Group prenatal care, most often the CenteringPregnancy model, has been evaluated in studies using experimental or quasi-experimental designs. The research discussed here meets our standards of evidence for being methodologically strong and allowing for causal inference, unless otherwise noted. Each strong causal study reviewed has been assigned a letter, and a complete list of causal studies can be found at the end of this review, along with more details about our standards of evidence and review method. The findings from each strong causal study reviewed align with one of our eight policy goals from Table 1. The Evidence of Effectiveness table below displays the findings associated with participation in group prenatal care (beneficial, null,i or detrimental) for each of the strong studies (A through P) in the causal studies reference list, as well as our conclusions about the overall impact on each studied policy goal. The assessment of the overall impact for each studied policy goal weighs the timing of publication and relative strength of each study, as well as the size and direction of all measured indicators.
Table 2: Evidence of Effectiveness for Group Prenatal Care
|Policy Goal||Indicator||Beneficial Impacts||Null Impacts||Detrimental Impacts||Overall Impact on Goal
|Access to Needed Services||Adequate Prenatal Care||G, J, L||Positive
|Healthy and Equitable Births||Preterm Birth||A, B, E, G||F, I, J, L, O||Mixed
|Low Birthweight||A, B, E||D, F, G, I, J, L, O, P||
|NICU Admissions||A, E||G, I, J, L||
|Parental Health and Emotional Wellbeing||Gestational Weight Gain||M, P||J||Mostly Positive
|Depressive Symptoms||C, H||J||
|Optimal Child Health and Development||Breastfeeding Initiation||G, N||I, J||Mixed
Access to Needed Services
The findings from three randomized control trials (RCTs) show that participation in GPNC improves the likelihood that pregnant women receive adequate prenatal care.G,J,L Adequacy of prenatal care is often determined using the Kotelchuck Index, which measures adequate prenatal care on two dimensions: month of pregnancy during which care was initiated (earlier is better) and the percentage of recommended visits received (higher is better).10 Among a sample of more than 1,000 young, mostly Black women ages 14 to 25 with low-risk pregnancies, approximately 10 percent more women in GPNC than in individual care received adequate prenatal care.G In a smaller (n=322) RCT in two military settings, 63 percent more women in GPNC than in individual care received adequate prenatal care.J Among a sample of Black women with high-risk pregnancies, those who were randomly assigned to GPNC (13.7; SD=3.8) attended an average of approximately two additional prenatal visits, as compared to women in individual care (11.9; SD=3.8).L
Healthy and Equitable Births
The impact of participation in GPNC, most often the CenteringPregnancy model specifically, on birth outcomes has been widely studied, but findings are mixed. Several strong causal studies found positive impacts on birth outcomes, whereas others found no impact, with no clear pattern by study design, model of GPNC, or study population to account for the differences. Further, the evidence on whether GPNC reduces racial disparities in birth outcomes is inconclusive.
Findings from studies of the impacts of group prenatal care on preterm births are mixed. In a two-site RCT of an enhanced model of CenteringPregnancy with a mostly Black sample of more than 1,000 young (ages 14 to 25) women with low-risk pregnancies, 9.8 percent of mothers in group care delivered preterm, compared to 13.8 percent of mothers in individual care (OR=0.67; p=0.045), equal to a beneficial risk reduction of 33 percent.G In contrast, a 14-site cluster RCT of the same enhanced model of CenteringPregnancy with a sample of more than 1,100 adolescents ages 14 to 21 with low-risk pregnancies, one-third of whom were Black and approximately 60 percent of whom were Latina, found identical rates of preterm birth (10.1) percent among both mothers in GPNC and those in individual care in the intent-to-treat analyses.I
Two smaller RCTs – one among a military population (n=322)J and another among a high-risk group of Medicaid-eligible Black women (n=619)L – similarly found no statistically significant differences in the rate of preterm birth between women participating in GPNC and the control group. The military study evaluated an enhanced model of CenteringPregnancy, whereas the high-risk study evaluated a non-specified model of GPNC. Both studies conducted power analyses that indicated sufficient sample size to detect significant differences between groups; however, the low prevalence of preterm births in both studies, and particularly in the military study (only 10 preterm births in the GPNC group and 7 in the individual care group), suggests that the studies may have been statistically underpowered to detect differences in preterm birth outcomes specifically.
