Search Results (240)

Retrieval-augmented generation (RAG) is increasingly deployed in safety-critical cybersecurity question answering, where an answer must be both factually correct and faithful to its sources. Evaluation in this space routinely conflates the two, treating grounding as a proxy for correctness. We show that this conflation is unsafe. We introduce SENTINEL-QA, a bilingual English and Macedonian cybersecurity QA benchmark of 70 hand-authored items over a curated corpus of authoritative source documents, and use it to compare four answer generation pipelines (zero-shot, Classic RAG, DSPy-structured RAG, and GraphRAG) across three generators, judged by a four-model LLM panel on both factual correctness and grounding. The two metrics decouple under graph-based retrieval augmentation: on all three generators our chunk-similarity GraphRAG makes answers look better sourced without making them more correct, and as the retrieval budget grows it can even degrade factual correctness while grounding holds steady. Structured, minimal prompting improves both qualities, at lower latency and cost; a four-cell ablation attributes the factual gain to pruning the instruction, while the declarative schema contributes on grounding. The judge panel mirrors the split, agreeing more on what is well grounded than on what is true, under absolute as well as rank-based agreement criteria. We conclude that grounding alone is an insufficient target for RAG evaluation and that pruning the instruction, rather than expanding retrieval, is what improves factual correctness, without eliminating the grounded-but-wrong regime. Full article

Polypharmacy in older adults is common, rising and inconsistently defined, and it is widely treated as a problem of how many medicines a person takes. This non-systematic narrative review argues that its harm depends more on the appropriateness of the regimen and the pharmacological burden the regimen imposes, with the number of drugs a partly confounded marker rather than the principal cause. Drawing on evidence published mainly since 2020, it traces the shift from counting medicines to weighting them; the prescribing cascade as a patterned and detectable source of inappropriate use; the explicit criteria (Beers, STOPP/START, EURO-FORTA) and implicit measures (the Medication Appropriateness and Drug Burden indices) that give appropriateness operational form; and the evidence that deprescribing is feasible and generally safe, although its benefit is conditional and depends on how it is done. Because these criteria can be applied to medication lists held in routine data, potentially inappropriate prescribing can increasingly be identified beyond the single encounter and at population scale, even as automated tools remain unreliable at judging whether a flagged criterion applies to the individual patient. The review concludes that medication review should prioritise appropriateness and the individual benefit–risk balance rather than the medication count, and that detection can increasingly be scaled even though the clinical decision and the patient’s agreement cannot. Full article

Background/Objectives: Type 2 diabetes (T2D) is a multifactorial disorder arising from complex interactions between genetic susceptibility and environmental exposures, including diet. This systematic review aimed to synthesize and critically evaluate evidence on gene–diet interactions in T2D-related phenotypes among adults without diagnosed T2D at baseline. Methods: The review was conducted according to PRISMA guidelines and registered in PROSPERO (CRD420251175793). PubMed and Embase were systematically searched for eligible studies published up to November 2025. Observational studies and intervention trials examining interactions between dietary exposures and genetic variants, haplotypes, or genetic risk scores (GRSs) in relation to incident T2D, insulin resistance, glycemic traits, or β-cell function were included. Risk of bias was assessed using study design-specific tools. Results: A total of 61 reports representing 27 unique studies met the eligibility criteria, comprising 30 reports of randomized trials and 31 reports of non-randomized studies. Across studies, 55 Single Nucleotide Polymorphisms (SNPs) in 40 genes were examined, while 11 studies evaluated GRSs. The most frequently studied loci included TCF7L2, ADIPOQ, FTO, IRS1, GIPR and PPM1K. Reported interactions involved dietary patterns, macronutrient composition, dietary fat quality, fiber/whole-grain intake and specific foods, influencing incident T2D, insulin-related traits and glycemic outcomes. However, findings were heterogeneous, frequently inconsistent, and often lacked independent replication. Most randomized and non-randomized studies were judged at high or serious risk of bias. Conclusions: Genetic variation may contribute to interindividual differences in metabolic responses to diet, although robust and reproducible gene–diet interactions remain limited. Larger, well-powered, standardized studies across diverse populations are needed to establish clinically meaningful applications of Precision Nutrition (PN) in T2D prevention. Full article

