AI for Ai
One-line solution summary:
A medical social network utilizing artificial intelligence to help improve outcomes for patients with rare autoimmune diseases.
Pitch your solution.
AI for Ai is a social network designed to support the millions with rare autoimmune disorders while providing critical data for scientists, healthcare providers, and pharmaceutical companies to pursue research for improved treatments and cures.
With over 100 different unique classifiers, autoimmune diseases are complex to diagnose and manage, and frequently patients with one autoimmune disorder will develop more than one over the span of their treatment. Over 80% of all autoimmune cases are diagnosed in women, disproportionately affecting racial minorities. Because of fragmented healthcare data and studies, many patients turn to online forums such as Facebook or Reddit for medical advice.
AI for Ai allows patients to share their symptoms and support others affected by autoimmune disease, while Natural Language Processing (NLP) and Computer Vision (CV) consolidates data and makes recommendations to healthcare providers and scientists on areas of exploration to track trends in disease progressions and causation.
Film your elevator pitch.
What specific problem are you solving?
Autoimmune disease affects 9.4% worldwide (Lerner, Jeremias, & Matthias; 2015), with an estimated 1% of the population suffering from rare autoimmune diseases. It is one of the leading causes of death and disability among young and middle aged women (Walsh et al, 2000).
Autoimmune illness research is extremely fragmented, leaving patients with chronic illnesses, expensive medical bills, and difficulty finding effective treatments. As a result of the lack of medical clarity, many patients must seek help elsewhere in the form of online forums. For example, the Kawasaki Disease Support Group on Facebook has 2,700 members, many of whom detail every aspect of their children’s illness out of desperation for help.
Sample sizes among current research in autoimmune trends are typically small, with some having as few as 20 participants. As a result, autoimmune treatments are extremely varied and causation is not clearly known.
Both of these factors play into AI for Ai’s design in a cyclical relationship. We provide the ability for patients to document their symptoms and medical treatments, which then supply researchers and healthcare providers with samples and data for their research. This research is then given back to the patients, and the cycle continues, until cures are achieved.
What is your solution?
AI for Ai is designed for use by two audiences: patients and providers. For patients, the app acts as a social support group, providing the ability to connect with and support other patients, track their symptoms and disease progression, as well as connect directly with their providers and read up on new research and findings.
The interface for providers (scientists, researchers, healthcare providers and pharmaceutical companies) allows these users to access data scrapers and NLP-based AI systems to identify promising areas of exploration in identifying disease progressions, environmental and genetic triggers, treatments and eventually cures. AI for Ai also scrapes and includes publicly available data from the NSF and NIH, which is the largest public funder of biomedical research worldwide.
Patients, providers, scientists and regulators create profiles and consent for the services provided by our platform individually. They can decide what information they want to be public on the platform (i.e. they can choose not to disclose their full legal name outside of verification purposes on the platform). All patients sign a Health Insurance Portability and Accountability Act (HIPAA) waiver in the United States and comparable waivers in other regions.
Who does your solution serve, and in what ways will the solution impact their lives?
To start, AI for Ai focuses on patients affected by rare, life threatening autoimmune disorders, estimated to be roughly 2.4 Million english speaking patients globally. This patient community costs more than $2.5B annually to treat with ongoing medications and diagnostic testing. As we scale, the platform will expand to include all individuals afflicted with autoimmune illnesses (Ai), as many patients with rare autoimmune diseases frequently are affected by more than one autoimmune disorder.
This community is underserved across a number of aspects. Autoimmune illnesses are a leading cause of death and disability among young and middle aged women in the United States (Walsh et al, 2000). According to the National Institute of Health, 80% of individuals with autoimmune illnesses are women, and conditions like lupus strongly affect people of color far more so than their white counterparts (Lim et al., 2014). Access to healthcare among racial minorities is far lower than for caucasians, indicating that vast swaths of the population with autoimmune illness do not have the access to care necessary for these diseases. Even more alienating can be the price of autoimmune illness medications-- one study found that Crohn’s patients spend roughly $30,000 on medications and related medical care in their first year of diagnosis alone (Park et al., 2019).