Two retrospective matched studies using propensity scores of women’s participation in CenteringPregnancy in South Carolina both found a reduced risk of preterm birth among GPNC participants compared to women receiving individual prenatal care across samples of varying risk. Among a sample of 15,000 women, the majority of whom were enrolled in Medicaid, that included women with high-risk medical conditions, the rate of preterm birth was approximately 30 percent lower among GPNC participants (7.5 percent) than among the comparison group (10.6 percent).A Among a sample of over 6,000 Medicaid-enrolled women that excluded high-risk mothers, participation in GPNC reduced the risk of having a preterm birth by 36 percent.E
In contrast to the consistent findings from South Carolina, findings from two retrospective matched studies using propensity scores in Tennessee were mixed. Both studies had large samples of over 6,000 women and included those with high-risk pregnancies. In one study that examined participation in one of two models of GPNC (CenteringPregnancy or Expect with Me), after adjusting for the number of individual care visits,ii the rate of preterm delivery was almost 80 percent lower among GPNC participants (1.8 percent) than among the comparison group (8.6 percent).B In the other Tennessee study, no significant differences were found between women participating in CenteringPregnancy and the matched comparison group (OR=0.83; CI: 0.61-1.12).O
Finally, an evaluation of the Strong Start for Mothers and Newborns initiative (Strong Start) for women covered by Medicaid and CHIP during pregnancy found no significant differences between GPNC participants across 11 sites and a matched comparison group on any birth outcomes, including rates of preterm birth.F Sites offering both traditional and group forms of care with opt-in policies that had low acceptance rates were excluded to reduce selection bias.
Findings from studies of GPNC on low birthweight are mixed, but most find null impacts. Both of the retrospective matched studies using propensity scores of women’s participation in CenteringPregnancy in South Carolina found a significantly reduced risk of low birthweight among GPNC participants compared to women receiving individual prenatal care. In the South Carolina study of 15,000 women, including those with high-risk medical conditions, the rate of low birthweight was 35 percent lower among GPNC participants (7 percent) compared to the comparison group (10.7 percent).A In the South Carolina study of 6,000 women with low-risk pregnancies, participation in GPNC reduced the risk of having a low birthweight infant by 44 percent.E Similarly, in the Tennessee retrospective matched study of GPNC with a sample of over 6,000 women that included high-risk pregnancies, the rate of low birthweight was 83 percent lower among participants in either CenteringPregnancy or Expect with Me (1.2 percent) compared to the matched comparison group (7.2 percent).B
These positive findings must be considered in the context of the other strong causal studies with varying designs and populations that all find no impact of GPNC on low birthweight. RCTs, both large (n>1,000) and small (n<300), both including and excluding high-risk pregnancies, and with samples of varying demographic characteristics, all find no statistically significant impact of GPNC (CenteringPregnancy, enhanced CenteringPregnancy, and unspecificed models of GPNC).D,G,I,J,L As with preterm birth, the smaller RCTsJ,L may have been statistically underpowered to detect significant differences in rates of low birthweight. For example, a small RCT (n=282) of adolescent women found a rate of low birthweight among GPNC participants that was approximately half the rate among individual care participants, but the difference was not statistically significant – a likely consequence of being statistically underpowered.D Further, for every retrospective matched study with propensity scores that demonstrated a positive impact on low birthweight, a similarly-designed study finds no impact on low birthweight, including the Strong Start evaluation.F,P,O
Studies have also examined the impact of GPNC on the likelihood that newborns are admitted to the Neonatal Intensive Care Unit (NICU). The findings on NICU admissions are mixed, with experimental evidence showing null impacts and quasi-experimental evidence from South Carolina showing beneficial impacts.
The South Carolina studies, which also found significant reductions in the likelihood of preterm and low birthweight births, found significant reductions in the likelihood of a NICU admission or stay. Among the sample that included high-risk pregnancies, the rate of NICU admissions was 39 percent lower for GPNC participants (6.2 percent) than for the matched comparison group (10.1 percent).A Among the sample with all low-risk pregnancies, participation in GPNC reduced the risk of having a NICU stay by 28 percent.E
In contrast to the findings from retrospective matched studies with propensity scores, four RCTs found no significant impacts on NICU admissions.G,I,J,L The RCTs ranged in sample sizes from over 1,000G,I to 322.J Similar to the other birth outcomes, the smaller RCTsJ,L may have been statistically underpowered to detect differences, given the low prevalence of NICU admissions.