Sport-related concussion (SRC) rehabilitation has advanced toward active, multidomain management, yet recovery may still be judged largely through symptom resolution and broad clinical indicators that may not fully capture persistent functional deficits. Emerging evidence suggests subtle disturbances in postural control, gait, and sensorimotor coordination may persist beyond apparent clinical recovery, raising the possibility that unresolved asymmetries represent an underrecognized dimension of dysfunction. This Viewpoint proposes symmetry restoration as a potential rehabilitative construct in SRC management and explores how biofeedback-guided approaches may provide a conceptual framework for identifying, monitoring, and retraining symmetry-related deficits. Drawing from concussion research, motor control theory, rehabilitation science, and biofeedback applications, this article discusses postural and movement asymmetries as possible markers of incomplete recovery, examines visual, wearable, neuromuscular, and auditory biofeedback strategies as potential mechanisms for symmetry-informed rehabilitation, and outlines clinical implications and future research priorities. Rather than proposing symmetry as a stand-alone determinant of recovery, this Viewpoint advances the conceptual proposition that symmetry-oriented approach may complement existing multidomain models by serving as an additional layer of functional assessment alongside symptom reporting, neurocognitive evaluation, vestibular and oculomotor examination, exertional testing, and routine clinical assessment. Within this framework, symmetry-related measures are envisioned not as independent clearance criteria, but as potentially informative indicators of residual sensorimotor function that may help guide rehabilitation progression and contribute to more functionally informed return-to-sport decision making through adjunctive measures such as center-of-pressure behavior, center-of-mass displacement, gait symmetry, stance and swing time asymmetry, limb-loading patterns, interlimb coordination, and muscle activation symmetry. Full article

Background: Recent technological advances, such as 3D laser scanning, digital photogrammetry and virtual autopsy, have significantly reshaped the acquisition and interpretation of scientific evidence in criminal proceedings. These tools enable highly detailed visualizations that can enhance factual understanding while also introducing potential cognitive biases for judges. This study aims to assess the legal, epistemological and forensic implications of 3D virtual reconstruction as evidence, examining its admissibility, scientific robustness, and associated risks within contemporary criminal trials. Methods: An interdisciplinary methodology was applied, integrating criminal procedural law, forensic medicine, digital forensics and legal epistemology. The study includes doctrinal analysis, a review of the literature from the last decade, Italian and comparative case law, and evaluation of European forensic guidelines (particularly ENFSI). Direct observation of laboratory practices and technical assessment of reconstruction tools were conducted to evaluate scientific validity, repeatability, traceability and transparency. Results: The findings reveal that 3D reconstruction provides high metric accuracy, durable digital preservation of the crime scene and improved communication of complex dynamics to the court. It allows experts to test alternative hypotheses, reducing ambiguity in technical explanations. Nevertheless, several critical issues emerged: the persuasive power of highly realistic imagery, dependence on non-transparent software processes, potential vulnerabilities in the chain of custody for digital data, and the absence of a unified Italian regulatory framework governing digital scientific evidence. Conclusions: 3D virtual reconstruction constitutes a powerful but epistemologically complex form of scientific evidence. Its probative value depends on transparent methodology, verifiability and strict compliance with adversarial safeguards. The study underscores the need for national technical standards, clear admissibility criteria and specialized training for legal professionals. A coherent regulatory framework is essential to ensure that digital technologies enhance, rather than distort, the pursuit of truth in criminal justice. Full article

