Investing in CenteringPregnancy™ Group Prenatal Care Reduces Newborn Hospitalization Costs

Women's Health Issues 27-1 (2017) 60–66

www.whijournal.com

Maternal Health

Investing in CenteringPregnancyÔ Group Prenatal Care Reduces Newborn Hospitalization Costs Amy Crockett, MD, MSPH a,*, Emily C. Heberlein, PhD a, Leah Glasscock, MS b, Sarah Covington-Kolb, MSW, MSPH c, Karen Shea, MSN b, Imtiaz A. Khan, DO d a

Department of Obstetrics and Gynecology, Greenville Health System, Greenville, South Carolina Anthem, Inc., Indianapolis, Indiana Greenville Health System, Greenville, South Carolina d BlueChoice Health Plan South Carolina Medicaid, Columbia, South Carolina b c

Article history: Received 15 October 2015; Received in revised form 19 July 2016; Accepted 19 September 2016

a b s t r a c t Objectives: CenteringPregnancyÔ group prenatal care is an innovative model with promising evidence of reducing preterm birth. The outpatient costs of offering CenteringPregnancy pose barriers to model adoption. Enhanced provider reimbursement for group prenatal care may improve birth outcomes and generate newborn hospitalization cost savings for insurers. To investigate potential cost savings for investment in CenteringPregnancy, we evaluated the impact on newborn hospital admission costs of a pilot incentive project, where BlueChoice Health Plan South Carolina Medicaid managed care organization paid an obstetric practice offering CenteringPregnancy $175 for each patient who participated in at least five group prenatal care sessions. Methods: Using a one to many case-control matching without replacement, each CenteringPregnancy participant was matched retrospectively on propensity score, age, race, and clinical risk factors with five individual care participants. We estimated the odds of newborn hospital admission type (neonatal intensive care unit [NICU] or well-baby admission) for matched CenteringPregnancy and individual care cohorts with four or more visits using multivariate logistic regression. Cost savings were calculated using mean costs per admission type at the delivery hospital. Results: Of the CenteringPregnancy newborns, 3.5% had a NICU admission compared with 12.0% of individual care newborns (p < .001). Investing in CenteringPregnancy for 85 patients ($14,875) led to an estimated net savings for the managed care organization of $67,293 in NICU costs. Conclusions: CenteringPregnancy may reduce costs through fewer NICU admissions. Enhanced reimbursement from payers to obstetric practices supporting CenteringPregnancy sustainability may improve birth outcomes and reduce associated NICU costs. Ó 2016 Jacobs Institute of Women's Health. Published by Elsevier Inc.

With the passage of the Affordable Care Act and mounting evidence of deficiencies in quality and efficiency in the costly United States’ health care system, managed care organizations (MCOs) and physician practices face increasing pressure to reduce costs and improve the quality of health care (Centers for Medicare and Medicaid Services, n.d.a; Grol & Grimshaw, 2003; Funding Statement: This study received no external funding and the authors have no financial conflicts of interest to disclose. * Correspondence to: Amy Crockett, MD, MSPH, Department of Obstetrics and Gynecology, Greenville Health System, 890 W. Faris Rd, Suite 470, Greenville, SC 29605. Phone: 864-455-5032. E-mail address: [email protected] (A. Crockett).

Institute of Medicine, 2001, 2013). The traditional feefor-service model incentivizes providers to increase the volume of services without addressing quality and outcomes; documented quality concerns include lack of adherence to guidelines and slow adoption of evidence-based practices, overuse and underuse of services, high rates of preventable medical errors, and disparities in service utilization and outcomes (Agency for Healthcare Research and Quality, 2013; Institute of Medicine 2001, 2002, 2013; McGlynn et al., 2003). Innovative payment models designed to improve clinical processes, patient outcomes and experiences, and structural aspects of care are increasingly used by public and private payers to address both quality and

1049-3867/$ - see front matter Ó 2016 Jacobs Institute of Women's Health. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.whi.2016.09.009

