Covengers AI: Arming doctors with Data

In a pandemic when the pathogen is novel and the vaccine development takes time, the only saviour is a drug. Therefore while different therapeutic strategies are being tried by clinicians across the globe, the reporting of this data is dependent on compilation and scientific publications, which is time consuming and not a priority for overworked clinicians. While bigger hospital chains share data amongst their clusters, such platforms are non-existent in smaller hospitals(40% in the US Hospitals) and in developing nations. Also, we observed effective therapeutic protocols from smaller government hospitals in India (mortality rate- 3 out of 600 hospitalised patients). Therefore we are creating this global, central, health data repository that can inform clinicians of best therapeutic strategies in real time. Further repurposed drugs are cheaper and widely available across the globe compared to novel drugs, for every life in every corner of the planet matters.

Statistically there is a wide range of variation (0.1%-23.7%) between nations of the number of serious cases, translating to mortality. Further a median average of 11% of all COVID19 mortalities were the frontline healthcare workers, the rane went upto 20% in certain hospitals. Considering the differences in health care infrastructure and genetics of the population would play a role in recovery from the disease, a conservative estimate of 30% attribution lies to the treatment options being administered to these patients. An knowledge and faster diemmination of an effective therapeutic strategy would have at least saved a 100,000 lives thus far. Similarly, the understanding of effective prophylaxis measures would have by the most conservative estimates prevented the loss of a third of healthcare workers, who are most needed in these times and are exposed to highest viral load. The lack of a central platform prevents us from quickly identifying and  combating any epidemic/pandemic situation now and in future.

We are developing a platform that acts as a global repository for treatment data during disease outbreaks/epidemics/pandemics. The data will be collected from hospitals as well as patients across the globe. Once this data is collected, it would be grouped, clustered and processed to deduce the outcomes via machine learning for different treatment options. This processed data will then be available to clinicians in real time to be able to quickly deduce the best treatment regimen for patients, saving lives. This system will also have the capability to generate recommendation and confidence scores for each patient based on their medical history.

Our target population is every human being on earth. However, our target population is to save those lives who cannot afford a COVID 19 treatment and where the doctors do not have access to all protocols being developed across the globe and an understanding of their effectivity. These in particular include clinicians working in developing and under-developed nations and those working in Refugee camps. 

We have done surveys with over 80 clinicians in United States, Europe, India, South America and Africa to understand how they decide on therapy protocols, what are their outcomes and will a platform such as ours help them. 80% of these clinicians and 100% in the developing and under-developed nations feel they need a platform like ours that could give them faster global insights into both prophylaxis and treatment. Their feedbacks are also being used additional features to the MVP, which are being implemented in creation of the final platform.

People in the lower economic strata in developed nations (1 in 7 Americans) and majority (70%) in developing and under-developed nations today cannot afford the costs on COVID19 hospitalization and treatment. Further Remdesivir, a drug which mildy improves recovery costs 4000-5000 USD per patient,in addition to the hospitalization costs. The patent of Remdesivir and all new drugs henceforth is viable until 2038. In contrast, Dexamethasone, repurposed for severe COVID patients costs only 8 USD per patient. Our platform relies on real world evidence for rapid identification and prioritization of such repurposed drugs, informing governments to prioritize and fund their trials.

Since humans are now armed with technology, this pandemic and every disease outbreak henceforth can be tackled more effectively by a global repository of treatment data which can disseminate information instantaneously across the globe, with statistical evidence. This approach saves lives faster, eliminates the effect of fake news (HCQ/bleach anyone?) and can streamline government expenditure and efforts in the right direction. To the best of our knowledge the only other similar initiative is, A Pandemic Initiative: https://www.apandemic.org/. They have partnerships only in the United States and UK and need collaboration with initiatives like ours working with developing nations. We are currently exploring ways to partner with them.

