SO-KAR empowers NON-LITERATE diabetics IN DEVELOPING COUnTRIES
Diabetes is a major public health problem worldwide. According to WHO, it is no longer a disease of predominantly rich nations, its prevalence is steadily increasing everywhere and most markedly in the world’s middle-income countries. What makes it a complex public health issue is that it leads to severe complications when poorly controlled. So the challenge is to come up with a solution to help patients, especially in developing countries, better control their disease.
I worked on this project as a UX / UI Designer.
I was responsible for user research, ideation, wireframing, interaction design, visual design, prototyping and user testing.
A data scientist has been consulted to understand more about artificial intelligence and deep learning technology.
I started the research by interviewing a diabetologist from Morocco to have a global idea about the disease and understand his patients’ struggles but also to understand how doctors are struggling in helping them treat their diabetes.
This interview revealed that there are two patients categories. First are educated people who understand their condition and second are from an underprivileged population, usually not educated, and therefore are unable to manage their diabetes properly and are unaware of the consequences.
The diabetologist I’ve interviewed gave me an overview of diabetes. He explained its 3 types, how it affects the body and how it can be managed. However, I needed to dig deeper in literature to better understand the matter and better prepare patients interviews.
At first, my focus was on addressing the broader issue of diabetes management. However, upon discovering a specific group of individuals who are both non-literate and faced with the challenges of managing a complex condition like diabetes, I made the conscious decision to direct my efforts towards assisting this particular demographic.
I conducted attentive listening sessions to gain a comprehensive understanding of the individuals in front of me. I sought to delve deep into their lives, learning as much as possible about their unique stories, the everyday obstacles they face, and how they navigate life while coping with the complexities of their disease.
Next, I embarked on an exploration of how these individuals engage with technology, specifically their usage of mobile phones and the applications they rely on. In addition to conducting interviews, I conducted extensive secondary research to gain insights into designing user interfaces for non-literate users. One particularly valuable resource was Gabriel White's presentation titled « Design at the Edges – UX Design for Developing Countries ».
At some point, I was lost in my own research findings and I didn’t know where to go next. Different patients had different problems and the diabetologist I've interviewed pointed out even more. I decided to follow the design thinking process and see where it leads. I started with the empathy map to sort out all the information collected.
This project was particularly challenging because the target user is different than me, different than people around me and different than people I usually design for.
Telling the persona’s story helped me get to know the end-users deeply, connect with them and be able to know what they really need and how the solution I will design will help them overcome their struggles.
During the research process, I not only interviewed individuals with diabetes from Morocco, a developing country, but also individuals from Europe. Their experiences highlighted the overwhelming nature of managing diabetes. Living with diabetes entails frequent blood sugar checks, making healthy food choices, staying physically active, adhering to medication and insulin doses, and constant concerns about low or high blood sugar levels and potential complications. These individuals were well-educated, had a comprehensive understanding of their condition, participated in diabetes education programs, and received regular medical care. Despite all the support they had, they still felt overwhelmed. This realization raises the question of how challenging it must be for someone who never received formal education, struggles with reading, and already faces daily hardships to effectively control diabetes and maintain their overall health.
Another thing that my research revealed is that non-literate people aren’t familiar with the common design paradigms. They use mobile phones only to make phone calls, audio messaging (WhatsApp) and for entertainment (Youtube).
They rely on memorizing the location of buttons to navigate without relying on reading labels or comprehending the meaning of icons. During an interview, one individual shared that she uses emojis to identify her contacts since she is unable to read. For example, she associates a flower emoji with her daughter, a moon emoji with her son, and a heart emoji with her sister. This insightful information highlights the unique strategies and adaptations non-literate individuals employ to navigate digital interfaces.
My vision was to take advantage of what technology can offer and equip non-literate people with a tool that will help them make informed health decisions and better manage their diabetes.
I envisioned an app that will help them follow their doctor instructions correctly without having to read medicines names, remember insulin doses, or self-interpret blood glucose testing results.
Non-literate individuals with diabetes heavily rely on their relatives to assist them in taking their medication. Their family caregivers need to have a comprehensive understanding of the doctor's instructions and be present consistently. In the absence of such support, patients often resort to guessing which pill to take and the appropriate insulin dosage to inject. However, they are aware of the potential fatal consequences of injecting the wrong dose of insulin, leading them to sometimes skip their treatment for that day. Unfortunately, this practice is detrimental to their overall health and well-being.
