The German pharmaceutical giant, Merck
Design Science Research, User Interviews, Wireframing, Prototyping, Android Development, Focus Groups, Usability Testing
Nairobi (Kenya) | Darmstadt (Germany)
Patients in developing countries like Kenya face many challenges obtaining quality and affordable healthcare. To address this, Merck launched the CURAFA™ initiative.
The CURAFA™ initiative consists of:
small point-of-care healthcare facilities offering medicines and basic clinical services
a set of digital health initiatives aimed at increased patient engagement
A CURAFA™ Point-of-Care Facility
Many patients in Kenya suffer from chronic non-communicable diseases (NCDs) like diabetes or hypertension, and these conditions are often caused by patients lacking awareness, accessibility, and affordability of medication or treatment.
In Kenya, due to the lack of qualified doctors and nurses, Community Healthcare Workers (CHWs) play a critical role in bridging the gap between patients and healthcare professionals.
Community Healthcare Workers:
are voluntary health workers
have limited standardized training outside the formal nursing or medical curricula
deliver a range of very basic health, promotional, educational and mobilization services within their communities
play a critical role in increasing patient engagement and health awareness in their communities
From the outset, due to CHWs' vital role in Kenya when it comes to patient engagement, they were to be the primary focus in the UX Design process.
How can we increase NCD patients’ health-seeking behavior through a mobile-based tool for Community Healthcare Workers (CHWs), managed out of Merck-sponsored point-of-care facilities in Kenya?
DESIGN SCIENCE RESEARCH
To design, test, and develop the optimal and most user-friendly mobile solution for CHWs and patients, I used a Design Science Research (DSR) methodology.
DSR follows multiple, iterative cycles to design, test and develop prototypes with users, and is used by renowned research institutions such as MIT’s Media Lab & Stanford's Centre for Design Research.
In this project, I conducted 5 comprehensive DSR cycles over the course of 6 months.
DESIGN CYCLE 1: DESK RESEARCH
Before heading to Kenya for user research and prototyping in the field, I needed a thorough background on healthcare delivery in Kenya, healthcare challenges faced by CHWs, current mobile health apps used by NCD patients, and any other information that might be helpful to understand mobile and health behavior of the potential user.
While the goal of any UX design project is to, naturally, design the optimal user experience, it is also necessary to have a proper background on the environment the user operates in.
CYCLE 2: BUILDING EMPATHY
THROUGH STAKEHOLDER INTERVIEWS
With a solid desk research background on CHWs and healthcare delivery in Kenya, it was time to validate this research with actual CHWs through field research in Kenya.
During a 3-week field trip to Nairobi and surrounding areas, I did extensive field research through user interviews with CHWs and other public health stakeholders.
FINDINGS & LEARNINGS FROM INTERVIEWS
Non-communicable diseases (NCDs) are very common
NCDs, specifically diabetes and hypertension, were the health conditions prevalent in households that were most mentioned by CHWs.
This insight was in line with the desk research - thus, not new news - but an indication that the problem being addressed is an actual one.
CHW training and experience levels differ significantly
In line with Learning #1 and Learning #6 from the desk research, CHWs in different areas and counties were trained by different NGOs and thus had varying levels of theoretical and practical knowledge. It cannot be assumed that they have a standardized background or standard certification (e.g. such as a professional nurse or clinician).
Health education is a key activity
In line with Learning #2 from the desk research, the majority of work that CHWs do in households is not curative of nature. More specifically, CHWs spend a great deal of time educating patients on dietary habits, general sanitational issues, chronic disease awareness and management, and family planning. Patients also noted that they appreciate educational content very much.
CHWs are responsible for household data collection
The county governments expect of CHWs to collect household data. The data is collected in large, paper-based logbooks with standardized fields. However, not only are logbooks archaic and cumbersome to carry between households, they also only collect YES/NO questions household "census-style" questions.
These logbooks are carried from household to household by CHWs to collect data, and they only collect high-level household data.
For example, for a household with a diabetic patient, it is only recorded that a household member has diabetes.
The patient's individual data - e.g. the blood sugar or blood pressure reading - is never recorded.
For chronic disease management, this is a major concern.
