Association of group prenatal care with postpartum visit attendance: a secondary analysis of the Centering and Racial Disparities randomized clinical trial.

OBJECTIVE: The American College of Obstetricians and Gynecologists recommends a comprehensive postpartum visit (PPV) within 12 weeks after birth.1 Nationally, an estimated 60% to 65% of patients attend this visit, and attendance is lower among patients with public insurance.2 Group prenatal care (GPNC) has been associated with higher PPV attendance in observational studies that could not account for self-selection.3 We evaluated whether GPNC, compared with individual prenatal care (IPNC), was associated with PPV attendance in a planned secondary analysis of the Centering and Racial Disparities (CRADLE) randomized clinical trial, with prespecified attention to race and ethnicity.

STUDY DESIGN: CRADLE (NCT02640638; institutional review board-approved; written informed consent obtained) randomized 2348 medically low-risk patients 1:1, stratified by race and ethnicity, to CenteringPregnancy GPNC (10 2-hour group sessions) or IPNC at a single Southeastern academic center between February 2016 and March 2020; enrollment ended early because of the COVID-19 pandemic.4 Conditions precluding nurse practitioner-led care (eg, pregestational diabetes, medicated chronic hypertension) were exclusions. We additionally excluded pregnancy loss before 20 weeks and transfer of care before delivery. PPV attendance was defined by billing code 0530F for a comprehensive postpartum visit. Prenatal care adequacy was categorized as inadequate, intermediate, adequate, or adequate-plus using the Adequacy of Prenatal Care Utilization (Kotelchuck) Index, on the basis of care initiation timing and visit count. The Patient Activation Measure, which assesses self-reported knowledge, skills, and confidence in managing their health, was administered at enrollment to a subset of patients. The primary analysis was intention-to-treat (ITT). Only 2 prespecified exploratory samples, modified ITT (mITT; ≥1 visits in the assigned arm) and per-compliance (PC; ≥5 visits), were not fully randomized; logistic regression adjusted for baseline variables that differed between arms. Relative risks (RRs) were estimated using log-binomial or robust Poisson regression (SPSS 29 [IBM]; 2-sided P<.05).

RESULTS: Of the 2047 participants analyzed (1018 GPNC, 1029 IPNC), 39.2% were Black, 23.0% Hispanic, and 36.5% White; the mean age was 25.2 years, 43.7% were nulliparous, and 94.5% were Medicaid-eligible. In the ITT analysis, PPV attendance did not differ between arms (68.8% vs 67.4%; RR, 1.02; 95% confidence interval [CI], 0.96-1.08; P=.55) (Figure). In the exploratory mITT (73.3% vs 68.8%; adjusted RR, 1.10; 95% CI, 1.03-1.17) and PC (76.7% vs 70.3%; adjusted RR, 1.14; 95% CI, 1.07-1.21) samples, GPNC was associated with higher attendance; however, exclusions were asymmetric (24% of GPNC vs 2% of IPNC participants were excluded from mITT). Among Black participants, GPNC was associated with higher attendance in all 3 samples, including the ITT subgroup (72.7% vs 63.9%; RR, 1.14; 95% CI, 1.03-1.25; P=.009), with no difference among Hispanic or White participants (Table). Among GPNC participants, PPV attendance rose with the number of GPNC sessions attended, from 54.9% among those attending none to 84.7% among those attending 9 or 10 (P<.001), a gradient that was present in each racial and ethnic group (Appendix 1 and 2). Adequate or adequate-plus prenatal care (72.7% vs 45.8%; RR, 1.59; 95% CI, 1.39-1.82) and nulliparity (73.9% vs 63.6%; RR, 1.16; 95% CI, 1.10-1.23) were associated with attendance; within each adequacy category, PPV attendance was similar for GPNC and IPNC. Baseline patient activation (n=529) was similar between arms (P=.92).

CONCLUSION: In the only analysis that preserved randomization, GPNC did not increase PPV attendance. The higher attendance in the mITT and PC samples and among Black participants should be regarded as hypothesis-generating because those samples had disproportionately greater exclusion of GPNC participants, subgroup comparisons were not powered a priori, and adjustment cannot exclude residual confounding by patient motivation, although baseline activation was similar between arms in the subset measured. The null ITT result raises the possibility that self-selection contributed to the higher attendance reported in observational studies.3.