Abstract
<sec> <title>BACKGROUND</title> <p>Intensive care unit (ICU) clinicians must detect and interpret changes across multiple physiologic data streams while managing interruptions, alarms, and competing tasks. Conventional monitors preserve exact values and waveforms but often require serial visual search and mental integration. The Visual Patient Avatar (VPA) uses direct information representation, mapping vital signs to patient-like phenomena such as body pulsation, lung motion, and cyanotic skin coloration. A critical synthesis is needed because the evidence spans laboratory experiments, simulation, qualitative studies, and early implementation work.</p> </sec> <sec> <title>OBJECTIVE</title> <p>This narrative review evaluates the human factors evidence for VPA in ICU-relevant monitoring, distinguishes supported effects from unproven clinical claims, and clarifies its distinctive role within the broader field of human-centered display design.</p> </sec> <sec> <title>METHODS</title> <p>PubMed/MEDLINE and Google Scholar were searched from January 2018 through July 16, 2026, using terms for Visual Patient, Visual Patient Avatar, avatar-based monitoring, and Visual Patient Predictive. Reference lists and forward citations were checked. The core evidence set comprised 22 peer-reviewed VPA-specific publications (21 empirical studies and 1 protocol) plus 1 preprint protocol. Selected literature on situation awareness, display design, alternative systems, and regulatory human factors informed interpretation. Findings were grouped by monitoring task, setting, outcome type, and proximity to patient outcomes. Reporting was informed by the Scale for the Assessment of Narrative Review Articles.</p> </sec> <sec> <title>RESULTS</title> <p>VPA showed its clearest and most consistent advantages when clinicians had to extract information rapidly or under constrained viewing. In the foundational study of 32 anesthesia professionals, participants recalled a median of 9 vital signs with VPA versus 5 with conventional monitoring; workload was lower (National Aeronautics and Space Administration Task Load Index score 60 vs 76). In a 5-center study of 50 ICU clinicians, VPA increased the overall correct-response rate (rate ratio 1.25, 95% CI 1.19-1.31). Diagnostic confidence also increased (odds ratio 3.32, 95% CI 2.15-5.11), while workload decreased (coefficient -7.62, 95% CI -9.17 to -6.07). Advantages were also reported for peripheral and distant viewing, standardized distraction, and short multi-patient tasks. In high-fidelity crisis simulation, split-screen monitoring was noninferior for critical-task performance. Noninferiority was not established for avatar-only critical-task performance, but avatar-only monitoring increased the probability of verbalizing the correct cause of the emergency (hazard ratio 1.78, 95% CI 1.13-2.81), while split-screen was noninferior for this outcome. Perceived workload did not differ, and lower immediate usability ratings were observed in participants with minimal prior avatar experience. Early first-exposure studies identified concerns about missing exact values, unfamiliarity, and occasional visual crowding when many variables were abnormal; later clinical implementation studies mainly favored split-screen use and emphasized individualized thresholds and local adaptation, with visual overload no longer emerging as a prominent recurring theme. No study demonstrated improved patient outcomes.</p> </sec> <sec> <title>CONCLUSIONS</title> <p>The evidence supports VPA as a clinically relevant overview layer for rapid perception and initial comprehension, complementing rather than replacing numerical values, waveforms, trends, bedside assessment, and clinical reasoning. Converging task-based findings support its fit with interruption-rich ICU work, although direct ICU evidence remains limited and proximal to care. Independent, pragmatic studies should test recognition time, alarm management, time outside target ranges, workload, use errors, and patient-centered outcomes.</p> </sec> <sec> <title>CLINICALTRIAL</title> <p>not applicable</p> </sec>