Role in project
Sprint Planning • Sketching • Research • Concepting •
Usability testing • User flows • Prototyping • Test planning
Teammates: Ashlynn Denny, Michelle Wang, Jacqueline Wijaya, Jason Wong
Tools: Axure, Sketch, Keynote
Deliverables: Research plan, research notes, application map, user workflow graph, wireframes, prototype
Neopenda is a medical device startup that gives patients in low-resource settings access to high-quality care. Our clients, Sona and Tess, are the CEO and CTO, respectively. They started Neopenda in school in 2015 when they had a professor who was the conduit and they started working with nurses to help improve newborn mortality.
Scenario: In Uganda, neonatal care is a predominant issue. The wards that care for neonatal infants are severely low in resources, where babies with hypothermia are treated with nothing but plastic bags to keep them warm. Our clients informed our team that the ratio of infants to nurses reaches 80:3. Because they are stretched so thin, the nurses are unable to consistently monitor infant vitals in the ward. Therefore, the nurses are forced to play a high-stakes guessing game regarding treatment. The product and platform we were designing for was a wearable device for neonatal infants. The device transmitted live vitals onto a tablet and allowed nurses to monitor vitals in real time.
Diving into each client-conducted field report helped me empathize with not only our stakeholders and the behemoth of an issue they’ve tackled, but with our potential users as well. I learned that there are 3 million newborn deaths every year, and 80% of those deaths are preventable, especially in low resource environments.
Initial client meetings
Sona and Tess put together a current platform. However, they both expressed that they felt it lacked the user experience element. After thorough discussion and shaping a research plan, we found that Tess and Sona's main goal was to make their platform more user-friendly, especially for ESL users. In addition, they wanted to become more user-focused in their operations moving forward and asked our team to create a research plan that they would be able to continue utilizing on their own after returning to Uganda.
Starting with the problem statement exercise, our clients gave us multiple options for nurses, doctors, and hospital admins, but we collaboratively narrowed down to:
Nurses need a way to know which patients are in the most danger because they want to improve care, save lives, and be efficient. Currently, nurses have too many tasks to manually take care of at once.
Considerations: It was disclosed to us that Ugandans were historically weary of western practices of medicine. Until they or their loved ones were in grave health, they would rather stick to African potions. We weren’t only crossing medical standards but cultural boundaries as well. Moving forward, our design and layout should make sense to Ugandan doctors and nurses. Sona and Tess chose an Android tablet format, due to the efficiency of bringing them from the US to Uganda. However, there were issues to address with the tablet format, such as the high tendency of theft of technological equipment in Ugandan wards.
It’s common for wards to have an average of 20 neonatal infants
The tablets needed to communicate:
Oxygen levels (SPO2)
How current nurses diagnose in neonatal care is still guesswork
Mothers stay with the infants at the ward while being compliant to the nurses’ guidance
Reports on paper aren’t efficient and the data is sometimes made up, due to the hectic environment and lack of time the nurses are given to debrief
Twins are a very common occurrence in Uganda. There are instances of wards consisting up to 30%-40% of twin infants, being named Twin 1 and Twin 2
Current format nurses and doctors prefer to help phrase interview questions:
APGAR score, a ranking from 1-10 that categorizes the baby’s health,
Alarm rings when a baby’s vitals drop or exceed a certain ‘stable’ parameter
Certain alarms go off so often, nurses begin to ignore them
Issues locating the baby in a packed NICU ward
After going through the onboarding material Neopenda had provided for us, we focused on how to present large amounts of data in an easily digestible format. Our goals with desk/domain research:
Learn about wearable technology monitoring vital signs for newborns and similar wearable vital monitors
Understand the current state of monitoring vital signs for neonates in Africa
Differentiate American vs. East African medical technology standards
We then divided specific emphasis of research. I focused on presenting large amounts of data and gathered inspiration from Edward Tufte, as well as practices in other markets like the stock market and airport FIDS. UX best practices and the iNICU dashboard view show the use of cards to display patient information as well as pertinent information about the patient without cluttering the screen.
Part of our onboarding materials was a competitive analysis conducted by our clients, which included different health monitoring organizations. We conducted a competitive analysis but looked outside of the healthcare industry for inspiration. We came across a similar platform that monitored vitals of livestock through wearable technology and first responder/EMT software. A key issue was displaying large amounts of vitals in an organized way. All platforms addressed the problem statement collectively.
