Discrepancy between perceptions and acceptance of clinical decision support Systems: implementation of artificial...

Area of Science:

• Clinical Pharmacy
• Health Informatics
• Artificial Intelligence in Medicine

Background:

• Artificial intelligence (AI) tools require clinician acceptance for optimal effectiveness.
• An AI-based clinical decision support system (CDSS) was developed to aid vancomycin dosing.
• This study explored clinician perceptions of CDSS implementation.

Purpose of the Study:

• To assess critical care pharmacists' perceptions of an AI-based CDSS for vancomycin dosing.
• To understand barriers and facilitators to adopting AI tools in clinical practice.

Main Methods:

• Qualitative study involving 13 semi-structured interviews with critical care pharmacists.
• Discussion of 104 clinical cases, comparing pharmacist empiric doses with CDSS recommendations.
• Analysis of pharmacist reactions and feedback upon revealing CDSS outputs.

Main Results:

• Pharmacists spend significant time individualizing vancomycin therapy.
• Most pharmacists believed a CDSS could enhance vancomycin dosing decisions (62-69%).
• Significant discrepancies between pharmacist and CDSS doses (85%); pharmacists often declined CDSS recommendations due to distrust and data limitations.

Conclusions:

• Pharmacists expressed enthusiasm for AI in drug dosing but showed reluctance towards clinical integration.
• Barriers include distrust, lack of dynamic analysis, and incomplete data integration.
• Further research is essential to optimize CDSS implementation for clinical acceptance and improved patient outcomes.