Skip to content

ISPE 2024: Breaking Barriers in Cancer Research

Published

September 2024

By

Blythe Adamson, PHD, MPH, Head, Outcomes Research & Evidence Generation, International

ISPE 2024: Breaking Barriers in Cancer Research

This year’s ISPE unveiled many insights for Flatiron Health's #RealWorldEvidence team to carry forward into our work to reimagine the infrastructure of cancer care. We’re sharing our top three below:

#1: #AI continues to be a critical tool to maximize the use of available real-world data (RWD):

The oncology landscape is constantly evolving, and thus relevant and impactful research questions may be different at any given time. Large Language Models (LLMs) and other generative AI models, for instance, can be employed to extract key data elements from unstructured fields in electronic health records, generate synthetic data, and enumerate confounders to control for in analyses more efficiently. 

It was a clear theme at ISPE 2024 that in pharmacoepidemiology space, AI is being used to drive scaled solutions necessary to answer critical research questions previously limited due to cohort size and/or longitudinality. And, these solutions have been validated across the data lifecycle—often using adaptive analyses to ensure data-quality while also matching the scale at which the industry is growing.

Here at Flatiron, ML unlocks over 4M patient records and 1.5B data points in Flatiron’s network to offer deeper insight for research across the patient journey. The value of this work was shown in Flatiron-led abstract ‘Characterization of novel longitudinal oncology real world datasets in Germany and the UK: Biomarker prevalence findings’ presented at this year’s conference, which used Flatiron’s ML-powered HorizonDB to identify cohorts a cohort of more than 1,000 patients with cancer in Germany and UK from 4M+ patients from around the globe including the US, Germany, UK, and Japan. The research resulted in a fit-for-purpose dataset for potential use in comparative-effectiveness research across countries. 

#2: Increasing robust, diverse datasets, using multi-national common data models, will support understanding of differences across different geographical locations and subpopulations:

Several HTA and regulatory bodies have emphasized the need for sources of local data across Europe and Asia. While in some cases learnings from data sourced in one country can be applied to another, there are considerable limitations when the standard of care differs between countries. After participating in the featured symposium “Addressing RWE Transportability Concerns in Health Care Decision-Making. How to Do It Right?” at this year’s conference, it’s clear this topic remains high-priority, and identifying when transporting data is appropriate is needed.

Our discussion showed that given limitations of transporting data across country borders, improving local data—curated through local EHRs—combined with methods to enhance data applicability across diverse populations, will be critical to unlocking pharmacoepidemiologic insights across the globe. Further, new technologies like Trusted Research Environments, also known as Secure Data Environments, allow researchers to access cross-country data safely, securely, and in compliance with local laws. 

#3: RWD remains critical to providing evidence in support of regulatory approvals and pharmacovigilance studies

There was a clear theme at ISPE 2024 on the positive impact of RWD and RWE’s inclusion in regulatory submission for oncology drug development and post-marketing commitments. However, validating the use of EHR-derived RWD will remain critical as the scope and applications continue to increase. 

Flatiron-led abstract ‘Feasibility of prospective real-world studies to fulfill regulatory commitments and requirements in oncology within the United States’ spoke the specific use case of RWD to support post-marketing commitments through prospective real-world studies (PrwS), designed to leverage a combination of routinely collected clinical data and intentionally captured data. The research showed that PrwS, powered by RWD, are a feasible addition to traditional clinical research methods, opening doors to accelerated, more efficient research while potentially lowering the cost of studies and enabling increased participation from sites and patients.


To learn more about how Flatiron’s Real-World Evidence team is transforming the possibilities of oncology research and tackling bold research questions, reach out today.

Share

Posted in