An app and a wearable ecosystem that helps pregnant people manage nausea
using machine learning
Lia helps reduce the impact of nausea on a pregnant person's lifestyle by helping them understand nausea, respond to it and receive support from loved ones
A 9-week grad capstone project advised by Artefact
Ideation, Concept Development, Wearable Design & Prototyping, Information Architecture, Wire-framing, Layout Design, Interface Design, Product Experience, Systems Design & Project Management
In the United States, around 5 million people experience nausea during pregnancy every year. Nausea is attributed to rising hormone levels and is most prevalent during the first trimester, but for some, it can last until the end of pregnancy.
Although it affects a significant number of people, nausea is overlooked by doctors as it is accepted as a normal part of pregnancy.
Nausea can significantly reduce the quality of life, increasing feelings of depression, and negatively impacting employment, household responsibilities, parenting, and family relationships.
Understanding nausea and finding suitable remedies is key to reducing its impact, but it is often a difficult and unreliable process.
This is where Lia comes in.
Lia is a machine learning-driven design concept that helps pregnant people manage nausea. It consists of a wearable that detects nausea and an app that surfaces the nausea pattern and recommends tailored context-specific remedies.
From the research phase of this project arose three key insights that guided the design of Lia. They are as follows:
The lack of a fine-tuned understanding
of nausea patterns makes it difficult to prepare for ahead of time.
Varied patterns of nausea between people cause them to seek solutions through trial-and-error, which can be time-consuming or lead to defeat
The mental toll on pregnant people is heightened by gaps in communication and support during the first trimester.
While research participants had a general idea of their nausea,
they expressed that it still felt unpredictable and opaque. I saw an opportunity to help pregnant people better understand and plan for their nausea.
For many of the participants, nausea felt like an insurmountable barrier that they resigned themselves to. I wasn't seeking to solve nausea altogether, rather I saw an opportunity to improve the process of discovering remedies in a personalized way.
Since nausea in most cases occurs before people announce their pregnancy, pregnant people rely heavily on a small circle of people for support. But at the same time, it can be difficult to communicate the support they need with their loved ones. I wanted the design response to include loved ones, to reduce the burden of nausea on the pregnant person.
Assist with understanding
& planning for nausea
Personalize the process of discovering remedies
Reduce the burden on
the pregnant person
Additionally, through the course of this project, I hoped to explore three specific interests through design. They are as follows:
Using data to empower human beings
Making data intuitive and actionable
Promoting human agency
To begin exploring these opportunity spaces
I led the ideation through a braiding session. The goal was to come up with as many ideas and directions as possible. During this process, I tried to explore a mix of physical and digital modes for the solution.
Ideation Sketches [Click to enlarge]
By tying these ideas to the research insights,
I identified 3 overlapping directions that also incorporated my design interests in data.
Integrating them together would create a more holistic system.
Design Directions [Click to enlarge]
To transform these high-level ideas into what would become Lia, I set out to answer 3 overarching questions:
How can I make Lia possible?
How can I make Lia ethical and inclusive?
How can I make Lia easy to understand and act upon?
Making Lia Possible
Tracking nausea requires logging information about their bouts of nausea, triggers, and further analyze this data. Research participants had indicated that tracking was too burdensome and combined with the effects of nausea it made it a lot more difficult. Therefore, an automated system was needed to collect, analyze, and project the nausea patterns.
USING MACHINE LEARNING
The core of Lia's abilities would be to recognize patterns. Machine learning is a technology that excels at this. By using machine learning, the system could connect different data points from individuals to a broader population, to create a detailed nausea pattern.
To determine what data was needed for the algorithm I looked
through research papers and patents. I discovered 4 sources that outlined how biometrics, combined with machine learning, could be used to detect nausea and other related conditions. Based on these sources, it can be estimated that this technology can be developed within the next 3-5 years.
Papers & patents about machine learning
for nausea [Click to read more]
There are 3 key types of biometric data for nausea detection:
heart rate, temperature, and galvanic skin response. This data can be used in conjunction with movement data from an accelerometer to improve accuracy and reduce false positives.
Lia combines this data with contextual data gathered through the app and data that the user inputs in order to construct a more complete picture of the user’s nausea.
Types of data needed for detecting
& understanding nausea
Lia needed a channel to collect as much of this data automatically as possible. I looked into various wearable trackers to collect the body data required, however, many of them do not contain a galvanic skin response sensor which is central to detecting nausea. Therefore I chose to design a tracker that could do this.
Before designing the form of the tracker I investigated all the possible placements that could effectively detect all three biometrics. I
down-selected to positioning it over the rib cage as it not only effective for detected but would also be discreet and unobtrusive.
Tracker placement exploration [Click to enlarge]
More on the design coming soon!