A. Crockett et al. / Women's Health Issues 27-1 (2017) 60–66

costs in the managed care environment, although evidence of their effectiveness is mixed (Felt-Lisk, Gimm, & Peterson, 2007; Institute of Medicine, 2013; James, 2012; Powell et al., 2012). In the United States in 2012, 11.5% of births occurred at less than 37 weeks’ gestation and 8% of babies were born at low birthweight (Martin, Hamilton, Osterman, Curtin, & Matthews, 2013). These high rates of adverse birth outcomes negatively impact immediate and long-term health and development outcomes of infants, and represent substantial healthcare costs to payers (Behrman & Butler, 2007; Petrou, Eddama, & Mangham, 2011). Inpatient costs associated with preterm births exceed $6 billion per year, representing one-half of all costs associated with infant births (Russell et al., 2007). Increasing access to traditional prenatal care services has not resulted in reductions in preterm birth and low birth weight, calling into question payers’ substantial financial investment of approximately $10 billion per year in prenatal care services (Childbirth Connection, Catalyst for Payment Reform, & Center for Healthcare Quality and Payment Reform, 2013; Curtin, Osterman, Uddin, Sutton, & Reede, 2013). The causes of adverse birth outcomes include complex and multifaceted social, psychological, and behavioral risk factors, which can be difficult to address in the brief medically focused traditional individual care model (Behrman & Butler, 2007; Lu, Tache, Alexander, Kotelchuck, & Halfon, 2003). Redesigning prenatal care is a potential strategy for improving quality and reducing costs associated with negative birth outcomes (Krans & Davis, 2012; Lu et al., 2003; Walford, Trinh, Wiencrot, & Lu, 2011). The CenteringPregnancy model of group prenatal care is a patient-centered model bundling the medical assessment of traditional prenatal care with comprehensive prenatal health education, consultation, and peer support facilitated by a credentialed health care provider within an empowering group environment (Rising, Kennedy, & Klima, 2004). The Centering Healthcare Institute (Boston, Massachusetts) maintains the curriculum, conducts training, and evaluates and certifies sites offering this trademarked model of group prenatal care (Rising, 1998; Rising et al., 2004). One randomized control trial and several cohort studies have demonstrated substantial reductions in the rates of preterm birth for women participating in CenteringPregnancy (Ickovics et al., 2007; Ickovics et al., 2003; Picklesimer, Billings, Hale, Blackhurst, & Covington-Kolb, 2012) and the Agency for Healthcare Research and Quality has identified CenteringPregnancy as a service delivery innovation with strong evidence (Agency for Healthcare Research and Quality, 2014). The CenteringPregnancy model of care requires substantial investment by provider practices in initial and ongoing training, redesign of clinic space, lost provider productivity, office support staff time, patient education materials, and annual certification fees (Picklesimer, Heberlein, & Covington-Kolb, 2015). Although practices bill for routine prenatal care for each participant in group care, and grant funding (most notably by the March of Dimes and Centers for Medicaid and Medicare Services’ Strong Start Program) has covered startup costs for some practices, few payers have developed enhanced reimbursement policies to cover the increased ongoing costs of providing CenteringPregnancyÔ. Increased costs to the practice may be small in comparison with the potential savings accrued to payers and the overall health care system, yet they remain a significant barrier to implementation and sustainability. Innovative incentive strategies and enhanced reimbursement to encourage practice adoption and patient retention in CenteringPregnancyÔ and the sustainability of CenteringPregnancy at the practice level are potentially a sound

61

investment for payers to reduce inpatient costs associated with adverse birth outcomes. Because the Medicaid program is the largest payer of maternity benefits in the United States, policies to improve access and quality of health care services provided to women and newborns in this time period are particularly relevant to public payers (Markus & Rosenbaum, 2010). Few peer-reviewed research studies of CenteringPregnancy have incorporated cost analyses at the practice or third-party payer levels. At the practice level, two studies suggest the financial impact of implementing group prenatal care may be neutral or positive if large group sizes are maintained (e.g., 10.6 enrollees per group), delivery volume is high, and groups are facilitated by certified nurse midwives or nurse practitioners instead of physicians (Mooney, Russell, Prairie, Savage, & Weeks, 2008; Rowley et al., 2015). An analysis of basic insurance billing data suggests that CenteringPregnancy participants do not incur additional prenatal care or delivery costs to insurers compared with individual care participants (Ickovics et al., 2007). Potential cost savings in newborn hospitalizations in neonatal intensive care units (NICUs) are considerable (Picklesimer et al., 2015) but have not been evaluated rigorously from the payer perspective. Published cost studies incorporating payer incentive strategies to support provider adoption and patient retention in CenteringPregnancy are also lacking. To investigate potential cost savings for payers when they invest in CenteringPregnancy patient retention and program sustainability, we evaluated the impact of a pilot between a Medicaid MCO and a large prenatal care clinic with an existing certified CenteringPregnancy program on NICU admission rates and costs. Methods Pilot Incentive Project This pilot project was part of a larger initiative instituted by the South Carolina Department of Health and Human Services, state Medicaid agency to provide enhanced reimbursement for CenteringPregnancy encounters. Beginning in July 2012, Medicaid provided $40 per patient per CenteringPregnancy session (up to $200 for five sessions) to the South Carolina Medicaid MCOs, of which $30 per patient per session (up to $150) was to be passed on to the provider. This reimbursement was contingent upon the provider being certified or under contract for certification as a CenteringPregnancy practice with the Centering Healthcare Institute. In this pilot project, the BlueChoice Health Plan South Carolina Medicaid MCO paid the prenatal care practice an additional incentive payment of $175 for women who participated in at least five CenteringPregnancy sessions. Providers could use $25 of the incentive to market CenteringPregnancy to patients. Participation in five sessions was identified as the minimum acceptable dose of CenteringPregnancy; other research has found a dose–response relationship between CenteringPregnancy participation and gestational age and birth weight (Ickovics et al., 2007). The MCO incentive payment amount was derived from additional costs incurred by the practice to provide CenteringPregnancy, including supplies, administrative time, lost provider productivity, ongoing training, and program certification. Interventions Individual prenatal care In the United States, individual care for women with uncomplicated pregnancies involves monthly provider visits for the

