Search Results (666)

Rwanda’s sustainable green economy agenda has placed irrigation and input-support schemes at the centre of efforts to raise tropical fruit productivity in the Mugesera region, yet the econometric analysis presented, in a Cobb–Douglas production function estimated by panel OLS and a Nerlove partial-adjustment model of agricultural supply response required the analyst to pre-specify a log-linear, constant-elasticity functional form. This paper asks whether machine learning methods that learn input–output relationships directly from data can cross-check those findings, and presents a documented Python version 3.10.12 pipeline for doing so. Using a synthetic panel dataset (six regions, 2009–2018, N = 346 farm-year observations) calibrated to match the elasticity magnitudes reported in the original Cobb–Douglas estimation, three models—multiple linear regression, random forest, and gradient boosting—are trained on land, fertilizer, seed, lagged price, and lagged output features, and are evaluated on held-out data. All three models explain a substantial share of variation in output (R² between 0.72 and 0.75 under a random split; for results under panel-aware validation), with tuned gradient boosting achieving the best accuracy under the random split (R² = 0.751, RMSE = 0.268). Permutation feature importance recovers the same relative ranking of input variables as the original regression—fertilizer, then seeds, then land area—without the model being given the Cobb–Douglas functional form in advance. Because the synthetic data were explicitly generated to reproduce these elasticities, this recovery demonstrates that the modelling pipeline behaves as intended; it is a calibrated methodological demonstration and not an independent empirical validation of the original findings. The results are consistent with lagged price and output operating mainly through farmers’ input decisions rather than as direct predictors of output once input quantities are known, although we present this as a hypothesis for future testing rather than a demonstrated finding. Re-running the identical pipeline on real Mugesera farm-level survey microdata is the necessary next step before these results can be treated as genuine empirical findings; this caveat applies to every quantitative result reported below, not only to the conclusion. Full article

The rapid digitalization of public administration has transformed auditing environments, especially in emerging economies expanding their electronic governance (e-governance) frameworks. This study identifies and prioritizes the key drivers influencing electronic auditing (e-auditing) development in Iraq over the next decade. Using a mixed-methods design, the research first identifies potential drivers through qualitative interviews and open-ended questionnaires. These drivers were then evaluated and prioritized via a two-round Delphi survey involving a purposive panel of 20 experts, including senior auditors, accounting academics, and IT-audit specialists with over 15 years of professional experience. The analysis identified 19 significant drivers categorized into four clusters: (1) emerging audit technologies, (2) information security and data quality, (3) e-governance and transparency, and (4) professional capabilities. Results highlight that technological innovations, specifically real-time monitoring and machine learning, are the most influential drivers. Furthermore, cybersecurity and transparent governance mechanisms are identified as essential pillars for digital auditing in the Iraqi context. By providing a foresight perspective in a post-conflict, emerging economy, this study offers a unique conceptual framework that integrates e-governance maturity with auditing evolution. The findings provide actionable insights for policymakers and regulatory bodies to modernize auditing practices in high-uncertainty environments. Full article

Research has documented the various effects of internet use on adolescent well-being. However, few studies have integrated the Displacement Hypothesis and the Stimulation Hypothesis to compare the effects of online and offline communication on the quality of interpersonal relationships among adolescents. Moreover, limited research has examined the role of age cohort in these relationships. The present study addresses these gaps by investigating the age-related differences in these associations. Specifically, we examined the influence of both online and offline communication on the quality of adolescents’ interpersonal relationships and explored the moderating role of age cohort. Data were drawn from the China Family Panel Studies, a nationally representative survey. A cross-sectional sample of 1486 adolescents aged 10 to 15 years (54.4% boys and 45.6% girls) was analyzed using correlation, hierarchical regression, and bootstrap-based mediation analyses. The results indicated that adolescents who engaged in more online communication also tended to engage in more offline communication. Furthermore, both online and offline communication are jointly related to the quality of interpersonal relationships among adolescents. Importantly, age cohort was found to moderate the effects of both online and offline communication on the quality of interpersonal relationships. Theoretical and practical implications are discussed. Full article

