Adaptive AI-based mobile simulation for drug calculation competency in nurses: a mixed-methods RCT

Medication dose miscalculations remain one of the most persistent patient safety concerns in clinical practice, while traditional teaching methods often fall short in ensuring proficiency. The integration of artificial intelligence (AI) based learning tools has the potential to enhance knowledge, clinical decision-making (CDM), and self-efficacy among nursing staff.

To evaluate the effectiveness of a researcher-developed artificial intelligence-based mobile learning application designed to simulate realistic hospital medication environments with adaptive feedback on nursing's drug calculation knowledge, decision-making, and self-efficacy, while exploring its feasibility and acceptability in routine training contexts.

Mixed-methods study with an embedded design, utilizing two-arm randomized controlled trial for quantitative evaluation and focus group discussions for qualitative exploration.

The study was conducted at a non-governmental hospital where 56 nurses, randomly assigned to intervention group (n = 28) or a control group (n = 28). The mobile learning application simulated realistic hospital medication administration scenarios incorporating adaptive corrective feedback and progressively complex drug calculation cases. Data was collected at baseline and post-intervention using a drug dose calculation knowledge assessment, a clinical decision-making scale, and a self-efficacy scale. Post-intervention focus group discussions with the intervention group were analyzed using thematic analysis.

The study group achieved significant improvements in knowledge, decision-making, and self-efficacy post-intervention compared to the control group with large effect sizes. Qualitatively, six themes emerged which discussed several aspects of the training such as; perceived clinical benefits, realism, increased nurses' confidence, ease of use and engagement, training efficiency and challenges and the potential for improvement.

This is one of the first randomized controlled trials (RCTs) conducted using mixed-methods and focusing on training nursing staff on drug-dose calculation with artificial intelligence (AI) based simulation. Results indicated that the intervention was associated with significant improvements in the measured outcomes and was perceived positively by the participants. However, its broader implementation across different contexts requires further investigation.

ClinicalTrials.gov Identifier: NCT07448259. Retrospectively registered on February 18, 2026. Registration Link: https://clinicaltrials.gov/study/NCT07448259 .