One of AI for Ai’s team leaders, Jenn Halweil, is a member of the community that we serve. As a person with Crohn’s Disease, whose immediate family is affected by both hereditary spherocytosis and inflammatory bowel disease, she provides an outlet into the common struggles that come with having an autoimmune illness.
Online forums and support groups for autoimmune patients have also been a beneficial outlet for identifying common needs and complaints. Due to the fragmented and inaccessible nature of autoimmune medical research, these forums include extensive descriptions of common struggles and symptoms that have been used to refine AI for Ai’s model and design.
Our target community needs access to research, medical professionals, and a community of patients to help identify more effective treatments and eventually cures. Using our NLP software, we can make this research and treatment possible, eventually refining the data to help uncover cures.
Total Addressable Market (TAM) = The number of individuals globally diagnosed with an autoimmune illness.
Serviceable Addressable Market (SAM) = English speaking patients with readily available internet access.
Serviceable Obtainable Market (SOM) = 2.4 Million patients affected by rare, life threatening autoimmune illnesses in english speaking communities with readily available internet access.
The above figures are the annual recurring costs of treatment for this patient community. Beyond the opportunity to help improve patient outcomes, we can also save patients and medical providers billions in per annum expenses by identifying more effective treatment protocols and cures.
Which dimension of the Challenge does your solution most closely address?Unlock collaboration among patients, scientists, and health care providers to improve patient outcomes
Explain how the problem you are addressing, the solution you have designed, and the population you are serving align with the Challenge.
AI for Ai is built upon the foundation of collaboration. Patients with rare autoimmune illnesses track their symptoms and medical information, allowing for scientists and providers to formulate research and suggestions.
Demographically, women and underrepresented groups have a history of being overlooked and harmed by the medical industry. Although many of these groups have been denied access to research and support due to systemic stereotypes and the rare nature of their diseases, AI for Ai unites these communities and creates the connections necessary to find effective cures, by for example, facilitating clinical trial diversity or new research studies.
In what city, town, or region is your solution team headquartered?New York, NY, USA
What is your solution’s stage of development?Prototype: A venture or organization building and testing its product, service, or business model.
Explain why you selected this stage of development for your solution.
AI for Ai is in the prototyping stage of development. We recently won the Women in AI (WAI) 2021 Global Hackathon providing us with funding as well as research and resource support from Intel.
Our team currently consists of a mix of industry leaders and product engineers, as well as data scientists focused on Natural Language Processing (NLP) and Computer Vision (CV).
MIT's Solve funding will enable us to construct and test the AI elements of our interface, before expanding into a regional New York City pilot for patients with Kawasaki's, Lupus, and/or Inflammatory Bowel Disease.
Who is the Team Lead for your solution?
Which of the following categories best describes your solution?A new application of an existing technology
What makes your solution innovative?
Online forums such as Facebook and Reddit have amassed tens of thousands of members seeking help for their rare autoimmune illnesses out of desperation. These individuals chronicle their symptoms and even give away their personal health data in the hopes of finding help and support.
AI for Ai is an innovative alternative to these forums because it analyzes this data to provide better treatment pathways and cures. This is a completely novel approach to a support group, as we offer not just social and medical support, but also an outlet for patients to express their concerns and receive concrete solutions.
The very nature of AI for Ai catalyzes the opportunity for future breakthroughs in the autoimmune field. Numerous studies on autoimmune disorders have small sample sizes, sometimes as low as 20 patients, and the ability to reach out to others on AI for Ai will help ensure that these sizes can grow, diversifying and increasing the power of autoimmune studies.
The platform will also help uncover new research opportunities. For example, our team leader, Jenn Halweil, was able to discover new treatment options for her illness after realizing she and her cat were on the same medication and diagnosed with the same disorder, and that hundreds of patients on a Facebook forum for Crohn’s Disease were similarly on the same medications as their pets, despite no research exploring this correlation.
Describe the core technology that powers your solution.
The backbone of AI for Ai’s system is data collected from disparate resources, aggregated and matched with the most efficacious audience. Natural Language Processing (NLP) is used in search functions, language translation, and chatbot surveys. Recommendation Systems will be used to suggest resources to the patients and providers. Matching Algorithms will be used to match clinical trials with the patients.
