Blog
Addressing the challenges of monitoring unstructured safety data
Anuradha Prabhakar, Associate Director, Product Management, Safety Technology
Nov 06, 2024

The amount of unstructured safety data is surging in the healthcare and life sciences industry. Numerous sources contribute to this data, but audio has emerged as a critical yet complex contributor. Medical audio data, including recordings from clinicians, amounts to a staggering 1.5 billion hours annually. This ‘rise in audio’ as a source of safety data presents a significant challenge to commercial and medical information call center teams who are required to analyze this information to identify adverse events (AEs) and product complaints (PCs). Traditional methods are not equipped to handle such a vast volume of information, often leading to missed safety events and delayed responses.

The challenges of monitoring unstructured safety data from audio

Audio data, in the context of pharmacovigilance primarily stems from patient support and medical information call centers. This data introduces several variables into safety operations:

  1. Complexity of audio data: A call is rarely perfect, and these audio files are no exception. Audio quality can vary due to factors such as background noise, technical glitches and inconsistent recording standards, making manual review tedious and error prone.
  2. Integration gaps: Managing multiple disconnected systems leads to poor data intake and subpar coordination between call center and pharmacovigilance (PV) workflows. This lack of integration results in an increased number of touchpoints, ultimately affecting the quality of intake and reporting.
  3. Manual burden on agents: The traditional process of listening, transcribing and documenting safety events is labor-intensive, diverting call agents’ focus from patient-centric activities and potentially leading to missed AEs and PQCs.

Given these persistent challenges, there is an immediate need for effective, scalable solutions to streamline safety operations, reduce manual intervention and minimize the risk of missed safety events.

Next-generation solutions for optimizing audio source data workflows

IQVIA has re-imagined the traditional call center workflow for safety. Blending a mix of automation approaches and a re-engineered process, are changing how organizations approach unstructured audio data. This combination of technology and workflow innovation has enabled batch processing of audio files, automated transcription and analysis, allowing for faster, more accurate AE reporting.

Key components of the digitally enhanced workflow:
  1. Voice-to-text transcription: Automated transcription converts audio recordings into structured text formats, allowing for rapid analysis.
  2. Safety ontology-enhanced NLP accurately identifies potential adverse event and product complaint terms and timestamps patient transactions of interest, achieving high accuracy even in native language translation and reducing the need for manual listening.
  3. RPA for data extraction and integration: RPA tools automate the extraction of safety data elements from text files and map them to predefined templates like CIOMS forms, E2B format, eliminating manual data entry errors and accelerating reporting times.
Real world impact: efficiency and quality gains

In a recent case study, IQVIA implemented this strategy to monitor a bot that was deployed to liaise with insurance company on behalf of patients. 100% manual review of calls would otherwise be required to ensure potential AEs were reported. By automating quality control of call reviews, the company managed to handle 60,000 calls annually with only 1.3 full-time employees, compared to the 22 FTEs that would have otherwise been required using traditional methods.

Similarly, in a project involving a major pharmaceutical company, this strategy resulted in the identification of 8331 initial events never previously reported by call center agents and 8510 follow ups containing significant updates.

Upstream and downstream benefits for safety operations

Upstream, call center agents are freed from the administrative burden of AE/PC reporting, allowing them to return their focus to the core of their jobs, patient engagement and care. Downstream, the automation of PV intake, data extraction and reporting reduces the time taken to book safety events into the safety system, enhances data accuracy and supports a more solidified safety surveillance system.

The future of audio data management in pharmacovigilance

As AI technology continues to evolve, its application in PV will extend beyond transcription and reporting to include more sophisticated capabilities such as real-time risk assessment, predictive analytics and proactive case management. This shift will enable life sciences organizations to harness unstructured data more effectively, ensuring comprehensive safety surveillance and improved patient outcomes.

With AI-powered solutions for unstructured audio data now available, organizations have the opportunity to transform their safety operations and achieve new levels of efficiency and quality. Embracing these advancements is key to staying ahead in a rapidly changing regulatory landscape.

To learn more about IQVIA’s next-generation solutions for applying AI in the PV workflow to rapidly review your data, reduce your resource needs and remove risk, visit us at our website or reach out at VigilanceDetect@iqvia.com.

Contact Us