Background: Antimicrobial resistance complicates the selection of appropriate regimens for urinary tract infections (UTIs), even when susceptibility data are available, particularly where infectious disease (ID) expertise is scarce. Machine learning clinical decision support systems (CDSS) may support prescribing, but evidence from Latin America is limited. The goal of this study was to evaluate the concordance between antimicrobial regimens selected by physicians and those recommended by a machine-learning-with-human-in-the-loop (ML-HITL) CDSS (OneChoice^®), and to assess CDSS appropriateness against an independent, blinded expert reference standard. Methods: In this cross-sectional, survey-based concordance study conducted in Lima, Peru, 194 verified physicians contributed 224 eligible evaluations across 42 real UTI case codes with complete culture and antimicrobial susceptibility data. Of the 224 evaluations, 70 were contributed by infectious disease specialists and 154 by non-ID physicians. Participants selected OneChoice^® and alternative antimicrobial regimens. Responses were compared with CDSS recommendations under three concordance definitions. Discordances were adjudicated by an external panel blinded to the source of the recommendation. Non-independence was addressed using cluster-robust methods. Results: First-choice, alternative, and general concordance were 50.9%, 40.6%, and 62.5%, respectively. ID specialists showed higher concordance than non-ID physicians (65.7% vs. 44.2%; 51.4% vs. 35.7%; 72.9% vs. 57.8%; all p ≤ 0.034). ID specialty was independently associated with concordance (adjusted OR 2.19–2.68; p ≤ 0.016). Among discordant evaluations, the external panel judged the CDSS recommendation to be preferable in 90.5–93.2% of cases. Physician–CDSS concordance was moderate and higher among ID specialists. The external adjudication findings indicate that the CDSS recommendations were frequently aligned with expert assessment when physician and CDSS recommendations differed; however, the study did not evaluate comparative clinical effectiveness or patient outcomes. Conclusions: Physician–CDSS concordance was moderate and higher among ID specialists, yet discordances overwhelmingly favored the CDSS on independent adjudication. These findings suggest the CDSS aligns with expert reasoning and may support antimicrobial selection in high-resistance settings. Full article

Background/Objectives: Interferon-free direct-acting antiviral (DAA) therapy cures chronic hepatitis C virus (HCV) infection, but its short-term effects on health-related quality of life (HRQoL) are captured differently by generic and disease-specific instruments. We examined longitudinal changes in utility scores and HRQoL in a multicenter Japanese cohort. Methods: Adults with chronic HCV completed the EuroQol 5-Dimension 5-Level questionnaire (EQ-5D-5L), the 8-Item Short-Form Health Survey (SF-8), and the Chronic Liver Disease Questionnaire (CLDQ) at baseline and at 12, 24, and 36 weeks after treatment initiation; 48-week data were included when available. Complete-case panels were analyzed for each instrument (SF-8, n = 112; CLDQ, n = 131; EQ-5D-5L, n = 128). Domain trajectories were summarized and compared with baseline. Results: By week 36, SF-8 general health improved significantly from 50.42 to 52.47, whereas vitality (50.73 to 52.55) and mental health (51.02 to 53.05) showed nonsignificant numerical increases. CLDQ showed improvements in worry (5.21 to 5.82) and total score (5.21 to 5.47). EQ-5D-5L utility values remained high and largely stable (0.913 to 0.920), suggesting ceiling effects in patients with relatively good baseline health status. External real-world evidence also suggested better on-treatment HRQoL with ribavirin-free regimens. Conclusions: In Japanese patients with HCV, interferon-free DAA therapy was associated with early improvements in symptom-proximal and mental domains of HRQoL. Generic utility scores changed little over the short term, indicating that disease-specific patient-reported outcome (PRO) instruments and utility measures should be used together for patient-centered assessment and cost–utility modeling. Full article

