OSLER AI

Democratizing access to healthcare through Artificial Intelligence and Blockchain for early disease detection, diagnosis, symptoms progression prediction and to connect patients with medical specialists.

The main objective of our project is to help save lives through technology, ensuring an assertive medical diagnosis and offering an economically accessible service for all people. OSLER AI is a software that uses image classification technology and predictive analysis to generate fast, accurate and low cost medical diagnoses using X-rays, CT scans and Convolutional Neural Networks algorithms for Deep Learning (a field of Artificial Intelligence), this to reduce the rate of medical error and help doctors to completely innovate the healthcare they provide. Likewise, the software has its own virtual medical assistant called WILLIAM which is available for free to the patients and which by using Natural Language Processing and Speech Recognition, is able to offer quality digital primary medical follow-up to a patient 24 hours a day, giving recommendations for care, prevention and soon, storing all the information on the symptoms and their progress in a virtual medical record protected by Blockchain technology. Finally, the user will be connected with medical specialists near their area according to their possible illness and individual needs, forming a health support network between registered doctors and patients and optimizing processes.

The world population is increasing, we are almost 8 billion people living on this planet, and if the COVID-19 pandemic has taught us anything, it is that the health systems were not prepared enough to face, control the virus and meet the demand for doctors needed to keep the processes running properly.

Chest X-rays are the most widely used imaging tool for detecting and managing pulmonary diseases such as Lung cancer, Pneumonia and COVID-19. However, did you know that ⅔ of the world’s population lacks access to medical diagnoses through their radiological imaging?

According to the World Health Organization, each year there are 2.6 million deaths in low-and middle-income countries alone due to unexpected events associated with errors in medical diagnoses and the issuance of incorrect prescriptions. All died as a result of a disease that could have been treated and cured if it had been detected on time. 

In our country, 33 million Mexicans lack access to the health system and some families cannot afford to keep a good medical follow-up, especially in rural communities, because they live far from hospitals and the resources in their area are too limited. 

Disparities in access to health care are one of the biggest problems in today's society, they can be caused by high costs, not being insured, geographic area, race, ethnicity and many other different aspects. Resulting in the decline in the quality of life of the people in a community and inequality in the individual development.

In OSLER AI we strongly believe that health should not be a luxury but a right. That is why our mission is to bring our technology to rural communities and low-income families to improve access to healthcare services in parts of Mexico and the world where qualified Health personnel are limited. 

Currently, we mainly have two direct groups of people that our project will help: Families and Doctors. Starting with Families; They can count on our service to easily receive a correct medical follow-up with William (our virtual medical assistant) and use the registered information as a tool to create a medical record and find the right specialist within their budgets and close to their area. While for doctors it allows them to rely on our Image Classification algorithms for a better effectiveness during the differential diagnosis stage and the Natural Language Processing technology as a way to track the progress on symptoms of their patients bringing innovation in their assistance and reducing for example: the rate of 12 million Americans that are misdiagnosed each year in the United States according to a new study published in the journal BMJ Quality & Safety.

As well, we want to partner with hospitals, universities and health and research centers to grow and validate our databases and expand the capacity of medical assistance and correct follow-up that can be provided to patients with chronic diseases to help minimizing the 71% of all deaths globally each year.

We definitely have a cause, we have been close to people who unfortunately have lost their lives due to misdiagnosis and malpractice. Additionally, we perfectly know how stressful it feels when we do not receive the medical attention we deserve for being tight with money. We realized that a tool is needed to improve accessibility and delivery of healthcare, so we got down to business right away.

When we begun with the software's development, we instantly understood how important the validation of future users is for a good digital solution, that is why we reviewed research studies and interviewed different doctors to ask them about detecting diseases at an early stage, what they believed was currently failing in the health system and what they thought was needed to solve the problem.

Since July 2021 we have partnered with the America Solidaria Foundation which aims to create community development projects for children and teenagers in vulnerable situations and we had the opportunity to align Osler AI with the UN Sustainable Development Goals.

On September 23, 2021, our project concluded the program: TecLean Explora, a business incubator for high-impact Startups by the Tecnológico de Monterrey, one of the most important universities in Mexico. During the incubator, we did polls for potential patients that showed a tentative trend in people wasting too much money monthly in healthcare without that many options of their budget and that the 66% of the respondents feel that the medical care provided by their last doctor could be improved.

In February 2021 we started doing research on Neural Networks and digital image classifiers with Deep Learning.

In March, we developed the first functional version of the prototype using the Streamlit’s framework, Python and ML / DL libraries. With this first version, we were able to differentiate pneumonia cases from normal cases by binary classification with a 96% accuracy.

Here is an example of our training process:

In April, the prototype was improved, making it possible to differentiate cases of viral pneumonia from bacterial pneumonia, with the help of databases of digital X-ray images from recognized universities.

In May, COVID-19 cases were included, different architectures of convolutional neural networks were tested, and models were trained to improve prediction levels and confusion matrices.

