A Conversational AI Chatbot is Helping New Mothers Keep Healthy (#1)
Research in Translation
Introduction
Welcome back to ‘From the Computer to the Clinic’ - a newsletter about computational biology and its contributions to biomedical research.
In this newsletter, we explore how computational biology research can drive clinical progress. By sharing success stories in one disease area or domain of research, we aim to inspire the use of these successful approaches for other diseases and research areas also.
If you haven’t already, you can subscribe to this newsletter, or share it with friends and colleagues
Part I
There is a lot of talk about the potential of artificial intelligence to dramatically change the way we live our lives. Some cite the power of AI to revolutionize the healthcare industry or make work more efficient. Others fear that AI will worsen existing inequalities and kill human creativity. But most claims about the potential of AI – good and bad – are more speculative than concrete. In this series, we explore a specific example of an AI tool with the power to improve the healthcare system and trace back its origins to a few fundamental ideas that emerged from academic research.
The tool in question* is designed to improve maternal health, specifically in the so-called ‘fourth trimester’: the months right after a new mother has given birth. In this time a new mother faces many potential health challenges. Infection, blood clots, and high blood pressure are common post-pregnancy complications. And just as challenging as the physical complications of pregnancy are the mental and emotional challenges of adapting to motherhood.
For new mothers in the US, a one-time visit about 6-12 months after giving birth is standard – but just one post-pregnancy visit is certainly not ideal, and many new mothers (around 50%) do not even attend this visit. This is especially true for new mothers of lower socioeconomic status, for whom a postpartum doctor’s visit is likely more difficult to balance among many other priorities like caring for the new child (plus other children) and re-integrating into the workforce.
Penn Medicine, the University of Pennsylvania’s multi-hospital system, has been working to address this problem since 2018, when they started their Healing at Home Program. The program’s initial focus was getting new mothers’ home from the hospital faster after they gave birth, minimizing cost and discomfort for new parents. Over time, however, the clinicians and other staff running the program came up with the idea of a text-message based support system for mothers to contact physicians and ask questions after they leave the hospital as a way to improve postpartum health.
.jpg){kind=link}
In theory this is a great idea, but there are a number of challenges in practice. New mothers need detailed, individual responses to questions about caring for their new child and taking care of their own health. This is infeasible for a hospital team of physicians, even at a large center like Penn Medicine, who are already overwhelmed with the job of taking care of patients still in the hospital.
One alternative is to create an automated chatbot that responds with pre-defined answers to questions. This kind of chatbot - which is often referred to as ‘rule-based’ or ‘decision-tree’ chatbot - has an internal map of possible questions and answers and is cleverly designed so that users can select certain questions or paths through a conversation and be guided towards appropriate, expert-curated responses (a lot like a choose your own adventure book). As you can imagine, however, this system is limited by the amount of responses that have been designed into it, and is likely to fall short in answering many of the specific questions that new mothers have. It is also rather impersonal - every new mother will get the same answers, and there is no ‘memory’ in the system of the mother’s unique medical history and personal background.
To address both of these shortcomings, the Penn Medicine team has partnered with the company Memora Health. Memora has developed a ‘conversational AI’ chatbot – like ChatGPT, but specifically trained to address the concerns of new mothers. The chatbot’s name is Penny, and Penny has the potential to improve postpartum care dramatically.
Stay tuned for part II, coming soon…
*The topic of this series was inspired by a recent article from Bill Siwicki, Managing Editor of Healthcare IT News