Improved Data Sharing, Collaborative Learning Could Enhance Real-World Data Analysis

Improving data-sharing with initiatives that encourage worldwide collaboration among pharma stakeholders could lead to significant improvements in evidence-based decision-making, according to the authors of a recent analysis published by JMIR Medical Informatics. The investigators noted that managing real-world data (RWD) for chronic diseases, such as multiple sclerosis (MS), carries a significant challenge due to current literature that oversimplifies RWD processes and lacks a holistic framework.

“Chronic diseases, such as (MS), present significant obstacles for research, primarily because of their limited prevalence, resulting in smaller study populations,” the study authors wrote. “The scarcity of the affected individuals is reinforced when considering the dispersion of (RWD) across diverse repositories. This scarce RWD, sourced during routine clinical care, further coupled with heterogeneity in formats, quality standards, and regulatory guidelines, make the comprehensive collection and extraction of meaningful clinical insights even more challenging.”

The lack of RWD for diseases with limited patient populations has created a gap that has a negative impact on the ability to identify significant disease patterns, patient experiences, and treatment outcomes, the study authors noted. They pointed to the beginning of the COVID-19 pandemic when research efforts incorporated innovative data acquisition strategies to overcome the dearth of data. These efforts brought to the forefront that the difficulties in gathering RWD are understated—pointing to issues such as diverse formats that lack a holistic framework—and emphasizes the need for a more interconnected system, according to the study.

“This fragmented focus points to the need for a more comprehensive strategy that neither compromises nor overlooks any part of the RWD management process,” the authors wrote. “The absence of a holistic framework, coupled with the growing diversity and volume of RWD sources, intensifies the challenges in health care data sharing and the conversion of RWD into actionable evidence, underscoring the need for standardized management.”

The researchers presented the Global Data Sharing Initiative (GDSI) as a potential solution, which uses a three-layer data acquisition framework that complies with legal and ethical standards. The analysis presented a comprehensive, research question–agnostic, multistakeholder-driven data analysis pipeline that addressed individual data sharing, core data set sharing, and federated model sharing.

The investigators used a demand-driven methodology for standardization across three streams of data acquisition: a data quality enhancement process, a data integration procedure, and a concluding analysis stage that delivers on RWD-sharing requirements. The researchers used the successful implementation of the COVID-19 and MS GDSI to show its efficacy.

“The COVID-19 pandemic underscored a pressing need to understand its effect on people with MS. Recognizing the criticality of solid evidence for disease management, a global strategy involving neurologists, patients, and registries was adopted,” the study authors wrote. “This collaborative approach paved the way for GDSI’s formation and the development of an end-to-end RWD analysis pipeline. Through this effort, GDSI emerged as the most comprehensive federated international cohort of people with MS impacted by COVID-19, becoming an invaluable resource for informed decision-making.”

However, the study noted that drawing conclusions from these streams necessitates the consideration of the limitations in observational study designs, which they said can provide unparalleled real-world insights; however, each study has specific, unique limitations that create challenges when examining post hoc analyses based on the findings.

“Although GDSI showcased significant advancements, challenges inherent to its structure and execution were encountered,” the authors wrote.

They found that the flexibility and adaptability of the GDSI data analysis pipeline combined with the disease-specific components can create a flexible tool that develops sturdy data architectures across the biomedical research landscape.

“Serving as a practical blueprint, GDSI addressed not only current health care challenges but also laid the groundwork for future initiatives,” the study authors concluded. “Its hybrid approach to data acquisition and analysis provided a scalable framework applicable to other health care sectors. In doing so, GDSI stands as a compelling example of how data sharing and collaborative learning can meaningfully advance health care research, going beyond the specific challenges of MS and COVID-19.”

Reference

Pirmani A, De Brouwer E, Geys L, Parciak T, Moreau Y, Peeters LM
The Journey of Data Within a Global Data Sharing Initiative: A Federated 3-Layer Data Analysis Pipeline to Scale Up Multiple Sclerosis Research. JMIR Med Inform 2023;11:e48030. DOI: 10.2196/48030. PMID: 37943585. Accessed November 13, 2023.