Parental Health and Emotional Wellbeing
Several studies have examined the impact of GPNC on parental health and emotional wellbeing, and the findings are mostly positive, with the exception of the study in two military settings, which found no significant impacts. Though the military study had a smaller sample size than the other RCTs, power analyses suggested the sample was large enough to provide reliable estimates for parental physical and emotional wellbeing.
Gestational Weight Gain
Mothers’ physical health during pregnancy has important implications for pregnancy and perinatal outcomes; both excessive weight gain during pregnancy and obesity are risk factors for pregnancy complications.M The social and peer support associated with the GPNC model, along with the time spent discussing nutritional choices and exercise and conducting self-assessments of weight and blood pressure, have been theoretically linked to a reduced likelihood of excessive weight gain during pregnancy.M,P,J
Secondary analysis of data from the 14-site RCT of an enhanced model of CenteringPregnancy showed that adolescent women in GPNC gained significantly less weight during pregnancy and retained less weight at 12 months postpartum compared to adolescent women in the control group.M Similarly, a retrospective matched study using propensity scores with a sample of almost 400 young women found that mothers in CenteringPregnancy were significantly less likely to have excessive weight gain.P However, a small RCT conducted across two military settings found no difference in gestational weight gain between GPNC participants and women receiving individual care.J
Maternal Depressive Symptoms and Stress
Perinatal depression is associated with negative birth, neonatal, and infant outcomes, and although GPNC theoretically may improve mothers’ emotional wellbeing, in part through increased social support,C findings are mixed. Secondary analysis of data from the 14-site RCT of an enhanced model of CenteringPregnancy demonstrated a positive impact on perinatal depressive symptoms – the rate of probable depression decreased by 31 percent between the second trimester and 12 months postpartum among adolescent women in GPNC, compared to a 15 percent reduction among adolescent women in individual care.C Secondary analysis of data from the two-site RCT of an enhanced model of CenteringPregnancy found no overall significant differences in rates of depression or stress over time between GPNC participants and women receiving individual care. However, high-stress women participating in GPNC reported significantly more decreased stress and lower depression than high-stress women in individual care.H High-stress women in GPNC decreased an average of 5.1 points on the Perceived Stress Scale (PSS) from study entry to the third trimester, compared to an average decrease of 2.8 points among high-stress women in individual care, though differences were no longer significant one year postpartum. High-stress women in GPNC decreased an average of 6.5 points on the affect-only component of the Center for Epidemiologic Studies Depression Scale (CES-D) from study entry to one year postpartum, compared to an average decrease of 4 points among high-stress women in individual care. In contrast, the two-site military study found no significant differences in either stress (measured by the PSS) or depression (prenatally, measured by CES-D or postpartum, measured by the Postpartum Depression Screening Scale) between women participating in GPNC and those participating in individual care.J
Optimal Child Health and Development
The impact of GPNC on the likelihood that mothers initiate breastfeeding is mixed. In the two-site RCT of an enhanced model of CenteringPregnancy with a young, largely Black sample, women in GPNC had almost twice the odds of initiating breastfeeding than women who received individual care (OR=1.73; CI: 1.28-2.35).G A similar difference in breastfeeding initiation emerged in a retrospective matched study using propensity scores with a sample of 800 women across four CenteringPregnancy sites in Tennessee.N In that study, women who participated in GPNC had more than twice the odds of reporting any breastfeeding at hospital discharge than women who received individual care (OR=2.08; CI: 1.32-3.26). However, no significant differences were found in the odds of exclusively breastfeeding at hospital discharge or the odds of any or exclusive breastfeeding at the six-week postpartum follow-up.
In contrast to the positive findings, two studies found no significant differences in breastfeeding initiation, though rates of initiation were high among all women in both studies.I,J In the 14-site RCT of an enhanced model of CenteringPregnancy with a sample of adolescent women, 88.8 percent of women in GPNC reported initiating breastfeeding compared to 87.2 percent of women in individual care.I In the smaller, two-site military study, 94 percent of women in both groups reported initiating breastfeeding, though rates of breastfeeding across groups had dropped to 56 percent by three months postpartum.J
- An impact is considered statistically significant if p<0.05.
- All group prenatal care participants received at least one individual visit – at intake, when a group visit was missed, or if another visit was deemed medically necessary.