We interviewed 6 users total including: OBGYN doctors, nurses who’ve worked in Uganda, neonatologists, Ugandan doctors, and one of Neopenda’s formal advisors. All users and SMEs were remote, so we were limited to phone and video calls via Skype, Zoom, and Facetime. As a team, we encountered various obstacles through the interview process. In order to solve for the limitations of testing over the phone, we emailed screenshots of the interface beforehand to get feedback on the Neopenda platform.
Our team wrote interview questions with the goal of clarifying the assumptions:
Do only doctors get access to changing threshold vitals for preterm and full term babies?
Do healthcare providers in Uganda have the same level of education/knowledge as doctors in the US?
Is English the predominant language?
Is there a universal baseline of knowledge of newborn care?
Is it too difficult to keep track of 20 vital cards at once?
Our clients confirmed that 20 different cards would be ideal due to the fact that all vitals need to be constantly monitored by the nurses for their platform. The rest of our assumptions would be verified or corrected through emails with our client and user interviews.
After interviewing users and subject matter experts of varying backgrounds, we mapped out their insights. This activity revealed specific themes around current processes and workflows, cultural considerations regarding the environment in NICUs, and pros and cons of the current Neopenda dashboard.
After our user interviews, we synthesized our findings with an affinity diagram. We split up interview notes and as a team combined insights. Through those insights we discovered four primary emerging trends:
1. Visual display over auditory: We showed our interviewees the current Neopenda dashboard and found that nurses preferred visual alarms over auditory ones, due to the distraction factor
2. False alarms: Babies tend to roll around so the device often slips off of them, which triggers many false alarms.
3. Hospital record-keeping: There were particular discrepancies between what visitors and Ugandan natives reported on patient record frequency and ease of access.
4. Location of babies: Nurses have a lot of difficulty locating and keeping track of the high number of neonatal patients in critical condition.
Multiple nurses stated that, “All four numbers are important, but it’s difficult to scan quickly.”
This is especially true when overseeing a large amount of neonatal infants. This validated assumptions in regards to the perception of the current platform. We also learned that if beds aren’t labeled in the room, the layout of vitals would not make sense. There were sometimes two or three babies to a bed or incubator, so we needed a flexible labeling and organizational structure for varying physical environments.
Due to that frequent annoyance, nurses would mute the alarms, especially the respiratory rate alarm. Upon asking the US doctors how they prevent false alarms, they explained that alarms don’t go off for a set period of time for breathing and heart rate. Both doctors and nurses stated that nurses typically know the equipment better, and are accustomed to setting parameters based on the neonate’s age.
Alarm parameters shouldn’t be easy to change by mistake, and nurses should change the alarm as the first responders.
The Ugandan doctor we spoke to mentioned that the charts are filled out regularly based on the severity of the babies’ condition and patient charts are kept nearby. However, US doctors visiting Ugandan hospitals mentioned that nurses in Uganda are too busy to record patient information. We learned that even in the US, vitals are recorded by hand. This depended on the unit but it was common practice for vitals to be recorded by hand and then transcribed onto a computer. In Uganda, there are generally no electronics available for that purpose.
From user interviews, we found that everything is manual, and that there is no access to printers in NICUs, which indicates the need for a digital platform.
We learned that patient capacity varies from ward to ward, which also influences the organizational layout of beds and equipment. Some nurses and doctors mentioned that up to five to six babies are placed on a single bed. Because of this, nurses try to organize babies by issue whenever possible, which means that the sickest babies are kept closest to each other. Overall, better organization is preferred but that isn’t always an option for nurses.
Many babies have come into the NICU wards with no name. This posed an interesting challenge for our design to accomodate a relatively disorganized ward and allow nurses to address neonates in critical condition as soon as possible.
After we synthesized our user interviews, we revised our initial problem statement to:
There is currently a lack of medical equipment, space, and staff in low-resource environments.
Neonatal nurses in low-resource environments need a way to efficiently scan vital monitoring information and find neonates in critical condition so that they can channel their energy to the child with the most need.
Our team revised our problem statement to narrow our scope on the nurses and the most immediate issue at hand: being able to scan neonatal infant vitals and find the neonates in the crowded ward. My team used four design principles to guide our designs moving forward:
My team and I performed a heuristic evaluation on the current Neopenda platform to gain insight on ongoing practices and to evaluate how we could improve their product.