Background: Pharmacist-led interventions in Gulf Cooperation Council (GCC) countries have been increasingly studied, yet their overall effectiveness across clinical and healthcare outcomes remains incompletely defined. To address this gap, a systematic review and meta-analysis of randomized controlled trials and quasi-experimental studies published between 2000 and 2025 was conducted, following PRISMA 2020 and Cochrane standards. Methods: Searches of PubMed/MEDLINE, Scopus, Web of Science, and CENTRAL identified 437 records, of which 20 studies met the inclusion criteria; 13 contributed meta-analyzable data as randomized controlled trials and seven as quasi-experimental studies. Results: Pooled random-effects estimates favored pharmacist-led care for HbA1c, fasting glucose, low-density lipoprotein cholesterol, diastolic blood pressure, and medication knowledge, although these estimates carried substantial, largely unexplained heterogeneity and were rated as low to very low certainty under GRADE. The effects on systolic blood pressure, total cholesterol, triglycerides, high-density lipoprotein cholesterol, unplanned healthcare use, and antimicrobial utilization were favorable but not statistically significant. Quasi-experimental studies consistently demonstrated reductions in mortality and readmissions, though hospital and ICU length of stay remained variable. Risk of bias was judged as some concerns for randomized trials and moderate to serious for quasi-experimental studies, with substantial heterogeneity observed across blood pressure and lipid outcomes. Conclusions: Overall, pharmacist-led interventions in GCC settings were associated with improvements in glycemic control and LDL cholesterol, with additional benefits in mortality and readmissions, although the certainty of evidence was low to very low, owing to substantial heterogeneity and the predominance of non-randomized designs for the inpatient outcomes. These findings underscore the need for standardized intervention models and outcome measures. Full article

The use of gamification is growing within the field of architectural heritage, with applications in conservation, exhibitions, educational programming, and other dissemination mechanisms for interpretation. However, existing applications remain terminologically scattered and formally heterogeneous, lacking clear conceptual definitions or consistent evaluative criteria, making relevant practices difficult to identify, delimit, and compare, while risking misclassification or inconsistent conceptual judgment. Architectural Heritage Gamification (AHG) is proposed here as a distinct research object. Through conceptual alignment, historical contextual analysis, and review-based examination of the literature, this paper clarifies the need for AHG and develops a working definition accompanied by inclusion and exclusion criteria. AHG cannot be identified solely by media form, product form, or interaction intensity, but should instead be judged in relation to architectural centrality, the presence of identifiable gamification mechanisms, the extent to which architectural logic shapes mechanism formation, and normative conditions such as heritage interpretation and public responsibility. AHG is defined here as a design practice in which gamification elements are selected, organized, and judged within architectural heritage contexts on the basis of architectural logic and under the constraints of heritage interpretation, in order to support protection-oriented public understanding, participation, and meaning-making. Full article

Alkaline water is increasingly marketed for musculoskeletal and recovery benefits, yet its relevance to healthy aging, sarcopenia prevention, and functional capacity in older adults remains largely unexplored. This systematic evidence map and translational appraisal examined whether the available comparative human evidence on alkaline water is applicable to aging populations and longevity research. Following PRISMA guidance, PubMed and Scopus were searched from January 2005 to September 2025. Eligible studies were controlled or comparative observational human studies reporting muscle strength, physical performance, or recovery outcomes. Risk of bias was assessed using RoB 2, ROBINS-I, and JBI criteria; evidence certainty was judged narratively using GRADE-informed principles. Ten studies met the inclusion criteria. Most enrolled young athletic populations; only two had partial relevance to aging cohorts. Crucially, no study included participants aged 65 years or older or assessed primary sarcopenia-relevant endpoints such as appendicular lean mass, gait speed, or chair-rise performance; this total absence of data in the target demographic represents the central limitation of the current literature. Risk of bias ranged from some concerns to serious. The most consistent signals were short-term improvements in lactate clearance and perceived exertion in young male athletes. Evidence for strength, functional performance, and safety in older adults was absent or indirect. Current evidence, rated low to very low certainty for aging-relevant outcomes, does not support alkaline water as an evidence-based strategy for healthy aging or muscle preservation in older adults. Age-appropriate trials using EWGSOP2-aligned outcomes are urgently needed. Full article

