Avatar-based patient monitoring improves information transfer, diagnostic confidence and reduces perceived workload in intensive care units: computer-based, multicentre comparison study

Patient monitoring is the foundation of intensive care medicine. High workload and information overload can impair situation awareness of staff, thus leading to loss of important information about patients' conditions. To facilitate mental processing of patient monitoring data, we developed the Visual-Patient-avatar Intensive Care Unit (ICU), a virtual patient model animated from vital signs and patient installation data. It incorporates user-centred design principles to foster situation awareness. This study investigated the avatar's effects on information transfer measured by performance, diagnostic confidence and perceived workload. This computer-based study compared Visual-Patient-avatar ICU and conventional monitor modality for the first time. We recruited 25 nurses and 25 physicians from five centres. The participants completed an equal number of scenarios in both modalities. Information transfer, as the primary outcome, was defined as correctly assessing vital signs and installations. Secondary outcomes included diagnostic confidence and perceived workload. For analysis, we used mixed models and matched odds ratios. Comparing 250 within-subject cases revealed that Visual-Patient-avatar ICU led to a higher rate of correctly assessed vital signs and installations [rate ratio (RR) 1.25; 95% CI 1.19-1.31; P < 0.001], strengthened diagnostic confidence [odds ratio (OR) 3.32; 95% CI 2.15-5.11, P < 0.001] and lowered perceived workload (coefficient - 7.62; 95% CI - 9.17 to - 6.07; P < 0.001) than conventional modality. Using Visual-Patient-avatar ICU, participants retrieved more information with higher diagnostic confidence and lower perceived workload compared to the current industry standard monitor.

Automated summary

Patient monitoring is crucial in intensive care medicine, but high workload and information overload can hinder staff's situation awareness. To address this, a virtual patient model animated from vital signs and patient installation data was developed to improve mental processing of patient monitoring data. This study found that compared to conventional monitoring, the virtual patient model led to higher accuracy in assessing vital signs and installations, increased diagnostic confidence, and reduced perceived workload.