After couple rounds of crazy 8s I came up with a solution that will help the non-literate patients take their medication on their own. It consists on the recognition of medicines according to doctors prescription. The app will also relieve the burden from the family caregiver.
Computer vision is a field of artificial intelligence. It involves working with digital images and videos to deduce some understanding of contents. Object detection is associated with Computer Vision and describes a system that can identify the presence and location of a particular object within an image using deep learning. This technique consists on a mathematical optimisation of a huge number of parameters that allow catching shapes, colors and dispositions configurations.
According to Gabriel White, it is hard for persons with low education to build a mental model of the system they are using. They are often struggling to successfully understand where they are in the system and how things are related to each other. So for this app, I tried to make the navigation as simple as possible without sub-categories to help users navigate easily.
First, the doctor scans her QR code to access prescription page and enter instructions (name of drugs, dose, how and when to take it etc). Then the patient will scan medicine boxes or insulin bottles to know if it’s the right one then get more instructions. The patient can also scan her meal to get a reminder one hour and a half later to check her blood glucose level. After that she can scan the meter so that she can get interpretations and the test results get saved.
I sketched wireframes and tried to explore designs that are simple, easy to understand and remember with the minimum of information. Having a clean interface with minimum visual clutter will help reduce users' confusion.
For this project it wasn’t about creating a harmonious color palette or a trendy interface. It was all about helping patients use the app in the most easy way. I used different colors to help them recognize buttons and remember the function of each one but also to understand if the medicine they are scanning is the one to take or not. I made sure that the colors have names in users common language to simplify the communication between patients and doctors.
I also used graphic illustrations to help them understand the small amount of text.
Another insightful information shared by Gabriel White is that people with low literacy use devices like TV, decoder or tap player and the reason why those devices are easy to use is because they have spatialized UI: each button has a distinct location, a distinct appearance and a single function.
I tried to design an interface where each button has a specific location and a specific color and does on task. That will allow users to use the spatial reasoning and easily navigate. For example, after understanding how to use the app they remember that by taping the orange button on top right corner they will get audio assistance.
The app opens the camera, so if the user opens the app by mistake she might freak out and want to exit right away. So I made sure there is a way out: I placed an exit button on the center of the bottom bar, so that the user can leave the app at any time.
People with low levels of literacy are usually nervous when it comes to technology. As products recognition and finding the doctor instructions can’t happen immediately, I added animations to indicate ongoing work. This will give constant feedback to users and will reduce their uncertainty.
As explained before colors are use in this project as a memory aid. First there is green and red which means right and wrong. They are used in the scanner frame that turns green when it’s the right medicine to take and red when it’s the wrong one. They are also used for blood glucose test interpretation: green when it’s normal, red when it’s too high or too low. Orange will be used for audio assistance and blue to start scanning
I prototyped two scenarios to test the app. First scenario, the user will try to find the pills they have to take and the the right bottle of insulin, get instructions and make sure they filled the syringe with the right dose of insulin.
The second scenario consists on scanning a dish to get a reminder one hour and a half after the start of eating to check the blood glucose level, then scanning the meter to document the results and get interpretations.
Unfortunately, due to lockdown I had to test the prototype remotely. As my users aren’t comfortable with laptops and video calls I asked a friend in Morocco to help with this task. He went to users, presented the product and explained how it works. He video called me so that I can watch them using the app.
As I didn’t want to disturb them while they where using the app and bias the test results, I kept texting my friend and sending him questions I wanted to ask and I listened to their answers.
At the end of the session I talked to them to get their feedback and know their thoughts about the app. I was glad to see how excited they were as soon as they understood how the app works.
So-kar is an app that will directly affect users' health so there is no room for errors. It is, for this reason, important to test it further with more individuals and also with health professionals to be sure that it will operate error free.
I started this project without having an idea about the final product. I only wanted to design an app for diabetics but had no prior knowledge. Therefor, I spent a lot of time asking questions and trying to find a problem worth solving. Finding a problem and people I care about helped me come up with a solution that will make a small, but meaningful difference.
People with low overall literacy have often a low health literacy and are unable to properly follow their doctor's instructions. Studies have shown that not having the needed skill to make health-related decisions affects a patient’s adherence to a treatment regimen which may decrease its benefits and this leads to poor health outcomes. This concerns patients with all kinds of diseases not only diabetes. So I think that extending this solution and adapting it to different diseases may have a not inconsiderable impact on health outcomes in general.