Background: The management of diabetic foot disease and knee osteoarthritis (OA) with smart orthotics holds significant importance during the early stages of these conditions, given their potential consequences, including functional impairment, chronic pain, and economic burden. Real-time monitoring of plantar foot pressure enables early detection of abnormal force distribution and gait biomechanics, allowing for the redirection of forces away from affected ulcers or arthritic joints. This is the first systematic review to synthesise clinical evidence for smart orthotics technology with real-time plantar pressure sensor biofeedback across both diabetic foot ulcer prevention and knee osteoarthritis management simultaneously. A search of the PROSPERO register confirmed no existing registration covers this specific combination. Objectives: To examine the clinical evidence for the use of standard and smart orthotics in the prevention and management of diabetic foot ulcers (DFUs) and knee OA, and to evaluate their impact on plantar pressure redistribution, ulcer recurrence, pain, biomechanics, and economic burden. Eligibility criteria: Studies published in English involving human adult participants (≥18 years) with a clinical diagnosis of diabetes mellitus (at risk of DFU or with peripheral neuropathy) or knee OA, where the intervention involved any orthotic device or smart/intelligent insole with clinical outcomes reported, were included. Studies on healthy individuals only, those not reporting participant age, and non-weight-bearing protocols not differentiated from weight-bearing were excluded. Information sources: Five databases were searched: CINAHL (EBSCO Information Services, Ipswich, MA, USA), PubMed Advanced (National Library of Medicine, Bethesda, MD, USA), Wiley Online Library (John Wiley & Sons, Hoboken, NJ, USA), Cochrane Library (Cochrane Collaboration, London, UK), and Google Scholar (Google LLC, Mountain View, CA, USA). Searches were completed in May 2026. Methods: We conducted a comprehensive literature review. This review was structured and reported with reference to the PRISMA 2020 statement (Preferred Reporting Items for Systematic Reviews and Meta-Analysis; University of Ottawa, Ottawa, ON, Canada) to guide transparency of reporting. It does not constitute a full Cochrane-style systematic review; risk of bias assessment was applied to key included studies and GRADE (Grading of Recommendations Assessment, Development and Evaluation; McMaster University, Hamilton, ON, Canada) certainty ratings were applied informally and narratively rather than as formal per-outcome evidence profiles. Five databases were searched yielding 92,637 records. After removal of 398 duplicates by Rayyan, 92,239 records remained. A subsequent automated keyword-based relevance filter applied within Rayyan (Rayyan AI, Doha, Qatar), prior to human screening, excluded 84,572 records that did not contain any terms related to orthotics, diabetic foot, or knee osteoarthritis, yielding 7667 records for human title/abstract screening. A narrative synthesis approach was adopted owing to the heterogeneity of study designs and outcome measures across included studies, which precluded meta-analysis. This review was not prospectively registered. A complete list of all 78 included studies, including those not individually discussed in the results and discussion. Results: The available clinical studies report promising findings for orthotics and smart orthotics in pain reduction, ulcer prevention, and potential reduction in economic burden, though conclusions are limited by small sample sizes, heterogeneity, and predominantly open-label designs. Recent research found that orthotics can be used to alter the gait pattern that influences knee OA by reducing excessive force on the affected joint. A randomised controlled trial demonstrated an 80% relative risk reduction in DFU recurrence (RR = 0.20; 95% CI: 0.06–0.79; p = 0.022), with absolute event rates of 6.3% in the intervention group versus 30.8% in controls (ARR = 24.5%); a second trial reported a 71% reduction in ulcer incidence over 18 months; and a third randomised controlled trial demonstrated statistically significant plantar pressure reduction (p < 0.01) in patients with diabetic neuropathy. Conclusions: The available evidence suggests that orthotics may be associated with improved pressure redistribution, reduced ulcer incidence, and benefit in the management of knee OA. Although the number of studies directly comparing smart orthotics with standard orthotics remains limited, the limited comparative studies suggested that smart orthotics showed promising results in reducing ulcer incidence, providing the patient with real-time feedback to offload via their electronic devices. These findings, while preliminary, highlight the potential of smart orthotic technology as an adjunct to standard orthotic care in reducing the overall burden of diabetic foot disease and knee osteoarthritis. Limitations: The primary methodological limitation of this review is the open-label design of all included smart orthotic trials, which precludes participant blinding and introduces performance bias. However, this limitation is structural and inherent to the wearable technology field—analogous to surgical trials—and is substantially mitigated by the use of objective primary outcome measures (plantar pressure and ulcer recurrence) across the three included RCTs, the consistency of effect direction across independent RCTs conducted in different countries, and a narrative sensitivity analysis confirming robustness of findings (Risk of Bias Across Studies Section). Formal per-outcome GRADE evidence profiles were not produced; overall certainty of evidence was assessed narratively with reference to GRADE domains and is judged to be low to moderate for smart orthotics in DFU prevention and low for knee OA management, consistent with the Level 2–3 evidence base and open-label study designs. Future adequately powered, multi-site RCTs with standardised outcome reporting, minimum 24-month follow-up, and integrated health economic modelling are the highest priority to extend these preliminary findings. Registration: This review was not prospectively registered. Full article