Background/Objectives: Multimorbidity is associated with substantial healthcare burdens, but its associations with health and economic outcomes among young adults remain poorly characterized. This study examined associations between clinically defined multimorbidity profiles and expenditures, catastrophic healthcare expenditure, health-related quality of life, and inpatient utilization across adult life stages. Methods: Medical Expenditure Panel Survey Panel 24 data were analyzed for 4331 U.S. adults followed from 2019 through 2022, including 942 young adults. Respondents were classified into five profiles: no included conditions, respiratory only, cardiometabolic only, any-cancer involved, and cardiorespiratory. Longitudinally weighted generalized estimating equations were fitted. Results: Profile distribution differed significantly across age groups. Among young adults, overall profile and profile-by-year tests were nonsignificant for total expenditures and most other outcomes. In 2019, the cardiorespiratory profile was associated with higher expected expenditures than no included conditions (expenditure ratio = 3.44, 95% CI: 1.59–7.46; p = 0.002). It was also associated with higher hospitalization odds across 2019–2022 (OR = 3.39, 95% CI: 1.28–8.99; p = 0.014), although the overall profile test was nonsignificant. Sex, insurance, and income were associated with several outcomes. Mental health scores were lower in 2021 and 2022 than in 2019 within the reference profile, without evidence of uniform changes across profiles. Conclusions: Overall profile effects were not consistently demonstrated among young adults. Parameter-specific cardiorespiratory estimates suggested a potentially elevated burden but remained exploratory because of sparse observations and nonsignificant overall tests. Socioeconomic vulnerability remained an important correlate of healthcare burden. Full article

Background/Objectives: Physical literacy (PL) is increasingly being recognized as an important factor associated with physical activity participation and health outcomes. However, evidence remains limited regarding the association between PL and quality of life (QOL) in adults, particularly the statistical indirect association through moderate-to-vigorous physical activity (MVPA). Therefore, this cross-sectional study aimed to examine the associations among PL, MVPA, and QOL in Japanese adults and to assess the statistical indirect association between PL and QOL through MVPA. Methods: Japanese adults aged 18–64 years were recruited from a commercial online survey panel using age- and sex-quota sampling in 2023. After excluding incomplete or inconsistent responses and participants with missing or incomplete MVPA data, 861 participants were ultimately included. PL, QOL, and MVPA were assessed using the Physical Literacy for Life tool, World Health Organization QOL-26, and International Physical Activity Questionnaire, respectively. Associations among PL, MVPA, and QOL were examined using Pearson’s correlations, multiple linear regression, and mediation analysis with PROCESS macro (Model 4). Results: PL was positively correlated with MVPA (r = 0.257, p < 0.01) and QOL (r = 0.416, p < 0.01); MVPA was positively correlated with QOL (r = 0.248, p < 0.01). PL remained significantly associated with QOL after including MVPA and covariates (β = 0.374, p < 0.01). The statistical indirect association between PL and QOL through MVPA was significant (b = 0.0011, BootSE = 0.0003, 95% bootstrap confidence intervals: 0.0005–0.0018), indicating a modest indirect association through MVPA. Conclusions: PL was positively associated with QOL among Japanese adults, and MVPA partially explained this association in a statistical mediation model. Full article

Navigational obstacles pose compound collision and pollution risks, yet conventional quantitative assessment models relying on data-driven “best-estimate” approaches suffer from “alarm masking”, whereby critical risk signals are diluted through averaging. This study develops a conservative hybrid risk assessment framework that preserves critical risk signals while systematically incorporating qualitative factors beyond the reach of quantitative data. The fuzzy inference rules of an existing integrated model were redesigned into a priority-stratified hybrid hierarchical–parallel fuzzy inference system (HHP-FIS); a qualitative evaluation matrix of four categories and 32 items was constructed through a two-stage expert procedure (a Delphi panel of eight officials and an analytic hierarchy process (AHP) survey of 59 experts with 34 valid responses); and a Red Flag Protocol was introduced as a fail-safe veto mechanism. The framework was verified through eighteen paired random-input simulations across two grid systems and a case study of a 68.9-ton drifting fishing vessel near Seongsan Port, Jeju Island. The model upwardly reclassified underestimated low-frequency, high-consequence scenarios, raised the case-study risk from Low (44.6 and 47.7) to Moderate (59.1 and 74.8), with an action level consistent with expert judgment, and was robust to rule-weight perturbations, providing a decision-support tool for obstacle-removal prioritization and marine pollution prevention. Full article

