搜索结果：Econometrics

Extreme heat and disease transmission triggered by global warming pose severe threats to human health, creating an urgent need to promote low-carbon development. It is widely recognized that mixed land-use strategies play a critical role in reducing carbon emissions and mitigating global warming. In contrast to existing literature that solely focuses on either land structure mix or land function mix, this research integrates both into a unified analytical framework for comparative analysis. This approach aims to clarify the key directions that mixed land-use strategies should prioritize. By employing the information entropy method, this study utilizes remote sensing imagery and Point of Interest (POI) data to characterize structural and functional mixed land use across 268 prefecture-level cities in China, and further analyzes their impacts on carbon emission intensity. The experimental results show that land structure mix significantly reduces the carbon emission intensity; in contrast, the land function mix increases the carbon emission intensity. The mechanism analysis shows that the former generates positive externalities by promoting economic agglomeration, offsetting the negative effects of traffic congestion, whereas the latter intensifies the urban traffic pressure and does not show significant agglomeration effects. The spatial econometrics results further show that a significant spatial spillover effect of functional mixing, which further amplifies the carbon emission pressure in the city. The article further proposes policy recommendations.

Moment condition models are popular in statistics and econometrics, as they provide a powerful and flexible framework for estimation. However, estimation procedures based on these models can be sensitive to misspecification or the presence of outliers in the data. In the present paper, we introduce a class of robust estimators for moment condition models, representing robust alternatives to minimum empirical divergence estimators. The estimators are constructed by using truncated orthogonality functions and minimizing divergences in dual form, allowing to limit the impact of outliers or model deviations. We give the expressions of the influence functions of the estimators and prove their robustness. We also prove that the estimators are consistent. These theoretical results together with numerical examples, based on Monte Carlo simulations, show that extreme observations do not disproportionately affect the final estimates.

High-dimensional data arising in genomics, econometrics, and clinical medicine often exhibit substantial heterogeneity across multiple sources. While existing methods address multi-source heterogeneity, they do not adequately accommodate the combined challenges of high dimensionality and between-source heterogeneity. To address this gap, we propose a scalable Bayesian framework for multi-source heterogeneous quantile regression with spike-and-slab priors for simultaneous parameter estimation and feature selection. To overcome computational challenges, we combine mean-field variational inference with Laplace approximation and introduce a novel low-rank variational correction strategy that substantially improves approximation accuracy and adaptability in high-dimensional heterogeneous settings. This low-rank correction effectively captures the underlying dependence structure, leading to more robust and efficient inference. For model assessment and diagnostic analysis, we further develop a Bayesian score test coupled with local influence analysis. Extensive simulation studies and an analysis of The Cancer Genome Atlas (TCGA) data from four cancer cohorts (ESCA, PAAD, PCPG, and READ) demonstrate the computational efficiency, scalability, and practical utility of the proposed method in high-dimensional heterogeneous applications. The proposed low-rank variational correction algorithms are implemented in the R package LRQVB, which is publicly available on CRAN.

Road injuries are a leading cause of mortality and morbidity worldwide. Years of international efforts have aimed to strengthen policy engagement, including the 2020 UN General Assembly's proclamation of the Second Decade of Action for Road Safety (2021-30), targeting a 50% reduction in road traffic deaths and serious injuries by 2030. The aim of this study is to provide estimates to monitor progress and identify intervention gaps. As part of the Global Burden of Diseases, Injuries, and Risk Factors Study 2023, we estimated incidence, mortality, and morbidity of road injuries for 204 countries and territories from 1990 to 2023. Four road injury types and 47 nature-of-injury categories were examined. Morbidity and mortality data from clinical records, vital registration, and police reports were harmonised using meta-analytic techniques to ensure consistency and correct for systematic bias. Incidence was modelled with the meta-regression tool Disease Modelling-Meta-Regression version 2.1 and cause-specific mortality with the Cause of Death Ensemble model, both incorporating location-specific covariates to support interpolation. Years of life lived with disability (YLDs) were estimated from the prevalence and severity of the nature of road injury, and years of life lost (YLLs) from the number of cause-specific deaths multiplied by the standard life expectancy at the age of death. Disability-adjusted life-years (DALYs) were the sum of YLLs and YLDs. All metrics were calculated with 95% uncertainty intervals (UIs). In 2023, there were 50·9 million (95% UI 46·1-56·1) road injury incident cases, 1·34 million (1·04-1·58) deaths, and 75·3 million (59·8-89·2) DALYs globally. Road injuries were the leading global cause of death among males aged 10-39 years. Between 1990 and 2023, age-standardised incidence decreased by 38·3% (95% UI 36·9-39·7) and mortality decreased by 32·3% (6·1-49·0), but progress varied widely by World Bank income group. Mortality in low-income countries (43·8 [95% UI 31·7-56·0] deaths per 100 000 population) was approximately six times higher than in high-income countries (7·5 [7·1-7·9] deaths per 100 000), despite the high-income countries showing the highest age-standardised incidence rates (858·1 [95% UI 781·9-947·1] cases per 100 000). In the past decade, many countries achieved notable reductions in road injuries, but others, including Ghana and the USA, saw increases. More severe injuries tended to occur in low-income and middle-income countries. Although global incidence, mortality, and DALY rates from road injuries have declined, progress remains uneven, with pronounced disparities across income groups reflecting systemic inadequacies in infrastructure, vehicle standards, enforcement, and post-crash care. Strengthening emergency response, improving road design, enforcing safety measures, and adapting policies to the evolving demographics remain essential. Gates Foundation.