The idea of using Machine learning to establish co-relations and identifying patterns in clinical data is not new. Our platform is an amalgamation  of multiple technological platforms in clinical data analysis like Amazon Sagemaker/IBM Watson to be able to de-identify patient data, retrieve data from physical medical reports and analyse image data- X-Ray's, CT Scan etc. Considering the questions around data privacy, we would also be using Blockchain technology to record each conversion and trace it in future. 

We have created an MVP with a synthetic data set to establish a proof of concept: http://covengers.s3-website.ap-south-1.amazonaws.com/admin/dashboard

Humans are treating COVID19 in the same manner as they tackled AIDS despite all technological developments. We believe the bottleneck is not in the technological or scientific innovation but in the willingness to co-operate. In our surveys and discussions while every government thinks this is a logical solution, it is only time with increase in disease burden that some of them have now decided to start discussing about implementing our solution. We will create case studies out these countries, the lives we were able to save, present it on every major forum, persuade and help the other countries with creation of long term data sharing partnerships between nations for emergencies like Pandemics. The availability of this platform and underlying partnerships, will come immediately into action to save lives, in similar emergencies becoming rampant with climate change. Co-operation is why human race evolved to tackle larger animals when we were primitive and it will be one attribute that would save us.

We have devised the following strategies:

1. We are collaborating with hospitals in our primary network but that being a slow process we have decided to reach out to government officials to champion our cause and help us obtain data from different hospitals/health data aggregators.

2. We are also reaching out to health data aggregators such as 1Up Health and Intersystems for a collaboration for COVID19.

3. We have been in discussions with Government of Netherlands and El Salvador as well as Government officials in India for putative partnerships.

4. For the FDA approval, we would release the platform in two versions: the first version (only global patterns) first, followed by the second version which would go towards patient specific recommendation and therefore will be submitted to the FDA.

5. Since health data security is our top concern too, we are fundraising to employ experts to build a robust system ensuring HIPAA compliance.

6. For the financial goals- we would be operating in a not for profit mode to be able to collect donations and apply to philanthropic funds. We also aim to offer the insights generation from our data on a subscription basis to pharma and research institutions.

Immediate Goals: Considering the COVID19 emergency-following is how we have divided our goals:

1. Formation of strategic partnerships to obtain data-ongoing to 3months

2. Development of the platform- ongoing to 3 months

Long term Goals:

1. Ready deployment of partnerships for any epidemic/pandemic or a major novel disease outbreak

2. Application to rare diseases 

3. Establishment of Clinical trial cohorts and partnerships to treat rare disease patients.

The following are the critical challenges we face:

1. Collaboration with hospitals 

2. FDA approval

3. Security of patient data

4. Financial

We have devised the following strategies:

1. We are collaborating with hospitals in our primary network but that being a slow process we have decided to reach out to government officials to champion our cause and help us obtain data from different hospitals/health data aggregators.

2. We are also reaching out to health data aggregators such as 1Up Health and Intersystems for a collaboration for COVID19.

3. We have been in discussions with Government of Netherlands and El Salvador as well as Government officials in India for putative partnerships.

4. For the FDA approval, we would release the platform in two versions: the first version (only global patterns) first, followed by the second version which would go towards patient specific recommendation and therefore will be submitted to the FDA.

5. Since health data security is our top concern too, we are fundraising to employ experts to build a robust system ensuring HIPAA compliance.

6. For the financial goals- we would be operating in a not for profit mode to be able to collect donations and apply to philanthropic funds. We also aim to offer the insights generation from our data on a subscription basis to pharma and research institutions.

We are a team of 4 people and 10 student volunteers who have developed the solution thus far. 

I am a postdoctoral researcher at Memorial Sloan Kettering Cancer Center. I  developed the literature review and data analysis pipelines to establish correlations.  I have created similar hacks from literature even while using natural compounds to treat my fathers metastatic cancer, increasing his lifespan from 3 months to 13 months (we used Vitamin C- last clinical trial by Nobel Laureate -Linus Pauling, 2020 trial- scheduled by Lewis Cantley). 

I also designed the concept and am responsible for negotiations and partnerships.I head the postdoc association Careers committee and I have an experience of discussions with a spectrum of people and creating initiaves in collaboration with the students and the leadership.