Background/Objectives: Multiple Sclerosis (MS) is a complex, chronic autoimmune neuroinflammatory disorder that significantly impacts patients’ health-related quality of life (HRQoL) and increases the burden of healthcare costs. However, evidence that quantifies the covariate-adjusted differences between the MS and non-MS populations is limited due to the presence of outliers, which frequentist approaches may not adequately address. Therefore, this study aims to examine healthcare expenditure and HRQoL in patients with MS compared to the non-MS population using a robust Bayesian approach. Methods: This retrospective cross-sectional study includes adults ($\ge$18 years) with MS and those without MS using the 2017–2022 Medical Expenditure Panel Survey (MEPS) data. The Bayesian quantile LASSO (BQL) is applied to examine the association between MS and different response variables under three quantile levels. Markov Chain Monte Carlo (MCMC) with Gibbs sampling was used to estimate the coefficients from the posterior distribution of model parameters. The convergence of the MCMC chain has also been assessed to ensure the reliability and stability of the posterior estimates. Alternative methods, including Bayesian LASSO and the multivariate Generalized Linear Models (GLMs), are also applied for comparison in both prediction and estimation. Results: The results of the BQL show that for the healthcare expenditures, the estimated total healthcare expenditure in patients diagnosed with MS is $29,860.11 (95% credible interval $27,826.96 to $31,825.63) more compared to those without MS under the quantile level 0.5. With the coefficient −3.16 (95% credible interval −4.31 to −2.07), MS is negatively related to the mental component of VR-12 under the median quantile. Compared with individuals without MS, patients with MS have a 13.40-point lower score on the physical component of the VR-12 (95% credible interval −14.56 to −12.19) at the 0.5 quantile. At the median quantile, BQL achieves prediction errors of 6723.86 for healthcare expenditure, 6.57 and 6.41 for the mental and physical components of the VR-12, respectively. Conclusions: BQL with a quantile level of 0.5 shows the lowest in-sample prediction error when examining the healthcare costs and HRQoL in MS. Full article

Background/Objectives: Long COVID (LC) has affected 7.2% of the population of the United States (US). Mental and physical health have been increasing in prevalence over the last few years. This study aimed to investigate the association between various characteristics and mental and physical health status among US adults with LC. Methods: The study was cross-sectional in design and used data from the 2023 Medical Expenditure Panel Survey (MEPS). We assessed predisposing, enabling, and need variables in US adults with MEPS-defined LC using multivariable logistic regression analysis. The data was weighted to produce nationally representative estimates. Results: It was determined that individuals with a low income level, a high degree of pain, and poor physical health were each associated with higher odds of poor mental health in US adults with LC. An age of 50–70+ was associated with lower odds of poor physical health in US adults with LC. Educational achievement up to and including high school, having a functional limitation, exercise participation, any pain, multiple comorbid conditions, and poor mental health were each associated with higher odds of poor physical health in US adults with LC. Conclusions: Several variables were associated with poor mental and physical health status among US adults with LC. Further research should be conducted to explore these variables in more detail and investigate possible interventions for healthcare providers. Full article

Background and Objective: Glucagon-like peptide-1 (GLP-1) receptor agonist therapy has intensified the policy debate over whether obesity pharmacotherapy increases short-term healthcare expenditures or reduces downstream medical costs (i.e., non-drug, acute-care expenditures). This study estimated same-year associations between GLP-1 use and medical spending among adults with observed obesity. Methods: Analyses used Medical Expenditure Panel Survey (MEPS) person-year files compiled from 2018 to 2022, with body mass index (BMI)-defined analyses restricted to 2019 and 2021. The primary analytic sample included 7144 person-years, of which 275 represented sustained GLP-1 users. Overlap-weighted Poisson pseudo-maximum likelihood models, utilization regressions, two-part decompositions, and a layered diagnostic framework were applied to assess spending patterns and same-year cost implications. Robustness was evaluated through sensitivity analyses across all-obesity, severe-obesity, and diabetes-excluded subsamples, as well as exploratory molecule-specific specifications. Results: Findings showed no evidence of same-year cost savings, but rather a distinct shift in spending composition. In fully adjusted primary models, GLP-1 use was associated with reduced non-drug medical spending (beta = −0.2494; marginal effect = −$2586 per person-year) and lower acute-care spending (beta = −0.5856; marginal effect = −$2019), while total medical spending was higher (beta = 0.3182; marginal effect = +$6483). Point E values of 1.89 (non-drug) and 2.99 (acute-care) indicated moderate-to-strong resilience to unmeasured confounding. Negative control models using dental spending and visits showed significant positive associations. Since this residual confounding operates in the direction opposite to the negative acute-care association, the headline compositional pattern is unlikely to be an artifact of upward health-seeking bias. Utilization models indicated fewer inpatient discharges (IRR = 0.686) and inpatient nights (IRR = 0.524). Two-part models further suggested that the reduction in acute-care spending was concentrated on the intensive margin among individuals with positive expenditures. Conclusions: The use of GLP-1 is not associated with overall short-term cost savings. This usage is related to a shift in the composition of observed medical spending during the same year. This change is leading to a decline in hospital admissions and acute care costs, in parallel with the rise in pharmaceutical spending. Non-drug and acute-care costs were lower among regular users. However, total spending was at a higher level, consistent with a trend in the budget records from that period in which drug costs were the dominant factor. These associations are consistent with a potential reallocation of healthcare resources from inpatient acute care toward pharmacy and ambulatory services, with corresponding planning implications for payers and possible workforce reallocation of nursing capacity from acute to outpatient settings, contingent on replication in post-2022 data. Full article

