Case Study: Maternal Monitoring
Maternal mortality in the United States is on the rise, particular among women of color, Native Americans, and women who live in rural areas. With an increasing number of hospitals shuttering their obstetric units and limitations posed by lack of transportation and poverty, could technology step in to help these women?
I designed a conversational and visual interface system to connect underserved women to healthcare services during the perinatal period.
WHAT I DID
Research, design research, user interface design, conversational interface design, prototyping
Two of my friends have lost their wives during the postpartum period. It was utterly devastating. What was supposed to be a joyous time, ushering in new life and growing a family in love, became something tragic. In 2017, while heavily pregnant with my son, I listened to this NPR report. Needless to say, I wasn’t particularly thrilled at the prospect of giving birth - in fact, I was scared. But I wondered, why were women dying and could technology provide a solution? Let’s explore.
Step 1: Defining the Problem
Alright, so as you can imagine, this is a HUGE, complicated, multi-faceted problem. There are so many factors (e.g. poverty, systemic racism, accessibility, assumptions, lack of training, etc.) that all play into what we are facing today. So in order to make designing for this even remotely possible, I really needed to define the problem and the first part of that was figuring out WHO was being affected. (Please forgive the change in colors with the images, these were taken from multiple iterations of the same project).
So I turned to the internet and looked up various sources in order to get a general idea of who is affected. In my research I found that the women most affected were rural, African American, and Native women. Now I had always assumed that the reason more women were dying was because they couldn’t afford healthcare but that really wasn’t the whole story or even the case for many women. What I did find out was that many hospitals in rural areas were closing down their obstetric units because they were just too costly to run. And many women were forced to travel 1-2 hours to each prenatal visit and travel that far when it was time to give birth. Many of these women, fearing that they were going to deliver their babies on the side of the road, were opting to have elective Caesarean sections or inductions. This in-turn was leading to more complications and many women were too far from the hospital, or too busy/overwhelmed to go. So they would wait it out at home. . . .
For poor women in urban areas, it was kind of the same story. Granted, systemic racism is a huge factor, but not one that I could really solve directly(more on this later). Like the rural women, it was also very hard for poor urban women to get to their appointments- lack of transportation and childcare being the most noted factors.
At this point it looks like we have a direction in the sense that I know we have three underserved groups of women - now it’s time to talk to some of them.
Step 2: The Interviews
The first thing I did was call up my buddy who is a nurse at the Cleveland Clinic. I asked her to give me some insights to what she felt were the major sticking points for our identified women. She pointed out that a lot of women don’t understand their symptoms and can’t figure out what’s normal. She then pointed me to the Post-Birth Warning Signs page where I got the inspiration for my first iteration of this project. (See more on this below in the Prototyping section).
I then went online and basically put out a plea/beg for my friends who fit the aforementioned groups to give me a jingle. I didn’t really know any Native Americans but it wasn’t too hard to find people from the other two groups. Here is a quick summary:
What I heard from the women I spoke to reiterated what the research was saying all along and I picked up themes of isolation, insecurity, and fear. Women were having to travel far for OB appointments and one opted for an induction in order to get some sense of control (again, what the research already pointed to). What surprised me was that I picked up on a theme of apathy and I began to wonder how apathy could play into patient outcomes.
Step 3: Ideation
Alright, I have the problem and some research. This is the time when I start to brainstorm. Here are sketches from this part of the process.
Step 4: Prototyping the First Iteration
For my first iteration, I simply created an app prototype of the POST-Birth checklist. I started by creating a user flow that uses a series of Boolean functions to help a women figure out if she needs to go straight to the ER or call her doctor’s office. I made both mobile and desktop versions of the app with XD and Invision respectively.
