Nursing Perceptions of the Intended Use of Artificial Intelligence to Prevent Medication Errors: A Qualitative Descriptive Study

ABSTRACT

Aim

To explore the perceptions of nursing professionals in high-demand healthcare services regarding the adoption of AI-based support systems for the prevention of medication errors.

Design

A qualitative descriptive study was conducted between November 2024 and March 2025.

Methods

Sixteen semi-structured interviews were held with nurses from emergency and intensive care units, guided by conceptual dimensions of the Technology Acceptance Model framework. Participants were recruited using purposive and snowball sampling. ATLAS.ti v.9 software was used for an inductive thematic analysis.

Results

Two major themes emerged: (i) professional reflections on medication safety and related risks; and (ii) integrating artificial intelligence into nursing practice to reduce such risks and prevent medication errors. While artificial intelligence was recognised as a promising resource to support clinical decision-making and reduce cognitive load, nurses identified barriers, including limited training, inadequate technological infrastructure, unreliable data sources, and ethical concerns that could compromise its safe implementation and thereby hinder its potential to prevent medication errors.

Conclusion

AI-based support systems are perceived as useful, but complex resources for addressing medication errors, which remain a critical challenge in healthcare. Its successful implementation depends not only on the availability of resources, but also on the organisational context and the ability to respond to the needs and concerns of healthcare professionals.

Implications for Clinical Practice

Integrating artificial intelligence into routine workflows to support clinical decision-making and reduce medication errors in high-demand settings requires more than infrastructure and technical training. Effective adoption demands participatory design, clear role delineation, and context-sensitive training aligned with medication-management processes. Lack of alignment may result in artificial intelligence increasing complexity instead of contributing to safer and more efficient medication administration.

Reporting Method

Methods and findings are reported following SRQR recommendations.

Patient or Public Contribution

No patient or public contribution.