Sun 6 Sep
8:00 - 12:00
Applied Generative AI for HEOR: Introduction
Session Type: Short Course
Topics: Methodological & Statistical Research
Level: Introductory
Separate registration required.
The rapid advancement in generative artificial intelligence (GenAI) presents an opportunity for transformative potential in the field of health economics and outcomes research (HEOR). This course provides an introductory understanding of generative AI models with a particular focus on large language models (LLMs), which are transforming the field of HEOR. Participants will be provided with an overview of the most appropriate ways to access LLMs, going beyond the use of chatbots. Further, they will be given insights into how to use prompt engineering, retrieval-augmented generation (RAG) and agents to conduct scientific research and gain an understanding on issues pertaining to privacy and security when using GenAI for HEOR. Participants will further explore specific applications of these models for conducting robust scientific HEOR research in, for example, systematic literature reviews (SLR) and economic evaluation. The course aims to equip participants with the knowledge to begin to use generative AI techniques for specific HEOR contexts and to appreciate how these innovative approaches can enhance HEOR activities. Practical exercises using Python and relevant AI frameworks will be incorporated for participants to follow along.
PREREQUISITES: Students should have a general understanding of common HEOR concepts such as SLRs and cost-effectiveness models. Knowledge of Python or similar programming languages such as R is considered a benefit but not required.
Speaker
-
Timothy Reason, BSc, MSc
Estima Scientific, United Kingdom
Tim Reason is co-founder of Estima Scientific and specializes in AI and evidence synthesis, having spent 15 years in the field of HEOR and technology. Tim is managing director of Estima, driving business activities, innovation and strategy for the company. Tim’s specializes in the intersection of HEOR, software development and AI to drive better outcomes for patients. Tim is the lead author on 2 seminal papers in AI for HEOR, showing that AI can be used to automate health economic modelling and NMA.
8:00 - 17:00
Budget Impact Analysis in Practice: A Hands-On Course in Strategic Conceptual Design, Model Building, and Communication
Session Type: Short Course
Topics: Economic Evaluation
Level: Intermediate
Separate registration required.
This hands-on, interactive course equips participants with the conceptual and practical tools needed to develop, adapt, and communicate budget impact analyses (BIA) for real-world decision making.
The course begins with a brief review of BIA concepts and a 6-step approach to strategically design these analyses. We will discuss practical applications, including customization to accommodate payer-specific data; balance of structural simplicity, accuracy, and face validity; interpretation of results; and development and communication of compelling value messages. Technical topics will include static versus dynamic budget impact models, good Excel model-building practices, considerations for device and diagnostic technologies, and incorporation of realistic features such as patient copayments and availability of generics.
Instructors will walk through 2 budget impact models programmed in Excel (1 static and 1 dynamic) and will work with participants during hands-on exercises to customize and adapt these models to specific real-world circumstances. These Excel-based models will be provided to participants in advance, both for use during the session and for later reference. Throughout the course, participants will have opportunities to network and to work in small groups to discuss key concepts, plan and implement model adaptations, and develop and communicate value narratives.
This course is designed for participants seeking to deepen their understanding of BIA, gain practical exposure to Excel-based models, and strengthen their ability to communicate BIA in ways that influence decision making. Registrants who wish to participate in the interactive portions of the course will need to bring materials for handwrtten activities.
PREREQUISITE: Participants are expected to have familiarity with basic BIA concepts and working knowledge of Microsoft Excel. A microcourse on the basics of BIA is available in the ISPOR Education Center.
Speakers
-
Stephanie Earnshaw, MS, PhD
Peritia, Morrisville, NC, United States
Stephanie Earnshaw has performed health outcomes and health services research for 30+ years. Her research focuses on applying decision-analysis techniques (eg, decision trees, Markov processes, simulation) to industry-related issues and health care problems. In addition to developing budget-impact and cost-effectiveness models to support health technologies for the pharmaceutical, biotechnology, and diagnostic and medical device industry, she has developed innovative mathematical models using these methods to determine pricing strategy, predict clinical outcomes, allocate resources, and cost care pathways particularly in support of medical diagnostics.
Dr. Earnshaw received her PhD in Industrial Engineering at North Carolina State University and is a member of ISPOR and the Institute for Operations Research and Management Sciences. She has presented her work at professional conferences and has published in several peer-reviewed journals. She has presented workshops and various courses on decision-analytic modeling techniques for pharmaceutical companies and organizations such as ISPOR, the Academy of Managed Care Pharmacy (AMCP), and the Centers for Disease Control and Prevention (CDC). Dr. Earnshaw has served on the ISPOR Board of Directors and as Chair of the Audit Committee and Educational Council. She has held an Adjunct Faculty appointment at the University of North Carolina’s Eshelman School of Pharmacy, Division of Pharmaceutical Outcomes and Policy, is honored as a Distinguished Alumni in Industrial and Systems Engineering at North Carolina State University and is one of the lead authors of “Budget-Impact Analysis of Health Care Interventions: A Practical Guide.”
