Cure Xchange Challenge: Health AI for Good
NiADA (Non-invasive Anemia Detection w/AI)
What is the name of your solution?
NiADA (Non-invasive Anemia Detection w/AI)
Provide a one-line summary of your solution.
Monere has a unique, patent-pending app for rapid anemia screening using only smartphone images.
If your solution has a website or an app, provide the links here:
www.monere.ai
What specific problem are you solving?
Today, the world sees one maternal fatality every two minutes. That's 262,800 deaths a year. Although most of these deaths occur in the developing world, this death rate is increasing in the United States for both African American and Caucasian women, and most of these deaths are preventable.
PPH (post-partum hemorrhage) is one of the leading causes for maternal death. Because anemia is a dangerous condition that reduces the oxygen-carrying capacity of blood, anemic pregnant women are at high risk for tissue hypoxia, morbidity, and death after PPH.
Anemia is a "silent killer" as testing is rare and its symptoms are non-specific -- often leaving it undetected until it is too late. The World Health Organization (WHO) estimates Iron Deficiency Anemia (IDA) puts 37% of pregnant women (45 M) globally at risk of PPH and maternal death.
Beyond pregnancy and childbirth, anemia strikes 571 M reproductive age women costing 4% of global GDP through the impact anemia has on overall health and illnesses.
It is estimated that 269 M children under five years of age are at risk of impaired cognitive development from anemia – both from conditions in-utero and after birth due to malnutrition. In the USA, purely anemia diagnoses account for 5,900 non-maternal deaths, 623,000 emergency visits, and 2.5 M primary care visits annually.
The cost of care for an inpatient with anemia is 1.5 times higher than for a patient without anemia. Cancer and chronic kidney disease (CKD) patients experience 30-90% upstaging due to anemia causing frequent hospitalization or death.
Today, anemia can only be detected when an individual meets with a medical professional and is tested for anemia (hemoglobin). All medical professionals lack tools for effective, accessible, and non-invasive screening. This lack of accessibility translates into a lack of proper prevention and monitoring and translates into unnecessary cost for health insurance and immense pressure on healthcare infrastructure.
If an effective, non-invasive and easy-to-use anemia screening and monitoring solution is available, anemia could be managed at home or at a point-of-care location, translating into better health and the prevention of serious illness.
What is your solution?
Monere is building and testing NiADA, a smartphone app, which provides a 30-second-screening and regular monitoring solution for anemia at scale. The app uses a custom computer vision deep-learning algorithm built 50,000+ connected datasets of 1) lower eyelid images, 2) medical history and 3) lab-tested anemia indicator (hemoglobin level) collected from our partner hospitals.
The solution has two main components: a mobile app named NiADA and a Data & AI platform named Andromeda, which is the backbone for NiADA.
Part one: NiADA - The user facing mobile app
- The user logs in and takes a photo of lower and inner eyelid for themselves/or other listed users in the account. Users can provide medical history which is stored in Andromeda (the HIPAA-compliant Data & AI platform described below).
- The captured image is run through NiADA's segmentation model to make sure the region of interest (ROI), the conjunctiva pallor, is present and usable.
- The segmented and validated image is sent to the anemia screening inference API in Andromeda.
- The result is received real-time and is stored for preventive monitoring.
- The user is provided with recommended course of actions based on predicted anemia severity and other relevant medical history known to NiADA.
- The user can return to NiADA to screen, monitor and manage their anemia.
Part Two: Technology note: NiADA is developed using react-native libraries so that it can be used with both android and iOS phones. TensorFlow Lite is used for embedded AI for ROI segmentation if a mobile device has enough computing resource present to use it.
Andromeda- Data & AI Platform
The Data & AI platform, Andromeda is a backend ecosystem.
- The Andromeda platform hosts de-identified sex, age and demographically segregated lower eye lid images; lab test results; other relevant medical history (i.e. pregnancy status) using standard encryption and following HIPAA standard.
- Data collectors, doctors and administrators use a limited-access mobile app to collect data and validate them, thus producing quality connected data.
- Quality-checked connected data is input to a machine learning training pipeline to produce a novel semi-supervised transformer-based regression model for anemia screening and is published as an API-endpoint.
