This systematic review evaluated natural language processing (NLP) approaches for detecting cognitive impairment in electronic health record clinical notes. Following PRISMA guidelines, we analyzed 18 studies (n = 1,064,530) that employed rule-based algorithms (67%), traditional machine learning (28%), and deep learning (17%). NLP models demonstrated robust performance in identifying cognitive decline, with median sensitivity 0.88 (IQR 0.74–0.91) and specificity 0.96 (IQR 0.81–0.99). Deep learning architectures achieved superior results, with area under the receiver operating characteristic curves up to 0.997. Major implementation challenges included incomplete electronic health record data capture, inconsistent clinical documentation practices, and limited external validation. While NLP demonstrates promise, successful clinical translation requires establishing standardized approaches, improving access to annotated datasets, and developing equitable deployment frameworks.

Objectives This study conducted a comprehensive probabilistic cost-effectiveness analysis comparing robotic exoskeleton therapy to conventional physiotherapy for stroke rehabilitation in Singapore, focusing on three patient groups categorised by their Functional Ambulation Category (FAC) scores. Design A probabilistic cost-effectiveness analysis was conducted alongside a non-randomised controlled study. Costs and Quality-Adjusted Life Years (QALYs) for both interventions were calculated and compared over a 6 month period. Setting The study was carried out at Alexandra Hospital, Jurong Community Hospital and St Luke’s Hospital in Singapore. Participants Individuals requiring inpatient gait rehabilitation from acute to subacute stages of stroke recovery, with FAC scores of 0–1, were included in the analysis. Primary outcome measure The primary outcome measure was QALYs, a composite measure combining the length and quality of life into a single value. Results Robotic exoskeleton therapy was found to be cost-effective compared with conventional physiotherapy across all patient groups, with Group 2 (FAC 0) showing the most favourable cost-effectiveness profile (incremental cost-effectiveness ratio (ICER): US$ 28 259.62 per QALY gained). The probabilistic sensitivity analysis demonstrated the robustness of the results, with QALY gains and the cost of the robotic exoskeleton having the largest impact on the ICER. Conclusion The findings suggest that robotic exoskeleton therapy is likely to be cost-effective for stroke rehabilitation in Singapore, particularly for patients with severe mobility impairments (FAC 0). The results have important implications for clinical practice, resource allocation and future research in the field of stroke rehabilitation in Singapore. Trial registeration number NCT05659121.

Artificial intelligence models analyzing speech show remarkable promise for identifying cognitive decline, achieving performance comparable to clinical assessments. However, their “black box” nature poses significant barriers to clinical adoption, as healthcare professionals require transparent decision-making processes. This challenge is compounded by regulatory requirements, including GDPR mandates for explainability and medical device regulations emphasizing AI transparency. Following PRISMA guidelines, we systematically reviewed explainable AI (XAI) techniques for speech-based detection of Alzheimer’s disease and mild cognitive impairment across six databases through May 2025. From 2077 records, 13 studies met the inclusion criteria, employing XAI methods including SHAP, LIME, attention mechanisms, and novel approaches across machine learning architectures. Models achieved AUC values of 0.76-0.94, consistently identifying acoustic markers (pause patterns, speech rate) and linguistic features (vocabulary diversity, pronoun usage). While XAI techniques demonstrate promise for clinical interpretability, significant gaps remain in stakeholder engagement, real-world validation, and standardized evaluation frameworks.