Connecting risk prediction to intervention: evidence from a randomized trial testing patient navigation and education to improve postpartum outcomes.
Kimberly B Glazer, Sarah Lindley, Angelina Malenda et al.
✦ AI-curated · Sources linked
In 30 seconds
This pilot randomized controlled trial evaluated risk-targeted postpartum education and patient navigation among self-identified Black individuals who gave birth at a Philadelphia academic medical center. A validated risk calculator identified participants with at least 5.5% predicted risk of postpartum hospital use (PHU); 220 were randomized to intervention or standard care. Within 30 days of discharge, PHU occurred in 20.0% of intervention participants and 16.4% of controls (risk ratio [95% confidence interval]=1.2 [0.7-2.0]). Engagement with educational materials was high, but preliminary estimates did not indicate reduced PHU.
Key findings
- Among 220 randomized participants, 30-day PHU incidence was 20.0% with intervention and 16.4% with standard care (risk ratio [95% confidence interval]=1.2 [0.7-2.0]).
- Among intervention participants, 79% found the educational materials helpful and 61% reviewed them more than once.
- Risk-based recruitment identified a socially vulnerable cohort: 50.0% reported annual household income <$45,000 and 14.2% were unhoused.
- Real-time risk ascertainment was challenged by time-intensive chart abstraction, variable-definition discrepancies, and missing data.
Why it matters
Postpartum hospital use is an important outcome amid persistent racial inequities in maternal morbidity and mortality. This trial shows that risk-targeted recruitment and navigation can engage participants facing substantial social vulnerability, while also highlighting practical barriers to applying prediction tools during recruitment. The preliminary outcome estimates do not establish benefit, but provide information relevant to designing larger evaluations.
What to keep in mind
This was a pilot trial at one academic medical center, and the preliminary effectiveness estimate was imprecise. Risk ascertainment also required time-intensive chart abstraction and was affected by discrepancies and missing data.