Care management was effective at reducing infant mortality among the most vulnerable infants enrolled in an Ohio Medicaid managed care organization.
Objectives: In 2012, the Ohio Department of Medicaid introduced requirements for enhanced care management to be delivered by Medicaid managed care organizations (MCOs). This study evaluated the impact of care management on reducing infant mortality in the largest Medicaid MCO in Ohio.
Study Design: Observational study using infant and maternal individual-level enrollment and claims data (2009-2015), which used a quasi-experimental research design built on a sibling-comparison approach that controls for within-family confounders.
Methods: Using individual-level data from the largest MCO in Ohio, we estimated linear probability models to examine the effect of infant engagement in care management on infant mortality. We used a within-family fixed-effects research design to determine if care management reduced infant mortality and estimated models separately for healthy infants and nonhealthy infants.
Results: Infant engagement in care management was associated with a reduction of 7.4 percentage points (95% CI, —10.7 to –4.1; P <.001) in infant mortality among the most vulnerable infants, those identified as not well at birth. This effect was larger in recent years and likely driven by new statewide enhanced care management requirements. Infant mortality was unchanged for healthy infants engaged in care management (coefficient = 0.03; 95% CI, —0.01 to 0.08).
Conclusions: This study provides evidence that care management can be effective in reducing infant mortality among Medicaid MCO enrollees, a population at high risk of mortality. Few infants were engaged in care management, suggesting to policy makers that there is room for many additional infants to benefit from this intervention.
Am J Manag Care. 2020;26(3):127-131. https://doi.org/10.37765/ajmc.2020.42637
Despite a 15% decline in the US infant mortality rate (IMR)—from 6.86 deaths per 1000 live births in 2004 to 5.82 in 2014—rates remain high for many vulnerable subpopulations.1 One particularly vulnerable group is the Medicaid population, whose IMR is twice that of the non-Medicaid population. One study reported that the IMR of Indiana newborns enrolled in Medicaid was 7.16 per 1000 live births compared with 4.19 in the non-Medicaid population.2 Elevated IMR among Medicaid enrollees is of particular importance because more than 40% of births in the United States in 2016 were covered by Medicaid.3 Moreover, nationally, 70% of Medicaid participants are enrollees in managed care organizations (MCOs),4 an alternative payment model in which there are financial incentives to reduce costs. Thus, any successful intervention focused on Medicaid MCO enrollees has the potential to avoid a large number of infant deaths and serve as a model for both the fee-for-service and non-Medicaid populations.5
Many state Medicaid programs have implemented care management requirements for MCOs or designed care management interventions to improve birth outcomes. Although care management has no strict definition, it is loosely described as a patient-centered, team-based approach to coordinating medical care with an emphasis on chronic conditions.6 Prior studies suggest that care management for pregnant women may lead to fewer preterm births,7 fewer low-birth-weight births,8,9 and reduced infant mortality.10 In 2012, the Ohio Department of Medicaid introduced requirements for enhanced care management to be delivered by Medicaid MCOs, including a minimum staff/member ratio, quarterly face-to-face contact with members, and multidisciplinary teams to care for high-risk members.11 For high-risk infants in the intensive care unit (ICU), the teams included nurses and social workers and were designed to manage unique infant needs (eg, connection to specialists). Although care management guidelines were not specific to infants, these changes provide a unique opportunity to conduct a case study examining the potential for care management to reduce infant mortality. As such, we conducted a quasi-experimental study to determine if enhanced care management reduced infant mortality among Medicaid managed care enrollees in an Ohio MCO.
We obtained individual-level enrollment and claims data from CareSource, a large Ohio Medicaid MCO that covers 60% of Ohio Medicaid enrollees. We combined these data with county characteristics from the Area Health Resources File from 2009 to 2015. In our preferred model, we studied 61,560 infants born from 2009 to 2015, enrolled in CareSource at birth, who had at least 1 sibling enrolled since birth in the MCO. Infant mortality within the first year of life was measured as a dichotomous variable. Infants were linked to mothers and siblings using a maternal identifier. Care management was offered to infants predicted to be high cost using a proprietary algorithm or who were identified as born from high-risk pregnancies. Infant engagement in care management was defined as a dichotomous variable indicating any engagement beginning in the first month of life.
