Postpartum Education Via Artificial Intelligence for Recovery and Loneliness: A Randomized Controlled Trial

The goal of this clinical trial is to learn whether a postpartum chatbot powered by generative artificial intelligence (genAI) can help new mothers get better pelvic floor health information and feel less lonely after childbirth.

The main questions this study aims to answer are:

• Does using the chatbot improve postpartum pelvic floor health knowledge?
• Does using the chatbot help reduce feelings of loneliness during the postpartum period?
• Does using the chatbot impact pelvic floor symptoms?

Researchers will compare standard postpartum care to standard care plus the chatbot.

Participants will:

Be assigned by chance (like flipping a coin) to standard postpartum care with or without access to the chatbot.

If in the chatbot group, participants will receive education and support via the chatbot over a 4-week period.

Both groups will complete questionnaires to measure their pelvic floor knowledge, pelvic floor symptoms, feelings of loneliness, depression, infant bonding, perceived social support, adverse childhood experiences, and peri-traumatic distress.

The chatbot was created by urogynecology experts in collaboration with UC San Diego computer science and biomedical informatics researchers. The chatbot is designed to give new mothers personalized, evidence-based information and support in real time.

Pelvic floor disorders (PFDs)-including urinary incontinence, fecal incontinence, and pelvic organ prolapse-are highly prevalent and debilitating conditions affecting approximately 25% of women in the United States.Pregnancy and childbirth are known risk factors for developing PFDs. Recent imaging studies show that injuries to the levator ani muscle, perineal body, and perineal membrane occur in up to 19% of first-time mothers. Urogynecologic conditions, such as urinary and bowel incontinence, are common after pregnancy and often result in a catastrophic impact on maternal quality of life not just because of the physical symptoms, but also because of the associated embarrassment, isolation, and negative stigma.Compounding this challenging recovery further, one in five women will experience a perinatal mood disorder, such as anxiety, depression, or PTSD. Despite the profound physical and psychological impacts of childbirth, postpartum care remains limited, inconsistent, and difficult to access.

It is well established that many postpartum individuals seek education, community, and emotional validation on various online platforms. When postpartum individuals are interviewed, many report feeling their postpartum health concerns are neglected and minimized, resulting in feelings of disempowerment through the help-seeking process. Further compounding these feelings, is the lack of timely or frequent postpartum care with many women having difficulty attending postpartum visits and often being only seen once postpartum. Furthermore, patients who experience a severe pelvic floor injury at delivery may not have access to specialty care clinic for peripartum pelvic floor disorders. Through qualitative interviews with postpartum individuals with obstetric anal sphincter injury (OASI), themes of seeking social support through online communities such as Facebook and Instagram were identified, though the effectiveness of these groups in meeting patients' needs is unclear. While social media platforms may serve as a support bridge for patients as they await their postpartum visit, much of the content on these platforms is anecdotal, humorous, and not produced by physicians. Furthermore, these platforms also expose recovering individuals to distressing anecdotal narratives or idyllic imagery that may further worsen mental health in this vulnerable patient population.

As postpartum individuals navigate the fourth trimester, feelings of loneliness and isolation are common. Loneliness is linked to adverse health outcomes, including cardiovascular disease, cognitive health, infectious disease, and all-cause mortality, yet its impact on pelvic floor muscle recovery remains largely unexplored.In the U.S. Surgeon General's 2023 advisory, the healing power of social connection and community was emphasized. Lunstad et al., acknowledge that despite mounting evidence healthcare has been slow to recognize that social relationships are health determinants. Postpartum individuals are not immune to loneliness, in fact in a systematic review, the prevalence of loneliness among peripartum women ranged between 32% to 100%. Identifying solutions to address the epidemic of postpartum loneliness is critical to improving the overall health of both birthing individuals and their infants.

Optimizing postpartum maternal mental and physical health is not only critical for the postpartum individual's recovery, but also profoundly influences infant development. The first weeks and months after delivery represent a sensitive window for neonatal brain development, attachment formation, and physiological regulation. When postpartum individuals experience persistent pain, incontinence, or limited mobility, it can hinder early caregiving, reduce breastfeeding success, and increase maternal-infant separation.Maternal mental health disorders have been associated with disrupted bonding, impaired infant stress regulation, and long-term cognitive, emotional, and behavioral challenges in children. Emerging research in developmental neuroscience demonstrates that early maternal-infant interactions shape the infant's stress response system and neurodevelopment trajectory through mechanisms involving cortisol regulation, oxytocin signaling, and epigenetic modifications. It is possible that through improved maternal knowledge of pelvic floor health, that postpartum individuals will experience improved self-efficacy, decreased loneliness, and possibly even improved physical recovery.

However, peripartum pelvic floor disorder knowledge is lacking. McLennan et al., evaluated the information that patients received during pregnancy, and found that the most neglected content area was education on PFDs. Furthermore, Reagan et al., found that most available patient education materials are above the reading level recommended by the National Institutes of Health for maximum patient comprehension. Numerous publications have concurred that there is a global lack of knowledge and understanding of urinary incontinence, pelvic organ prolapse, and bowel incontinence educating patients on PFDs, with many authors calling for increased emphasis of pelvic floor health and development of educational programs to inform patients.Rutledge et al., performed a randomized controlled trial assessing peripartum pelvic floor health knowledge as measured with the Prolapse and Incontinence Knowledge Questionnaire (PIKQ) among women randomized to receiving written handouts only or written handouts plus an interactive workshop. Results of this study demonstrated that both groups improved in their knowledge, but interestingly patient perspectives on preferred education was discovered. Themes surrounding the delivery of educational content were identified, such as a preference for reviewing informational materials on their own time and having the opportunity to receive real-time responses to questions. In addition, themes of preferred educational content developed, including desire to understand the etiology of PFDs, comfort with discussing PFDs, and relief in learning treatment options.

While in person support groups, text based platforms, and telehealth coaches have attempted to address this unmet need for postpartum support and education, conversational agents (chatbots) have recently emerged as a promising tool for scalable, real-time health education and support in this patient population. Chatbots have the potential to deliver just-in-time adaptive intervention (JITAI), a novel intervention design that adapts the provision of support (ie type, timing, intensity) to delivery support at the moment and in the context that the user needs it most. JITAI increases user receptivity to support and is increasingly being used to help support health behavior changes, including mental health. Using natural language processing, chatbots interact with users to provide education, emotional support, and symptom guidance. Their availability on smartphones has made them a widely used tool for health information retrieval and mental wellness. Prior studies evaluating chatbots in the peripartum period have demonstrated successful engagement of individuals, improved user knowledge, and reduction in depressive symptoms. Furthermore, outside of the peripartum patient population, chatbots are known to reduce loneliness through direct companionship and social interactions. Kim et al., explored the mental health potential of AI social chatbots over four weeks and found that UCLA loneliness scores started to improve by week two among individuals interacting with the chatbot. Thus, we propose a similar study intervention time frame (4 weeks).

There is an urgent, unmet need for comprehensive, real time postpartum pelvic floor health education and support in the immediate postpartum period. To address this unmet need, we have developed a genAI postpartum chatbot to bridge the gap for patients to their providers. The chatbot was built using UCSD Health approved Large Language Models (LLMs) made available within the UCSD Health Cloud. In order to provide accurate curated information in response to patient questions in the chatbot, the chatbot implements retrieval augmented generation (RAG).

In the setting of rising maternal morbidity, coupled with limited healthcare access, understanding the impact of a postpartum educated AI chatbot on maternal loneliness, knowledge, and physical recovery is critical to improve maternal and infant health.