Vibesdoc

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

Although countries from the region share cultural ties, each has very different healthcare systems and different life standard conditions. 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 effectively handled the pandemic more effectively than the rest. 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 in 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. Current patient engagement tools focus on a small part of the patient journey, such as optimising for SEO searches, 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 diagnoses and appropriately directing patients to further testing or specialist from a place of learned experience and not hunched guess. Thus, Vibesdoc's 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 tests (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 promote such offerings first with their loved ones and then in their communities, as testing is heavily needed in the region.

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 regional governments to fall back 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 when needed, as those are relatively low cost and easily shippable. Pairing such results with our survey would increase their reliability and enable further integration with patient medical records.

During the pandemic, companies and countries made available large data sets to researchers for studying and potentially helping with the sudden surge in coronavirus cases. Our team leveraged large unidentifiable data sets available to us from such organisations to feed an AWS AI/ML instance.  This instance was further tweaked as a Bayesian Algorithm by our data scientists (Claire, who teaches at Chicago and Nina, who holds a PhD from Stanford). Our Bayesian algorithm has proven successful in increasing the accuracy of covid lateral flow assays as our team of Data Scientists collaborated closely with our team of Physicians. They designed the survey questions that gather all relevant patient information that might impact patients' exposure risk to the Virus given their experience and the Data Set.

Our team of PhDs in Data Science, Microbiology, and Medicine published a white paper to share our experience building the Bayesian Algorithm powering Vibesdoc's core AI/ML. As available in such proceedings, our algorithm has proven to raise the accuracy of lateral flow assays by ~16%, from a base of ~70% accuracy to a final range of results at ~86%. Lateral Flow assays are a newer form of rapid testing, but one which gained popularity during the covid19 pandemic and which technology is  FDA approved for several diseases, health markers, and further screening.

Such paper is available for download and online reading at the following link.

http://proceedings.mlr.press/v...

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 proves 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 this tool to fulfil such capabilities, more unidentifiable data from further collaborations and tweaks in code to comply with regulations and expectations would be needed.