- A separate and custom natural language processing (NLP) model is used to provide recommendation for managing anemia based connected metadata.
Technology, architecture & possibilities
Andromeda is built on cloud (both AWS and Azure) infrastructure with GPU compute and Python as programming language. It facilitates all backend operations required for NiADA to work. Every single feature on NiADA starting from login to logout interacts with different API-endpoints inside Andromeda ecosystem. These API-endpoints are protected under security keys only to be used by known subscribers.
So, these API-endpoints are designed to be available to subscribe for integration from other app/websites who want to add regular anemia severity indicator to their existing health and wellness app/website/ecosystem.
Who does your solution serve, and in what ways will the solution impact their lives?
1. Our beachhead target customers are pregnant women. The most common cause for maternal death is postpartum hemorrhage (PPH) for which anemia is the underlying cause. An estimated 37% percent of pregnant women are anemic and hence are at risk of PPH. This can be prevented if anemia is detected during pregnancy. The absence of regular screening and monitoring of anemia makes PPH management at child birth hard, leading to one maternal death every two minutes globally.
Beyond death, unmanaged anemia during pregnancy can result in premature delivery, low birth weight in infants, and can cause developmental delay in children. Today, maternal deaths are rising in USA --where we see the highest rate of maternal death among all developed countries.
Future Data's solution, NiADA, enables for the first time non-invasive anemia detection & monitoring. Those at risk -- or with a known case of anemia -- can monitor at home their hemoglobin to facilitate proper nutritional planning and medical intervention to increase iron and thereby prevent PPH.
2. Beyond childbirth, anemia plagues reproductive age women and children under five all over the world. Globally 571 M women (WHO, 2023) and 270 M children under five are anemic. Iron deficiency anemia leads to diminished oxygen-carrying capacity in red blood cells, which in turn diminishes energy efficiency, causes loss of ability to focus during school years, reduces work capacity and productivity, leading to an estimated cost of billions in lifetime income - mostly in reproductive age women. In the absence of a regular anemia monitoring system to diagnose and facilitate the appropriate public health nutritional supplement program and treatment strategy, anemia is only diagnosed when it is too late.
3. Growing senior population all over the world and about 16 M in the USA are living with anemia and decreased quality of life. Unmanaged anemia in seniors causes frequent visit to hospital (2.8 M visits due to anemia only) and 6,000 deaths with anemia as primary cause.
NiADA is a must-have solution to detect and monitor anemia at home or at a point-of-care health center for proper timely intervention and nutrition planning for a better quality of life.
4. Anemia is a common condition in cancer and chronic kidney disease patients which compromises life expectancy and requires prompt diagnosis and regular hemoglobin level management for improved clinical outcomes. Without proper anemia management, there can be a 65% increase in mortality in cancer patients. But only 40% of cancer patients get treated for anemia. NiADA will be extremely useful for monitoring anemia just with a cellphone app.
5. Blood transfusion patients - Nearly 21 M blood transfusions happens in a year -- in the US alone. Unfortunately, 31% of all blood transfusions occurs without a hemoglobin test because a test in many hours or days is simply not useful. NiADA will help medical practitioners take more informed decision for blood transfusion with real-time anemia (hemoglobin) testing.
6. Each year, an estimated 6.8 M people in the U.S. donate blood. All of these donors are tested before donating for hemoglobin with a fingerstick test that pricks the finger to extract blood. Unlike the fingerstick process, NiADA is non-invasive, real-time and painless. This will improve pre-donation hemoglobin screening by reducing risk of infection, cost and pain, while improving accuracy.
How are you and your team well-positioned to deliver this solution?
Two of our three cofounder are reproductive age women and personally experienced negative effect of anemia. A huge number of our female friends reports tiredness, occasional dizziness but attributes them to usual overworking of women in general. They know that their female children are also affected by anemia as they hit puberty without an easily available tool to manage it at home. The intervention waits for severe and unfortunate events. Our journey for developing NiADA started with the extreme experiences faced by of two women who are Mou's close childhood friends. One of then, Evelyn, toughed it out until she passed out and became immobile and rushed to hospital special care unit for blood transfusion . She had a hemoglobin of 5, acute anemia for months. The woman , who is high level government official , was experiencing extreme tiredness to do anything and obesity for years, only to find out severe anemia when she visit a doctor for operating two tumors .