In AI for Ai, we use NLP to “mine” patient medical information. When patients enter their recent symptoms, medications, treatments, medical tests, or clinical trials, our NLP recognizes the language and identifies statistically relevant trends. This allows for the medical information to be sifted into components that can be utilized by researchers and companies to conduct research and develop medications.
This NLP distinguishes us from a typical forum for sharing medical advice. On sites like Facebook or Reddit, where patients often turn to for help, the massive amounts of data are often of no use to medical professionals due to sheer volume and accessibility issues. In employing an NLP, we are able to put this information to use to benefit researchers, and more importantly the health of patients. As we scale we intend to additionally deploy computer vision technologies to assess medical records with visual data.
Provide evidence that this technology works. Please cite your sources.
The primary driver of our breakthrough is derived from Natural Language Processing (NLP).
NLP has been used since the early 1960s as a “chatbot,” with scientists seeking to create advanced AI capable of fooling humans into thinking it was a human too (Bassett, 2018).
In more recent years, NLP has been explored as a way to recognize voices and spoken language. Google’s recent real-time closed caption service, Live Caption, uses a form of NLP to both understand speech as well as record the speech as words on devices (Google). This NLP technology has also been extremely promising in real-time translation services, such as Google Translate. New live translation features allow for an individual to speak in one language, and the device to “speak” it back into another language using automatic speech recognition natural language processors.
The most similar NLP to AI for Ai’s NLP is a text mining natural language processor. Text mining uses NLP AI to transfer text to data that can be used for scientific data analysis and research (Rajman & Besançon, 1998). Text mining is used by the government, insurance companies, retailers, and more to sift through expansive data.
Our NLP uses text mining to convert patient symptoms and records into a usable form for research and innovation. As we scale beyond our pilot, we intend to also build upon existing computer vision frameworks to mine visual medical data such as colonoscopy photos, skin lesion photos, and eye imaging. Computer vision also dates back to the 1960s and has since been employed extensively across many industries including medical devices and social networks.
Please select the technologies currently used in your solution:
Does this technology introduce any risks? How are you addressing or mitigating these risks in your solution?
Patient privacy is essential to the platform. Users do not have to use their real name for their User ID, and can choose what information they share with other users on the platform. Those user classes are extremely varied so what a patient shares with a doctor can be different than what they share in a forum discussing diet with other patients for example. All of the information needs to be heavily encrypted and data security will be a significant cost moving forward. We may explore utilizing blockchain technologies to create distributed ledgers that are more digitally defensible.
A significant focus of the platform is creating content to educate patients and medical providers to improve outcomes so that patients do not expect a holy grail from sharing their data while providers continue to research cures. This will likely include discussions on inclusion and building trust with marginalized groups, who have frequently been harmed or discriminated against within dominant healthcare systems.
Select the key characteristics of your target population.
Which of the UN Sustainable Development Goals does your solution address?
In which countries do you currently operate?
In which countries will you be operating within the next year?
How many people does your solution currently serve? How many will it serve in one year? In five years?
Current Number: 528.
This is the number of people who so far have agreed to be beta testers for the application.
One Year: 42,000.
Our pilot will start in New York City. Year one our focus will be expanding this pilot to english speaking urban areas in the US, United Kingdom, Canada, Australia and New Zealand.
Five Year: 2.4 Million globally including non-native english speakers.
This represents 1% of all autoimmune sufferers globally and the percentage that is affected by rare autoimmune illnesses.
The above numbers are all based on patient / provider access. Additionally, our technology will help patient families as patient symptoms are improved and cures are uncovered, and it will save the pharmaceutical and health insurance providers billions in treatment and advertising costs that they can reinvest towards curing a wider array of autoimmune disorders.
Five years from now, our grand ambition is to have identified causality and cures for certain rare autoimmune disorders, but in the absence of that outcome we would settle for being able to measure and document improved patient outcomes and reduced disabilities while we chase cures. Our autoimmune patients will be able to supply the data necessary for scientists to examine potential markers for causation within populations, which still are not completely understood in the world of autoimmune illnesses.