Herpes zoster (HZ) is common in Finland, leading to long-term postherpetic neuralgia (PHN) especially with ageing. HZ reduces productivity, placing an economic burden on society. An integrated actuarial and macroeconomic model assessed the return on investment of recombinant zoster vaccine (RZV) in Finland for adults ≥ 50 years versus no vaccination. A validated economic model determined HZ cases and complications avoided, health gains (quality-adjusted life-years, QALYs), and healthcare cost-savings. Actuarial techniques projected population trends incorporating employment factors. Lost productivity averted through vaccination provided fiscal revenue for the government, and increased gross domestic product (GDP) for society. QALY gains were monetised using 1-3xGDP per QALY. RZV was administered for 20 years, with lifetime impact assessment. RZV vaccination of around 790,000 50year-olds was estimated to save 3,779 QALYs and €52 million (M) in healthcare costs. Averted productivity losses generated €152 M in fiscal revenue and €367 M in GDP, while vaccination costs were €236 M. From the fiscal perspective, 90% of vaccination costs were offset by fiscal and healthcare cost savings (without considering QALY gains). From the societal perspective, for every €1 invested in vaccination, €2.60-4.10 was saved (depending on QALY value). The benefit-cost ratio (BCR) remained positive across sensitivity analyses. The highest BCRs were achieved when vaccinating adults aged 50 years; or 65 years when excluding productivity benefits. RZV vaccination in working-age adults yielded economic benefits in Finland, generating €2.60-4.10 for society per €1 invested in vaccination, and with fiscal benefits offsetting 90% of vaccination costs. WHAT IS THE CONTEXT?: Shingles (herpes zoster, HZ) is a common illness in Finland, especially in older adults. It can cause long-term nerve pain (postherpetic neuralgia, PHN) and other complications. HZ also reduces work productivity, creating economic costs for individuals and society. This economic study assessed the return on investment of introducing vaccination for adults aged 50 years and older with the recombinant zoster vaccine (RZV) versus no HZ vaccination (current situation). WHAT WAS DONE?: An economic model estimated the benefits of vaccination over 20 years, considering avoided HZ and PHN cases, healthcare cost-savings, and economic gains from reduced productivity loss, such as tax revenue and increased economic output (gross domestic product, GDP). The study also calculated the return on investment by comparing vaccination costs to financial benefits. The financial value of health benefits (quality-adjusted life-years, QALYs) gained through vaccination were also considered. WHAT WAS LEARNED?: Over the analysis timeframe, around 790,000 adults aged 50 years were vaccinated, saving 3,779 QALYs and €52 million in healthcare costs. Reduced productivity loss generated €367 million in GDP gains, including €152 million in tax revenue. The benefit-cost ratio showed that every €1 spent on vaccination returned €2.60–4.10 to society. The highest returns were seen when vaccinating at age 50, or at age 65 when productivity gains were excluded. CONCLUSION: RZV vaccination in working-age adults is a positive investment for Finland, benefiting both public health and the economy.

