Digital Tools for Mild Cognitive Impairment: A Systematic Review and Meta-analysis of Diagnostic Accuracy and Methodological Challenges

Abstract Mild cognitive impairment (MCI) represents an intermediate stage between typical aging and early cognitive decline. As such, an early and accurate diagnosis is essential in making timely interventions. Digital tools, including mobile applications, web platforms, wearable devices, and artificial intelligence-driven systems, have been developed and validated to capture multidimensional data, offering innovative screening solutions. This meta-analysis aims to evaluate the diagnostic accuracy of digital tools for MCI detection in different populations and settings, with a particular focus on three key issues: (i) the overall diagnostic performance of digital tools, (ii) the influence of methodological quality of studies, and (iii) the impact of demographic factors and familiarity with technologies on diagnostic accuracy. This meta-analysis assessed diagnostic accuracy across 32 studies, reporting pooled sensitivity (0.808, 95% CI: 0.775–0.838) and specificity (0.795, 95% CI: 0.757–0.828), but with considerable heterogeneity ( I 2 = 71.5% sensitivity; 84.0% specificity). The HSROC analysis revealed significant intrinsic variability ( τ ₐ = 0.807) and minimal threshold variability ( τθ = 0.291). Meta-regression indicated that applicability concerns significantly reduced specificity ( p = 0.037), with older age also predicting lower specificity ( p = 0.029). Thus, implementing standardized protocols, rigorous validation processes, and targeted adaptations are crucial steps for enhancing the effectiveness of digital tools in detecting MCI.