Many health conditions are still very stigmatized
Many patients with chronic NCDs such as diabetes and hypertension are hampered to seek proper healthcare due to misplaced social stigmas. These stigmas are unfortunately usually caused by a combination of common myths, traditional and religious beliefs, and inaccessibility of proper primary healthcare and health education. To combat this, patient support groups help a lot.
Patient support groups are quite common in Kenya. The value is that it creates a safe platform for patients in reassuring them that they are not alone in fighting their disease. They also have the opportunity to discuss common problems and side-effects of medication. Some support groups also purchase medication together and assist needy members.
A support group in Nairobi
Patients value in-person contact for medical advice
While all patients have access to a mobile phone - whether it's a smartphone or a feature phone - and are comfortable in using these devices, all patients and CHWs noted that for medical advice, in-person contact with a medical professional is preferred. This complements the previous insight of in-person support groups.
CHWs have access to smartphones, but patients do not
In most cases, CHWs had access to low-end Android smartphones. Also in most cases, these smartphones were actually provided by NGOs responsible for training CHWs. While the patients interviewed were definitely not a representative sample of the population, none of them had a smartphone. Most CHWs also reiterated that most patients in households they serve do not have smartphones.
CHWs visit patients about once a month
Depending on the sparseness/density of the area, CHWs visit households about once a month, and visit about 30-40 households a month (!). Since they are not paid for their work, they usually go by foot.
In very rural areas, they usually travel by "boda-boda" (motorcycle taxi).
In areas like this with very isolated households, CHWs noted that it can take 1-2 hours of travel time between each house (by foot) or 15 minutes by boda-boda.
CYCLE 3: WIREFRAMES, MOCK-UPS & DESIGN CONCEPT VALIDATION
With the insights gathered from the stakeholder interviews, I decided to prototype a three-part socio-technical concept addressing the needs of both the patients and the CHWs.
3-PART SOCIO-TECHNICAL CONCEPT
As most CHWs have Android smartphones, I created mock-ups of a simple app to record basic patient vital data such blood sugar and blood pressure.
This would enable CHWs to record patient data which they could not do through the archaic and limiting hardcopy logbooks.
I created the concept of a SMS-based group chat where patients can talk with one another about their condition and where educational info can be sent a CURAFA™ pharmacist / clinician).
As most patients do not have smartphones, an app for them would not make sense. A digital forum (group chat) where they can connect without having to travel far addresses this problem.
I created the concept of a combined monthly support group session at a CURAFA™ point-of-care facility, combined with a formal educational session of a CURAFA™ pharmacist or clinician.
While patients are comfortable with using mobile technology, in-person contact is still highly valued when it comes to discussing personal health-related issues.
CHW USER JOURNEY
Based on the interviews with CHWs and accompanying them when they visit households to collect patient data in the hardcopy logbooks, I created a user journey for collecting a patient's vital data through the proposed Android App.
A typical CHW user journey
ANDROID APP WIREFRAMES
Based on the user journey detailed above, I created low-fidelity wireframes of the NCD Vitals App using moqups. Unfortunately, due to limited time in the field in Kenya, I could not test the wireframes in person with CHWs. However, due to my in-detail and many interviews with CHWs, I was confident that this app would be in the right direction, and decided to pursue the development of the app.
Low-fidelity CHW app wireframes
DESCRIPTIVE SCENARIO TESTING: MOCKUPS & CONCEPT VALIDATION
In addition to the app for CHWs, I also wanted to validate the concepts of the SMS Group Discussions and the in-person Support Group Sessions. In order to do this, I created low-fidelity mock-ups showing SMS messages I proposed to send on the SMS Group Discussions.
I recruited 7 patients with whom I did descriptive scenario testing. In short, this type of testing consists of explaining concepts to testers (in my case, with visual aids), and then gathering feedback.
Screenshots of the descriptive scenario testing guide
DESIGN CYCLE 4: ANDROID APP USABILITY TESTING & SUPPORT GROUPS
ANDROID APP DEVELOPMENT
Based on the User Journey Mapping and the Wireframes I developed in the previous cycle, I worked with an Android developer to build a functional Android App to test in the field with CHWs.
First version CHW App
I recruited CHWs across 3 different regions where CURAFA™ facilities are located. At each facility, I recruited 5 CHWs to test the app, thus doing usability testing with 15 CHWs in total.