We revisit the issue of Intimate Partner Violence (IPV) during the COVID-19 pandemic asking three questions: whether IPV worsened with lockdown, what pandemic-specific 'shocks' had the greatest impact, and how the results change when different measures of IPV are used. The large telephone survey we leveraged for this purpose was conducted in 2021 in the Italian region of Tuscany as part of a mixed-method research project on IPV during the first lockdowns, and it is, to our knowledge, the only (locally) representative survey on IPV during lockdowns conducted in Italy. Subjective evidence from the survey shows that, on balance, IPV worsened in frequency or severity or both. Econometric evidence suggests that parental overburden due to the presence of minors had the largest influence, followed by job loss, whereas we were not able to discern a significant influence for confinement in crowded space lacking privacy. Finally, and unsurprisingly, we found that using a fuzzy measure of violence outcomes that accounts for severity and intensity as well as prevalence of violence allows us to discern between shocks like job loss that primarily influenced the occurrence of violence without significantly influencing its harshness. Our empirical strategy principally relies on the exogeneity of pandemic-specific shocks to attribute causal interpretation to our estimates. However, our dependent variables (IPV outcomes) are binary or fractional, and endogeneity of some control covariates cannot be ruled out. To address these issues, we estimate average marginal effects using a two-step Control function (CF) approach combined with a quasi-likelihood method.

Artificial intelligence (AI) is reshaping national economies, yet country-level AI-macro relationships remain poorly understood. Using annual data for the United States and China, 1980-2020, we develop a four-layer triangulation framework-Pattern Causality, Granger causality, VECM-based cointegration, and lead-lag correlation-to map directional associations between AI activity indicators and macro aggregates through weighted networks and heatmaps. Three patterns recur. First, nonlinear AI-macro dependence is moderate and mostly positive, making AI indicators useful monitoring signals rather than stand-alone decision triggers. Second, publications and patents carry short-run predictive content in Granger tests, making them candidate early-warning indicators for macro surveillance. Third, cointegration places AI indicators mainly on the adjustment margin: Macro fundamentals condition long-run AI-macro co-movement more than AI indicators lead it. The international layer adds an important qualification. United States-China collaboration variables raise AI node centrality, especially for China, but a mechanical-expansion null benchmark shows that most of this increase is expected from enlarging the network; beyond-mechanical collaboration evidence concentrates in the cointegration layer. Overall, the US pathway is patent-oriented and selective, whereas China's is denser and more collaboration-intensive. The framework supports AI-macro risk monitoring and hypothesis generation, not structurally identified causal claims.

A smoke-free society by 2040 is a key target for Dutch municipalities, mandated by law. Smoking remains a public health issue in Northern Netherlands, where approximately one in four adults still smoke. This study evaluates the content of municipal health policies related to tobacco control and their correlation to regional smoking rates. We used open-access datasets to assess municipal health policies using a quantitative scoring model inspired by the Dutch Public Health Act. We developed the model using elements from the logic of events theory, which assesses policy impact across four dimensions: Obligations, Goals, Opportunities and Resources. We examined differences across municipalities and regions, exploring correlations between health policy scores, smoking prevalence and municipal-level factors, including socioeconomic status. We created an interactive policy map to visualise the results. Municipalities in the Groningen region achieved the highest health policy score (28 of 33), yet the region also had the highest smoking prevalence (24.68%), compared with Drenthe (20.51%) and Friesland (22.49%). Among the four policy impact determinants, Opportunities differed across regions. At the municipal level, socioeconomic status was negatively correlated with the total health policy score (r=-0.38), with specific correlations for Opportunities (r=-0.36) and Resources (r=-0.33). We developed a reproducible approach to assess regional disparities in tobacco control efforts. Despite high policy scores in some municipalities, smoking prevalence remained high, suggesting barriers to effective implementation, including socioeconomic factors, governance and financial planning.

Excessive sugar consumption is associated with substantial health and economic burdens. Public policies aimed at reducing sugar intake, such as education campaigns and product labelling, have shown limited effectiveness, prompting growing interest in fiscal measures such as sugar taxes. While taxes on sugar-sweetened beverages have demonstrated reductions in consumption, evidence remains limited regarding their extension to other high-sugar products. This study assesses the potential impacts of sugar taxation policies in France, the United Kingdom, and Spain on three key product categories that contribute substantially to sugar intake: non-alcoholic beverages, biscuits, and dairy desserts. Using nationally representative scanner data and a structural econometric model, we estimate demand, model firm pricing behavior under oligopolistic competition, and simulate the effects of a two-tiered sugar-based tax. Results indicate that firms generally over-shift the tax to prices, leading to significant reductions in purchases and sugar intake, with the largest impacts observed in non-alcoholic beverages and French dairy desserts. The tax is particularly effective among households with overweight or obese adults. Although consumer surplus and firm profits decline, these losses are outweighed by fiscal revenues and reductions in the social costs of excessive sugar intake. Overall, our findings suggest that extending sugar taxes to other food and drink products can reduce sugar consumption and generate positive welfare effects. We then provide valuable evidence for policymakers considering broader fiscal measures to address diet-related health challenges.

