Hit records

Medical data management must mean a role for AI

The review and analysis of medical records are critical tasks in the pharmaceutical and life sciences industries, crucial for validating clinical research, supporting epidemiological studies and ensuring regulatory compliance.

But the volume of medical records to be evaluated can be overwhelming, especially given the increasing complexity and regulatory requirements.

To address these challenges, many organisations are turning to AI and advanced cloud-based data analytics solutions.

Key documents include patient charts, diagnostic reports, medication histories and test results. The comprehensive nature and volume of these records’ present significant challenges, as manually reviewing them is time-consuming and often takes experts away from more strategic work.

Traditional manual approaches are increasingly proving inadequate. The reliance on time-intensive review cycles consumes valuable time and resources while posing risks of error and inconsistency. Furthermore, records often come in various formats including PDFs, image scans, typed and handwritten text, checkboxes, and charts, making it difficult to synthesise and consolidate information efficiently.

This process can be a major bottleneck, diverting resources that could be better spent on core activities such as developing new therapies, conducting clinical trials and ensuring regulatory compliance.

AI-driven technologies, such as those incorporating advanced document vision techniques, can streamline and automate the processing and interpretation of medical data.

These solutions handle diverse document types, including handwritten and form-based records, and tailor data extraction to specific needs.

AI technology declutters and summarises critical medical information, enabling reviewers to make determinations quickly and accurately. By contextualising forms and layering nuanced text analytics, AI ensures that essential information is readily available, significantly reducing the time spent searching for critical details.

In our previous column, we explored how large language models (LLMs) are expediting document creation and summarisation in clinical trials.

Similarly, AI solutions for medical record review aims to streamline complex processes, improving data accuracy and efficiency.

The impact of AI in medical record review has been shown in various real-world applications. For instance, a collaboration between SAS and a major US healthcare group resulted in a 70% increase in efficiency over two years. This translated to labour savings equivalent to 80 full-time employees or $12 million annually. AI solutions processing millions of images daily demonstrate their scalability and reliability. This is just one example of how advanced technology is revolutionising medical record review.

The core of these innovations lies in their technical capabilities. Demonstrations of AI technology show the ability to identify medications, patient visits, diagnoses and histories, presenting this information in a standardised, easily navigable format.

Users can analyse results on a larger scale using real-time dashboards, synthesising extracted data alongside structured information.

AI-driven solutions for medical record review represent a significant advancement in pharmaceutical efficiency.

By optimising workflows and reducing the time spent on data sifting, these technologies enable reviewers to focus more on applying their expertise to deliver high-quality results.
As the industry evolves, leveraging advanced technologies will be essential in driving progress and achieving impactful outcomes in pharmaceutical and life sciences research.

Kayt Leonard is Global Health and Life Sciences Advisor at SAS. Go to sas.com