Fundamentally we are creating a new model of medical research, whereby the patient is an active participant in improving systemic outcomes and uncovering breakthrough treatments.
What are your impact goals for the next year and the next five years, and -- importantly -- how will you achieve them?
There are approximately 200M+ people in the world who suffer from autoimmune illnesses. Our initial focus is on the 2.4 Million who have rare autoimmune diseases and then we expand that out over several years to more common autoimmune illnesses. Frequently patients with an autoimmune disorder will have more than one, so this should not be a large jump from our initial product offering and go-to-market strategy for scale. Our initial funding will be from grant programs, competitions, and challenges such as this one. After that we’ll begin to onboard pharmaceutical and insurance clients, and raise VC funding to expand the platform. Once we have maximized our impact for autoimmune illnesses and rare diseases, we could potentially expand the technology into support for diseases more broadly like cancer and diabetes, but this would be a very long-term goal, and would only occur if we scaled the company to be a global player in the market, i.e. a “google of healthcare” that is a public-private partnership with significant global support. The fundamental breakthrough here is consolidating and marrying the wealth of patient journaling that occurs on social networks with medical data and use natural language processing and artificial intelligence technologies to help begin to understand disease progression, treatment improvements, and hopefully eventually causality and cures for these diseases. We are choosing to apply it first and foremost to critical, life-threatening rare autoimmune diseases with the intention of expanding it to aid additional patients as we grow.
How are you measuring your progress toward your impact goals?
Patients having improved treatment outcomes including significant reduction in symptoms, improved quality of life, improved access to cutting edge treatments and medical programs, and extended life span.
Identifying cures to these diseases.
Reduced medical costs for patients and providers to meet the outcomes described above.
As 80% of autoimmune sufferers are women, and the majority are from underrepresented groups, we are also tackling the UN SDG Goals of 5) Gender Equality - women have a longstanding history of being neglected by institutional medical providers in terms of participation in research studies and codified biases such as hysteria. (10) Reduced Inequalities - similarly many underrepresented groups have been neglected in medical trials and overlooked in medical data or lack proper healthcare access and support. (9) Industry, Innovation, and Infrastructure - Healthcare and pharmaceutical manufacturing are fundamental infrastructure that is key to the wellbeing of our global society and we would be saving these sectors money and expanding their impact. Overall we are focused on (4) Good health and well -being and intend to prioritize (17) Partnerships for the Goals - rapidly scale our platform across geographic borders, as our founding advisory team is international in origin.
What type of organization is your solution team?
For-profit, including B-Corp or similar models
How many people work on your solution team?
At present we have 5 part-time staff, with 4 advisors. Our WAI Accelerate program meets weekly evenings and weekends.
How long have you been working on your solution?
Under 1 year
How are you and your team well-positioned to deliver this solution?
Jenn Halweil - Team Lead - Electrical Engineer turned Story Engineer and Serial Entrepreneur. As an autoimmune illness sufferer herself, she is focused on attracting patients to the platform, attracting resource support, as well as uncovering new treatment approaches and cures.
Camille Eddy - Project Co-Lead - Mechanical Engineer turned Product Engineer, and globally recognized speaker focused on diversity and inclusion. Camille met and introduced President Obama at a keynote function, and intends to lead operations overseeing product and team development, as well as attracting diverse advisors to the project.
Stephanie Bell - Biomedical Engineer turned UI Designer. She is a member of Ladies that UX, and is leading Design and Interface for our patient portal.
Jennifer Lin - User Experience Analyst and Content Strategist. She is leading design and interface for our provider portal.
Suparna Pawar - Technology Stack and Data Science Analyst. She is leading the backend architecture and business intelligence aspects for AI for Ai, including HIPPA and WCAG 2.0 compliance.
Additionally, we are being mentored by Jenn Glenski, an AI machine learning scientist who focuses on Natural Language Processing, Dr. Sunny Shuoyang Zhang, who is focused on innovation adoption and building more inclusive AI programs, Maria Parysz, ethical AI leader of a global data science community: Kaggle Days, and Elizabeth Haines McGee, Director of Innovation and Engagement at Intel.
What is your approach to building a diverse, equitable, and inclusive leadership team?
