeliness and adequacy of prenatal care rates; and significantly lower birth costs, NICU costs, and maternal and infant costs in the first year after birth.

Conclusions: These results provide insight into how payers and care providers can partner to improve maternal and infant outcomes among Medicaid members and subsequently experience cost savings.

Am J Manag Care. 2025;31(12):In Press

Takeaway Points

In this study, we investigated how implementing a payer-led supported value-based care (VBC) program can lead to improved maternal and infant outcomes and reduced costs among members enrolled in Medicaid managed care. We examined outcomes among members with a care provider participating in supported VBC compared with members whose care provider did not participate. The former group had the following outcomes:

• Infants had similar neonatal intensive care unit (NICU) admission rates but spent 2 fewer days there.
• Nearly 10% more women received timely prenatal care.
• Mean birth costs decreased by $450, and NICU costs decreased by $472.

Maternal mortality and morbidity in the US are worsening, with significant disparities in outcomes for non-White people, older expecting mothers, and those on Medicaid, who face higher rates of low birth weight (LBW) and preterm births.1-3 Medicaid plays an important role because approximately 2 of every 3 adult women enrolled in Medicaid are of reproductive age, and Medicaid currently finances approximately 41% of all births in the US.4 Improving maternal health among those enrolled in Medicaid is a priority for government agencies, health care organizations, care providers, and advocates. Using value-based care (VBC) is one way the health care system can improve maternal and infant outcomes.5

VBC models implemented by Medicaid programs for obstetric (OB) care providers are not traditionally designed to change how maternity care is delivered during the prenatal, labor, and postpartum periods. In states with Medicaid managed care, it is often the responsibility of the managed care plans to meet those quality metrics. Plans can support VBC-attributed OB care providers to further improve the impacts on maternal and neonatal outcomes by providing additional resources via provider enablement to OB care providers. This combination of VBC plus additional customized support (ie, supported VBC) can bolster the providers’ ability to meet their goals to improve outcomes, close disparities, and reduce costs.

Our objective was to assess the impact of a payer-led, supported VBC initiative among Medicaid OB care providers in 16 states with an Elevance Health–affiliated Medicaid managed care plan. We included only care providers partnered with provider enablement support staff within their state. We compared maternal and infant outcomes and costs among members attributed to an OB care provider participating in supported VBC with those of members who were attributed to an OB care provider not participating in supported VBC.

METHODS

Program Description

Elevance Health, one of the largest health insurance providers in the US and with a substantial presence in the Medicaid sector, created the Obstetric Quality Incentive Program (OBQIP), which is an incentive-only VBC program that rewards OB care providers for improving the quality of care among their members with Medicaid. Under the OBQIP, OB care providers can receive a bonus payment if they meet a predetermined set of performance measures tied to the program’s goals of improving outcomes for both mothers and infants. Under the model, provider supports include provider enablement staff known as OB practice consultants (OBPCs). OBPCs are experienced maternal health clinicians, most of whom are nurses with previous OB experience, who serve as clinical liaisons to support and enable OB care providers through bidirectional communication, providing evidence-based resources and sharing timely and relevant data to inform clinical decision-making.

Design

Our primary analysis included the OBQIP care providers who worked closely with and were supported by OBPCs (ie, participating in supported VBC) and therefore examined how the combination of provider incentives and support could improve quality, cost, and outcomes. This study used a cross-sectional design to compare outcomes for deliveries occurring between July 2020 and June 2022 among providers contracted under the OBQIP and receiving OBPC support (the intervention group) with those for deliveries occurring among similar providers not contracted under the OBQIP and who also did not have OBPC support (the control group).

