Is accuracy enough? trust and barriers to AI-based clinical decision support in clinical neurophysiology

Abstract

Objective

To investigate attitudes, trust, expectations and concerns among patients and healthcare personnel regarding the use of AI-based clinical decision support systems (AI-CDSS) in clinical neurophysiology.

Methods

We iteratively designed a questionnaire and conducted a national survey of patients and healthcare personnel in Norway. Patients were recruited from a quality registry, while healthcare personnel and technicians were recruited from hospital laboratories. Items included binary and Likert-scale questions addressing attitudes, expectations, concerns and trust, as well as a scenario-based trust assessment. Between-group comparisons were performed using non-parametric and categorical statistical tests.

Results

Both groups reported positive attitudes toward AI-CDSS. Patients more often expected AI-CDSS to improve diagnostic accuracy (p = 0.005), while healthcare personnel anticipated improved equity of services across regions (p = 0.017). Key concerns in both groups were related to responsibility, doctor-patient relationships, AI literacy, overreliance on AI, and potential deskilling. The scenario-based question revealed lower trust when AI-CDSS directly influenced decision-making without physician involvement. Trust in public institutions was high, whereas trust in commercial or industry use of health data was consistently low.

Conclusions

Attitudes toward AI-CDSS in clinical neurophysiology are positive, but diagnostic accuracy isn’t enough: Acceptance is conditional on human oversight, delineated responsibilities, and steps taken to improve transparency and minimize deskilling. Patients want to share health data for AI purposes, but the commercial sector needs to build trust.

Significance

AI-CDSS in clinical neurophysiology is promising, but design and implementation need to address key concerns shared by both patients and health care personnel.