This prompted us to run surveys among our female friends ( who are all reproductive age women) to understand the anemia prevalence reported by WHO as a global burden of disease , in a personal context , inside and in the periphery of our our own community.
We are long time participator-leaders who work on social issues related to women's rights for a better life and also on different community science projects through a non-profit we founded few years ago. Our friends, many of whom are public health professionals, run many community driven organizations which hold health camps , awareness-building workshops among women and other underrepresented communities in which we occasionally participate or fund.
Friends & these organizations , during our primary market research period provided us with valuable input for the community driven solution design for our product , NiADA.
Based on research and feedback , our solution includes a "moderated community" feature for providing support and anemia awareness for willing NiADA users, especially for pregnant women and patients fighting chronic disease like cancer where they can exchange their experience on our platform or share those in other social media.
Which dimension of the Challenge does your solution most closely address?
In what city, town, or region is your solution team headquartered?
Lehi, Utah, USAWhat is your solution’s stage of development?
Prototype: A venture or organization building and testing its product, service, or business model, but which is not yet serving anyoneIn which of the following areas do you most need partners or support?
Who is the Team Lead for your solution?
Mou Nandi
What makes your solution innovative?
Our solution approaches the Anemia surveillance problem in a novel way. Our solution encompasses all of the following properties. These make our solution first of its kind to be available at scale.
- Novel AI technology - NiADA uses novel AI algorithm built based on thousands of sex-segregated data collected from hospitals with eye images and corresponding hemoglobin level and other medical history for precise screening. Future Data's NiADA employs custom denoising algorithms for eye images collected using consumer smartphone devices and a dedicated app for proper data collection.
- Easy to use - a smart phone app that takes a photo of inner eyelid to show the result.
- Non-invasive - does not require blood test for Anemia screening.
- Accessible - usable in both internet and no-internet areas using mobile friendly model.
- Convenient - no travel to lab is required just for an anemia screening.
- Inbuilt monitoring - shows history and trend at patient level and also at location or program level and at the demographic group (age, sex, pregnant, menopausal) as applicable for the customer/user.
- Community oriented - featured with bringing, motivative & connecting users in preventive programs online.
- Scalable, extendable & low-cost - grows with use as backend ecosystem is based on a separate Data & AI platform. Scalable technology supports low-cost service.
- Integration-ready - Every feature of the solution is implemented as an API endpoint, making it suitable for integration-based subscription for other health and wellness platforms and insurance apps.
It is most desirable by the pregnant and reproductive age women, cancer, CKD and other chronic disease patients as we put anemia screening and monitoring in the palm of their hand, empowering them to take charge of their quality of life(QOL) with effective preventive care, addressing the underlying root cause for most post partum hemorrhage (PPH). It is a preferred way for doctors, insurance, schools, NGOs and govt agencies as it promotes preventive healthcare, better nutrition as medicine, lowers cost of healthcare , better patient outcome by reducing unwanted side-effects with anemia drug use and reduces pressure on our already strained healthcare infrastructure.
"Future Data" pioneers access to medical screening at home with the use of smartphone and novel AI technology starting with providing an effective anemia management & care solution. This is our first product. Our underlying Data & AI platform is architected to extend beyond anemia management and unlock the potential of preventive healthcare by bringing medical screening in the palm of our hand.
How does your solution address or plan to address UN Sustainable Development Goal 3 for Good Health and Well-Being?
Our solution directly addresses the problem of one maternal death every two minutes.
The no.1 cause for maternal death is PPH (post-partum hemorrhage) for which Anemia is both the underlying cause and effect creating a vicious cycle.
Anemia, frequently resulted from iron deficiency, is a dangerous condition that reduces oxygen carrying capacity of blood. WHO estimates, Iron Deficiency Anemia (IDA) 37% of pregnant women (45 million) globally and all of these 45 pregnant women are at risk of PPH.
All maternal deaths due to hemorrhage with underlying cause of anemia can be be prevented in the first place.
We provide an effective, non-invasive and easy-to-use anemia screening and monitoring solution to manage the condition at home or at a point-of-care location to facilitate proper nutrition and birth planning and avoid death.
