Utu: Patient-Driven Self Care

Having developed alopecia areata at the age of six, my sister waited years to receive an accurate diagnosis. In even the largest city of Tanzania, doctors did not have a lot of experience with the condition and continued to misdiagnose it as a fungal infection. On top of that, the lack of understanding and awareness of alopecia areata in the community made it difficult for us to find support. This ultimately delayed access to treatment and management plans and left my sister feeling isolated as she navigated this complex world on her own. 

After an accurate diagnosis, my sister and family felt helpless due to the lack of guidance, support, and information. We weren’t sure what actions could be taken to support her. Over time she started noticing patterns of how and when patches of hair would appear and disappear, but all of the doctors shrugged it off as random insignificant occurrences. As she got older, and increasingly frustrated with the scant and extremely technical research available, she started tracking her condition and its progression through photos, videos, and notes - like a journal - in the hopes of self-identifying ways that she could manage and improve her condition. 

After collecting data on various sporadic interventions, and over a long period of time, it became exponentially difficult to associate the causes and effects of her efforts and almost impossible to compare her experience to others because of the lack of community. She wanted to be able to share her own data to support others living with the condition. 

On the other side of the problem, the data she was collecting is invaluable for alopecia researchers. People like my sister - with a personal commitment to finding a solution - currently have no way to share that data in a privacy preserving way. Researchers of rare conditions often have a hard time accessing research participants, have biased participant groups, and are unable to leverage data quickly to see the effects of interventions. 

We recognize that this doesn’t start and end with the 2% of the global population that suffers from alopecia areata. Patients suffering from other rare conditions face similar challenges: 

1. Patients with rare diseases often do not have access to the information, guidance, and recommendations they need to optimally manage their conditions.

2. Researchers - who are in a position to know what to do - do not have enough data to help them effectively move forward. 

We know from working in healthcare, and health technology, that patients and researchers have the same goal. We believe we can bring them together to both improve the quality of life for patients with rare diseases and advance research on conditions that affect tens of thousands of people around the world. 

Utu is an open source, two-sided care platform for patients with rare diseases and researchers.

For patients, Utu is a smart and personalized care journal. The platform: 

1. Provides an AI-powered, personalized health assistant that supports condition management to improve quality of life. We achieve this by:

   1. Fully guaranteeing patient data privacy by working offline and encrypting data

   2. Using algorithms to consider all self-documented condition progression and external factors (interventions) to identify likely patterns and recommend behavioral/lifestyle changes

2. Involves patients in solution finding by allowing them to contribute to research (if they choose to) through:

   1. Sharing their data with research institutions and getting feedback, including recommendations and advancements that will help them manage their condition

   2. Sharing their data and experiences with other patients with the same condition in order to build supportive communities around them

As a fully decentralized platform, patient data is entirely patient owned and managed on Utu. Patients can choose how much to share, with whom to share, and the levels of anonymity and privacy they wish.

For researchers, Utu is a direct link to the populations they wish to serve. The platform:

1. Pools data from willing and extremely committed patients for research purposes

2. Allows researchers to respond and communicate with patients by asking them questions, encouraging (constructive) behaviors, and even answering their concerns

3. Allows researchers to contribute to the core, open source technology and even leverage it to fit their specific needs or condition of interest

The platform is delivered through mobile applications (android/ios) for patients and through a web application for the researchers and their institutions. For technical contributors, the open source platform will be available on Github.

Utu serves two populations:

1. Patients with rare conditions

2. Researchers studying rare conditions

Patients

We believe everyone has the right to access high quality healthcare, regardless of where they live or what condition they have. Utu holistically supports the journey of a patient with a rare condition like alopecia areata. The platform improves quality of life by providing personalized care and clinical decision support - powered by the patient’s own data. As patients self-document their condition, as well as its progress, triggers, and outcomes, they are able to get powerful insights and recommendations from AI algorithms. For patients who feel isolated navigating the complex world of their condition, the platform allows them to connect with others suffering from the same condition, share helpful tips, and feel a sense of community. 

Critically, the platform also allows patients to take ownership of the problem and action towards solutions by contributing to research and collaborating with researchers. Data on Utu is entirely self-owned and self-managed by the patient, giving them control over how much they share. 

Researchers 

Researchers of rare conditions benefit from the Utu platform by getting access to a pool of highly motivated patients and families that come from diverse backgrounds. Researchers are able to engage these individuals in projects that help advance their understanding of conditions affecting individuals around the world. For alopecia areata researchers in New York City, leveraging data generated and shared by patients globally could have a significant impact on their knowledge of the condition and the effects of various interventions. 

In addition, as an open source platform, researchers are able to contribute to, and benefit from, the constantly growing technology to expand their research to other conditions. 

Our team is passionate about using technology for social good and is committed to ensuring that everyone, everywhere has access to high-quality healthcare. 

First, we have a deep connection to patients living with alopecia areata. Our team has personal connections with family members who suffer from the condition and have spent their entire life navigating a broken care system that wasn’t built to support them. We understand the challenges faced by patients with the condition and have been able to learn what works for long-term care and management. 