Shaishav Vashi- an Engineer by qualification from one of India's top Engineering Institution (National Institute of Technology) generated the synthetic dataset and is involved in volunteer management.

Manish Gupta and Dileep Kanumuri- are software developers from India's leading technical schools (Indian Institute of Technology), having worked for Goldman Sachs- Built the platform (MVP).

Student volunteers are conducting informational interviews with clinicians/hospitals and professionals from Adobe , IBM , etc. are helping with development of machine learning pipelines.

We are a nascent initiative and currently negotiating on establishing partnerships with initiatives and governments across the globe.

Our key customers and the revenue model for each is as follows:

Pharmaceutical Industry: Data insights generated from our platform will be available to the Pharma Industry to better plan clinical trials, design better analogs for disease and also establish Clinical trial patient cohorts- All of these services would be provided to them at a cost.

The novel correlations obtained from this data would also be helpful to Biotech startups and Research institutions, giving them interesting data based directions to pursue. These would be provided to them at a subscription cost.

1. Funding via grants and donations

2. Subscription of data insights for Pharmaceutical and Biotech Companies and Research institutions.

3. Establishment of Clinical trial cohorts for pharma industry.

This is our first venture into a social enterprise and since MIT Solve has mentored and supported a diversity of such initiatives across the globe we believe this is the best mentorship platform for a venture like ours. We believe getting into this program will also help us better plan and execute our venture which is a major, never before therapeutic initiative. Lastly the connections and networks with fellow cohorts and alumni would help us open the opportunities we haven't been able to yet.

We want to create systems and partnerships in countries where drugs and the protocols to treat them are not readily available during pandemics. We would like to partner with local NGO's to include every healthcare provider get the information they need in the remotest corners of the planet.

We also want to convince every country and government the need of the hour to implement this solution and help us use this time in building infrastructures so that humanity has a fighting chance against the epidemics/pandemics arising in future. 

Bill and Melinda Gates Foundation

World Health Organization 

and every small and large foundation who might have a niche in remote areas in under-developed and developing nations.

We are the first global platform that aims to have in built machine learning algorithms provide real time data analysis and correlations on different therapies being tried for a disease world wide. Since Artificial Intelligence is the core of this technological platform we believe we qualify for the prize.

If we were awarded this prestigious prize, it would be used to provide digital infrastructures such as tablets/smartphones to clinicians in remote corners of under-developed and developing nations to ensure they receive access to our platforms. We would also use it to establish partnerships to deliver the most effective drugs (from our insights) to these regions for prophylaxis and treatment.

The amount would also be used to expand our resources in applications to diseases beyond COVID, immediately which are rampant in these vulnerable populations. We would be working to understand effective therapies against Rare diseases, Drug resistant Tuberculosis (TB), HIV- TB co-infections where the mortality is 0.2 million people per year, a majority of whom are in Africa and India and suffer from poor immunity. 

Refugees are one of the most vulnerable population and we believe the prophylactic measures and treatment insights generated from our platform would be immensely useful to save their lives. Currently there are 70 million refugees, Asylum seekers and forced migrants, the conditions in which they live are very conducive to spread to the virus. Tens of thousands of refugees have been infected per region globally and a few thousand have succumbed to death.

Given this prize, we would spend it on partnerships, providing infrastructures like tablets for doctors to access our platforms and medicines which are most effective in our data to be readily available to these populations.

Since our solution aligns with multiple Sustainable Development Goals of the United Nations and covers humanity as a whole, we believe we qualify for this prize. As with all the prizes above we will be using this amount to ensure the following in remote, under-developed and developing areas of the world:

1. Implementation of infrastructures like Tablets for clinicians to access our platforms.

2. Establish partnerships with organizations working in remote areas and agencies ensuring the best drugs from our data reach these regions for a successful implementation of our mission- saving every life.

3. Establish partnerships with Refugee groups for infrastructure, drug delivery and implementation.