Harnessing AI for Healthcare Research: Transparency and Accuracy in Data Cohorts

Harnessing AI for Healthcare Research: Transparency and Accuracy in Data Cohorts



In a groundbreaking study conducted at Charité – Universitätsmedizin Berlin, Briya has demonstrated the potential of its AI-driven platform, AIRE, in advancing research methodologies by enhancing transparency and ensuring accuracy in handling complex healthcare data. Briya's AIRE platform serves as a bridge, transforming fragmented healthcare datasets into reliable and defensible evidence. The recent evaluation conducted at one of Europe’s largest university hospitals highlighted how query-level transparency in the AIRE platform enables researchers to validate AI-generated cohort results effectively.

The Study: Validating AI Cohort Results



The study aimed to assess the accuracy of AI-generated results compared to traditional hand-written SQL queries. Out of 23 responses generated by AIRE, 22 were found to be correct or could be traced back to specific clauses in the originally generated SQL. This remarkable accuracy is especially significant as it covers a variety of cohort analyses that are often considered challenging by benchmark studies. The researchers pointed out that without the clear visibility into the underlying queries, distinguishing accurate results from plausible-sounding inaccuracies would be nearly impossible.

Dr. Dmitriy Desser from the Department of Neuroradiology at Charité remarked, "Routine hospital data serves as a robust foundation for research, yet minute definitional variances can drastically alter cohort compositions. The query transparency provided by AIRE enabled us to pinpoint and elucidate these differences. Without such transparency, the appraisal we conducted would not have been feasible."

Revamping Research Methodologies



Traditionally, assembling patient cohorts for clinical studies is a time-consuming process often spanning several weeks, requiring custom database programming. Yet, the AIRE platform reveals how natural language can expedite this workflow. Utilizing AIRE, the researchers converged on 25 natural-language questions regarding adult patients diagnosed with malignant brain tumors and subsequently refined 12 of these questions in a new dialogue stream. The AI was juxtaposed against the hospital’s original SQL data queries, operating entirely within Charité’s network and adhering to its stringent data privacy regulations.

The capability of AIRE to build analytical methodologies in tandem with researchers allows for automated SQL generation, which is displayed alongside analytical results. This means when uncertainties arise, AIRE seeks clarifications, thereby fostering a comprehension of the analytical process. The authors also acknowledged AIRE’s potential to propose alternative definition sets and to reveal limitations when the data does not support the intended queries.

The Path Forward for AI in Research



Dr. Or Shaked, Executive Director of RWE Medical Research at Briya, elaborated on the company’s approach: "Our continual testing throughout AIRE's development highlighted the importance of traceability in AI-generated analyses and the necessity of constant monitoring. Charité’s real-life evaluation took our insight to new heights by scrutinizing our platform in a true clinical setting. It underscored that a number is only as dependable as the visibility of the underlying query."

This emphasis on transparency is echoed by Jason Hein, Chief Commercial Officer at Briya, who stated, "Our pride lies in the results of this study and the methodological rigor behind it. Trust in AI for scientific inquiry is established through transparency and the vigilance of the scientists employing it. By consistently introducing AIRE to customers and peer institutions like Charité, we are embracing scrutiny, which in turn reinforces our commitment to transparency. This is our strategy for leading in this market."

The preprint detailing this research can be accessed at medRxiv. Further discussions on the findings will be presented in a webinar scheduled for November 18, 2026, with registration information available on Briya’s official website.

Conclusion



As healthcare research increasingly integrates AI technology, the lessons from Briya’s project stand out. The precise engineering of methodologies equipped with clear audit trails can revolutionize how researchers interact with data and derive insights. By substituting traditional methods with AI tools like AIRE that maintain accuracy and transparency, Briya is positing itself at the forefront of next-generation healthcare research solutions, reaffirming the significance of meticulous data governance and scientific accountability.

Topics Health)

【About Using Articles】

You can freely use the title and article content by linking to the page where the article is posted.
※ Images cannot be used.

【About Links】

Links are free to use.