Data

Our study primarily relied on medical claims data from Elevance Health–affiliated Medicaid managed care plans operating in 16 states, covering regions from the Southeast to the Midwest and parts of the Northeast and Southwest, reflecting varied demographics and health care needs. Data systems contained established, deidentified linkages between mothers and babies derived from managed care plan identifiers to link the claims for mothers to their infants. Deliveries were attributed to a provider with at least 1 prenatal and delivery claim filed under the same participating OB provider tax identification number. Deliveries were identified as OBQIP+OBPC when the attributed provider participated in the OBQIP and had an OBPC consultation at least 1 month but no longer than 12 months prior to the delivery date (the intervention group). Control group deliveries occurred when the attributed provider did not participate in the OBQIP and had never had OBPC support. We excluded deliveries that did not meet these criteria in accordance with the OBQIP program description.

Historical provider-level measures were based on deliveries from 2019. Member-level covariates were calculated using data from a 5-year prepregnancy period. Data collection was for quality improvement purposes under health care operations, exempting it from institutional review board review.

Outcome Variables

Cost and quality outcomes were compiled at the delivery level, and subsequently, where appropriate, proportions were calculated to monitor population-level trends. Table 1 provides how these outcomes were defined.

To ensure the quality of the claims data, we excluded members for whom maternal claims could not be linked to claims for an infant and members whose delivery costs were in the bottom 1%, amounting to below $2420 (see eAppendix A for exclusions [eAppendices available at ajmc.com]). To address cost outliers, we capped delivery and neonatal intensive care unit (NICU) costs at the 99th percentile ($61,396 for deliveries and $231,589 for NICU stays). To control for the impact of substantial unit cost variation across states and facilities, we calculated our cost outcomes on median unit costs across the managed care plan’s Medicaid book of business. The median unit cost was calculated by delivery type (vaginal vs cesarean delivery) and by NICU days. These unit costs were then combined with utilization metrics at the member level to derive an estimate of delivery costs unconfounded by regional variations in prices and state reimbursement policy.

Statistical Analysis

We used a propensity-balanced multivariable regression model to estimate the impact of OBQIP+OBPC participation on delivery cost and quality outcomes. To account for underlying differences in risk factors between intervention and control groups, we created inverse probability weights based on the predicted probability of exposure to OBQIP+OBPC. These probabilities were estimated using a random forest predictive model and included a battery of maternal risk factors (eAppendix B) and providers’ historical practice patterns; inputs into the random forest model were selected based on an imbalance of unweighted characteristics between the intervention and control groups (eAppendix C). We assessed the weighted balance between the OBQIP+OBPC and non–OBQIP+OBPC deliveries using standardized mean differences (SMDs) less than 10% as a threshold for successful weighting. All analyses were conducted using SAS Enterprise Guide 7.1 (SAS Institute Inc). Intervention and control group balance was assessed on a variety of covariates, including maternal age, race, state, multiple gestation, chronic conditions, predicted risk of cesarean delivery based on predelivery characteristics, modeled probability of NICU admission based on predelivery characteristics, prior pregnancy outcomes (eg, prematurity, LBW, cesarean delivery), benefit structure differences, COVID-19 treatment, and participation in other known maternity case management programs. Recognizing that there is variation among OB practices in the overall risk level of their populations, we also controlled for providers’ historical rates of cesarean delivery and NICU admissions. Because the program’s introduction was staggered across markets, we ensured that both intervention and control groups were drawn from the same market and that deliveries happened during the same year and month to accommodate any variations in the duration of an OB care provider and OBPC working together.

Any covariates that remained unbalanced after applying the inverse probability weights were included in our final weighted outcome models; each model also included state, maternal age, and race. For each outcome, we measured the risk-adjusted difference between OBQIP+OBPC and non–OBQIP+OBPC. Using the GENMOD procedure in SAS, generalized linear models were estimated specifying a normal distribution and identity link function for continuous outcomes, and logistic regression models were estimated specifying a binomial distribution and logit link function for binary outcomes. All models were propensity score weighted.