Objectives: The aim of the present systematic review and meta-analysis is to compare bracket adhesion failure rates between resin-modified glass ionomer cement (RMGIC) and conventional resin adhesives (CRA) during fixed orthodontic treatment (OT), based on evidence from randomized controlled trials (RCTs). Methods: The research question is “Is there a difference in bracket adhesion failure rates between RMGIC and CRA?” The study was performed in accordance with the PRISMA guidelines. A comprehensive literature search was performed across multiple databases without time or language restrictions through February 2026. Keywords were used in different combinations using Boolean operators. Hand searching was performed and disagreements were resolved via discussion. The risk of bias (RoB) and certainty of evidence (CoE) were assessed using the Cochrane risk of bias tool and Grading of Recommendations Assessment, Development and Evaluation approach, respectively. Quantitative data synthesis was conducted using a random-effects model to calculate pooled odds ratios and 95% confidence intervals. Results: Seven RCTs met the inclusion criteria. Bracket failure rates ranged from 5.95% to 15.0% for RMGIC and 3.4% to 25.0% for CRA. The pooled meta-analysis revealed no statistically significant difference in bracket failure between the two adhesive types (OR = 1.00; 95% CI: 0.60 to 1.67), although substantial statistical heterogeneity was observed (I² = 69.0%, p = 0.0065). One included trial demonstrated significantly improved retention for RMGIC when combined with a specific enamel deproteinization conditioning step prior to bonding. Three studies had a low RoB and the remaining were judged as having “some concerns”. The overall CoE was low. Conclusions: Based on the currently available randomized evidence, no statistically significant difference in bracket adhesion failure rates was observed between RMGIC and CRA during fixed OT. However, given the low CoE, substantial heterogeneity among studies, and relatively short follow-up periods, these findings should be interpreted with caution. Further well-designed randomized controlled trials with longer follow-up are needed to provide more definitive conclusions. Full article

Large language models (LLMs) are increasingly being used in radiology-related workflows, but their application to reference, regulatory, and methodological queries remains limited by hallucinations and the static nature of model knowledge. This study aimed to develop and evaluate a retrieval-augmented generation (RAG) system for radiologists designed to provide grounded responses to such queries. A knowledge base was created through a survey of practicing radiologists and expert validation of sources, resulting in a corpus of 1049 documents. The system incorporated structured document parsing, a two-level parent–child vector database, hybrid dense–sparse retrieval, reranking, and a local large language model. Performance was assessed through functional testing, automated LLM-as-a-judge metrics, and multireader expert evaluation by 16 radiologists using 400 technical queries. No hallucinations were detected in the 77-query functional testing set during expert review. On the full technical dataset, automated Contextual Precision, Contextual Recall, and Answer Relevancy were 0.735, 0.881, and 0.890, respectively. Expert evaluation showed high response accuracy (mean, 4.53/5) and high expert-assessed Contextual Precision (0.886). Inter-expert agreement was substantial to excellent for most Likert-scale criteria. These findings suggest that a hierarchical RAG architecture can provide reliable access to radiology-specific reference information, although external validation and automated updating of the knowledge base remain necessary. Full article