This study examines whether climate smart agriculture (CSA) practices reduce technical inefficiency among smallholder vegetable farmers in Northwest Ethiopia. Using cross sectional data from 550 households producing onion, potato, and tomato, the study applies a stochastic frontier analysis to estimate technical inefficiency and derives inefficiency scores for second stage analysis. To identify the determinants of inefficiency, three econometric models, beta regression, Fractional Logit, and Two-limit Tobit, are employed and compared. Model selection based on the loglikelihood, Akaike Information Criterion, and Bayesian Information Criterion confirmed the superiority of the beta regression model. The results reveal that average technical inefficiency levels are 0.170 for onion, 0.195 for potato, and 0.244 for tomato, indicating substantial scope for improving productivity through better resource utilization. The findings consistently show that the adoption of CSA practice combinations significantly reduces technical inefficiency across all vegetable crops. In particular, the integrated adoption of soil and water management practices reduces the largest inefficiency part, followed by climate related and soil based practices. These results highlight the importance of complementarily among agricultural technologies to reduce inefficiency. Socioeconomic and institutional factors, such as education, livestock ownership, credit access, and market proximity, significantly influence inefficiency levels. This study contributes to the literature by explicitly integrating CSA practices into the inefficiency framework and showing that the bundled adoption of climate smart practices enhances production efficiency. These findings provide important policy insights for promoting integrated and context specific CSA interventions to improve smallholder productivity and resilience.

Day-to-day fluctuations in heart rate variability (HRV) are widely used to infer autonomic recovery in endurance athletes. However, the extent to which HRV can be forecast one day ahead from readily available external and internal training-load metrics remains unclear. In this study, we evaluated whether machine learning models can predict next-day HRV in competitive cyclists using the two load descriptors most commonly collected in practice: external load quantified as total mechanical work in kilojoules (kJ) and internal load quantified as session rating of perceived exertion (RPE). Seven male competitive endurance cyclists were monitored daily for sixteen weeks, yielding 590 athlete-days of longitudinal data (seven independent time series). Two machine learning approaches-support vector regression (SVR) and extreme gradient boosting (XGBoost)-were compared with a conventional autoregressive model with exogenous inputs (ARX) as a traditional time-series benchmark. Each model was trained individually per athlete under two predictor scenarios (using past HRV-only or past HRV plus kJ and RPE) and across multiple lag orders (1, 4, 7, 10 and 14 days), with forecasting accuracy expressed as root mean squared error (RMSE). Across all athletes, adding kJ and RPE to the past HRV produced only modest reductions in RMSE relative to HRV-only models. XGBoost achieved the lowest one-step-ahead RMSE at short lag, while all models converged at longer lag orders. Predictive accuracy differed markedly between athletes, reflecting the well-known individual nature of autonomic responses. These findings suggest that the two routinely collected load descriptors examined here-total work (kJ) and RPE-add limited information beyond recent HRV history for forecasting next-day HRV, and that broader contextual variables are likely required to meaningfully improve athlete monitoring.

Large-scale ecological restoration projects (ERPs) have become crucial strategies for fragile regions to mitigate degradation and enhance ecosystem services. However, conventional evaluations of ERP effectiveness often rely on a single model, which is insufficient to capture nonlinear relationships and the interactive effects of multiple driving factors during the restoration process. To address this limitation, this study develops a comprehensive analytical framework that integrates the Generalized difference-in-differences, vine copula model, and piecewise structural equation modeling. By combining spatial analysis with econometric techniques, this framework not only identifies the direct and indirect effects of vegetation change but also characterizes its complex nonlinear response mechanisms. When applied to the karst peak-cluster depression in southwest China, a representative fragile ecosystem, the results reveal that ERPs significantly enhanced regional greening across geomorphic units from 1991 to 2020, though the effectiveness varied markedly among regions. Furthermore, the combined effects of extreme drought and heat greatly increased degradation risks, underscoring the region's high sensitivity to climate fluctuations. Human activity intensity displayed a nonlinear relationship with greening, characterized by the most favorable restoration conditions occurring in low-disturbance zones such as the urban-rural fringe. Moreover, soil thickness was identified as a key constraint on vegetation recovery in karst landscapes. Overall, this study provides a novel methodological framework that integrates spatial analysis, econometrics, and probabilistic modeling to unravel the complex driving mechanisms of vegetation dynamics. The framework not only deepens understanding of karst ecological restoration but also provides transferable guidance for ecological planning and adaptive management in other fragile ecosystems.

