Document Type

Article

Publication Date

3-30-2026

Abstract

Background: Advancements in medical technology, particularly in perfusion practices such as ECMO and heart-lung bypass, have improved patient care but also introduced novel risks. Technological errors—ranging from hardware malfunctions to software failures and AI model inaccuracies—can directly threaten patient safety.

Objective: To identify current trends and risks associated with technology in perfusion practices and synthesize evidence on the frequency, nature, and outcomes of technology-related errors.

Results: Evidence indicates substantial deficiencies in medical devices, predictive models, and health IT systems. AI-based predictive models frequently failed to detect critical health deterioration, missing 66% of injuries in some in-hospital mortality simulations. Adverse events linked to medical device hardware and software contributed to over 13,500 deaths and 1.55 million injuries over a 3.5-year period. Perfusion-related incidents occur in approximately 1 per 138 cases, with serious injury or death in 1 per 1,453 procedures. Common perfusion-related hardware issues included heater/cooler failures, pump stoppages, and defective clamps. Software-related errors, including touchscreen resets, sensor misreads, and false alarms.

Conclusions: Technology in perfusion practices is a double-edged sword. While it facilitates clinical care, persistent vulnerabilities in devices, software, and predictive models pose ongoing patient safety risks. Strengthened monitoring, standardized reporting, rigorous testing, and enhanced human–machine interfaces are essential to mitigate preventable harm.

Keywords: e-iatrogenesis, perfusion-related, medical devices, artificial intelligence

Download

Share

COinS