Backgrounds: As one of the important liquid biopsy methods in recent years, circulating tumor cells (CTCs) have been proven to be valuable in judging metastasis and prognosis in various tumors. With the advent of immunotherapy, the expression of programmed death-ligand 1 (PD-L1) on CTCs has also attracted researchers’ attention, but its value in bladder cancer has not been fully explored. Methods: This study enrolled patients diagnosed with urothelial carcinoma who were treated in the Department of Urology, The First Affiliated Hospital of Nanjing Medical University from 2020 to 2023. Peripheral blood samples of the patients were collected to detect PD-L1+ CTCs. Meanwhile, the patients’ basic clinical characteristics, pathological data, and follow-up information were collected. The Kaplan–Meier method was used to estimate the overall survival (OS) and progression-free survival (PFS) of the patients, and the log-rank test was employed to evaluate the statistical differences. COX regression analysis and Firth penalized Cox regression analysis were adopted to assess the risk factors. Finally, according to the clinical or pathological characteristics, the effects of PD-L1+/PD-L1− CTCs on different subgroups of the population were evaluated, respectively. Results: According to the inclusion and exclusion criteria, a total of 109 patients were finally enrolled in this study, including 95 males and 16 females, with a median age of 68 years. The expression of PD-L1 on CTC was as follows: 19 patients were PD-L1+ CTC and 90 patients were PD-L1− CTCs. The results showed that the pathological grade of patients with PD-L1+ CTCs was significantly higher than that of patients with PD-L1− CTCs (p = 0.003). With a median follow-up time of 48 months (IQR: 45.8–49.5), prognostic analysis indicated that PD-L1+ CTCs were a risk factor for OS in bladder cancer patients (HR = 4.696 (1.477–13.596), p = 0.011). Subgroup analysis revealed that in patients with non-muscle-invasive bladder cancer (NMIBC) and in the subgroup of patients with high pathological grade, those with PD-L1+ CTCs exhibited significantly poorer prognosis in terms of PFS and OS compared with patients with PD-L1− CTCs. Conclusions: The presence of PD-L1+ CTCs in patients with bladder carcinoma may be closely associated with high-grade disease and poor OS. Full article

The increasing use of large language models (LLMs) in enterprises creates a need for routing mechanisms that select models according to both technical performance and organizational preferences. This article proposes a multicriteria decision-support framework for enterprise LLM routing that combines AHP-based criterion weighting with SAW-based prompt-level model selection. The framework evaluates prompts according to criteria related to required accuracy, business risk, reasoning depth, cost sensitivity, response-time sensitivity, standardization, and creativity. The empirical evaluation was conducted on 500 heterogeneous business prompts, using GPT-5-nano as the prompt-scoring router, GPT-4o-mini as the cheaper response model, and GPT-5 as the stronger response model. Costs were calculated from actual input and output token counts, including routing overhead. Response sufficiency was assessed using a structured LLM-as-a-judge protocol with three evaluator profiles. The proposed SAW routing variant with confidence margin and risk veto achieved a sufficiency rate of 94.4%, compared with 94.6% for the always-strong strategy and 86.8% for the always-cheap strategy. Relative to always-strong routing, it reduced total cost by 37.4%, with only a 0.2 percentage-point decrease in sufficiency. The framework was also compared with keyword-risk, token-threshold, TF-IDF centroid, logistic-regression, multiplicative-SAW, and TOPSIS baselines. The results indicate that an interpretable multicriteria router can achieve near-strong-model response sufficiency at substantially lower cost while preserving auditability and alignment with enterprise decision criteria. Full article