Additionally, the first version of WILLIAM was developed, a virtual medical assistant that uses Speech Recognition, Natural Language Processing and a corpus of medical recommendations related to COVID-19 and other diseases.

Currently, work is being done on the integration and final deployment of the web application and in the development of a mobile application in order to be able to bring OSLER AI to those rural or marginalized areas that lack internet or the service is limited.

OSLER AI uses Convolutional Neural Network algorithms for image classification and disease detection based on X-rays, CT scans and previous trained models. Additionally, it incorporates an AI Healthbot which by using Natural Language Processing and Speech Recognition, allows users to interact with a virtual medical assistant to monitor and follow-up on the symptoms or medical conditions and to receive recommendations for their care. Patients will be able to have personalized electronic medical records encrypted with Blockchain technology and have the possibility to share that information through a smart contract with their treating physician via our platform. The application will allow patients to be connected with specialist doctors close to their area of interest with Geolocation Technology, and if they require it, they will be able to receive medical attention through telemedicine and electronic prescription. 

Some of the technologies that are part of the project are mentioned below:

• Image classification and disease detection: Tensorflow, Keras, OpenCV, PyTorch, Transfer Learning and Core ML.
• Natural Language Processing and AI Healthbot:  Natural Language Toolkit (NLTK) library, Tensorflow, and BERT.
• Speech Recognition:  PyAudio library, and Google Speech Recognition API.
• Prediction of diseases and symptoms: NumPy, Scikit-learn, and Facebook-Prophet.
• Bigdata and real-time data:  PySpark, and Koalas library.
• Prototype: Streamlit framework, Python, Flask, HTML, CSS and JavaScript.
• Mobil Technologies: Kotlin (Android), and Swift (iOS).
• Storage: MongoDB.
• Blockchain: Ethereum Blockchain. However, we are exploring different blockchain technologies to reduce storage costs and smart contracts.
• Geolocation Technologies: GPS, Google Maps Geolocation API, iOS Maps and Core Location API.

In Mexico there are 128 million people and 33 million of them lack access to the health system. In addition, according to the Secretaria de Salud Federal (Federal Health Secretariat), in our country, there are currently 277,287 doctors practicing their profession, this means that there are only 2.1 doctors per thousand inhabitants.

Our solution will have great potential to accelerate processes, make them feasible and provide doctors with the necessary digital tools to offer a good medical follow-up and facilitate connections with patients.

By the end of 2022, we expect to have 30,000 patients on the platform and 100 doctors actively offering their services. By the end of 2023, we plan to expand and reach 1,000,000 registered patients and 3,500 active doctors with specialities such as Radiology, Oncology, Pulmonology, Neurology and Cardiology.

Our main goals for 2022 are to launch the commercial version of our software, validate it in the market, create a marketing campaign, grow in social media, having at least 30,000 active users on the platform and raise 60,000 dollars through a crowdfunding campaign, seed capital and what we obtain from competitions.

In June 2022, are expecting to release the commercial version of the application. We are determined to empower families to take control of their health and well-being, and in the long term, to be the number one digital platform in Mexico and in the world for connecting doctors with patients by involving Artificial Intelligence technologies to facilitate health care delivery which is very needed today.

Our impact metrics for measuring the project's progress are:

• Compliance with SDGs goals: 3 - “Good health and well-being” (3.3: Fight communicable diseases, 3.4:  Reduce mortality from non-communicable diseases and promote mental health and 3.8:  Achieve universal health coverage) and 10 - “Reduced inequalities” (10.3: Ensure equal opportunities and end of discrimination).
• Compliance with Key Performance Indicators (KPIs).
• Customer satisfaction surveys.
• Customer acquisition costs.
• Number of monthly downloads.
• Number of monthly user registrations.
• Growth in social networks.
• Financial indicators, such as Internal Rate of Return, Net Present Value, Return of Investment and Operating costs.
• Reduced mortality rates.

1. Financial barriers: Resources are required to hire specialized physicians in order to validate the databases. 

2. Our age can also be another barrier: We have seen that in the world of social entrepreneurship many times you are judged beforehand for being young and people think that because of this reason, you cannot consolidate your projects. That is an erroneous view of the past, today there are many digital platforms that allow us young people to expand our knowledge about new technologies, consolidation of startups and that empower us to be agents of change.

3. The medical databases that we currently have are public, however, it is necessary to expand them to be able to detect diseases such as some types of cancer, which is why reliable information is required from sources such as: clinics, hospitals, universities and research centers, that in most cases are private or difficult to access data.

4. In addition, halfway to the finish line, we still have to finish programming the official version that is going to be available to the public and also, to build a strong brand image on social networks.

Click on the names to see our personal portfolios :D

Mariana Gonzalez - She loves programming and has always been interested in solving problems that arise in her community with a sustainable approach. She developed her first video game at age 12 about fighting the Global Warming and has been learning about new technologies ever since. She has participated in different social innovation competitions and was among the winners, has taken leadership workshops for women, pitch, design thinking, STEM and entrepreneurship and collaborated with organizations with the aim of supporting children and teenagers in vulnerable situations.

