Impact of a mobile decision-support tool on the efficiency and accuracy of CIED data retrieval: a comparative pilot study

Background/Introduction

Rapid access to technical information on cardiac implantable electronic devices (CIEDs) is essential during implantation, follow-up, and troubleshooting. Conventional sources, such as manuals or static PDF documents, are often time-consuming and prone to errors. Digital tools and AI-based search engines may streamline information retrieval and support clinical workflows, although their accuracy for highly specialised content has not been thoroughly assessed.

Purpose

The aim of this multicentre pilot study was to compare the efficiency and accuracy of retrieving CIED-specific information using a mobile-based reference app, AI-assisted search, and conventional methods.

Methods

A comparative study was conducted in 18 hospitals, involving 24 physicians and 13 technicians/nurses, all experienced in CIEDs. Participants were asked to retrieve answers to four structured questions, each focusing on a specific technical aspect of cardiac implantable electronic devices: (1) identifying the device connector type, (2) determining the theoretical device longevity, (3) verifying the MRI compatibility of the device–lead system, and (4) retrieving information on an official advisory.

Each question was answered using three methods: (1) conventional sources, (2) AI-assisted search, and (3) a dedicated mobile reference database specifically structured for CIED data. For each question, retrieval time was measured, and accuracy was evaluated by comparison with the manufacturers’ official technical documentation.

Results

Across all participants, the mean time per question decreased from 264 ± 277 s using conventional methods to 56 ± 29 s with AI and 53 ± 44 s with the app (−79.9 % vs. baseline; p < 0.001). The percentage of correct answers increased from 56.3 % (baseline) to 80.0 % with AI and 95.7 % with the app. Improvements were observed consistently in both physicians (53.4 % to 97.6 %) and technicians/nurses (60.7 % to 92.9 %).

While the AI tool achieved 80.0 % accuracy, errors were due to incorrect or incomplete information. The mobile application reached 95.7 % accuracy, with residual errors attributable to user interpretation rather than content inaccuracy. This finding highlights the importance of accuracy in device-related information sources, particularly when such data are used to support healthcare professionals.

Conclusion(s)

The use of a mobile CIED reference database significantly improved both efficiency and accuracy in retrieving device-specific information compared with conventional methods. While AI-assisted search reduced retrieval time, the lower accuracy and frequent occurrence of incorrect or incomplete information highlight potential risks when using generic AI tools for specialised technical content. Domain-specific, structured reference tools may provide valuable educational and informational support to healthcare professionals, optimising information retrieval in clinical workflows.Figure 1Figure 2