62

A. Crockett et al. / Women's Health Issues 27-1 (2017) 60–66

first 28 weeks of pregnancy, every 2 to 3 weeks until 36 weeks, then weekly (usually 10–15 visits). Visits include an initial medical and psychosocial history, ongoing medical assessment, and patient education on pregnancy and prenatal care, options for delivery care and educational programs, breastfeeding, and pediatrician selection. Women receive routine screenings as well as specialized tests, interventions, and referrals depending on risk factors and the course of pregnancy (American Academy of Pediatrics & American College of Obstetricians and Gynecologists (2012)). Prenatal care visits are usually short (10–15 minutes) and focused on identifying medical risks, with limited opportunity for counseling and support (Novick, 2009). CenteringPregnancy In the CenteringPregnancy model, groups of 8 to 12 pregnant women due within the same month attend up to 10 sessions. Each session begins with a brief individual medical assessment followed by a 90-minute facilitated group discussion. CenteringPregnancy curriculum topics include pregnancy, labor, and delivery; nutrition; stress management; infant care and breastfeeding; and healthy relationships. Facilitators are trained to adapt the content of each session to address women’s needs and questions, and the open discussion format promotes information sharing among women, social support, and the involvement of significant others (Rising et al., 2004). At the study site, facilitators were nurse practitioners or certified nurse midwives. Women had additional individual care visits if needed to adhere to the recommended individual care visit schedule. Sample Selection All CenteringPregnancy patients in this pilot study began their prenatal care between March and August 2013 at the same high-volume hospital-affiliated obstetrics clinic in an urban setting in the southeast. The clinic has offered CenteringPregnancy since 2008 and has received annual certification from the Centering Healthcare Institute, indicating high-quality and consistent provision of the CenteringPregnancy model. To be eligible for CenteringPregnancy at the clinic, patients must have begun prenatal care before approximately 20 weeks’ gestation; medical exclusions included pregestational diabetes or hypertension, multiple gestation, and a body mass index (calculated as weight [kg]/[height (m)]2) of greater than 45 kg/ m2. Approximately 75% to 80% of the clinic population is typically eligible for CenteringPregnancy, with 9% ineligible for late entry to care, 5% for body mass index exceeding 45 kg/m2, and 2% each excluded for pregestational diabetes, hypertension, or multiple gestation. All eligible patients were offered CenteringPregnancy for their prenatal care, with approximately 30% opting for CenteringPregnancy instead of individual prenatal care. Three to four CenteringPregnancy groups were available each month for every new patient who opted for group care. The clinic provided the MCO with the patient list, number of CenteringPregnancy sessions attended, newborn hospitalization type (i.e., well-baby or NICU admission), and identifiers to link maternal and infant records. The individual prenatal care comparison group criteria were 1) known pregnancy as of March 2013, 2) known delivery date, 3) completed risk assessment screening, 4) not enrolled in the MCO’s obstetrics case management program, 5) received at least four individual prenatal care visits, and 6) restricted to known ethnicity in the enrollment data of White, African American or Hispanic. The pregnancy risk assessment score, a proprietary

