Vibesdoc

Vibesdoc is a patient engagement tool for prompt diagnosis using patients' case information gathered from a survey & rapid home tests with an AI backend. This tool computes the probability of a diagnostic given a patient's case variables leveraging large data sets to increase accuracy.  Although initially focused on covid, the same code with few alterations can also improve other rapid tests' accuracy for other contagious diseases.

With Vibesdoc's survey results matching patient symptoms with diagnosis, we would promptly identify diseases and provide users with suggestions on practitioners and institutions for treatment. It would also suggest further tests to verify diagnosis possibilities and an alert stating the urgency of treatment required. The results would also be displayed in an explainable visual manner to help people understand what their symptoms might be correlating and help them understand their physicians' suggestions, thus increasing their participation in treatment.

In the past decades, Latin America has experienced a surge in infectious diseases such as Dengue, Cholera, Venezuelan Equine Encephalitis, Leptospirosis, Swine Flu, and most recently, Covid-19.

Although countries from the region share cultural ties, each has very different healthcare systems and diverse life standard conditions that differentiate them. In 2020, at the start of the pandemic, Costa Rica, Panama, and Chile, carried out the most testing and thus appeared to have more cases than their neighbours. These countries have been much more effective than the rest in handling the pandemic. Nations that conducted fewer tests seemed to have fewer cases at first but still struggle to get a hold of such disease.

Nine out of 20 countries conducting the lowest disease screening per capita are in Latin America. It's worrying that these countries have a sharp rise in infectious diseases and chronic conditions concurrently from more developed populations living at their urban centres. Thus a cultural and institutional approach to patient testing and monitoring is paramount. Currently, getting a medical test in this region is troublesome. As there are not enough places to acquire it, they are prone to high costs and inaccuracies in diagnosis.

Vibesdoc is a tool leveraging big data to aid in the diagnosis, using existing tools such as EHRs/EMRs, rapid testing, cloud computing, and mobile internet access through a smartphone.  Although the application of such algorithms is not common, it is not a new technology, and neither are symptom checkers to recommend patients for further actions.

That said, the application for which Vibesdoc is intended is innovative in itself. Current patient engagement tools focus on a small part of the patient journey, such as optimising for SEO searches, or aiding in easier appointment booking, or simply allowing for a vast library of medical knowledge to be available online for patients to research.

Vibesdoc intends to merge these technologies and become more of an automated tool that will optimise general practitioners' role in identifying potential diagnosis and appropriately directing patients to further testing or specialist from a place of learned experience and not hunched guess. Thus, Vibesdoc power lies in the tool's synergistic nature, tailored by practitioners and field experts through enhanced collaborations with leading institutions to build a genuinely working tool that makes medicine understandable and actionable for people to participate voluntarily.

The current target market would be the low hanging fruit for the capabilities of our solution. Nevertheless, because many women from Central America & Caribbean urban centres are already used to taking rapid tets (for pregnancy, for example), then providing them with tools to get tested for more conditions would have a higher probability of catching on. From our survey results, we believe that these initial adopters will also promote such offerings first with their loved ones and then in their communities, as testing is heavily needed in the region regardless.

As institutions, practitioners, and patients start seeing local success examples of prompt disease detection, the solution would gain popularity and impact the entire healthcare system. An autonomous solution that would handle massive testing would also benefit governments from the region to fall back to in case of a future pandemic.

Utilising already FDA approved lateral flow assays would also provide an option to provide testing in remote areas of the Americas for when needed as those are relatively low cost and easily shippable. Pairing such results with our survey would increase their reliability and utilise the ability to be further integrated with patient medical records.

To better understand our offering's market perception, we carried out a survey answered by more than 250 people anonymously. 

From this research, we perceived that physicians and patients do not trust the power of an AI enough to rely on it solely for diagnosis, but would use it as an additional tool to get more information from the patient before their visit.

Currently, people go to hospitals, labs or clinics to get guidance and testing for potential diseases. Patients have to rely on self-research on symptoms, practitioners, and tests before getting the needed aid to get treatment.