Initially, our team discussed whether or not we would improve their platform, or start from scratch. However, as we proceeded, we realized that Neopenda had built their platform based on market research and the way vitals are presently being tracked.
So when our users expressed that the screens made sense to them, we decided to pivot and improve the platform rather than eliminating it altogether, leading us to concept new ideas while solely focusing on nurses within our scope.
Concepting and testing
In previous client projects, concepting was a struggle in terms of keeping the focus on the scope of the project without being feature-focused. This time around, utilizing 6-8-5s and leveraging inspiration from our client’s progress were key parts of my process. Guided by our research, problem statement, and design principles, we began to draft our initial ideas and inspirations. My goals with concepting were to address the problem statement at hand. I mapped out four different concepts within an hour time span with the hopes of solving our client’s biggest concern: how to monitor 20 neonates vitals, while efficiently locating infants in the ward of varying sizes.
Our team completed five out of the six concept tests we had originally scheduled with our users. Just like with our user interviews, we were presented with challenges including: scheduling between Ugandan and US time zones, technological constraints with concept tests relating to screen sharing capabilities, unfamiliar medical jargon used in a healthcare ward, and making sense of the differing user perspectives. We were able to make changes and continued to iterate as we began prototyping. Designing based on assumptions (not testing with Ugandan nurses) was an uncomfortable step, which since then, has changed. Recognizing that Neopenda would continue usability testing on their own, our team would come with equipped our clients with a test plan to follow.
My concept that eventually reached the concept testing round with our users is what I called, Queue Up.
What we tested
Utilized card format for baby vitals and highlighted neonatal vitals that are outside of “safe” parameter ranges
Infants who currently don’t need immediate assistance will be placed in a section labeled “stable” and with 4-5 cards on each row, rotating every 10 seconds
“Find” button that would trigger a visual alarm on the wearable device
Color and number naming system of infants
Tracking particular trends of specific infants using graphs
What we learned
Users found that the “queue” category is a useful and quick way to notice babies in need
“Find” button would be helpful for nurses locate specific infants
“Stable” category, which would be constantly changing, doesn’t allow for spatial memory
Nurses wouldn’t have time to look at or take notes on the trends page
Nurses gave feedback that the word “Queue” and “Stable” weren’t commonly used
With the word “Find” a nurse mentioned, “You don’t have to find the babies, they aren’t lost!”
What we tested
Platform displaying the layout of NICU ward
Nurses would be able to hold and place the babies represented by icons
Tapping into the individual baby icon would expand into the baby’s vitals
What we learned
This would create additional work for the nurses on the floor
Would be too complex to implement in different wards because not one NICU ward is formatted the same
Due to the lack of feasibility and efficiency, which goes against our design principles, we removed this concept after two tests.
What we tested
Card format platform that would present the four vitals important to nurses, as well as a color and numbering system
Timer that starts when an alarm goes off, tracking the time it takes for a nurse to respond to it, for accountability
Graph of the selected vital depicting trends, while also displaying the other three vitals, allowing the user to monitor the other vitals simultaneously
Tooltip which a user could click into, saving real estate and giving options to view additional information
What we learned
Users mentioned that they prefer the standard card view
Doctors preferred to have timers on the alarm to keep nurses accountable, however the nurses disagreed
Not all alarms would trigger the same response from nurses
Certain vitals cause more of a reaction and that vitals in general are correlated
Users like the tool tip to save real estate on the card
What we tested
Neopenda platform with all infant patients that fits on a single screen
Vertical orientation with the horizontal view of vitals, while utilizing sparklines
Clicking into patient row would expand details
What we learned
User feedback was polarizing
We had to take in additional context where the insights were coming from
US doctor and nurse that liked this format stated they did so because it was what they were familiar with
US nurses with experience in Uganda mentioned it would be difficult to scan in a high pressure environment and with the nurses always running around
Referencing our problem statement and design principles, making the platform easy for nurses to scan was a main focus for our design.
Synthesizing our concept tests was the greatest challenge of our project with Neopenda. After the majority of our concept tests, we realized that there wasn’t a clear direction in which our users showed patterns of preferences. To overcome the lack of direction, we synthesized other insights, utilized our domain research, along with design assumptions and ideated our fifth concept.
What we tested
Quick alert bar on the top of the tablet screen indicating which babies are in immediate need
Drop down location of the baby
Which format they are placed in (i.e. bed, incubator, etc.)