algorithm developed by Anthem, Inc., is derived from a logistic regression model incorporating factors associated with adverse pregnancy outcomes, including a history of preterm labor, preterm birth, low birth weight, diabetes, sexually transmitted disease in the prior year, and current high blood pressure. The structure of the BlueChoice Health Plan South Carolina Medicaid data systems impacted the geographical location and use of prenatal care criteria available for identifying the comparison group. Because the MCO’s data system for Medicaid enrollees in South Carolina did not have the capacity to link maternal and infant records, the comparison group was drawn from Anthem’s Medicaid markets in 13 other states with this linking capacity. The use of bundled prenatal care codes only allows for identifying patients with three or fewer versus four or more prenatal care visits, rather than the five visits required for enhanced reimbursement. Because using billing codes was the most appropriate way to identify the frequency of individual prenatal care attendance, the analysis included individual care and CenteringPregnancy participants with four or more visits or groups. Institutional Review Board approval was not necessary because this study was an analysis of the MCO’s membership data. Propensity Scores and Matching Because randomizing patients to treatments was not possible in our retrospective study, we used the available characteristics of study participants to develop propensity scores (e.g., the odds of participating in the treatment), which were then used to create matched pairs of CenteringPregnancy and individual care participants. Propensity scores reduce observable differences and allow for a more rigorous comparison between two groups (Hale, Picklesimer, Billings, & Covington-Kolb, 2014; Rosenbaum & Rubin, 1983; Rubin, 2007). Propensity scores were created using logistic regression with age, race (African American/Hispanic or White) and components of the pregnancy risk score (as described) as predictor variables for participation in CenteringPregnancy, the outcome variable. The predicted probability of being a CenteringPregnancy patient represents the propensity score. CenteringPregnancy and individual care participants were then matched on propensity score, age, race and risk (e.g., CenteringPregnancy patients with high blood pressure and history of preterm labor were matched to individual care patients with the same risk profile) (CocaPerraillon, 2007; Rubin, 1980). Through this matching process, CenteringPregnancy program eligibility standards of excluding women with pregestational hypertension or diabetes were applied to the individual care cohort; we were not able to restrict the individual care cohort using the CenteringPregnancy body mass index of greater than 45 kg/m2 criteria. The four-visit minimum was used as a reasonable proxy for timing of entry to prenatal care. Each CenteringPregnancy observation was matched to five randomized individual care patients using a caliper of 0.0001, representing 0.025 of the SD of the mean propensity score without replacement, generating 5 samples of treated (CenteringPregnancy) and untreated (individual care) observations. Using calipers limits the range of possible matches to provide greater assurance of ideal matching and balanced cohorts (Hale et al., 2014). One-to-many matching, as opposed to 1:1 matching, increases the number of comparison group patients included in the analysis and the precision of treatment effect estimates when the study population size is small. All statistical analyses

A. Crockett et al. / Women's Health Issues 27-1 (2017) 60–66

were conducted using SAS Enterprise Guide 7.1 (SAS Institute Inc., 2014). Outcomes To identify newborn hospitalization admission type for the individual care cohort, maternal and infant records were linked

Pregnant women with valid risk score entering group care March-August 2013 N=107

63

using authorizations and claims data, and hospital deliveries resulting in well-baby admissions versus NICU admissions were identified using revenue codes (NICU: 0172 through 0175, Well baby: 0170, 0171, 0179). The newborn hospitalization admission types for CenteringPregnancy patients provided by the practice were verified with MCO claims data for final categorization as well-baby versus NICU admissions.

Pregnant women with valid risk score and Medicaid eligible as of March 2013 N = 22,064

Women enrolled in MCO obstetrics case management program N=1,343

Women with unknown delivery date or invalid baby match N=6,460

Women with unknown delivery result (NICU or well baby) N=2,659

Women with <4 Centering Sessions N=22

Women with <4 Individual Care Sessions N=1,282

Women with race other than White, Black or Hispanic N=1,532 85 group care cohort

Delivery resulƟng in NICU admission N=3

8,788 individual care cohort

Delivery resulƟng in well baby admission N=82

Delivery resulƟng in NICU admission N=1,176

Delivery resulƟng in well baby admission N=7,612

1:5 Caliper Matching on propensity scores, age, race and clinical risk score

Matched group care final cohort N=425

Delivery resulƟng in NICU admission N=15

Matched individual care final cohort N=425

Delivery resulƟng in well baby admission N=410

Delivery resulƟng in NICU admission N=51

Figure 1. Study population and exclusion criteria.

Delivery resulƟng in well baby admission N=374

64

A. Crockett et al. / Women's Health Issues 27-1 (2017) 60–66

Table 1 Age and Risk Scores of the Study Population before and after Propensity Score Matching Before Propensity Score Matching

n Age, mean  SD Risk score, mean  SD Race (White), mean  SD

Centering Pregnancy

Individual Care

85 24.5  4.35 22.57  3.33 0.45  0.50

8,788 25.5  5.70 20.28  7.72 0.44  0.50

After 1:5 Propensity Score Matching

p Value*

Standardized Difference y

.182 <.001 .955

0.20 0.38 0.01

CenteringPregnancy

Individual Care

p Value*

Standardized Difference y

425 24.50  4.33 22.57  3.31 0.45  0.50

425 24.50 4.4 22.57  3.33 0.44  0.50

.932 .966 .783

0.00 0.00 0.02

Abbreviation: SD, standard deviation. 2 * c test for independence. y Standardized difference ¼ difference in means or proportions divided by standard error, imbalance defined as absolute value greater than 0.20 (small effect size).