To ensure that the combined effects of OBQIP+OBPC were not driven by one component alone, we also performed a separate supplemental analysis to measure the impact of OBQIP alone and OBPC alone, each compared with non–OBQIP+OBPC (eAppendix D).

RESULTS

Our study included 19,913 members who had a delivery with an OBQIP provider supported by an OBPC (intervention) during the study period (resulting in 20,154 infants) and an additional 103,374 members as controls (105,231 infants). Table 2 shows the maternal and provider characteristics of the intervention and control groups before and after inverse probability weighting, along with their SMDs.

Overall, the weighting improved the balance between the intervention and control groups as demonstrated by the decreases in the SMD across all characteristics after weighting. Even after weighting, several provider characteristics remained different between intervention and control groups; OB practices performing deliveries in the intervention group tended to perform more deliveries annually (991 vs 894), have a higher baseline cesarean delivery rate (36.4% vs 33.1%), and have a lower baseline NICU admission rate (8.8% vs 10.1%). These factors were included as covariates in the subsequent outcomes analysis.

The modeled quality measure outcomes and sample size (denominator) for each outcome among the intervention and control groups are provided in Table 3. Overall, the intervention group performed significantly better in 5 of the 9 quality outcomes. The intervention group had significantly lower NICU length of stay (9.95 days vs 12.22 days; P < .0001), preterm birth rates (7.03% vs 7.90%; P < .0001), and LBW rates (7.13% vs 8.57%; P < .0001). The intervention group also experienced significantly better timeliness of prenatal care (76.89% vs 66.98%; P < .0001) and adequacy of prenatal care (59.78% vs 57.02%; P = .0006).

The control group had significantly higher vaginal birth after cesarean delivery (VBAC) rates (2.40% vs 2.16%; P = .0057) and significantly lower cesarean delivery rates (30.47% vs 32.88%; P < .0001). There were no significant differences in NICU admission rates or primary cesarean delivery rates.

The modeled cost outcomes for the intervention and control groups are provided in Table 4. The intervention group experienced significantly lower costs across all categories. Mean birth costs per delivery were $449.57 less per birth in the intervention group than in the control group (P < .0001). Mean NICU costs among NICU admissions per delivery were $472.22 less per birth in the intervention group than in the control group (P < .0001). Mean maternal costs per delivery in the year after birth were $99.85 lower in the intervention group than in the control group (P = .0114), and mean infant costs per delivery in the first year after birth were $458.24 lower in the intervention group (P < .0001).

When we considered the impacts of OBQIP and OBPC administered separately, we found that OBQIP alone was associated with $213.82 savings and OBPC alone with $237.74 savings per delivery (see eAppendix D for additional results for OBQIP alone and OBPC alone).

DISCUSSION

This study is the first to investigate a multistate, Medicaid managed care OB VBC program supported by provider enablement. Previous research on VBC’s effectiveness for improving maternal and infant outcomes while reducing costs has had mixed results and was often limited to single-state programs.6

Our study results show that the quality of care for women with a care provider participating in OBQIP supported by a dedicated OBPC was significantly better than that for women who did not have a care provider participating in supported VBC. Supplemental analysis confirmed that when OBQIP and OBPC are administered together, they result in more favorable impacts than when either is provided alone, suggesting that their impacts were additive and improved outcomes along different pathways.

Including both prenatal care timeliness and adequacy as VBC metrics is important because both late onset of care and fewer visits can significantly increase risks associated with premature birth or LBW, such as smoking continuation during pregnancy, premature rupture of the membranes, and other issues.7 Study findings also indicate that 80% of pregnancy-related deaths are preventable, with quality prenatal care being a factor.8,9

Our study demonstrates significantly lower preterm birth and LBW rates when women had a care provider participating in OBQIP+OBPC. Preterm births have been increasing in recent years, with the highest rates among infants with non-Hispanic Black mothers and among births covered by Medicaid compared with private insurance.10,11 This trend continues with LBW rates, in which births covered by Medicaid have higher LBW rates compared with births covered by private insurance.12 This study is therefore timely in demonstrating that supported VBC programs that include specialty provider enablement resources for maternity care providers can improve infant outcomes in the Medicaid population.