What we learned
Users thought the alerts were a bit crowded, however appreciated the ever-present alarms
Preference to see alarms for specific vitals
Specifying infant format was not useful
“Sometimes there are 3 babies in an incubator” - Nurse
With the influence of combined insights from concept tests and aligning with our clients, our team converged the initial prototype. As a team, we prototyped the nurses’ main task, scanning and finding a neonatal infant in immediate need of attention.
Our team decided to map out the workflows of every possible scenario that a user would encounter with Neopenda, which led to three workflows: New Patient, Monitor Patient, and Discharge Patient.
Prototyping and testing
Aside from direct feedback of the concepts, our users also gave us valuable client specific insights for future productions on the Neopenda device.
Using different band colors for different types of patients such as: full term, preterm, and kangaroo care babies.
Updating hardware to light up specifically for the “Find” feature
Tracking Apnea because with respiratory rates the alarm would constantly be out of parameter (babies constantly temporarily stop breathing or their hearts start beating quickly).
The Apnea Alarm would signal if the baby has stopped breathing for more than 15 seconds.
Taking into account the head size circumferences for sizing full term and preterm babies.
My team and I then took that feedback and iterated on our concepts heading into prototyping in Axure. After initially building our prototype, we began usability testing. However, technical roadblocks and difficulties caused a lack of end user perspective. We ended up interviewing four users, three nurses and one doctor, through video call. Our team also made a visit to a prominent hospital in Chicago.
Despite the constraints of remote usability testing, our team made it a priority to communicate to the users the necessity of screen sharing platforms. Suggesting solutions such as Skype, Zoom, and Google Hangouts to ensure success in usability testing. And due to the hectic schedules of the nurses and doctors, I applied my client management skills to send emails and calendar reminders to follow up.
We approached our usability test with a set of tasks and interacted with our users with probing and empathizing questions. The goals for usability tests were to:
Gather user feedback on whether or not the updated dashboard helps users scan and find vitals
Assess if the sidebar hinders user workflow
Determine how users respond to temperature and other vital alerts
Task 1: Monitoring Vitals
Task 2: Add A Patient & Set Parameters
Task 3: Viewing Trends
Task 4: Patient Discharge
Additionally, our clients asked for a list of recommendations of changes based on our research. I took this as a sign that they trusted our process, research, and ultimately, our ability as designers to guide the immediate next steps of developing their process. So a teammate and I, established a list of recommendations ranked “immediate”, “soon”, and “later” along with our final prototype.
Walk through the neonatal nurse’s journey below:
High fidelity dashboard:
Neopenda test plan: now and later
A big focus of our process with Neopenda was preparing a test plan for future usability testing on the platform. The challenge was explaining design terminology and the testing process to individuals who are not familiar with design. The goal of the test plan was to be flexible & modular for future use.
In order to accomplish that, my team and I included:
A “happy path video” with PDF instructions on how a usability test is run in a best case scenario
Set up instructions of free recording software, software to install, and setting up the prototype on the tablet with an offline viewing option in case WiFi was an issue
Explaining roles in a usability test and giving examples of different question types
Collecting data by sorting outcomes, tasks, scenarios, recommendations, etc.
By implementing our future recommendations, we were able to provide nurses in low-resource environments a way to quickly notice and locate neonatal infants in need. And as a result, Neopenda now has an improved platform to help save lives. Building from what our team was able to accomplish in just three weeks, Neopenda now has a strong foundation to continue testing and iterating. In the future, they can expand the platform to cater to doctors and hospital admins, going beyond being a monitoring platform and becoming an informational hub of neonatal wards.
My Neopenda experience
Working with Neopenda, I learned the importance of aligning initiatives with my clients from day one. Once bridging the gap of communication and developing the relationship, I learned that the product and solutions following, become more purposeful and intentional in the outcome. I applied my strengths in building rapport with clients and users, presenting to clients, and well as concepting.
Areas in which I grew:
Furthered my comfort level with having a lack of users
Designing on assumptions, as long as there will be testing with iteration in the future
Conducting user interviews, concept tests, and usability tests remotely
Introducing and teaching my design process to my clients
Iterating designs through user feedback and continuing to test those changes
In just three weeks, my team and I were able to conduct our design process and overcome technological setbacks through continued iteration and basing our decisions on user insights and domain research. We were able to make a positive impact and changed the trajectory of the Neopenda direction to better help save lives in Uganda.
My client relationship feedback: