Return-on-Investment Analysis of an Enhanced Community Doula Program: Pre- and Post-COVID-19 Considerations.

Background: Community doulas are perinatal health workers who provide peer education, resource navigation, and support during pregnancy, childbirth, and the postpartum period. Evidence suggests that doulas improve the experience of care, reduce cesarean birth, and improve breastfeeding outcomes. However, people with low incomes cannot access affordable community doula support in most states due to lack of insurance reimbursement. To determine the affordability of Medicaid reimbursement for doula services, there is a need to fill a gap in research that employs real-world data and a return-on-investment (ROI) analysis approach.

Methods: We conducted a ROI analysis from the healthcare perspective of an enhanced community doula intervention, Best Beginnings for Babies (BBB). Healthcare and program cost data were collected alongside clinical outcomes from a randomized controlled trial of routine maternity care at Boston Medical Center with and without BBB. ROI was calculated as the net healthcare savings divided by the investment costs. Post-COVID-19 program costing estimates were also performed.

Results: Average healthcare costs per patient were $18,969 for the BBB group compared with $20,121 for routine care, a savings of $1,152. BBB program costs were an average of $971 per person. There was an 18% ROI. Lower costs for the birth hospitalization and NICU stays accounted for the largest areas of savings. Per-person program costs using proposed MassHealth fees produced a positive return on investment, although 2023 hospital doula program wages and salaries did not.

Discussion: The BBB-enhanced community doula program was cost-saving to payers and increased access to doula support for low-income people. Even with post-COVID-19 increases in program costs, analysis demonstrated doula support was still financially feasible. This study should reassure budget-conscious Medicaid payers that doula services are affordable.