Step 5: Testing
I created a prototype and I tested it with some people. Now when I say “test” I didn’t use a beautifully scripted test in the field- I just put it in front of people and said, “try it out”. Not your most robust testing and now, looking back, I am a little embarrassed that was how I did it. 🙈 It seemed to work fine but I really wanted to see what nurses thought. I wrote to a variety of nurses, including the nurses who developed the POST-Birth checklist for feedback and NONE responded. I was disappointed but I decided to continue to truck on through. We’ll eventually test this portion of the app later.
Step 6: Addressing the Apathetic User
At this point, that I felt that the POST-Birth app was just too simplistic and didn’t really address the bigger issues. The POST-Birth app is really just a prototype that was designed to help people make some sense of confusing symptoms and. . . that’s it. It does not touch on the real “meaty” issues like apathy and lack of access. I guess you could say I was, like my mamas, feeling a little apathetic too. So it was back to the drawing board (see literal drawing board).
It was at this time that I stumbled upon the Remote Pregnancy Monitoring Challenge offered through the Health Resources and Services Administration. The HRSA is an off-shoot of the Department of Health and Human Services that is the “primary federal agency for improving health care to people who are geographically isolated, economically or medically vulnerable”. Interesting. . . I put this contest in my back pocket and continued to plug away at the app.
Step 7: Building a Conversational Interface
It was at this point that two fellow designers joined me with the project and we decided to build a conversational interface because it very well might be the solution needed to address the apathetic user. During this group project portion I was in charge of overall design of the model, coming up with scenarios for the chatbot, user testing with the chatbot, developing personas, and making empathy maps.
You might be wondering why we went the way of the chatbot (aren’t they annoying?) but we did have a few compelling reasons.
81% of Americans text regularly and 97% of American adults text weekly
Many women are familiar with the way text messaging works. Text messaging is so ubiquitous these days that its use transcends class, race, and many cultures. In fact, a popular meme is that many young Americans are annoyed by phone calls and prefer to text. We were aware that there may be barriers to smart phone access but “the vast majority of Americans – 95% – now own a cellphone of some kind” and a chatbot does not even need a smartphone. It could work with a basic cellular phone or even a regular landline phone through voice recognition technology.
About 50% of adults 18-24 say “text conversations as just as meaningful as a phone call”
Again, it may seem strange to some that we would propose a chatbot as part of our solution but there is something about the app reaching out to patients and checking in on them. In our informal interviews we noticed that women sometimes felt apathetic and discouraged when dealing with their healthcare providers and having an app regularly check-in and having nurses available to answer questions and provide support could do much to help women and their families. Since we are looking at women of childbearing age, we should have fewer issues with on-boarding when compared to a more mature demographic. True, nothing can match a personal interaction but we know that financially this can be incredibly difficult to fund. With the chatbot, benign issues and questions can be handled by the chatbot and more serious concerns can be handled by the remote nurse.
Accessibility is already there.
Finally, the beauty of text messaging services is that accessibility features have been already included and would not need to be redesigned. Awesome!
Here is photographic evidence of the process. I made the slides using Adobe Illustrator and all work featured below is solely my work.
Step 8: The Contest
The issue with school projects is that they often lack constraints that you see in the real world. Fun? Yes! But that’s not how design works. Feeling desperate and in dire need of cash, I started googling UX competitions and found this contest. Interesting! I had already done most of the work and this competition has constraints (the group partners are now gone but I did ask a friend in the nursing program to help me to find nurses to look it over). The contest stipulates that the solution has to be “affordable” and “scalable”. Now affordable is a pretty dubious distinction. Affordable to whom? The government? The insurance company? The user? Scalable makes a little more sense. The solution has to work on various size of groups. I had only five pages to work with and I wrote out my solution to best meet the parameters of the project. I made visuals with Adobe Illustrator and essentially answered the questions the website posed.
Step 9: AND More Testing
So where are we now? It’s been months (I think we’re on month seven?) and this project just keeps growing. It has been submitted to the HRSA and we hope to hear back from them soon. The mobile app is currently under testing and I will be sure to report on those findings as soon as I finish.
Thank you for reading!