The intervention group spent significantly less time in the NICU by almost 2.5 days, which is unsurprising considering the intervention group’s lower rates of preterm births and LBWs. Conversely, there were no significant differences between the intervention and control groups in NICU admission rates. Further research is warranted to understand the national increase in NICU utilization and how supported VBC could influence NICU utilization.

Our findings show that when monitoring population-level trends, the intervention group had significantly lower primary (first-time) cesarean delivery and VBAC rates. Interestingly, overall cesarean delivery rates were significantly higher in the intervention group. The resultant cesarean delivery and VBAC rates could be associated with 2 programmatic characteristics. First, overall cesarean delivery rates were not an OBQIP measure because the program focused on decreasing primary cesarean deliveries. Second, the program focused on care providers who are lower quality performers as a strategy to help improve performance and quality; therefore, cesarean delivery rates may already be elevated among these care providers compared with non–OBQIP+OBPC care providers.

As previously mentioned, the combination of OBQIP+OBPC was more cost-effective than administering either program individually. Costs were significantly lower across all studied categories in the intervention group compared with the control group; this was likely influenced by the findings demonstrating better prenatal, delivery, and infant outcomes and shorter NICU stays among the intervention group. Although our analysis did not look at mother and infant costs beyond a year after birth, other studies have established that there are substantial long-term costs associated with infants born preterm or with LBW, including costs to not only the health care system but also to the education and social services sectors.12 Additionally, O’Neil et al calculated the medical and nonmedical costs associated with maternal morbidity, finding that health complications that result from pregnancy and delivery cost the US billions of dollars a year and have multiyear impacts on the health of both the mother and infant.13 Therefore, it is possible that there could be accumulative savings over time within our intervention group.

Limitations

There are several limitations to note. Although we adjusted for observable differences between the intervention and control groups, unobserved factors might still introduce selection bias. The cross-sectional design prevents us from inferring causality, although we controlled for key influences on birth outcomes and costs. Additionally, because our data relied on Medicaid managed care claims, we may have missed prenatal care initiated before joining a Medicaid plan, potentially undermeasuring prenatal care adequacy. Postpartum care data quality varied by state due to different billing practices, limiting our ability to assess program impact on postpartum care. Lastly, our analysis only included births attributed to a single provider, excluding cases in which care providers changed during pregnancy, which may influence outcomes.

Future research should integrate fee-for-service and managed care claims to comprehensively assess care utilization and include postbirth outcomes such as postpartum visit rates and depression screenings.

CONCLUSIONS

Improving outcomes among mothers and infants through VBC is a critical step toward addressing disparities in maternal and prenatal care across populations. This study demonstrates that partnering with and supporting OB care providers with clinically trained provider enablement staff, such as OBPCs, can transform care delivery. These OBPCs, embedded within local Medicaid managed care plans, help overcome challenges faced by OB care providers and promote the use of evidence-based care.

Our analysis, although focused on a 2-year frame of the OBQIP+OBPC program, builds on several years of positive results in terms of maternal, infant, delivery, and cost metrics. Importantly, combining OBQIP with OBPCs enhances outcomes beyond what each could achieve individually. This success underscores the potential to expand this supported VBC model to private insurers and other clinical areas such as oncology and cardiology. These findings suggest that payers, policy makers, and health care programs should model VBC that empowers care providers, thereby achieving the triple aim of improved patient care, health outcomes, and cost efficiency. 

Author Affiliations: Elevance Health (BH, LR), Washington, DC; Carelon, Richmond, VA (KA, SV), and Indianapolis, IN (BW, TI).

Source of Funding: None.