Beyond childbirth , anemia plagues reproductive age women (15 -49 years old) 571 million women's by jeopardizing their health and wellbeing, lifetime income and overall quality of life (QOL) around the world. NIADA can provide an effective way for these women measure and monitor anemia and to have a personalized nutritional plan through government and non-governmental organizations with the support of a digital community provided by NIADA.
Describe the AI components and underlying data that powers your solution.
NiADA uses multiple AI components to support a successful user journey for screening and monitoring anemia real-time resulting in a personalized nutrition plan to achieve higher quality of life (QOL).
Proprietary AI models:
1. Validation of captured photo of inner part of the lower eyelid - photo is validated using a proprietary AI model to extract two regions of interest (ROI), palpebral conjunctiva and sclera part of the eye. Palpebral conjunctiva( inner lower eyelid) is used for hemoglobin level detection , and sclera( the white part of the eye) is used for white balancing.
2. Detecting hemoglobin level and anemia - The extracted region of interests along with other medical data ( age , sex, pregnancy status, medical history) are passed through the hemoglobin(Hb) inference AI model to predict Hb level .
3. Recommendation on next steps - The Hb result is used with medical history through a recommendation AI model to provide users with possible next steps.
4. Data augmentation model is used to generate synthetic data to remove imbalance collected data based on hb level , age and sex categories.
Proprietary data sets for training the models:
NiADA is the user facing app that is built on and uses the proprietary data platform which hosts the data we collect from our partner hospitals .
We have built a dedicated data collection ecosystem , Andromeda, which includes a separate smartphone app to collect and validate data in an organized way to reduce human error in the process and a data access and storage platform to provide selective access through API. The platform hosts both organic and derived data created as part of the whole model building and recommendation process .
We have a data collection operation team in place to collect data daily from hospital outdoor and indoor using the data collection app . We currently host more than 45000 connected datasets and it is growing at a rate of 250 datasets daily.
Each data unit consists of at least two lower eyelid images ( left and right eye) , lab-tested hemoglobin value , sex, age , pregnancy and menopausal status , some contextual information like outside temperature , last water intake , time travelled to lab , travel medium and other medical history lie cancer/kidney disease existence and any medication taken by the patient.
Andromeda is used by data collectors and doctors.
Several automatic validation steps are part of the app to ensure all required data components are collected for each patient.
We also employ human-in-the-loop data validation process to produce curated data . Doctors validate collected data using the Andromeda app to accept or reject the data collected . This doctor's validation process is gamified using a test & reward model where the doctors predict the hemoglobin level before they see the actual lab-tested value for accepted data. This data enables us to test model accuracy wrt human doctors.
This rejected data is used as adversary examples for data augmentation.
How are you ensuring ethical and responsible use of AI in your work? How are you addressing or mitigating potential risks in your solution?
At Monere we are committed to build an ethical and responsible solution by employing data privacy and security rules following the guideline of HIPAA, GDPR and Digital Personal Data Protection Act, 2023 of India.
NiADA enables user to use our system in two different ways, securely and privately:
All data are encrypted while in transit and at rest.
1. Spot checking of hemoglobin level feature does not require storing of user data in the system. Data used for hemoglobin estimation is de-identified.
2. Monitoring and management of anemia requires users information to be stored for trend analysis (either email id or unique ID to separate user data from others) for personalized services. All information stored by the users are encrypted while in transit (https protocol) and encrypted at rest on AWS cloud storage. Every cloud service can only be accessed with proper authentication keys. We have followed the guidance "Architecting for HIPAA Security and Compliance on Amazon Web Services, 2023 " provided by AWS to implement PHI protection in AWS.
3. Recommendation based on medical history data uploaded by users feature uses AWS search on its own virtual private network, separated from other instances.
4. For data collection for model building - We collect individual consent for each data unit collected using the app. We have gone through partners hospital's ethics committee review for our informed consent for data collected for building the model and we ensure anonymity in collected data by design.
What are your impact goals for the next year and the next five years, and how will you achieve them?
Monere will transform anemia screening and monitoring, thereby reducing dramatically its prevalence and illnesses caused by undetected and treatable anemia.
Our beachhead market is India given our deep connection in the community and high prevalence of anemia. We are we are working on a proposal that would allow us to screen 480 million women in India through the Anemia MUKT (Free) Bharat (India) government program.
