emma thumb.png

EMMA: A MATERNAL HEALTHCARE COACH

Design Engineering II  |  Martin Bechthold (GSD) +  Woodward Yang (SEAS)  |  2017-2018
Advisers Jock Herron (GSD) and Peter Stark (SEAS)
In collaboration with Julie Loiland (MDE '18)
*Awarded Harvard's Master in Design Engineering Thesis Prize 2018

INTRODUCTION

Taking care of a newborn child is challenging and multifaceted. Information needed to provide the necessary care as a first time mother is uncurated through online resources, books, conversations with friends, and doctors’ opinions; expensive in both time and money for professional consultation, especially between different dimensions of care such as personal nutrition, fitness, and health; and is often limited to personal intergenerational experience with parents sharing information. The primary focus on care for the newborn ignores the care that the new mother should be taking to recover properly into a new way of life. With so many physiological and psychological changes, the new mother is at risk for mental illness and unhealthy habits around sleep, stress management, nutrition, fitness, and social aspects, which could lead to longer term health complications of overweight and obesity. This is an important point to intervene because it can negatively influence child growth and development.

Our solution, Emma, is a conversational maternal health coach focusing on postpartum health for mothers and their children.  Emma is not a replacement for doctors or medical specialists, but acts as a first point of contact to encourage conversation and answer questions backed by experts, culminating with a consultation with a qualified professional if needed.  The platform integrates machine learning, behavioral science, and user experience design to help new moms adjust after childbirth as well as promote good habits in the home for child-rearing.

PROCESS

Completed as a part of the first thesis class for the Master in Design Engineering program, the design and development process for the project was messy. 

timeline.png

SYSTEM DIAGRAMS

In terms of resources and support, it takes a village to raise a child. There is more support for the child after birth than there is for the mother, especially for first time new-moms or for new-moms with less access for support because of limited time or money.

Systems-Diagram.gif
Systems-Stakeholders.gif

STAKEHOLDERS

In order to understand varying incentives, current service innovations, and intricacies of environments of care, we surveyed stakeholders and precedent programs across clinical, public health, and personalized health care spaces. This revealed a need for a multi-dimensional approach to bridge traditional support, such as with the Circle Program at Brigham and Women's Hospital. A repeat theme was the need to make a solution that was not resource intensive on the clinic as well as one that could be delivered to the new-moms in their home, such as a digital platform similar to Virta Health. 

emma stakeholder logos.PNG
Postpartum care should become an ongoing process, rather than a single encounter, with services and support tailored to each woman’s individual needs.
— New clinical guidance published by the American College of Obstetricians and Gynecologists (ACOG) May 2018 issue of Obstetrics & Gynecology journal
precedent.PNG

USER INTERVIEWS

Initially, we conducted in depth discovery interviews with new-moms to better learn about the intricacies of postpartum care. The interviews highlighted the need for a personalized approach as there is no one size fits all solution for being a new mother as well as the desire for new moms to learn quickly on the job, which requires an accessible and reliable information source to help field disparate questions. A follow up survey with over a hundred participants confirmed the gap in care information support in the home. Below are a few quotes: 

“Everyone focuses on the child, I wish I had better information on how to still make myself a priority after birth.”
"I really needed more info on recovery. The recovery was worse than the delivery for me, both emotionally and physically, and that would have been helpful to know to expect."

"The reality of what women feel after having a baby. You hear about pregnancy, but no one tells you the physical effects of delivering a baby. Yes, you’re happy, but your body goes through a lot and no one talks about it."

new moms matter.PNG

PROBLEM STATEMENT

Taking care of a newborn child is challenging and multifaceted. The primary focus on care for the newborn ignores the care that the new mother should be taking to recover properly into a new way of life. With so many physiological, psychological, and social changes, the new mother is at risk for mental illness and unhealthy habits around sleep, stress management, nutrition, fitness, and social aspects, which could lead to longer term health complications, such as obesity. This is an important point to intervene because it can negatively influence child growth and development.

SOLUTION

Emma is a conversational digital health coach focusing on postpartum health for mothers and their children. Through integrating a chat bot virtual assistant, a community of experienced mothers and professions, as well as an on demand telemedicine solution, Emma provides many layers of curated information for various means of in home support. 

solution.PNG

Drivers of design include the product being intuitive through ease of use and intentional flow of services provided; accessible through a sustainable model of service with a low barrier to entry; and trustworthy through an established community of new moms and cited sources. 

front end.PNG

For the first viable prototype, we aggregated data around the main categories of interest, including breastfeeding, postpartum depression and health, as well as around age of the child. To train Emma, we used a word2vec model to assemble dimensions of relevance that she could provide the top three most relevant summaries and links to any question asked in the moment. Over time, she would get smarter based on user feedback and in application transcripts. 

back end.PNG

To test the first prototype, we had a group of local new moms in Boston keep a question journal for two weeks with all of the different daily inquiries they had as well as the way in which they did (or did not) look for answers. Then, they ran the questions through Emma and rated their satisfaction with her help. She was most suited for the information she was trained on, such as questions around milk and napping recommendations, but she did not work well for geographic specific inquiries or those focused on older children. In time, this information can be added. In conjunction with testing the first version of Emma as a chatbot, we ran through front end testing as well as general concept feedback with the group and here are some of the quotes: 

“Really inviting, interesting, and inclusive.  I love this idea.  I feel like I’m not being judged when I ask Emma questions.”
“I would recommend this app to my friends.  It’s so applicable to what moms need and it’s not something I’m seeing from other apps and sites.  You’re on to something here.”
“I’m into it.  The app offers a holistic solution for a very underserved group.  And the idea of connecting directly to specialists is so convenient.”  

CONCLUSION

While the initial reception of the prototype was positive, there is still much development work to do in order to build out a more comprehensive database before proper efficacy testing can be done. We furthered our research with the ethical implications of integrating an algorithm into complex decision making situations, and ultimately, we designed for an emphasized hybrid model always augmenting the human touch for community members, professionals, and especially new moms. Emma does not look to replace professional, clinical care, but she can help fill the current gap in the health care system with immediate in home support. As the initial proposal is for the U.S. health care system, Emma can also be implemented as a complementary service for countries with different support systems already in place by augmenting the time midwives and other means of support are not readily available at home or help reduce the intangible bias prevalent.