Cost savings were calculated by taking the difference between the resulting CenteringPregnancy NICU rate and the casematched individual care NICU rate to determine the avoided NICU admissions. Cost estimates for well-baby and NICU admissions were calculated using the MCO’s average cost per newborn and NICU admission from the CenteringPregnancy clinic’s delivery hospital in the first half of 2013. Because of the variability in contracted rates for hospital deliveries and NICUs across the MCO’s markets, costs specific to the delivery hospital were used to calculate final cost savings. MCO payments to the practice for CenteringPregnancy visits were deducted from savings to arrive at the net savings for the pilot. Statistical Analysis Differences in age, race, and clinical risk between the CenteringPregnancy and individual care cohorts were evaluated before and after matching using a c2 test for independence. For the resulting sampled and propensity-matched dataset without replacement, age, race, risk score, and treatment assignment (CenteringPregnancy or individual care) were used in multivariate logistic regression models, with the dependent variable as the delivery result, well baby versus NICU. Receiver operating characteristic curves were compared across models in predicting a NICU admission to determine the model with the best fit. The distribution of the sampled CenteringPregnancy and individual care patient populations were graphed, and a two-tailed t test was used to evaluate the difference in mean NICU rate for the sampled populations. Results During the study period, 85 CenteringPregnancy patients participated in four or more sessions, all of whom had complete clinical data used to calculate pregnancy risk scores and verified newborn records. Eighty-two of these patients (96%) participated in five or more visits, for which the practice received the incentive payment. After initial exclusions for missing data,

Table 2 Unadjusted NICU Admission Rates on Matched Sample after Sampling through 1:5 Matching on Propensity Score, Age Risk, and Race

n (samples/observations) Mean NICU admission rate* 95% confidence interval

CenteringPregnancy

Individual Care

5/425 3.5% 1.22%–5.84%

5/425 12.0% 9.83%–14.17%

* Difference in mean neonatal intensive care unit (NICU) admission rate statistically significant at p < .05 with 2-tailed t test.

restrictions to race/ethnicity groupings in the treatment group, MCO obstetrics case management participation, and having fewer than four prenatal care visits, 8,788 individual care patients were identified for propensity score analysis (Figure 1). Before propensity score matching, CenteringPregnancy participants had significantly higher clinical risk scores (Table 1). After matching of the individual care cohort to treatment group using 1:5 propensity score matching with a caliper of 0.0001, the CenteringPregnancy and individual care groups showed no differences in age, race, or clinical risk score (age, p ¼ .932; risk, p ¼ .966; race p ¼ .783). Standardized differences for age, race, and risk had an absolute value less than 0.20, indicating a balance between the two populations (Yang & Dalton, 2012). NICU admission rates on the matched cohort after sampling indicated a significant difference based on the two-tailed t test (p < .001), with 3.5% of CenteringPregnancy newborns having a NICU admission compared with 12.0% of individual care newborns (Table 2). Multivariate logistic regression analyses, including clinical risk score, race, age, and prenatal care type, indicated that participation in CenteringPregnancy may be an effective intervention for preventing a NICU delivery result. The odds of a CenteringPregnancy participant’s newborn having a NICU admission were approximately 27% of the odds of an individual care participant’s newborn having a NICU admission (Table 3; odds ratio, 0.27; 95% confidence interval, 0.15–0.49). To calculate MCO cost savings resulting from averted NICU admissions, the proportion of CenteringPregnancy NICU admissions was subtracted from the proportion of individual care NICU admissions to derive the projected number of averted NICU admissions (Table 4). The mean costs of NICU admissions and wellbaby admissions from the MCO’s claims data for the delivery hospital from the first half of 2013 and the actual MCO costs of the CenteringPregnancy payment incentive were used in the cost calculations. Cost savings reflect the number of estimated averted NICU admissions, multiplied by the average NICU admission cost, deducting CenteringPregnancy incentive and well-baby admission costs. Investing in CenteringPregnancy for 85 patients ($14,875) yielded a net savings for the MCO of $67,293 in NICU costs; when actual costs of the three NICU admissions for CenteringPregnancy were used instead of mean costs, the net savings were $80,586 (not shown). Discussion This is the first study to examine the savings from CenteringPregnancy to payers in NICU costs when an MCO enhanced reimbursement payment is used to defray the additional practice costs and to reward the practice for retaining women in

A. Crockett et al. / Women's Health Issues 27-1 (2017) 60–66 Table 3 Multivariate Logistic Regression for Delivery Resulting in Neonatal Intensive Care Unit Admissions Parameter

Estimate

Pr > c2

Intercept Age Risk score Race (White) Centering (yes)

2.22 0.001 0.01 0.071 1.31

0.0616 0.9717 0.7874 0.7933 <0.0001

Odds Ratio (95% Confidence Interval) 1.001 1.01 0.933 0.269

(0.944–1.061) (0.938–1.089) (0.558–1.562) (0.148–0.486)