Social sustainability and inclusive development require attention not only to the size of the middle-income group but also to households’ capacity to maintain middle-income status over time. This study examines which middle-income household-years in China face a higher probability of moving to low income at the next survey wave. Using household panel data from the China Family Panel Studies, the analysis follows middle-income origins at wave t to observed destinations at wave t + 1. It distinguishes the relative position within the middle-income interval, prior status switching, and maximum observed consecutive middle-income duration. The observed next-state transition sample contains 23,931 household-years, of which 28.2% move to low income. In joint models, a higher relative position and longer observed persistence are associated with lower subsequent downward-mobility risk, whereas prior switching is positively associated with risk but is less stable across waves. A descriptive three-item score shows a monotonic inverse-probability-weighted risk gradient, from 4.04% at score 0 to 50.79% at score 3. These findings describe longitudinal risk associations rather than causal effects. The study approaches social sustainability through one bounded empirical dimension, the durability of the household economic position, rather than as a comprehensive measure of social sustainability or household resilience. Middle-income expansion alone is therefore an incomplete indicator of social sustainability: assessments of middle-income consolidation should also distinguish current members’ observed histories and destination-specific risks. Full article

This paper asks whether the digital economy is becoming the new economy: not because traditional economic functions disappear, but because they are increasingly reorganised through digital infrastructure, digital finance and algorithmic capability. We introduce an open-data Digital Architecture framework and build a fully reproducible panel of 38 OECD economies for 2010 to 2024 from World Bank sources. The framework constructs transparent composite indices for digital infrastructure, FinTech, AI/ML capability, the traditional economy, and a clean new economy. Using two-way fixed-effects panels, we provide a descriptive answer: in OECD economies, the digital economy coincides with the new economy mainly through digital infrastructure. First, digital infrastructure is the load-bearing pillar of the digital economy’s association with the new economy: it accounts for about 24 per cent of within-country variation in new-economy outcomes, although its coefficient is imprecise under country-clustered errors. Second, the Digital Architecture Index is useful as a benchmarking construct, but it does not improve the within-country fit because the digital pillars are collinear and infrastructure carries most of the signal. Third, the evidence does not support a robust three-way configuration effect among digital infrastructure, FinTech and AI/ML capability; the association is sensitive to error structure, dynamics, outcome definition and true survey-year estimation. A future-dated placebo loads for the infrastructure block as well, so even the descriptive associations cannot be cleanly separated from slow-moving common trends. The paper therefore contributes a reusable open-data measurement framework and a clear empirical benchmark: in OECD economies, the digital economy aligns with the new economy primarily through digital infrastructure, the pillar that carries the strongest within-country association with productivity and entrepreneurship. The AI/ML capability proxy is externally validated against Eurostat enterprise-AI adoption data, with a cross-country correlation of 0.80. Full article