By the end of 2024, we aim to get our NiADA app into the hands of state health and education departments in the most populous state and districts in India. We hope to have screening available to 105 M women in one year -- 20% of the anemic women in India.
In five years we aim to cover most of India and will be established in the US market for obstetrics and -- ideally -- be an app sponsored by insurance companies as part of their cost containment and preventive care. We will continue to sell directly to governments, and will focus on countries with a high prevalence of maternal mortality.
With the increase in screening, more women will know they are at risk -- or already suffering from preventable and hard-to-diagnose anemia.
What type of organization is your solution team?
For-profit, including B-Corp or similar models
How many people work on your solution team?
Total - 23 (18 fulltime + 5 parttime).
Founding team - 3 members (1 engineer, 1 data scientist and 1 medical doctor, Pathology, MD)
India founding team - 1 member.
Advisor - 1 member
Engineer - 3 members
Data scientist - 1 member
Cloud System Administrator - 1 member
Data Collector - 8 members
Sales Manager - 1 member, part-time
Lawyer - 1 member, parttime
A medical board - 3 doctors, parttime
How long have you been working on your solution?
We have been working on our solution for more than 10 months now.
I started exploring the idea in last December 2022 and put together my founding team of three.
We started PMR (Primary market research) in January 2023.
We established partnership with hospitals in India and started collecting data in April 2023.
We published our first model for anemia detection on September 1st, 2023.
We started doctor's validation of our model on September 4th, 2023.
We started limited pilot testing for our product in October 2023 at one hospital.
What is your approach to incorporating diversity, equity, and inclusivity into your work?
Our leadership team comprises of 60% women, 80% person of color.
As a longtime proponent and practitioner of DE&I at workplace, we strive to hire women in the team. Currently we have 6 women in the team, out of total 20.
While hiring through the HR agencies, we instruct them to advertise our passion for bringing more women to high paying, hi-tech, challenging and interesting jobs. While sorting resumes, we make sure we pay more attention to unconscious bias before discarding a woman's resume.
We are solving a problem that affects mostly women, women's perspective is such a valuable tool for our business to succeed.
What is your operational model and plan?
Culture:
At Future Data, our vision is to be a global leader in Digital Health & Wellness by leveraging cutting-edge AI models to revolutionize medical screening and improve healthcare efficiency, reduce costs, promote access to preventive healthcare and enhance patient care by harnessing the power of connected health data. Our mission is to use AI & technology for social good starting with innovative smartphone based non-invasive anemia screening. At Future Data, we harness the value of open communication and innovative & inclusive ideas , foster curiosity and prioritize teamwork to create impact.
Distributed yet connected team:
Our team is decentralized, globally distributed yet connected using the power of technology while keeping the operating cost low to achieve positive cashflow fast.
Two-third of our leadership and founding senior engineers are in USA , which gives us strategic advantage of access to wider investors and network for global market expansion.
The rest of our leadership and our technology team are in India , making it possible for us to run continuous operations around the clock.
Partnerships for quality solution:
We are partnering with our prestigious university alumni groups to hire best talents for our technology and S&M team.
Our initial data collection partnership with multiple hospitals was formed in India given that the contracts are easier to avail and giving us the opportunity for smoother transition into paid piloting mode in the same hospitals.
We also partner with doctors to create a medical board to ensure quality of source data to deliver the best result.
On the Sales & Marketing side, we are partnering with private hospitals , insurers and other health and wellness applications and websites to create revenue channels beyond our own app, NiADA. Our solution is designed to integrate with other solutions seamlessly.
Effective prioritization for timely delivery:
We prioritized automating the data collection process from hospitals at the very beginning of starting our operations. This enabled efficient utilization of human resources and asynchronous validation and curation of collected data into connected data to create our novel AI algorithm.
Advantageous community ties in India , product market fit :
While pregnant women, globally, are our beachhead users, India is our initial market . More than 50 -80% of reproductive age Indian women are anemic. With proper nutritional planning , the condition is preventable. And yet an at home or at point-of-care screening result based management of anemia is intractable. Government , NGOs, private healthcare providers are in dire need for an solution like NiADA to help them manage this public health crisis. Choosing India at our initial market is already showing success through multiple paid piloting contracts with state ministries.