Recyclers worldwide face a common bottleneck: incoming mixed plastic bales are chemically opaque, yet the choice between mechanical recycling, chemical recycling, and energy recovery hinges on contaminant levels that cannot be judged by visual inspection alone. This study develops and validates a tiered analytical decision-support framework that translates standard laboratory measurements into explicit, actionable go/no-go routing criteria for any mixed polyolefin waste stream. The framework is organized into three successive analytical tiers of increasing specificity: Tier 1 uses FTIR and DSC for rapid polymer identification and thermal subclass confirmation; Tier 2 applies TGA/DTG for thermal stability assessment and filler quantification; and Tier 3 deploys ICP-OES, WD-XRF, CIC, and TG–MS for targeted heavy metal, halogen, and evolved gas profiling, triggered only when Tier 1/2 flags are raised. This staged logic minimizes unnecessary testing while ensuring that contaminant-relevant information is captured where it matters. The framework is demonstrated on nine blind mixed plastic waste streams (P1–P9) supplied by an industrial recycling facility without prior disclosure of polymer identity, filler content, or additive history—conditions that replicate the uncertainty encountered at any sorting plant globally. Application of the tiered protocol identified dominant polymers (HDPE, LDPE, PP), quantified inorganic fillers (CaCO₃ up to ~38 wt%), and detected hazardous contaminants, including chlorine (up to ~1900 ppm), lead, chromium, and titanium, enabling each stream to be assigned to a specific recycling route with defined contaminant thresholds. Because the method relies exclusively on commercially available, vendor-independent instrumentation and follows a reproducible, rule-based decision logic, it is directly transferable to recycling facilities in any geographic context without site-specific calibration. The proposed framework thus provides a practical, scalable decision-support tool for feedstock-level quality control under emerging regulations such as the UNEP Global Plastics Treaty. Full article

Background/Objectives: Hand hygiene is a cornerstone of infection prevention, yet the extent to which multimodal institutional hand hygiene interventions translate into measurable reductions in healthcare-associated infections (HAIs) remains uncertain. This systematic review aimed to evaluate the association between hospital-wide or multi-ward multimodal hand hygiene interventions and clinical HAI outcomes in acute care hospitals. Methods: A structured literature search was conducted in PubMed, Scopus, Embase, and Google Scholar using a combination of Medical Subject Headings (MeSH) and free-text terms related to hand hygiene, healthcare-associated infections, hospital settings, and intervention strategies. Eligible studies were quasi-experimental designs, including before–after, controlled before–after, and interrupted time-series studies, evaluating multimodal hand hygiene interventions implemented at hospital-wide or multi-ward level and reporting clinical HAI outcomes. Two reviewers independently assessed risk of bias using the ROBINS-I tool, and certainty of evidence across major outcome categories was summarized using GRADE. Results: twelve studies met the inclusion criteria. Overall, multimodal hand hygiene interventions were generally associated with favorable directional trends in clinical outcomes. Reductions were most consistent for broader institutional HAI measures and some device-associated infections, particularly central line-associated bloodstream infections. In contrast, organism-specific outcomes, including methicillin-resistant Staphylococcus aureus, vancomycin-resistant Enterococcus, and Clostridioides difficile, were more heterogeneous across studies and settings. All included studies were judged to be at serious or critical overall risk of bias, primarily because of confounding, lack of contemporaneous controls, co-interventions, and phased implementation. Conclusions: Multimodal hand hygiene programs in acute care hospitals may be associated with improvement in selected clinically relevant HAI outcomes, particularly at the institutional level. However, the overall certainty of evidence remains low to very low, and the strength of inference is limited by the non-randomized nature of the available studies and the difficulty of isolating the independent effect of hand hygiene within complex infection-prevention strategies. Full article