We used linear probability models to examine the relationship between infant engagement in care management and infant mortality. We first used a cross-sectional approach, estimating models controlling for infant, maternal, and community characteristics reported in the Table [part A and part B], including infant sex, race/ethnicity, maternal age at birth, a multiple birth indicator, number of primary care providers in the county, and unemployment rate. Indicators of maternal health conditions were constructed from claims data using Expanded Diagnosis Clusters to identify mothers who ever had a substance use disorder or serious mental health issue. These models also controlled for time-invariant county-level characteristics (eg, rural/urban county status) and birth-year-cohort invariant characteristics (eg, recessions) through the inclusion of county and birth-year fixed effects.
We also used a within-family approach, which adds controls for family-invariant characteristics (eg, genetic predisposition or parental education) through the addition of maternal fixed effects. The within-family model is our preferred design because it minimizes the potential for selection bias to have an impact on our treatment effect estimates. The simple difference between the mortality rate of those infants engaged in care management and those not engaged cannot be considered the true effect of care management because engagement was not randomly assigned. To overcome concerns related to nonrandom assignment, we used a quasi-experimental sibling-comparison research design and examined the subpopulation of maternal siblings who have the same health status at birth. For a detailed explanation of the sibling-comparison research design, see the work of D’Onofrio et al, who used this approach to remove confounding factors when estimating the effect of preterm birth on infant mortality and morbidity.12 Importantly, the sibling-comparison approach removes confounding variables and risk factors common to siblings, such as the material genetic environment and any time-invariant maternal characteristics.
Because this approach cannot remove confounders that are unique to any single infant, there still exists the possibility that the nonrandom assignment of care management could lead to biased estimates if those infants engaged in care management are healthier than those not engaged. If this were the case, then the estimated effect would confound differences in underlying health with the true treatment effect. However, this is of limited concern for our analysis because infants engaged in care management (n = 2524) were less healthy at birth (17.94% considered well at birth) than those not engaged in care management (n = 161,193, with 78.00% considered well at birth). To mitigate this concern, we reestimated our preferred model on the subpopulation of those infants considered not well at birth. This specification compares siblings who were both designated as not well at birth by the MCO. Well at birth is a designation constructed by the MCO using a proprietary algorithm that identifies those infants considered a healthy newborn (eg, absence of serious diagnoses or an explicit diagnosis of normal newborn), who are not admitted to a nongeneral room (eg, ICU), who are not transferred to another facility, and who stay in the hospital for less time than their mother and less than 7 days.
Finally, we examined our specification of interest excluding time periods to assess if the effect of care management is driven by the period after 2012, when Ohio implemented regulations for enhanced care management.
Analyses were conducted using Stata 14 (StataCorp; College Station, Texas). Significance tests were 2-sided, with P <.05 indicating statistical significance. Standard errors were clustered at the birth county for all models. RAND’s institutional review board approved the study. The study used deidentified, observational data; thus, patient consent was neither required nor obtainable.
In our cross-sectional design, we examined 163,717 infants, of whom 77% were considered well at birth. We also estimated the treatment effect using the within-family design for those with siblings (n = 61,562), as well as the sample in which all siblings were well at birth (64.45% of all infants with siblings) and the sample in which all siblings were not well at birth (11.4% of all infants with siblings).
Both the cross-sectional and within-family research designs illustrated that enhanced care management was associated with reduced mortality among high-risk infants considered not well at birth (Figure). The cross-sectional results using the full study period (2009-2015) suggest that infant engagement in care management is associated with a significant reduction in mortality for all (coefficient = —2.0; 95% CI, –2.3 to –1.7; P <.001), which is driven by the effect on infants considered not well at birth (coefficient = —2.5; 95% CI, –3.0 to –2.1; P <.001).
Our preferred model uses the within-family design to ensure that our findings are not driven by unmeasured differences across families. Across the full study period, we found that infant engagement in care management was associated with a significant reduction in mortality for infants considered not well at birth (coefficient = —7.4; 95% CI, –10.7 to –4.1; P <.001) but not for infants considered well at birth. This is a large reduction in mortality, as the predicted (ie, regression-adjusted) mortality rate for this group in the absence of care management is 7.5%.
We also estimated the preferred model for different time periods, one before enhanced care management (2009-2011) and another after (2012-2015). We found that the significant effect of care management on infants not well at birth was driven by the time period after enhanced care management was introduced.
Nationally, most Medicaid recipients are enrolled in MCOs, a payment model that uses care management to reduce costs and promote health. Prior research has documented the benefits of care management for pregnant women,7-10,13,14 and our case study demonstrates that care management directed at high-risk infants can reduce infant mortality.