Violeta Barrera - She considers herself a cooperative and creative person with a lot of interests. Some of the main ones are designing and editing as well, healthcare related topics, is a balance. She was selected to participate in a summer program on pharmaceutical chemistry at UNAM, leaving a growing curiosity about medicine topics, reflected in Osler AI. Editing is one of her passions and thanks to various Adobe courses, she has been able to help in Osler's social media and bring our message to others.

Maria Jose Flores - She is a passionate and hardworking person. She tends to find solutions fast, is very creative and knows when to ask for help when she needs it. She has always liked to fight for what she wants and believes in, and even if Maria Jose encounters obstacles along the way, she doesn't give up. She is very passionate about medical sciences and in the future she sees herself working at Osler AI leading the Medical Team. She is currently studying French and Chinese and took a course on computational thinking at MIT. As well, she is going to continue learning more languages and about other cultures to expand the project internationally.

• America Solidarity Foundation

After having participated in the national meeting CONCAUSA 2030, our project is part of its network of young activists, whose objective is the consolidation and impact on the Sustainable Development Goals of the United Nations Organization and support children and adolescents in a situation of vulnerability with community development projects.

• Socialab Mexico, Disruptivo TV, Samsung Responsabilidad Social México (Samsung Social Responsibility Mexico)

These three organizations are allies and collaborate with us through advice, mentoring, workshops and feedback.

• Zona EI (Innovative Entrepreneurship) and Instituto de Emprendimiento Eugenio Garza Lagüera

We are part of their community of entrepreneurs. They provide us with mentoring, advice and feedback.

We are fully aware that in life there are two options: let problems remain as they are or do something about them. We want to be part of the Change by integrating our passion and knowledge in technology to the health sector to create a sustainable network that benefits even those who unfortunately, are marginalized from receiving proper medical healthcare, like for example: The “Chontal Community” in our State.

OSLER AI is led by three 16-year-old girls, passionate about STEM and with a strong social conviction. Our main objective is to reduce inequalities, bring equity and digital health to all families regardless of their social status, especially oriented to marginalized, rural, sub-urban communities and vulnerable or low-income populations.

In 2021 we participated in 5 entrepreneurship competitions, such as: Enactus Hub Puebla (BootCamp), ChangeMaker Day - Social Innovation BootCamp by the Tecnologico de Monterrey, TEC Lean’s Explora Pitch Night, CONCAUSA MÉXICO 2021 and Samsung Solve for Tomorrow Mexico 2021, and in each of these opportunities we were among the winners, even though the fact that on some occasions we competed with university projects or established ventures. Therefore, we have not missed any potential opportunity to grow, learn, lead and strengthen our project.

https://drive.google.com/file/...

Winning The HP Girls Save the World Prize would allow us to take OSLER AI to the next level, letting us to hire medical specialists to help us validate and certify the databases and additionally, improve the Text Corpus for NLP and medical guides in such a way that we can give more accuracy to the predictive models as they are certified by qualified health sector personnel.

OSLER AI is a social impact, sustainable and inclusive entrepreneurship with the purpose of revolutionizing the world of medicine and bringing digital health to all families. Through the strategy of prevention, prediction, participation, personalization and precision.

A study published by  Frost and Sullivan concluded that 90% of women are responsible for making decisions regarding the health of their family environment. In addition, they have a representative force of 70% within the medical sector, being also the main caregivers of children and the elderly. Therefore, we infer that our largest number of users and beneficiaries of the application will be made up of women who, in many cases, are the pillars of the home. Additionally, we have contemplated being able to detect diseases such as: Breast Cancer and Ovarian Cancer, among others, with our AI and in a timely manner.

In Mexico, as well as any other country in Latin America, one of the greatest challenges facing any government and its community is public health, coupled with the consequences of the pandemic. This directly impacts the social and economic aspects of the population. Mexico has only 2.1 doctors per 1,000 inhabitants, and by September 2020 it was the country with the highest rate of deaths of workers in the health sector globally with a total of 1,320 workers as a result of their exposure to the SARS-CoV-2, even with 30% more than the USA.

Therefore, OSLER AI will help the public and private health sector in the prevention, diagnosis and fight against diseases, offering medical specialists the technology for less exposure to diseases by optimizing the interaction between patients and treating physicians. At the same time; reducing costs, waiting lists to be attended and improving the quality of life of the population.

On the other hand, we will create a foundation to educate and train women and girls to empower them in STEM disciplines and in entrepreneurship in order to provide them with tools that allow them to improve their social condition and turn them into the future leaders of the digital economy, and the most outstanding may be incorporated into OSLER AI, if they so wish.

Winning The Pozen Social Innovation Prize will allow us to hire software engineers, data scientists and Marketing in order to accelerate the development of the platform and be able to contribute quickly so that families can take control of their health and well-being.