Author Disclosures: Dr Howell and Ms Ramirez are employed by Elevance Health and hold stock grants and options. Ms Austin, Ms Varner, Ms Wynn, and Dr Inglis are employed by Carelon, which is owned by Elevance Health, and own company stock in Elevance Health.

Authorship Information: Concept and design (BH, KA, SV, TI); acquisition of data (BH, KA, SV, TI); analysis and interpretation of data (BH, LR, KA, SV, TI); drafting of the manuscript (LR, KA, BW, TI); critical revision of the manuscript for important intellectual content (BH, LR, BW, TI); statistical analysis (KA, SV); administrative, technical, or logistic support (BH, LR); and supervision (BH, BW, TI).

Address Correspondence to: Leah Ramirez, MS, Elevance Health, 1001 Pennsylvania Ave NW, Ste 710, Washington, DC 20004. Email: leah.ramirez@elevancehealth.com.

REFERENCES

1. Hoyert DL. Maternal mortality rates in the United States, 2021. National Center for Health Statistics. March 16, 2023. Accessed September 15, 2023. https://www.cdc.gov/nchs/data/hestat/maternal-mortality/2021/maternal-mortality-rates-2021.htm

2. Shanahan KH, Burdick KJ, Lee LK; Pediatric Policy Council. Reducing infant mortality: maternal health is infant health. Pediatr Res. 2022;92(3):623-625. doi:10.1038/s41390-022-02142-4

3. Anum EA, Retchin SM, Strauss JF III. Medicaid and preterm birth and low birth weight: the last two decades. J Womens Health (Larchmt). 2010;19(3):443-451. doi:10.1089/jwh.2009.1602

4. Births financed by Medicaid. KFF. Accessed September 15, 2023. https://www.kff.org/medicaid/state-indicator/births-financed-by-medicaid

5. Pierce-Wrobel C, Green K. To help fix the maternal health crisis, look to value-based payment. Health Affairs Forefront. July 16, 2019. Accessed September 15, 2023. https://www.healthaffairs.org/content/forefront/help-fix-maternal-health-crisis-look-value-based-payment

6. Value-based payment for maternity care in Medicaid: findings from five states. Medicaid and CHIP Access Commission. September 2021. Accessed September 15, 2023. https://www.macpac.gov/wp-content/uploads/2021/09/Value-Based-Payment-for-Maternity-Care-in-Medicaid-Findings-from-Five-States.pdf

7. Yan J. The effects of prenatal care utilization on maternal health and health behaviors. Health Econ. 2017;26(8):1001-1018. doi:10.1002/hec.3380

8. Pregnancy-related deaths: data from maternal mortality review committees in 36 US states, 2017-2019. CDC. Accessed September 27, 2023. https://www.cdc.gov/maternal-mortality/php/data-research/mmrc-2017-2019.html

9. Chang J, Elam-Evans LD, Berg CJ, et al. Pregnancy-related mortality surveillance--United States, 1991-1999. MMWR Surveill Summ. 2003;52(2):1-8.

10. Martin JA, Hamilton BE, Osterman MJK. Births in the United States, 2021. National Center for Health Statistics data brief 442. August 29, 2022. Accessed September 26, 2023. https://dx.doi.org/10.15620/cdc:119632

11. Medicaid and CHIP Beneficiary Profile: Maternal and Infant Health. CMS; December 2020. Accessed September 15, 2023. https://www.medicaid.gov/sites/default/files/2020-12/mih-beneficiary-profile_0.pdf

12. Petrou S, Sach T, Davidson L. The long-term costs of preterm birth and low birth weight: results of a systematic review. Child Care Health Dev. 2001;27(2):97-115. doi:10.1046/j.1365-2214.2001.00203.x

13. O’Neil S, Platt I, Vohra D, et al. The high costs of maternal morbidity show why we need greater investment in maternal health. The Commonwealth Fund. November 12, 2021. Accessed October 1, 2023. https://doi.org/10.26099/nz8s-4708