Before I joined the CHWs in testing the app during house visits, I briefly introduced the app and highlighted its intended purpose. Each CHWs did a “mock-testing” on one another before we headed out in the field to test with real patients.
Each CHWs was asked to:
• Register a new patient
• Measure and record blood pressure data twice
• Navigate the app to view previously recorded data
Usability testing the NCD Vitals App in diabetic's patients home
Usability testing the NCD Vitals App in diabetic's patients home
POSITIVE USER FEEDBACK
As intended, users noted that the app worked and navigated very easily. Users had no problem finding the functionalities and options which they needed to access to accomplish tasks. The fact that there were only three vital data fields was noted to further increase the simplicity of the app.
Desired data needs met
Users noted that the app managed to address the gap that the hardcopy logbooks lacked - namely to collect individualized patient vital data relevant to diabetic and hypertensive patients.
Ability to see data history sorted by patient
The logbooks "sorted" data entries by date visited, thus each time a CHW visited a household the household/patient data is entered anew.
The users noted that since you can scroll through entered patients and then view a patient's vitals data history, it is a major improvement on the logbooks.
A hardcopy logbook inherently have no way to protect confidential patient data as the data is open for anyone to view.
Thus, an app where the CHW must log in and which is password protected was noted as a major improvement by the users.
SOME IMPROVEMENT SUGGESTIONS
Complex medical terminology
Feedback: as not all CHWs have the same educational background, some medical terms like systolic/diastolic blood pressure were a bit too confusing.
Solution: additional user-friendly terms like top reading and bottom reading were suggested.
Data capture confirmation
Feedback: users mentioned that it was unclear to them after they entered blood pressure data that the data was captured successfully.
Solution: I redesigned the confirmation screen/notification that it's even clearer after data was successfully captured.
Feedback: since many CHWs visit rural households, it is not a given that data recorded can immediately be synced online.
Solution: offline functionality for the app is an important requirement, and was included in the app by the developers
Context-specific input format & validation issues
Feedback: certain entry fields and formats like date of birth and mobile number were entered differently in Kenya than anticipated.
Solution: I recorded users' preferred input formats and worked with the developers to modify the input fields.
PATIENT SUPPORT GROUPS & SMSs
Patients receiving an educational session by a CURAFA™ clinician
Signing up diabetic and hypertensive patients for the SMS Group Discussions
SMS GROUP DISCUSSIONS SAMPLE MESSAGES
At 2 of the CURAFA™ facilities, I recruited diabetic and hypertensive patients with the help of CHWs. The aim was to sign patients up for the support groups, send messages on the group over the course of a week, and then gather feedback from them again through focus group sessions.
Each session when patients visited the CURAFA™ facility consisted of the following:
an educational session where the CURAFA™ clinician gave a short talk on diabetes and hypertension management
a session where I signed-up patients for the SMS groups, and showed them how they can send message to one another on the group
a session where I showed patients which educational content and reminders they will receive over the next week
SMS Group Discussion Messages
DESIGN CYCLE 5: APP ITERATION, USABILITY TESTING & FOCUS GROUPS
Based on the feedback of the usability testing, I worked with the developers to incorporate user feedback into the app. I recruited CHWs at 2 CURAFA™ facilities - in total 11 CHWs for this round of testing.
Again, each CHW was asked to:
Register a new patient
Measure and record blood pressure data twice
Navigate the app to view previously recorded data
Screenshots of the iterated app can be seen below, showing some improvements such as:
different input formats for certain fields (left)
addition of more user-friendly terminology (center)
clearer confirmation after successful data entry (right)
ANDROID APP ITERATION & USABILITY TESTING
Selected improvements of the CHW App
PATIENT FOCUS GROUPS
Once one week passed since signing up patients for the SMS Group Discussions and SMS content, I again recruited patients (15 patients in total) at 2 CURAFA™ facilities with the help of CHWs. In total, I conducted 4 confirmatory focus groups testing sessions. I also asked CHWs to sit in on the focus groups in order to assist me with translating language feedback from patients.