In our study population, few infants were engaged in care management, suggesting potential for many infants to benefit from it. Extending care management to more infants and strengthening existing care management activities by requiring minimum staffing ratios, in-person encounters, and specialized teams for high-risk members may also help reduce infant mortality. Although not the focus of this study, MCOs could also consider maternal health when making decisions about engaging infants in care management. A prior study of a similar population of infants in an Ohio Medicaid MCO found increased odds of maternal mortality among healthy-weight infants born to mothers with severe mental illness or substance use.15
The MCO changed a number of care management strategies over the time period we studied; thus, we cannot identify the exact components of care management that contributed to the mortality decline. However, through extensive conversations with the MCO’s leadership and staff, we hypothesize that a key driver of the decline may have been better management of specialist care for those infants admitted to the neonatal ICU. Due to enhanced care management, the MCO we studied began connecting infants to specialists, ensuring appropriate and timely specialist follow-up visits, assessing need for specialist care, and encouraging collaboration and communication between specialists and/or the primary care physician. Although we do not have data that allow us to evaluate this hypothesis, we believe that the results presented here are still of importance because they highlight the potential mortality reductions that could occur if care management was extended to more infants who are not well at birth and if existing care management activities were strengthened.
Furthermore, because Medicaid-enrolled mothers and infants are highly vulnerable and may struggle with food insecurity, lack of stable housing, and other challenges related to poverty and discrimination, research is needed to determine the extent to which case management services may reduce infant mortality by addressing medical and/or nonmedical needs. Although this care management program was not specifically focused on addressing nonmedical social determinants of health, a systematic review of activities undertaken by Medicaid MCOs to address nonmedical needs found several studies focused on pregnant women and infants but little information about key characteristics of care management programs.16 Future studies should seek to identify the specific characteristics of care management programs that are most effective at reducing infant mortality.
This study has several limitations. First, it is not a nationwide analysis but a case study examining the effects of 1 large MCO in a single state. It is possible that the results here are not generalizable to other settings. Second, participation in care management was not randomly assigned. Although we used a quasi-experimental design and sought to minimize bias, it is still possible that selection into care management may be biasing our estimates. Third, the exact features of care management driving the large mortality reductions are unknown. Although we speculate that the primary driver is the connection to specialist care, this may not be the case. The effects could be driven by another feature of care management or by a combination of efforts—and this remains an important area for future research. Finally, the MCO’s use of proprietary software to help identify infants to be offered care management may make it challenging for other organizations to replicate this intervention. We encourage organizations to consider use of open source algorithms or seek approval to share details about proprietary algorithms to facilitate the adoption of effective interventions.
Care management reduced infant mortality among the sickest infants enrolled in a Medicaid MCO in Ohio. This finding is of importance because an increasing number of births in the United States are covered by Medicaid MCOs and currently just a small number of infants are enrolled in care management. State regulations should follow the best available evidence and incentivize care management participation for those most likely to benefit. Moreover, this result is of general interest to the entire population, as best practices in care management can be applied to those births not covered by an MCO, helping to reduce the US IMR.
The authors thank Coady Wing, PhD (School of Public and Environmental Affairs, Indiana University, Bloomington, IN), for helpful comments on the manuscript. Dr Wing was not compensated for his assistance. They also thank Angela Snyder (CareSource) for her technical assistance with the data and extract creation.Author Affiliations: O’Neill School of Public and Environmental Affairs, Indiana University (AJH), Bloomington, IN; RAND (AJH, AMK), Arlington, VA; Claremont Graduate University (DF), Claremont, CA.
Source of Funding: This work was supported by CareSource, a Medicaid managed care organization headquartered in Dayton, OH. CareSource worked with the authors to develop the research question. CareSource provided the data, constructed variables under the guidance of the authors, and provided technical assistance. CareSource had no role in the study design or conduct of the study. CareSource was not directly involved in the analysis of the data. CareSource did not assist in the preparation of the manuscript or in the decision to submit the manuscript for publication.
Author Disclosures: Drs Hollingsworth, Kranz, and Freund received contract funding through CareSource to examine factors related to infant mortality in its member population.
