Reflections on the September 2024 Workshop at MGH Institute for Technology Assessment, Harvard Medical School
By Turgay Ayer, PhD
Introduction
The Health Economics and Outcomes Research (HEOR) and Market Access (MA) communities are increasingly under pressure to generate quick insights by assimilating vast amounts of information from multiple data sources, identifying evidence gaps, and developing strategic solutions. Time constraints are a significant challenge in this environment. Generative AI is already revolutionizing industries such as the computer software sector and has immense potential to drastically transform how we work in the HEOR and MA spaces.
In early September 2024, I had the privilege of teaching a two-day workshop at the MGH Institute for Technology Assessment, affiliated with Harvard Medical School, focused on generative AI’s transformative potential within HEOR and Market Access. The workshop was organized by Dr. Jag Chhatwal, Director of MGH Institute for Technology Assessment.
My motivation for leading this workshop stemmed from my long-standing engagement with AI. I published my first AI paper in 2008, which focused on identifying suspicious mammographic images using neural networks to assist in cancer diagnoses. Remarkably, our neural-network algorithm outperformed seven out of eight radiologists.
Since then, I have been deeply involved in the AI space, utilizing and extending various AI and machine learning algorithms in my research, including decision trees, support vector machines, random forest models, and ensemble learners. With the advent of the generative AI revolution over a year ago, I invested heavily in this area by taking courses, reading books, and listening to podcasts. As my knowledge base in generative AI deepened, I recognized the vast impact it could have on the HEOR and MA communities. My personal motivation for teaching at this workshop was to share what I’ve learned over the years and give back to the community.
Use Cases in HEOR and Market Access
The morning sessions began by exploring the power of AI in systematic reviews, meta-analyses, and economic modeling. During interactive discussions, participants identified practical applications of AI in the field. One of the highlighted use cases was the rapid generation of Health Technology Assessment (HTA) insights using generative AI. This capability enables professionals to quickly synthesize vast amounts of data from multiple sources, identify evidence gaps, and develop strategic insights more efficiently than ever before.
Traditionally, professionals spend a significant amount of time identifying sources of information, synthesizing evidence, augmenting data, and generating insights. These tasks are not only time-consuming but also require meticulous attention to detail. With the integration of AI and generative AI, these processes can be significantly accelerated. Dr. Rachael Fleurence from the National Institutes of Health showed participants how AI can automate literature searches, extract relevant data, and even summarize findings, freeing up valuable time for more critical analysis.
Another notable use case, discussed by Dr. Chhatwal, was building early economic models with AI and parameterizing them using generative AI to aid in early discussions and pricing analysis. Often, there is a need for a quick yet sufficiently robust early economic model to inform strategic decisions. AI can assist in constructing these models swiftly, allowing for rapid adjustments and scenario testing. This accelerates the decision-making process and provides a solid foundation for strategic planning in market access.
Practical AI Demonstrations
A central part of the workshop was dedicated to live demonstrations and hands-on activities with foundational AI models, specifically those developed by OpenAI. We showcased AI-based tools for evidence generation, gap identification, and economic modeling. Participants engaged in interactive sessions, exploring how these AI models can be applied to real-world HEOR and Market Access tasks.
One of the highlights was the demonstration of ValueGen.AI, our innovative platform designed to streamline various HEOR processes. ValueGen.AI has multiple capabilities and modules, including synthesizing and summarizing evidence, identifying evidence gaps, generating smart insights, asking detailed questions and getting answers quickly by navigating through vast amounts of information sources, and building health economic models from scratch. It curates and augments data from multiple key sources, allowing users to navigate through vast amounts of information and obtain detailed answers promptly.
What sets ValueGen.AI apart is that, to our knowledge, it is one of the first and most comprehensive platforms to integrate HEOR and MA knowledge in generating insights and results. By incorporating domain-specific expertise, ValueGen.AI enables users to obtain more accurate and relevant outputs tailored to the unique needs of the HEOR and MA fields.
The platform adopts a human-in-the-loop approach, where the user specifies their objectives and requirements. ValueGen.AI then collects relevant information and, in the background, builds transparent models tailored to those needs. This process allows users to maintain full visibility into the modeling without getting bogged down by complex programming tasks.
Moreover, ValueGen.AI enables users to download code, generate comprehensive reports, create model summaries, and even draft abstracts. This suite of features significantly accelerates information synthesis, identification of evidence gaps, and the generation of smart insights. It empowers HEOR and MA teams by enabling them to interact with a powerful AI tool that provides immediate responses to their inquiries, effectively putting critical information at their fingertips. Furthermore, the HEOR Navigator suite within ValueGen.AI empowers users to develop robust models efficiently, facilitating early discussions and pricing analyses that are crucial in HEOR and Market Access.
Participants reacted positively to the demonstrations, noting the potential for these AI tools to automate labor-intensive tasks and enhance productivity. The ability of ValueGen.AI to provide comprehensive, domain-specific insights was particularly well-received. Attendees appreciated how the platform bridges the gap between complex AI functionalities and practical application, allowing professionals without extensive programming expertise to leverage advanced AI capabilities effectively.
Conclusion of Part 1
These practical demonstrations highlighted just how significantly AI can enhance HEOR and Market Access processes. The potential to automate labor-intensive tasks, generate rapid insights, and streamline workflows is transformative. But there’s even more to explore. In Part 2, we’ll delve into the advanced AI skills that can further unlock this potential, discuss the challenges we face in integrating AI into our workflows, and share key takeaways from the workshop that point towards the future of AI in our field. Stay tuned!
