JMIR Publications

ABSTRACT

Background:

The 42 days after delivery are a high-risk period for birthing individuals and newborns, especially those who are racially and ethnically marginalized due to historical and current structural racism. We deployed two chatbots – one for birthing individuals and one for newborn caregivers – to provide timely and trusted information about post-birth warning signs to birthing individuals and newborns. The chatbots also included relevant educational information (e.g. nutrition) and links to resources (e.g. organizations that may offer transportation for medical appointments) to support birthing individuals and newborn caregivers.

Objective:

Using mixed methods, we identified strategies to improve chatbot engagement and to evaluate user perceptions and preferences for these two chatbots.

Methods:

4,679 individuals received the newborn and/or postpartum chatbot outreach between September 1, 2022 and December 31, 2023. For interviews and surveys, we sampled patients from the hospital discharge lists based on prenatal care location, age, type of insurance, and race/ethnicity. We conducted surveys and interviews in English and Spanish to understand the user perceptions of the chatbot, as well as identify areas for improvement. We analyzed quantitative results using descriptive analyses in IMB SPSS Statistics and qualitative results using deductive coding in Dedoose.

Results:

Overall, 2,859 individuals opened the chatbot messaging (61%) and 100 patients who engaged with the chatbot completed the survey; of those, 40% identified as Black, 27% Hispanic/Latina, and 18% completed the survey in Spanish. Payer distribution was 55% public insurance, 2% uninsured, and 39% commercial insurance. Approximately 80% of survey participants indicated that chatbot messaging was timely and easy to use, and 59% and 66% found the reminders to schedule the newborn visit and postpartum visit useful. We also conducted 23 interviews; 78% of interviewees engaged with the chatbot, 61% of whom identified as Black, 17% identified as Hispanic/Latina, and 9% of interviews were conducted in Spanish. Interviewees provided positive feedback on chatbot use and content and also provided recommendations for improving the outreach messages.

Conclusions:

Chatbots are a promising strategy to provide information to birthing individuals and newborn caregivers about postpartum recovery and newborn care. Future work should measure the impact on specific health outcomes and reduce disparities. Clinical Trial: N/A