Aqueous two-phase systems (ATPSs) provide a versatile, fully aqueous platform for probing solute–water interactions and protein structure. This review first surveys the diversity and phase behavior of biphasic aqueous systems formed by polymers and salts. We describe how phase diagrams characterize ATPS formation and composition and how both polymer chemistry and salt identity, rather than molecular size alone, govern phase separation by modulating the solvent properties of water. Building on a modified binodal model, we show that phase separation and solute partitioning can be understood in terms of changes in aqueous solvent dipolarity/polarizability, hydrogen-bond donor/acceptor properties, hydrophobicity, and electrostatics, quantified via solvatochromic probes and homologous solute series. These measurements underpin solvent interaction analysis (SIA), in which the partition coefficients of small molecules and proteins across panels of ATPSs are used to generate “structural signatures” that sensitively report on amino acid substitutions, conformational changes, aggregation, ligand binding, osmolyte effects, and post-translational modifications, independent of protein size. We discuss how SIA can be implemented in vial-, plate-, and microfluidic formats and combined with diverse analytical readouts (HPLC, MS, colorimetric assays, and immunoassays), and we contrast this structure-focused approach with conventional concentration-only proteomic and biomarker strategies. Particular emphasis is placed on structure-based biomarker discovery, where disease-relevant shifts in proteoform distributions—especially glycosylation changes—are often more informative than bulk protein levels and where SIA can complement or simplify complex glycomics and top-down proteomics workflows. As a case study, we describe the recently FDA-approved IsoPSA assay, which applies SIA principles to prostate-specific antigen by measuring cancer-associated structural alterations in circulating PSA via its partition behavior in a proprietary ATPS. IsoPSA generates a single index that discriminates between high-grade prostate cancer and benign and low-grade conditions. Prospective, longitudinal, and MRI-integrated clinical studies demonstrate that IsoPSA improves pre-biopsy risk stratification, reduces unnecessary biopsies, and provides robust negative and positive predictive values within the PSA “gray zone.” Collectively, the data support aqueous solvent interaction analysis as a broadly applicable, mechanistically grounded technology for protein characterization, drug–protein interaction studies, and structure-centric biomarker development, exemplified by the clinical translation of IsoPSA. Full article

Fusarium wilt, caused by Fusarium oxysporum f. sp. melonis (FOM), is a main disease of melon (Cucumis melo L.). FOM 1.2 is the most widespread and detrimental variant of FOM, causing substantial economic losses under severe disease conditions. Current information suggests that resistance to race 1.2 (FOM 1.2) is controlled by multiple recessive genes and is strongly influenced by the environment. Therefore, identifying genetic polymorphisms within diverse melon populations is essential to elucidate the loci and putative candidate genes associated with resistance. The objective of this investigation was to identify single-nucleotide polymorphism (SNP) and structural variant (SV) markers associated with FOM 1.2 resistance utilizing a panel of 160 genotypes through a genome-wide association study (GWAS). Phenotypic evaluation was performed two weeks after sowing, at the first-true-leaf stage, on 2400 individual plants inoculated by the root dip method with a concentration of about 10⁶ spores/mL. Biochemical and disease-related traits, including area under disease progress curve (AUDPC), disease severity index (DSI), standardized AUDPC (SAUDPC), latent period (LP), catalase, peroxidase activity, and ascorbate peroxidase activity were measured 35 days after inoculation. PCA identified eighty-three individual melon plants with a broad range of disease-response variation. Genotyping-by-sequencing (GBS) was conducted on these plants, resulting in the identification of 737,435 SNPs and 75,133 SVs. Evaluation of the population structure outlined four genetic groups, including one associated with germplasm highly resistant to FOM 1.2. We used SNP data to describe linkage disequilibrium (LD), which was estimated to decay at 14 kb, on average. A GWAS was performed using the Bayesian information and linkage-disequilibrium iteratively nested keyway (BLINK) method, which revealed nine SNPs significantly associated with several disease indices, namely ascorbate peroxidase activity, AUDPC, catalase, peroxidase activity, rAUDPC, and SAUDPC. Also, eight SVs were associated with AUDPC and relative area under disease progress curve (rAUDPC), including translocation and deletion types. In addition, GWAS using the fixed and random model circulating probability unification (FarmCPU) method unveiled thirteen SVs associated with rAUDPC, peroxidase activity and ascorbate peroxidase activity, including translocation and inversion types. According to the performed models of GWAS, several significant SNPs and SVs, associated with putative candidate genes, including multidrug resistance-associated protein 6 (MRP6), LOB domain-containing protein 15 (LBD15), phosphomannomutase, and NADH-ubiquinone oxidoreductase B8 subunit, which may be involved in FOM 1.2 resistance. However, these findings represent a preliminary genome-wide survey and require further validation using high-coverage or long-read sequencing approaches. The results provide remarkable insights into the genetic control of FOM 1.2 resistance and valuable information for the implementation of the putative molecular markers identified in this study in melon breeding programs. Full article