CenteringPregnancy. We provide evidence that investing in CenteringPregnancy can produce better outcomes and lower costs in NICU admissions. Our study shifts the perspective of implementing this evidence-based practice from that of an outpatient obstetric practice to the broader view of managed care and policymakers, particularly Medicaid. The potential for selection bias and the small sample size of this CenteringPregnancy pilot are the primary limitations of our study. Although limitations in the MCO dataset prevented a direct matching using the five-visit threshold from the performance incentive, the NICU admission rates were very similar for patients with four versus five Centering visits (3.5% vs. 3.7% NICU admission rate), indicating a minor impact on results. We did not estimate the impact of CenteringPregnancy when women receive fewer than four visits. The retrospective nature of the study and use of secondary MCO data restricted our ability to incorporate additional sociodemographic or other factors associated with program selection (e.g., patient activation). Matching on clinical risk factors known to impact preterm birth reduced selection bias (Goldenberg, Culhane, Iams, & Romero, 2008); using a one-to-many matching increased the precision of the treatment effect estimates with our small sample. The cost impact of our pilot study is limited to cost-savings resulting from predicted decreased NICU admissions at the delivery hospital for the pilot project. Our study aim was to quantify payer savings in one critical outcome compared with the cost of paying provider incentives for implementing an evidence-based model of care; future cost-effectiveness research should involve

Table 4 Estimated MCO Cost Savings for NICU Admissions for CenteringPregnancy Cohort Item Estimated averted NICU admissions for CenteringPregnancy cohort (N ¼ 85) 1. Baseline NICU admission rate for individual care 2. Expected number of NICU admissions (85 * row 1) 3. Actual number of NICU admissions 4. Number of averted NICU admissions (row 2  row 1) NICU and well-baby admission mean cost per delivery (at delivery hospital) 5. Mean cost of NICU admission 6. Mean cost of well-baby admission Cost-savings calculation 7. Expected NICU admission costs (row 2 * row 5) 8. Actual cost of NICU admissions (row 3 * row 5) 9. Cost savings from averted NICU admissions (row 7  row 8) 10. Normal delivery costs for averted NICU admissions (row 4 * row 6) 11. Total cost of CenteringPregnancy incentives Cost savings (row 9  row 10  row 11)

Value

12.0% 10.2 3 7.2

$12,403 $990 $126,507 $37,208 $89,299 $7,131 $14,875 $67,293

Abbreviations: MCO, managed care organization; NICU, neonatal intensive care unit.

65

a full accounting of all health care costs during pregnancy, labor and delivery, and the postpartum period for both mothers and newborns. Another study limitation is our lack of statistical control for other sociodemographic factors associated with preterm birth and NICU admission rates in our analyses. Because South Carolina has higher rates of preterm birth and low birth weight than the United States’ rates (Martin, Hamilton, Ventura, Osterman, & Matthews, 2013), comparing our results with a matched sample drawn from 13 other states may underestimate the cost savings when CenteringPregnancy is provided in settings with relatively high rates of adverse birth outcomes. Further, comparing two cohorts enrolled in Medicaid provides some control for socioeconomic status. Our study has several strengths. Propensity score matching reduced selection bias and increased the reliability of results with our small CenteringPregnancy pilot sample size. Use of a multidimensional algorithm to represent maternal clinical risk factors with documented associations with adverse birth outcomes allowed for a rigorous comparison between the cohorts. Comparisons with published data on NICU admissions indicates the individual care cohort rate of 12.0% is comparable with national estimates of NICU rates from the Pregnancy Risk Assessment and Monitoring System (11.9% in 2011, based on 25 states reporting this measure; Centers for Disease Control and Prevention, 2015). Although evidence supporting CenteringPregnancy’s effectiveness in reducing preterm birth is promising, the underlying mechanisms accounting for its impact remain poorly understood. Additional research assessing its effects on maternal health behaviors, stress, patient activation, and cost implications for practices and payers in a variety of practice settings and reimbursement structures is needed to build a greater understanding of the range of positive outcomes and cost implications associated with CenteringPregnancy. Implications for Practice and/or Policy This study represents one opportunity for partnership between obstetric care providers and third-party payers in redesigning prenatal care, and is emblematic of the paradigm shift that is occurring in the broader health care field. The traditional, volume-driven, fee-for-service model of reimbursement is shifting toward a “value-driven” model in which patient outcomes are increasingly important. Investment in programs like CenteringPregnancy, which break down traditional silos of inpatient versus outpatient costs and obstetric versus pediatric outcomes, is a natural evolution in which third-party payers can help practices to deliver the highest quality of care by providing incentives to move to improved care models such as CenteringPregnancy. In 2012, the Centers for Medicare and Medicaid Services selected 15 different clinical sites across the United States to take part in a national demonstration project to evaluate the comparative effectiveness of group prenatal care (Centers for Medicare and Medicaid Services, n.d.b.). Undoubtedly, comprehensive data about the cost-effectiveness of group prenatal care models will be forthcoming from this large and diverse prospective cohort which is anticipated to include more than 20,000 women. The state of South Carolina is also evaluating its statewide Medicaid incentive program. If the model continues to demonstrate robust cost savings, the next challenge will be program replication in diverse practice environments and with a wide variety of payers.

