Highlights from Halifax: Six Themes from ICPE 2023

Last week in Halifax, the International Society for Pharmacoepidemiology (ISPE) held its annual conference, ICPE 2023. Attendees presented research findings and panelists discussed advancements in epidemiology methods, new regulatory developments, and the latest in health equity assessment and promotion. Here are just a few of the major themes that piqued our interest.

1. The often complex intersection of regulatory science and epidemiology. In Friday’s plenary session, Brian Strom, chancellor, Biomedical and Health Sciences at Rutgers University, moderated a discussion on regulatory decision-making as it pertains to pharmacoepidemiology. The FDA, EMA, PMDA, Health Canada, Israel MOH, and MHRA were all on a global panel that followed the fireside chat, and all agencies have released guidelines and collaborated on drug shortages through ICMRA. It was made clear that establishing consolidated guidelines and recommendations on the use of real-world evidence (RWE) is a challenging task, particularly given the case-by-case nature of determining acceptability. However, the opening plenary also made it apparent that the global regulatory community sees RWE as a valuable asset for evaluating safety and efficacy, and that they are committed to supporting its use in pharmacoepidemiology evidence generation.

   Also mentioned in the session was that the conversation among researchers should not be “randomization vs. RWE” (i.e., how can you make RWE look just like a randomized trial), but rather, researchers should thoughtfully consider their study design and select the best method and data source for it. While randomization is the best tool for balancing known and unknown confounding variables, in cases where large scale randomized clinical trials are not feasible, such as in rare disease, RWE may be a preferred option. RWE should be used to answer the right questions, and transparency and communication is key. Pamela Aung-Thin, Assistant Deputy Minister, Health Canada discussed the agency’s R2D2 (Regulatory Review of Drugs & Devices) program as an example of using RWD and RWE to provide broader and faster access to products that are better aligned with Canada’s healthcare system needs.

2. Enablers of study replication and transparency. Another theme of the conference was the importance of replication and transparency in ensuring consistent results. Our own Caitlin Dodd presented Panalgo’s research on the use of methotrexate and the risk of Alzheimer’s disease among persons with rheumatoid arthritis. Although methotrexate (MTX) did not appear to be protective against AD among U.S. patients, this work was notable given that it replicated an analysis among European patients that found a statistically significant protective effect that increased with the duration of MTX use.

   Another session focused on thoughtful code list creation, reporting transparency, reproducibility, and validity in code lists and computable phenotype construction. Code lists are a crucial component of real-world studies because they serve as the foundation for phenotypes that define study criteria and operationalize medical concepts. A poorly constructed code list may lead to misclassification of exposures, outcomes, covariates and reduce external validity (generalizability) of the results.

   The session’s advice on code list creation included checking to see if a code list already exists before you start from scratch (via FDA library, Sentinel public code lists, the Chronic Conditions Data Warehouse, payor (CMS) guidance documents and published literature); validating (through clinician review, comparing the prevalence of the phenotype in your data source to an established benchmark, and running sensitivity analyses); and keeping up with guidances as they evolve.

3. Machine learning: ripe for collaboration among all RWE scientists. Now that pharmacoepidemiology has become more familiar with the tools and capabilities of machine learning, there’s a desire to use the techniques for moving the field forward, rather than simply providing the best outcome prediction. Experts at ICPE recommended machine learning for bringing structure and clarity to observational data, as well as collaboration across data science, safety, and clinical teams to ensure the usefulness of results in clinically relevant applications. For improved collaboration, the ability to iterate analyses, transparently share methodological assumptions, and reproduce results are key for the acceptance and advancement of RWE.
4. Health equity, a highly nuanced and global topic. A potpourri of presentations covered disparities in the delivery of care and treatments and how we can pinpoint areas of deprivation or social disadvantage to direct more resources to those in need. Social (personal) and structural (environmental) determinants of health (SDoH) are some of the newer data sources used by pharmacoepidemiologists. The language around, and the indices used to assess, SDoH are rather interesting. For instance, a Danish presenter who was looking at adherence to tamoxifen vs. socioeconomic position (SEP), talked about SEP including information on “cohabitation” rather than “marital status,” because in Denmark, 23% of couples are living together (often long-term) but are not married (in the U.S., some states recognize this as “common law” marriage).

   A UK presenter investigated prescribing patterns vs. the Index of Multiple Deprivation (IMD), which is based on postal code and includes nearly 40 measures of deprivation. A U.S. presenter comparing the relative distribution of two types of meningococcal vaccine used the social vulnerability index (SVI), a CDC tool that was developed for natural disasters such as hurricanes. All presenters talked about the nuances in defining deprivation or vulnerability and how much it can vary by country, database used, or even study design.

5. Patient involvement in the development and the safe use of medicines. During Saturday’s plenary session, panelists discussed patient involvement in the development and safe use of medicines from the findings of CIOMS’ 2022 report, calling it a “landmark document” for patient involvement across the product lifecycle, and sharing how to “make big data bigger by bringing in patient perspectives.”

   Panelists reviewed five key principles from the report: incorporating the patient’s voice into the decision-making process; realizing that patients have expert knowledge of their diseases and conditions (just as important as that of scientific and medical experts); compensating for patients’ time in research; maintaining patients’ independence; training stakeholders for patient engagement activities; and the need for balanced, transparent information in written agreements. Although pharmacoepidemiologists and regulators often think they know what’s important to patients, there are always surprises, panelists noted, and involving patients more in product development and use can improve their quality of life.

6. Sustainable drug use and climate change. Environmental pharmacoepidemiology (EnvPE), which has several monikers, including pharmaco-environmentology and eco-pharmacovigilance, is a nascent field, but a fascinating one. Panelists discussed how healthcare systems’ carbon emissions and inefficient waste disposal can contribute to climate change. For instance, Viagra has been detected in numerous aquatic species. They noted that wastewater monitoring (used to map COVID-19 infection waves) can also be used to evaluate inappropriate drug disposal. EnvPEs leverage brilliant data visualization techniques to overlay as many 10 data sources to map hot spots of overprescribing, waste, or contamination.

For more information on how Panalgo can support your team’s selection of RWD and transparent generation of RWE, contact us today.