Focus group testing
The main aim of the focus group testing was to:
gather user feedback of the SMS Group Discussion service
gather user feedback of the reminders SMSs sent to patients on the group
gather user feedback on the concept of a monthly in-person support group and educational session at CURAFA™
SOME FOCUS GROUP LEARNINGS
Educational content is highly valued
Feedback: all focus group participants noted the educational content provided by the CURAFA™ clinician was extremely valuable to them.
Next step: work with the CURAFA™ personnel how this in-person educational content can be properly translated into digital communication channels (SMS).
Differing technological capabilities
Feedback: all patients were able to receive SMSs, but some did not how to reply and required help from family members (especially older patients).
Next step: I suggested tech-able patients to assist older patients during support group sessions on how to use the SMS Group Discussion.
Monthly visit frequency
Feedback: patients noted that they preferred in-person Support Group sessions to take place once a month, as some have to travel far to reach the clinic.
Next step: to increase contact with patients, I encouraged CHWs to visit patients two weeks after the support group session to maximize contact.
Differing language abilities
Feedback: even between neighboring regions, local languages differed significantly and not all patients were comfortable with English.
Next steps: work with CHWs and CURAFA™ personnel to translate SMS content into local regional languages.
PROPOSED PATIENT JOURNEY
Based on the usability testing during the house/patient visits with the CHWs, as well as focus group testing with the patients, I drafted a proposed patient journey involving both CHWs and patients, incorporating the 3-part solution described earlier:
Monthly support group sessions at CURAFA™
NCD Vitals App to be used by CHWs during house visits
SMS Support Group Discussion for patients to use whenever they want
Proposed monthly patient journey
CONCLUSION & IMPACT
CHALLENGES & FINAL LEARNINGS
The market which CURAFA™ operates in make for very thin margins, and many customers have extraordinarily little (if any) disposable income. This complicates the validation of a business model for solutions designed. Thus, from a user-centered design perspective, you must ensure that the products and services that you design and develop are absolutely needed and desired by the patient in order for the patient to place trust in them. The upside of this is, of course, that you really drill down to the core desire and needs of your user and simplify the user experience as much as possible.
#2 Limited access to mobile technology
Given that the CURAFA™ initiative specifically operates in the underserved and low-resource Kenyan market, potential users in the target market naturally do not always have access to smartphones. The embracing of technological solutions is there, but very often the access is unfortunately not. From a UX design and research perspective, this often limits the “status quo” solution of “there’s an app for that”.
#3 Older patients
Since diabetes and hypertension are often caused by subpar lifestyle habits over a longer period of time – unbalanced diets, non-adherence to medication, etc. – meaning that diabetic and hypertensive patients are often elderly. As the elderly are often more uncomfortable with technology, optimizing user experience becomes even more challenging.
#4 Language barrier
While most users are comfortable in English, many interviews and sessions with patients had to be conducted with the help of a translator. Especially elderly patients (see #3) tended to be more comfortable with their own local dialect. However, all patients spoke Swahili (Kenya’s lingua franca), so should service be translated, Swahili provides an easy “generalizable entry point”.
#5 In-person contact
Building on Interview Insight #6, patients are often more comfortable to discuss medical issues in-person with one another and with a health care professional. Thus, while digital-only products and services are more scalable and preferable from a business perspective, this is not a user-centered approach.
CONCLUSION & IMPACT
I’m very proud to say that:
my project was awarded the Social Impact Award of 2019 by the Technical University of Munich School of Management Alumni Association (my alma mater)
the research results are also currently under final review for academic publication in a top-tier academic journal
I was also offered by Merck to stay on with the company to conduct the further implementation of the project, as well as to take on other UX and Business Intelligence-related tasks (see here) within the greater CURAFA™ project. After this project, I stayed with Merck for about another year.
The vast majority of testers, both patients and CHWs, were very enthusiastic about using the newly-designed offerings on a more permanent basis. The support groups sessions at CURAFA™ were embraced heartily by the patients and CHWs, and were adopted in CURAFA™’s formal offerings to patients. They continued to be held once per month. I also continued to develop SMS-friendly educational content for patients and CHWs, which was handed over to CURAFA™ and included in their offerings.
In conclusion, this was a challenging, yet extremely rewarding project. Working, designing, and testing in the field in Kenya really showed me how big a difference simple digital solutions can make to increase patients’ quality of life, and to empower CHWs in their great day-to-day work they’re already doing.