66

A. Crockett et al. / Women's Health Issues 27-1 (2017) 60–66

References Agency for Healthcare Research and Quality. (2013). 2012 national healthcare disparities report. Rockville, MD: Author. Agency for Healthcare Research and Quality. (2014). Group visits focused on prenatal care and parenting improve birth outcomes and provider efficiency. Available: https://innovations.ahrq.gov/profiles/group-visitsfocused-prenatal-care-and-parenting-improve-birth-outcomes-and-provider. Accessed: May 7, 2015. American Academy of Pediatrics & American College of Obstetricians and Gynecologists. (2012). Guidelines for perinatal care (7th ed.) Elk Grove Village, IL: American Academy of Pediatrics. Behrman, R. E., & Butler, A. S. (2007). Preterm birth: Causes, consequences, and prevention. Washington, DC: National Academies Press. Centers for Disease Control and Prevention. (2015). Pramstat. Available: www. cdc.gov/prams/pramstat/index.html. Accessed: May 8, 2015. Centers for Medicare and Medicaid Services. (n.d.a). Quality of care & delivery systems. Available: www.medicaid.gov/AffordableCareAct/Provisions/ Quality-of-Care-and-Delivery-Systems.html. Accessed: May 7, 2015. Centers for Medicare and Medicaid Services. (n.d.b). Strong start for mothers and newborns initiative: General information. Available: http://innovation.cms. gov/initiatives/strong-start/index.Html. Accessed: August 27, 2013. Childbirth Connection, Catalyst for Payment Reform, & Center for Healthcare Quality and Payment Reform. (2013). The cost of having a baby in the United States. Available: http://transform.childbirthconnection.org/reports/cost/. Accessed: May 8, 2015. Coca-Perraillon, M. (2007). Local and global optimal propensity score matching. SAS Global Forum, 185, 1–9. Curtin, S. C., Osterman, M. J. K., Uddin, S. F., Sutton, S. R., & Reede, P. R. (2013). Source of payment for the delivery: Births in a 33-state and District of Columbia reporting area, 2010. Natl. Vital Stat. Rep, Vol. 62. Hyattsville, MD: National Center for Health Statistics. Felt-Lisk, S., Gimm, G., & Peterson, S. (2007). Making pay-for-performance work in Medicaid. Health Affairs (Millwood), 26(4), w516–w527. Goldenberg, R. L., Culhane, J. F., Iams, J. D., & Romero, R. (2008). Epidemiology and causes of preterm birth. Lancet, 371(9606), 75–84. Grol, R., & Grimshaw, J. (2003). From best evidence to best practice: Effective implementation of change in patients’ care. Lancet, 362(9391), 1225–1230. Hale, N., Picklesimer, A. H., Billings, D. L., & Covington-Kolb, S. (2014). The impact of CenteringPregnancy group prenatal care on postpartum family planning. American Journal of Obstetrics and Gynecology, 210(1), 50.e1–50.e7. Ickovics, J. R., Kershaw, T. S., Westdahl, C., Magriples, U., Massey, Z., Reynolds, H., & Rising, S. S. (2007). Group prenatal care and perinatal outcomes: A randomized controlled trial. Obstetrics and Gynecology, 110(2), 330–339. Ickovics, J. R., Kershaw, T. S., Westdahl, C., Rising, S. S., Klima, C., Reynolds, H., . Magriples, U. (2003). Group prenatal care and preterm birth weight: Results from a matched cohort study at public clinics. Obstetrics and Gynecology, 102(5, Part 1), 1051–1057. Institute of Medicine. (2001). Crossing the quality chasm: A new health care for the 21st century. Washington, DC: National Academy of Sciences. Institute of Medicine. (2002). Unequal treatment: Confronting racial and ethnic disparities in health care. Washington, DC: National Academy of Sciences. Institute of Medicine. (2013). Best care at lower cost: The path to continuously learning health care in America. Washington, DC: National Academies Press. James, J. (2012). Health policy brief: Pay-for-performance. Health Affairs (Millwood). Available: http://www.healthaffairs.org/healthpolicybriefs/brief. php?brief_id=78. Accessed: May 8, 2015. Krans, E. E., & Davis, M. M. (2012). Preventing low birthweight: 25 years, prenatal risk, and the failure to reinvent prenatal care. American Journal of Obstetrics and Gynecology, 206(5), 398–403. Lu, M., Tache, V., Alexander, G., Kotelchuck, M., & Halfon, N. (2003). Preventing low birth weight: Is prenatal care the answer? Journal of Maternal, Fetal, and Neonatal Medicine, 13(6), 362–380. Markus, A. R., & Rosenbaum, S. (2010). The role of Medicaid in promoting access to high-quality, high-value maternity care. Womens Health Issues, 20(1), S67–S78. Martin, J. A., Hamilton, B. E., Osterman, M. J. K., Curtin, S. C., & Matthews, M. S. (2013). Births: Final data for 2012. National Vital Statistics Report, Vol. 62. Hyattsville, MD: National Center for Health Statistics. Martin, J. A., Hamilton, B. E., Ventura, S. J., Osterman, M. J. K., & Matthews, M. S. (2013). Births: Final data for 2011. National Vital Statistics Report, 62(1). McGlynn, E. A., Asch, S. M., Adams, J., Keesey, J., Hicks, J., DeCristofaro, A., & Kerr, E. A. (2003). The quality of health care delivered to adults in the United States. New England Journal of Medicine, 348(26), 2635–2645. Mooney, S. E., Russell, M. A., Prairie, B., Savage, C., & Weeks, W. B. (2008). Group prenatal care: An analysis of cost. Journal of Health Care Finance, 34(4), 31–41. Novick, G. (2009). Women’s experience of prenatal care: An integrative review. Journal of Midwifery and Womens Health, 54(3), 226–237.

Petrou, S., Eddama, O., & Mangham, L. (2011). A structured review of the recent literature on the economic consequences of preterm birth. Archives of Disease in Childhood Fetal and Neonatal Edition, 96(3), F225– F232. Picklesimer, A., Heberlein, E., & Covington-Kolb, S. (2015). Group prenatal care: Has its time come? Clinical Obstetrics and Gynecology, 58(2), 380–391. Picklesimer, A. H., Billings, D., Hale, N., Blackhurst, D., & Covington-Kolb, S. (2012). The effect of CenteringPregnancy group prenatal care on preterm birth in a low-income population. American Journal of Obstetrics and Gynecology, 206(5), 415.e1–415.e7. Powell, B. J., McMillen, J. C., Proctor, E. K., Carpenter, C. R., Griffey, R. T., Bunger, A. C., . York, J. L. (2012). A compilation of strategies for implementing clinical innovations in health and mental health. Medical Care Research and. Review, 69(2), 123–157. Rising, S. S. (1998). CenteringPregnancy: An interdisciplinary model of empowerment. Journal of Nurse Midwifery, 43(1), 46–54. Rising, S. S., Kennedy, H. P., & Klima, C. S. (2004). Redesigning prenatal care through CenteringPregnancy. Journal of Midwifery and Womens Health, 49(5), 398–404. Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41–55. Rowley, R. A., Phillips, L. E., O’Dell, L., El Husseini, R., Carpino, S., & Hartman, S. (2015). Group prenatal care: A financial perspective. Maternal and Child Health Journal, 1–10. Rubin, D. B. (1980). Bias reduction using mahalanobis-metric matching. Biometrics, 293–298. Rubin, D. B. (2007). The design versus the analysis of observational studies for causal effects: Parallels with the design of randomized trials. Statistics in Medicine, 26(1), 20–36. Russell, R. B., Green, N. S., Steiner, C. A., Meikle, S., Howse, J. L., Poschman, K., . Petrini, J. R. (2007). Cost of hospitalization for preterm and low birth weight infants in the United States. Pediatrics, 120(1), e1–e9. SAS Institute Inc. (2014). SAS enterprise guide 7.1. Cary, NC: Author. Walford, H. H., Trinh, S., Wiencrot, A., & Lu, M. C. (2011). What is the role of prenatal care in reducing racial and ethnic disparities in pregnancy outcomes? In Handler, A., Kennelly, J., & Peacock, N. (Eds.), Reducing racial/ethnic disparities in reproductive and perinatal outcomes (pp. 151–179) New York: Springer. Yang, D., & Dalton, J. (2012). A unified approach to measuring the effect size between two groups using SAS. SAS Global Forum, 335–2012, 1–6.

Author Descriptions Amy Crockett, MD, MSPH, is Associate Professor with the USC School of MedicineGreenville and the Medical Director of the Greenville Health Obstetrics Center, is a maternal-fetal medicine specialist. Dr. Crockett is the principal investigator for several studies on the CenteringPregnancy model.

Emily C. Heberlein, PhD, Consultant with Greenville Health System. Dr. Heberlein’s research interests involve evaluating program and policy strategies to improve the health of pregnant women and mothers. She was a post-doctoral research fellow with Clemson University when this publication was developed.

Leah Glasscock, MS, Anthem, Inc. Ms. Glasscock holds a Master’s Degree in Applied Statistics and is currently a Senior Cost of Care Consultant in Health Care Analytics of the Medicaid Business Unit, Anthem, Inc.

Sarah Covington-Kolb, MSW, MPH, is the CenteringPregnancy group prenatal care coordinator at the OB Center, Greenville Health System, Greenville, South Carolina. She also coordinates the consortium of practices implementing CenteringPregnancy in South Carolina, providing training, technical assistance, and implementation monitoring.

Karen Shea, MSN, Vice President, Maternal Child Services, Anthem, Inc., provides strategic leadership and operational oversite for enterprise maternal and child services and programs, including utilization management, case management, health promotion, reimbursement policy, contracting, quality, and advocacy.

Imtiaz A. Khan, DO, BlueChoice Health Plan South Carolina Medicaid, is a Family Medicine Physician and currently the Chief Medical Officer for Blue Choice Health Plan Medicaid.