stub Matt Michelson, President of Genesis AI at Genesis Research - Interview Series - Unite.AI
Connect with us


Matt Michelson, President of Genesis AI at Genesis Research – Interview Series




Matt Michelson, is the President of Genesis AI at Genesis Research, an international HEOR and Real-World Evidence research organization with offices in the US and UK which supports the life sciences industry.

An established leader in evidence strategy, development, and communication, they support clients with database analyses, strategic and tactical health economics and outcomes research, literature reviews, economic modelling, scientific publications and market access strategy.

As experts in the development of innovative technological solutions, they provide a range of customized platforms, dashboards, data portals, and a unique artificial intelligence technology assisted review platform EVID AI, to support the identification, monitoring and extraction of precisely targeted evidence.

Could you share the genesis story behind Genesis Research?

In 2009, after time in academia and utilizing his deep quantitative background, our CEO Frank Corvino and his business partner, a clinician and health economist, combined their expertise to develop a new business model that supported life sciences companies with a more data driven approach gleaning insights from big data in consistently transparent, reproducible, and highly proficient ways.

Genesis Research was initially built on a commitment to think outside of the box to deliver value demonstration supported by Real-World Evidence (RWE). Since then, the company’s services have expanded to incorporate market access capabilities and strategy, Health Economics and Outcomes Research (HEOR), scientific communications, and advanced digital capabilities such as the fit-for-purpose use of artificial intelligence (AI).

The company developed into the leading international RWE and HEOR organization we are today because we listen to our clients and then use a diverse, highly skilled team to meet their changing needs at any stage of the product life cycle, employing a flexible “enhanced partnership model” that allows us to operate seamlessly as an extension of a client’s team.

Vast volumes of papers are published each year. Could you discuss how fast the rate of publishing is accelerating?

It’s difficult to pinpoint exactly, but estimates range from a 4 to 10% increase each year, with 1-2M articles currently published in health care per year. That's a massive number, especially if you estimate that it takes a person five minutes to read an abstract. PubMed itself (the public version of the Medline healthcare paper search engine) has a collection of more than 30 million articles.

Why are traditional ways of reviewing papers no longer working?

These tasks boil down to the process of finding the evidence, in particular papers, that answer your scientific questions. And it's not that manual methods can't do this. Rather, because the process is so onerous, it means the task isn't as flexible. No one wants to go through the labor to do it all again if something changes, and you can't search as many sources as you might want, given the huge time investment. And people will sometimes miss articles since it's hard to concentrate after reading a few hundred papers. In contrast, a trained machine never tires, and the efficiency gains mean that AI-based approaches are more flexible, since you can just run the next search, and you can do as many as you need.

What is EVID AI and how does it simplify the process for medical researchers to identify and comb through vast amounts of research?

EVID AI is the only medical literature database that uses machine learning to produce more than 80 million data points ─ from pre-clinical, clinical, economic, epidemiology, patient reported outcomes and review focus areas ─ and allows users to filter search results to the most relevant, current information needed for further analysis. It is the only platform with the capability to turn data embedded in multiple articles into structured tables that present relevant, requested data points in a clear format. This patented approach distills evidence into usable data, so that researchers can more easily develop graphs and dashboards to share with stakeholders without having to read a large volume of papers.

Genesis Research recently made enhancements to EVID AI ─ which is now the world’s largest current medical literature research platform ─ that help pharma teams and other healthcare decision makers find high quality, targeted results faster, more efficiently and from an unprecedented breadth of sources.

Using EVID AI, medical literature search tasks can now be done 59 times faster than manual efforts, with significantly more relevant evidence per search and 15 times fewer irrelevant articles. The patented machine-learning format, programmed through exposure to tens of thousands of training data points, is faster and more comprehensive than ever and will cut research time from months to weeks or days.

One benefit of EVID AI is that it allows researchers, and government regulators to trace the AI data back to its source. Why is this important and how does it work?

A key issue with many AI systems is that they are opaque – sometimes called “black boxes.” And this is true in the sense that we don't always understand why the AI is doing what it does. For instance, if it takes an article and breaks out all the results from the text, it can't necessarily tell you exactly why it chose those phrases and results, it can just show them to you. However, we put an emphasis on being transparent and providing data provenance (e.g., showing you where it comes from), so that users can always trace a result back to the sentence it came from, the article that contains that sentence, and then the original location of the article. That way, there is always a mechanism to trace results back to their source. In addition to providing better science, this is important for regulators, since if a pharma company is making a claim based on evidence from our AI, the regulator can verify that the data is correct and the source is valid.

Could you share a use case or case study of researchers using EVID AI?

Absolutely. There are many to choose from, but here are two that are useful, since they show how the tool can be used for both larger, budgeted projects and daily, ad-hoc tasks. In the first case, we had a pharmaceutical company engage a team to do a literature review update for oncology. This is a large undertaking because the specific field in oncology is large, the literature changes quickly, and the scope of the project is big. The original review included an analysis of the landscape (e.g., all the main drugs) and their results in terms of how effective they are, how safe they are, and the economic impacts. It included not only clinical trial reports, but also articles about observational studies, where scientists track the performance of a drug prescribed in the “real world” instead of just in a controlled, trial environment.

EVID AI helped this company gather all the new and updated results for this literature task, with tremendous savings. In contrast, we had an example where a scientist had built an economic model of the costs associated with switching across different mental health drugs called “budget impact models”. The challenge is pulling good estimates of how often such patients are switching drugs. When the scientist initially built the model, he spent an entire day scanning articles to find the result he wanted. With EVID AI, he found it in a few minutes.

What is your view of the future of human and AI collaborating in medical research?

As medical research gets more and more comfortable with AI, it will permeate areas from drug discovery to clinical trial recruitment to data analysis and reimbursement. Every aspect of the development of new therapies will benefit from AI, and the results will become embedded in the workflow. Rather than relying on separate tools that require context switching from a bench-science task to an AI task, it will become as intuitive as using GPS to get to a new restaurant. There won't even be a second thought. However, especially in the pharma industry, we’ll still need highly experienced people, like the teams at Genesis Research, to determination the relevance of data and initiate further analysis to aid decision-making.

Is there anything else that you would like to share about Genesis Research?

Genesis Research’s rapid growth is due to its ability to take on new challenges, ask the right questions, assemble the right teams to access and analyze the right data, and deliver solutions that move life science initiatives forward. An innovator in developing RWE-based solutions, the company is entirely data agnostic and works closely with clients to identify optimal data sources. We are proud to be an established leader in evidence strategy, development, and communication, as well as experts in the development of innovative technological solutions.

A founding partner of unite.AI & a member of the Forbes Technology Council, Antoine is a futurist who is passionate about the future of AI & robotics.

He is also the Founder of an investing website, the generative AI platform, & is he is currently working on launching a platform that will offer users the ability to configure and deploy autonomous agents by breaking prompts into sub-tasks.