Planning for expansion & follow-on market early:
As we complete penetrating into the Indian market , we are working on getting funding to expand the team go beyond. Outside investment will enable us working on GTM planning for USA and North American market as these markets are substantially different.
What is your plan for becoming financially sustainable?
Using to our deep network , it was easier to get approved by hospitals in India to collect the data and setting up partnership for piloting. Anemia a huge problem in India and Govt of India has special budget for combatting anemia at population scale. So we embark on marketing & paid piloting of NiADA in India first through govt ministries, private hospitals and schools( health camp) using a B2B and B2G revenue model.
We have excellent traction in Indian State Ministries where they have dedicated budget for reducing anemia prevalence among women , adolescent girls and children under five years of age. We are currently working with 5 states ( Ministry of health in West Bengal , Maharashtra , Haryana , Andhra Pradesh , Ministry of Education in Arunachal Pradesh and West Bengal , Ministry of Women and Children Welfare in West Bengal ) to design the paid piloting program in public school and public hospitals and rural primary care centers . These pilots only will create a substantial revenue stream of estimated USD 250k by quarter 2, 2024 .
We have also solid leads in two biggest hospital chains and national cancer institute in India for piloting NiADA in their outdoor and inpatient settings , especially in pregnancy ward and in nephrology departments.
We plan to break into USA market with B2B ( partner with insurances and other health wellness app) and also experiment with direct to consumer model targeting pregnant and reproductive age women initially. We are in process of getting funding to hire appropriate human resources to plan our USA GTM policies efficiently.
We are bootstrapping our project currently and in process of raising angel investment that will help us go through this initial piloting phase and concluding a few key hires to expand and attain cashflow break-even by the end on 2024.
What are your current operating costs, and what are your projected operating costs for the next year? Please include human capital estimates.
We strategically keep a considerable part of our development team in India to keep cost down. The technology leadership and a few key positions are located in USA.
Our current burn rate is USD $10k/month
Our requirement for scaling R&D and legal expenses will bring that to USD $18k/month burn rate for the first two quarters of the next year .
Next year Engineering, Product & Infrastructure + Data Science ( 30% USA, 70% offshore ) human resource need -
1 Head of product development( 10 -15 years startup experience)
1 Product lead ( previously engineer)
1 Product analyst
1 Engineering lead
4 front end developers react.js
4 Aws/azure engineer (platform)
3 Backend API engineer
3 GenAI/Search/ML engineer
1 senior(10 year) , 2 junior data scientists
Applicants can request and receive funding at a minimum of 50k and maximum of $100k. How much funding are you seeking to continue your work in 2024, and how did you select this number? What would you use this funding for? Funding is limited; please consider carefully the right amount to request.
We are seeking $100k in funding for the year 2024 to
- Hire key persons as state in the previous question
- Cover legal expenses
- Attain cashflow break-even
- Experiment with consumer app
- Raise Series A to fund scale-up
We are currently in fund raising mode and reaching out to angels for a total of $500k in funding for the next year . The funding scene for women and women of color has always been tough and it is tougher now.
The Cure Residency will provide winners with seed funding, mentorship, lab space, mentorship, educational programming, and networking opportunities. How do you imagine this opportunity will help support your work? Which aspects of the Cure Residency would you be most excited about?
We are hoping that funding , mentorship to help us mature our go-to-market policy and networking within the health-tech community will help us grow in United states.
We would
- be part of a diverse and vibrant network for any help building effective GTM strategy for global market.
- be part of an innovative peer group where we can participate in meaningful exchange for improving the solution .
- be able to attract exceptional and passionate talent to work with us.
- get us some regulatory help for the process of registering our App as SaMD(Software As Medical Device) in FDA classified list, this classification is quite new, so that it could be comparatively easier to introduce in North American market.
- connect ourselves to potential funding opportunities which are dedicated to solve women's health issue problem as Anemia is one.
- get funding for the pilot phase to get us to revenue generating stage and attain cashflow break-even.
Being part of a strong network excites us most as that will open up opportunities for other aspects like further funding , business development help to set us up for success in near future.
Solution Team
-
Ms. Mou Nandi Cofounder & CEO, Monere
to Top
What is the name of your organization?
Monere