Although we have completed our tool to increase the accuracy of covid antigen lateral assays, the solution proposed is much more than that. Our initial MVP is proof of how rapid testing can be improved when paired with big data AI. With a tool such as Vibesdoc, we could unlock the power of rapid testing using lateral flow assays and then genetic markers to identify infectious diseases and chronic conditions in development. To allow for this tool to fulfil such capabilities, more unidentifiable data from further collaborations would be needed and tweaks in code to comply with regulations and expectations.

This powerful solution and core technology were then taken by our front-end developers and presented as a reactJS web app. Using such a framework enabled the creation of a user-friendly, visually pleasing, yet robust solution that would be able to handle massive data requests through our Golang API and AWS & NGINX hosted servers.  We are currently modifying such a tool to work as a patient engagement tool to be subsidised by pharma and EHR companies to allow for an interactive tool to help them secure their market share by providing their clients with an edge for competing with others in the market. We believe that this algorithm's power should be further exploited by a complete evaluation of diseases, not just covid-19. Still, based on potential clients' feedback, we have landed on such a strategy to guide the continuation of such a tool.

Therefore the most impactful innovation, beyond a well-built tool, through which experts in respective fields pitch in to achieve it, the invention that Vibesdoc propose goes beyond the technology it epitomises. The possibilities to build on top of such a tool are endless. Eventually, even health and medical records could be personalised and given to individuals to manage so that treatment would adapt to a person's genetics and complete case information.

Our solution has not generated revenue or received funding. The last team, which started with CODI that is now Vibesdoc, did win a couple of Hackathons for small monetary value, but this was spent in development expenses and in covering for the time invested in building such a solution. Under Roberto Baldizon, the organisation is now seeking private partnerships with organisations like Novartis and public ones with the ministry of health of Costa Rica. These are still in negotiation, but transactions of this nature that we envision would feed the venture if needed for self-sustainability in the upcoming years.

Roberto Baldizon, has vast experience in business, engineering, and informatics. Freddy holds a PhD in microbiology. Nina, Claire, and Felicia all have PhDs in mathematics and are data scientists. Anurag is an MBA, Tal is a developer with experience in healthcare. And Jesse is a last-year resident in Osteopathic Medicine.

Our covid focused app (http://testforcov19.org) and the core team that made it possible (http://homecovidtests.org) remain online and involve in the project at least as advisors and believers in the mission.

Our team has already achieved success in increasing the accuracy of Covid lateral flow assays. We increased the accuracy of tests from 70% to 86% using big data to correlate patient answers from a survey to unidentifiable case data. For this project, we would shoot for doing the same for all rapid tests for contagious diseases. The algorithms developed and tweaked in this project will recommend patients appropriately to the right practitioner to alleviate their symptoms.

Currently, we developed the initial MVP using publicly available data from a nonSQL database hosted in NGINX. The ReactJS front-end survey and Golang API live in cloud solutions from Amazon and Google. The ML/AI algorithm is in R and Python. Most of the project is hosted in AWS and uses tools such as Sagemaker and S3 there.

At the beginning of this project, we intended it to be a wholly private and nonprofit venture to avoid potential hurdles involving HIPAA and GDPR. Therefore, with our covid app, we are not collecting any identifiable data nor tracking users in any way. For this new iteration of the tool as Vibesdoc, we would enhance the tracking and monitoring capabilities for long term sustainability. Currently, the only metrics to measure usage and performance are those provided by AWS, displaying a constant but still low use of the application, averaging 3 per cent of its full capabilities for our 8 GB EC2 instances.

Vibesdoc is designed for scale with a robust architecture utilising Facebook's react framework for the frontend and Google's Golang for the API connectivity with a Cloud NoSQL database. The architecture is built for future democratisation as the API would handle requests from other frontend surveys besides the one we have developed and the rest we will create. Currently, we have started talks with Novartis and Costa Rica's ministry of health to explore a potential partner through which we could reach millions of users quickly. If successful, the same type of partnership would be sought in the rest of Central America and the Caribbean and perhaps globally. Through Vibesdoc, we envision users constantly screened and tested for conditions whenever they feel unwell and want to seek help. This tool will be a great add on for practitioners, clinics, insurance companies, medical labs, and the entire supply chain alike. Vibesdoc optimises the whole patient flow for efficiency with the help of passed medical cases.