Our founding team is already extremely diverse representing Caucasian, Asian, African, and Latin demographics. Jenn Halweil and Camille Eddy, the project leads, are globally recognized leaders in the diversity and inclusion space. They met originally at NASA doing social media coverage for the Jupiter Juno probe insertion. Jenn Halweil is a Luminary Fellowship recipient to advance women-led entrepreneurship, DLT Talent recipient to advance inclusion in blockchain, and a recipient of the Forbes Next x 1000 award for her work scaling a social impact educational media company focused on diversity and inclusion in tech innovation. Camille Eddy has received numerous awards including having met and introduced Obama for a keynote presentation. As we continue to expand our team, diversity and inclusion will be a fundamental marker of our KPIs, rather than a press talking point like many existing companies. It is also extremely important that our user base have varied demographics as one of the main challenges of existing medical research and pharmaceutical trials is often the small, homogeneous demographic of participating patients.
Is your team led or managed by a person with a rare disease?
Our team is led by a patient with autoimmune disorder. Her immediate family (mother, brother, aunt, and grandmother) suffer from a rare form of autoimmune anemia known as hereditary spherocytosis that requires the removal of the spleen and/or ongoing blood transfusions for life, and leads frequently to additional complications like bilirubin gallstones or coincides with additional autoimmune disorders including inflammatory bowel disease.
Do you primarily provide products or services directly to individuals, to other organizations, or to the government?Organizations (B2B)
Why are you applying to Solve?
Our team lead, Jenn, experienced first hand the frustration of navigating healthcare systems through four different doctors and initial misdiagnoses, resulting in her eventual hospitalization. As a team of engineers, we believe in the power of technology to unlock improved pathways for rare disease patients who suffer from autoimmune illnesses.
Autoimmune illnesses are chronic, disproportionately affect women and underrepresented groups, and are a leading cause of death and disability in young and middle aged women. As such, we are excited about building solutions faced by these disinvested communities. AI for Ai promotes community and connection among rare disease patients and their advocates, and unlocks collaboration among patients, scientists, and healthcare providers to improve patient outcomes and identify cures.
The Horizon Prize found us through the aid of Instagram's advertising algorithm, after our team won the Women in AI 2021 Global Hackathon. So it seems only fitting that a community presently leveraging big data and the power of social networks, help elevate a medical social platform that is leveraging big data to improve patient outcomes. We're thrilled at the opportunity to work with some of the brightest minds in innovation to combat increasing rates in autoimmune illness worldwide.
A horizon is defined as "the limit of a person's mental perception, experience, or interest." Winning this prize will enable AI for Ai to improve and expand the horizon for patients suffering from chronic, debilitating, and life threatening autoimmune disorders, to create a more healthy and inclusive world for us all.
In which of the following areas do you most need partners or support?
Please explain in more detail here.
Human Capital - Expanding our board of advisors to include industry leaders in healthcare and ethical AI.
Financial - Seeking mentors as we develop our pitch materials and begin our search for a CFO or interim-CFO. We also need to identify the most effective payment and payroll services.
Legal - Identifying advisors to help us navigate HIPPA, privacy and disclosure regulations and disability compliance.
Technology - Adding server architects and additional software development and data analysis roles to our team. Choosing which technology platforms we could utilize to prevent us from re-inventing the wheel as we roll out our beta test.
What organizations would you like to partner with, and how would you like to partner with them?
CyberSecurity at MIT - To help protect patient data from hacking exploits.
The MIT Center for Collective Intelligence - To help us harness crowd data to address the complex challenge of identifying more effective treatments and outcomes for autoimmune patients.
The MIT Sloan Center for Information Systems Research - To help us better understand the implications of a digital transformation of medical data.
We also anticipate with other Horizon Prize solution participants such as this one: Digital Services for autoimmune patients by Saad Hasan in Bangladesh.
Noubar Afeyan, who serves as an Advisor to MIT Solve and is the Co-Founder and Chairman of Moderna, would be one of our top choices to advise our project. As would Eric Schmidt, Board Member of Alphabet Inc, whose subsidiaries include Verily Life Sciences, utilizing technology to better prevent, detect, and manage disease.