-
Thor-Henrik Brodtkorb, PhD
RTI Health Solutions, Ljungskile, Sweden
Thor-Henrik Brodtkorb, PhD, is executive director in Health Economics at RTI Health Solutions (RTI-HS). He holds a PhD in Health Technology Assessment from the University of Linköping and has been with RTI-HS for 12 years. He has been teaching courses in decision-analytic modeling at Linköping University as well as presented workshops and short courses on decision- analytic modeling techniques for organizations such as Pharma Industry Sweden, Swedish Agency for Health Technology Assessment and Assessment of Social Services (SBU), and ISPOR.
At RTI-HS, Dr. Brodtkorb leads the development of cost-effectiveness, cost- utility, cost-consequence, and budget-impact models for pharmaceutical, device, and diagnostic technologies. These models have been used to support reimbursement decisions in more than 15 European countries including NICE in UK, SMC in Scotland, TLV in Sweden, and NOMA in Norway. He has developed models and analyses in the areas of oncology, alcohol dependence, major depressive disorder, Alzheimer’s disease, dermatology, multiple sclerosis, cardiology, orthopedics, and asthma. His research has been presented at professional conferences and published in peer-reviewed journals. He is also a co-author of the book “Budget-Impact Analysis of Health Care Interventions: A Practical Guide.”
-
C. Daniel Mullins, PhD
University of Maryland School of Medicine, Baltimore, MD, United States
C. Daniel Mullins, PhD, is a professor at the University of Maryland School of Pharmacy. He is founder and executive director of the University of Maryland PATient-centered Involvement in Evaluating effectiveNess of TreatmentS (PATIENTS) Program. He received his BS in Economics from M.I.T. and his PhD in Economics from Duke University. His research and teaching focus on comparative effectiveness research (CER) and patient-centered outcomes research (PCOR). Dr. Mullins has received funding as a Principal Investigator from AHRQ, FDA, NHLBI, NIA, NIMHD, Patient-Centered Outcomes Research Institute (PCORI) and various patient advocacy organizations and pharmaceutical companies. He is the lead for the Community & Collaboration (C&C) Core of the University of Maryland Institute for Clinical and Translational Research (ICTR) and co-lead of the C&C Core for Johns Hopkins’ CTSA.
Professor Mullins is 1 of 2 editors-in-chief for Value in Health and is author of over 325 peer-reviewed articles and book chapters. At the University of Maryland Baltimore (UMB), he received the Dr. Patricia Sokolove Outstanding Mentor Award and the Dr. Martin Luther King Jr. Faculty Diversity Award. He was named Researcher of the Year at UMB and was awarded a University System of Maryland Wilson H. Elkins Professorship. He is a past recipient of the Dr. Daniel D. Savage Memorial Science Award, the Association of Black Cardiologists’ most prestigious annual award, and the ISPOR Marilyn Dix Smith Leadership Award.
13:00 - 17:00
Practical Applications of Large Language Models for Real-World Evidence Generation and HEOR
Session Type: Short Course
Topics: Methodological & Statistical Research
Level: Intermediate
Separate registration required.
Examine large language models (LLMs) from industry leaders such as OpenAI, Anthropic AI, Google, and Meta, focusing on their application in real-world evidence generation and HEOR. The course covers technical LLMs, including their architecture, embeddings, context window, tokenization and token cost, hallucinations, fine-tuning vs prompting vs retrieval augmented generation, risk-based frameworks, and current task-specific live benchmarks used for model assessment.
Participants will learn prompt engineering through hands-on, practical examples, empowering them to utilize commercially available LLMs with a token efficiency mindset. These examples include scientific literature retrieval, PICO extraction and processing, extracting and handling numerical data and summarizing tables and figures.
Upon completing this in-depth course, participants will gain the competencies needed to use LLMs responsibly for practical applications in RWE and HEOR, while remaining mindful of regulatory obligations. The course will include an introduction to agentic AI. To participate in practical exercises, attendees are required to bring a personal laptop and have access to a personal or corporate LLM account with file upload functionality.
PREREQUISITE: General knowledge of chat-based LLMs (GPT, Claude, etc) is essential. This is an intermediate course, and students should have prior knowledge of AI and have used chat based LLMs in a professional/work setting.
Speakers
-
Manuel Cossio
Cytel, Dubendorf, Switzerland
AI Engineer and Head of AI Solutions at Cytel with 13+ years of experience in HEOR. I lead the development of AI-driven solutions for evidence generation, economic modeling, and HTA landscaping—including EU JCA and market access. With expertise across both pharma and CRO consulting, I’m committed to advancing patient care through smarter, AI-enabled decision-making.
-
Benjamin Bray, MD, MBChB, MSc, FFBCS
Lane Clark and Peacock, London, United Kingdom
Dr. Ben Bray is a medical doctor and epidemiologist and is Evidence Generation lead at LCP Health Analytics. He has been working in health data science and epidemiology for over 12 years and has extensive experience in the development and validation of machine learning models and in applications of AI using health data. He has authored over 60 publications including in The Lancet, BMJ and PLOS Medicine and has co-authored multiple reviews on the use of AI and machine learning in various therapy areas. He holds an Honorary Senior Clinical Lecturer post at King’s College London, focusing on research into machine learning analytics using large health databases.