Authorship Information: Concept and design (AJH, AMK, DF); acquisition of data (AJH, AMK, DF); analysis and interpretation of data (AJH, AMK, DF); drafting of the manuscript (AJH, AMK, DF); critical revision of the manuscript for important intellectual content (AJH, AMK, DF); statistical analysis (AJH); obtaining funding (AMK, DF); administrative, technical, or logistic support (AJH); and supervision (AJH).
Address Correspondence to: Alex J. Hollingsworth, PhD, School of Public and Environmental Affairs, Indiana University, 1315 E 10th St, Bloomington, IN 47405. Email: firstname.lastname@example.org.REFERENCES
1. Mathews TJ, Driscoll AK. Trends in infant mortality in the United States, 2005-2014. NCHS Data Brief. 2017;(279):1-8.
2. Reducing infant mortality in Indiana. State of Indiana website. in.gov/isdh/files/Infant_Mortality_Report.pdf. Published December 2014. Accessed December 20, 2018.
3. Births financed by Medicaid. Kaiser Family Foundation website. kff.org/medicaid/state-indicator/births-financed-by-medicaid/?currentTimeframe=0&sortModel=%7B"colId":"Location","sort":"asc"%7D. Accessed January 18, 2019.
4. Medicaid managed care penetration rates by eligibility group. Kaiser Family Foundation website. kff.org/medicaid/state-indicator/managed-care-penetration-rates-by-eligibility-group/?currentTimeframe=0&sortModel=%7B"colId":"Location%20","sort":"asc"%7D. Accessed January 18, 2019.
5. Martin JA, Hamilton BE, Osterman MJK, Driscoll AK, Drake P. Births: final data for 2016. Natl Vital Stat Rep. 2018;67(1):1-55.
6. Farrell TW, Tomoaia-Cotisel A, Scammon DL, Day J, Day RL, Magill MK. Care management: implications for medical practice, health policy, and health services research. Agency for Healthcare Research and Quality website. ahrq.gov/sites/default/files/publications/files/caremgmt-brief.pdf. Published April 2015. Accessed January 15, 2019.
7. Roman L, Raffo JE, Zhu Q, Meghea CI. A statewide Medicaid enhanced prenatal care program: impact on birth outcomes. JAMA Pediatr. 2014;168(3):220-227. doi: 10.1001/jamapediatrics.2013.4347.
8. Hillemeier MM, Domino ME, Wells R, et al. Effects of maternity care coordination on pregnancy outcomes: propensity-weighted analyses. Matern Child Health J. 2015;19(1):121-127. doi: 10.1007/s10995-014-1502-3.
9. Kroll-Desrosiers AR, Crawford SL, Moore Simas TA, Rosen AK, Mattocks KM. Improving pregnancy outcomes through maternity care coordination: a systematic review. Womens Health Issues. 2016;26(1):87-99. doi: 10.1016/j.whi.2015.10.003.
10. Meghea CI, You Z, Raffo J, Leach RE, Roman LA. Statewide Medicaid enhanced prenatal care programs and infant mortality. Pediatrics. 2015;136(2):334-342. doi: 10.1542/peds.2015-0479.
11. Managed care agreements (January 1, 2013-June 20, 2013; March 1, 2013-June 30, 2014; July 1, 2014-June 30, 2015). Ohio Department of Medicaid website. medicaid.ohio.gov/Managed-Care/For-Managed-Care-Plans. Accessed August 20, 2019.
12. D’Onofrio BM, Class QA, Rickert ME, Larsson H, Långström N, Lichtenstein P. Preterm birth and mortality and morbidity: a population-based quasi-experimental study. JAMA Psychiatry. 2013;70(11):1231-1240. doi: 10.1001/jamapsychiatry.2013.2107.
13. Newman RB, Sullivan SA, Menard MK, et al. South Carolina Partners for Preterm Birth Prevention: a regional perinatal initiative for the reduction of premature birth in a Medicaid population. Am J Obstet Gynecol. 2008;199(4):393.e1-393.e8. doi: 10.1016/j.ajog.2008.07.047.
14. Piper JM, Mitchel EF Jr, Ray WA. Evaluation of a program for prenatal care case management. Fam Plann Perspect. 1996;28(2):65-68. doi: 10.1363/2806596.
15. White SE, Gladden RW. Maternal mental health and infant mortality for healthy-weight infants. Am J Manag Care. 2016;22(11):e389-e392.
16. Gottlieb LM, Garcia K, Wing H, Manchanda R. Clinical interventions addressing nonmedical health determinants in Medicaid managed care. Am J Manag Care. 2016;22(5):370-376.