In addition to our team’s health experience, we also have vast knowledge and expertise in digital health and clinical decision support. We have been working for the last four years to bring open source, AI-powered healthcare to clinicians and patients across the African continent. We understand deeply the challenges that exist in developing and delivering solutions that aim to improve healthcare delivery and services. Our current team has technical experience in software development, system design and architecture, data science, and artificial intelligence. 

Our team also has strong partnerships with medical experts, researchers, and health stakeholders across East Africa. We are working closely with Kilimanjaro Christian Medical Center (KCMC), the largest dermatology clinic in Tanzania. Additionally, we have an ongoing agreement to collaborate with Ifakara Health Institute (IHI), the largest health research institute on the continent, and have agreements with multiple district governments to pilot and deliver public health solutions. 

Our team has been working for a number of years to advance digital health across Africa by supporting and optimizing decisions for healthcare providers. As we have expanded our reach, three things have become increasingly clear: 

1. Patients with rare conditions, particularly in low-resourced and underserved communities, are suffering. Healthcare providers are currently ill-equipped to support them on their care and management journey.

2. Patients are rarely in control of their own health data. 

3. Researchers are missing out on large amounts of information about patients with rare diseases. 

We are applying to the Challenge because we believe there are solutions to these problems and want to expand and extend our work to holistically support patients with rare conditions, empower patients to take action, and contribute to the advancement of science. We’re also deeply excited about the potential partnerships we can learn from and collaborate with through applying to the Challenge. We recognize that by working with a globally minded organization, we can make an impact not only in East Africa, but also across the world.

Our approach is built from the perspective of the patient and all technology is built around empowering the patient to take the best care of themselves and participate in solution finding for a problem they know so well. It is fortunate that the patients are equally, or even more, motivated to find a solution and to become more than a statistic on global health charts. We are able to leverage years of working in low resource/data settings to build personalized and adaptive AI algorithms that can support patients even when they can’t get access to the specialists they need. This allows them access to a 24/7 on demand support system that gives undivided attention to its single patient, while learning from all other similar patients.

Additionally, being open source will allow researchers to expand the application of the solution through additional features, or even additional diseases that interest them and their organizations.

We’re starting by focusing on the rare disease alopecia areata. In a year, we plan to be serving 500 patients across three countries with the condition, and we plan to have at least one research organization leveraging the network. 

Over the next 5 years, we plan to expand to five other rare diseases and support 20,000 patients through the platform. Our impact goals are to: 

• Improve health outcomes and quality of life for people living with alopecia areata

• Improve health outcomes and quality of life for people living with five other rare conditions

• Lower the barrier for researchers to access and engage patients with rare conditions, and lower the barrier for patients to share their data with researchers

• Increase the impact of new research on alopecia areata (and other rare conditions) by facilitating sharing of information between researchers and patients

• Increase the support network for patients with alopecia areata (and other rare conditions)

We will achieve these goals by working closely with the most critical stakeholder: those who are affected by rare diseases. We will also establish relationships with additional stakeholders, including health organizations focused on rare diseases, research groups, Ministries of Health, etc. We regularly monitor and evaluate the impact of our solution against desired metrics, and communicate the outcomes of our work to stakeholders. This helps ensure that we stay on track.

The Utu platform relies on a few core technologies that help power and support the various stakeholders. Key technologies & techniques are:

• A peer-to-peer network of devices that allows for decentralized and patient-owned /managed data (patients are the peers in this case). The patient can choose how much data to share, if at all.

• To learn from data we use reinforcement & online learning algorithms (currently using Hidden Markov Models and Bayesian Multi-armed Bandits) to tease out cause and effect, and recommend the correct interventions.

• For personal recommendations and learning using patient data on a device, a federated learning approach helps power the sharing of the learnings from individual patients to support similar patient algorithms anonymously - without ever moving their data.

For patients who wish to share their data, they can choose to share “similar but not real data” that accurately represents them. If the patient chooses to remain anonymous, this data is synthetic and can even be differentially private.

Since we are in the prototyping phase for this solution, we are still working on clarifying and validating our business model. 

Currently, the Utu platform is free for patients to use as a personal care assistant. As the network grows in numbers and value, we intend to leverage it to generate revenue through researchers and/or research organizations. 

With an open source ecosystem, we typically take a “trusted vendor” approach and provide paid services to companies, organizations, and researchers who are to integrate clinical decision support and digital health solutions into their programming. We believe this model will also carry over to our work with Utu.

We have a multi-stream approach for financial sustainability: 

1. Our work has been, and will continue to be, partially funded by grants and/ or public financing. Given that we are primarily working in a context where healthcare is publicly funded (we are working on solutions that leverage emerging technologies that require time and validation to be successful) we plan to seek grants to facilitate research and development, evaluation, and implementation. 

2. Our current business model is built around a “trusted vendor” approach, where we provide paid services and custom solutions to organizations (such as UNICEF, etc). We will generate future revenue through long-term contracts with these types of organizations.

3. We plan to leverage the growing value of the network to generate revenue from researchers and/ or research institutes.

As an organization, we have received grants from Foundation Botnar, UNICEF Innovation Fund, and MCC PEPFAR to fund the development and piloting of AI and digital health solutions in East Africa. We have not yet received funding for this specific solution, nor have we received private investment funding.