Search

Clinical decision latency, defined as the temporal interval from the moment actionable clinical data becomes available to the initiation of a corresponding therapeutic or diagnostic action, constitutes an under-recognized yet critical safety variable in contemporary healthcare delivery. Prevailing patient safety paradigms predominantly concentrate on categorical errors of commission or omission while largely treating time as an exogenous operational factor rather than an intrinsic propagative risk element capable of independently driving harm. This conceptual systems article reframes clinical decision latency as a primary, quantifiable, and governable safety variable. It proposes the clinical latency oversight lattice (CLOL)—an original infrastructural framework specifically designed to detect, quantify, assign accountability for, and interrupt harmful temporal delays across care pathways. Drawing on a targeted synthesis of literature that collectively addresses clinical decision support limitations, diagnostic uncertainty propagation, consequences of treatment delays, health IT-induced temporal vulnerabilities, and AI integration challenges, the manuscript argues that latency functions not as mere logistical inefficiency but as a dynamic, modality-sensitive, context-dependent risk multiplier. The CLOL architecture organizes temporal accountability into four interdependent lattice layers linked by a bidirectional feedback topology that enables real-time drift monitoring, explicit actor/system responsibility mapping, safety-variable score propagation, and orchestrated mitigation responses. Three interpretive mathematical expressions capture core dynamics: risk propagation across pathways, exponential decay of decision confidence under accumulating latency, and cumulative governance/monitoring burden. By institutionalizing latency as a traceable safety variable within a closed-loop accountability structure, CLOL offers healthcare analytics and AI system designers a theoretical and architectural foundation for shifting from retrospective error analysis toward prospective temporal harm prevention in high-stakes clinical environments.

The integration of artificial intelligence into clinical decision support systems offers improved diagnostic accuracy and efficiency, but the opacity of many machine learning models raises concerns about trust, accountability, and regulatory compliance. Explainable artificial intelligence (XAI) has been proposed to address this by making model predictions interpretable to clinicians; however, its true clinical value remains uncertain, and evaluation has not kept pace with methodological development. This systematic review aimed to identify XAI methods used in clinical decision support systems, assess how they are evaluated with clinicians, and determine whether explanations improve diagnostic accuracy, trust, mental models, and efficiency. Following PRISMA guidelines, we searched PubMed, Web of Science, IEEE Xplore, ACM Digital Library, and Scopus for studies published between 2017 and 2024. Eligible studies included original research evaluating XAI in clinical decision support systems with clinician participants and reporting quantitative or qualitative outcomes. Risk of bias was assessed using adapted QUADAS-2 and ROBIS tools, and findings were synthesized narratively with subgroup analyses. From 2,847 records, 68 studies were included. The most common XAI methods were SHAP-based feature attribution (38%), saliency or heatmap methods (29%), concept-based approaches such as TCAV (15%), and counterfactual or example-based explanations (12%). Radiology was the dominant field (54%), followed by dermatology (18%) and pathology (12%). Evaluation approaches were highly inconsistent, with few validated instruments and most studies relying on Likert-scale trust measures or qualitative feedback. Only 16% of studies showed improved diagnostic accuracy with explanations, 67% showed no significant effect, and 17% reported reduced accuracy due to over-reliance or misinterpretation. Although 82% of studies reported increased clinician trust, trust rarely correlated with actual diagnostic performance. Overall, while XAI methods are widely studied in clinical decision support, their evaluation is inconsistent and their benefits are limited. Explanations tend to increase clinician trust without reliably improving diagnostic accuracy, and may sometimes worsen performance, highlighting a trust–accuracy gap that poses important safety concerns for clinical deployment.