Declines in the energy intensity of national gross domestic product cannot be simply taken as evidence of a country's contribution to global decarbonization, notably when they come from structural changes that relocate energy-intensive production abroad. Here we analyze the role of offshoring in shaping energy intensity trends in a panel of 15 countries of the Organization for Economic Co-operation and Development between 1970 and 2021. Using both a decomposition analysis and a structural econometric model, we show that shifts in the composition of national output not mirrored by equivalent changes in domestic consumption patterns significantly and persistently reduce national energy intensity. These findings support the need to move beyond production-based climate metrics and to incorporate global supply chains for a more reliable assessment of national decarbonization pathways.

Our objective is to document major healthcare resource use and associated costs in the care of patients with bronchiectasis, with a particular focus on the costs incurred by exacerbations. We use a unique dataset from the Bronchiectasis Observational Cohort and Biobank UK. The study includes a baseline cohort of 1119 patients with a primary diagnosis of bronchiectasis, followed for up to five years. Data were extracted and linked to centrally held National Health Service (NHS) resource use data. The average age of the cohort was 64 years and 62% were female. The most common bronchiectasis aetiology was idiopathic or post-infectious. Across follow-ups, exacerbations became less frequent: the proportion of patients with no events increased for all settings, while recurrent (≥2) events declined markedly and single-event categories remained largely stable. Exacerbations were predominantly managed in primary care settings. Based on exacerbation frequency and type and applying 2024/2025 NHS reference costs, the estimated total healthcare exacerbation costs for the study population ranged from £2 million to £3.2 million. Disaggregated estimates for the highest cost scenario include: general practitioner (GP)-only exacerbations (n=2824 events; £479 945), emergency department-managed exacerbations (n=311; £260 126) and inpatient-managed exacerbations (n=680; £2 428 291). The average estimated cost per patient over the full period was £1943. This corresponds to a weighted annual average of £495 per patient (including patients with no exacerbations), with most costs attributable to inpatient-managed exacerbations. Specifically, the average cost was £2980 for individuals with one or more exacerbations, £4865 for those with two or more and £5875 for those with three or more. The findings highlight the economic burden of bronchiectasis exacerbations and demonstrate that targeted management strategies to reduce exacerbation rates should increase health benefits and substantially reduce healthcare costs associated with bronchiectasis.

Data on long-term outcomes of surgical management of rectal endometriosis are scarce and generally provided by cohort studies where choice of surgical technique is based on various uncontrolled factors with significant lost-to-follow up patient rates. Based on the cohort of women enrolled in the ENDORE randomized trial, closely followed up over 10 years, we aimed to assess the long-term outcomes of surgical management of rectal endometriosis by nodule excision or segmental resection. Ten-year follow-up of a randomized controlled trial cohort enrolled in one center from May 2011 to October 2013 at Rouen University Hospital. 55 patients were managed for deep endometriosis infiltrating the rectum up to 15 cm from the anus, with more than a 20 mm-diameter area over the muscular layer and maximum 50% of rectal circumference. Patients randomly received either nodule excision (shaving or disc excision) or segmental colorectal resection. The primary endpoint of both the randomized trial and the present study was the number of patients experiencing at least one of the following symptoms: constipation (1 stool/>5 consecutive days), defecation pain, frequent bowel movements (>= 3 stools/day), anal incontinence, bladder dysfunction (at least one of the last three questions of the Urinary Symptom Profile score >=1). Secondary endpoints were values taken from the Knowles-Eccersley-Scott-Symptom Questionnaire (KESS), Gastrointestinal Quality of Life Index (GIQLI), Wexner scale, Urinary Symptom Profile (USP), pregnancy rate, recurrences and reoperation rates. Fifty-five patients were enrolled, and 5 patients stopped follow-up prior to 10 years (9.1%). The primary endpoint was present in respectively 74.1% vs. 71.4% of patients (OR 0.88, 95% CI 0.27-2.9, P=0.83), while 59.1% vs. 58.3% of patients subjectively reported normal bowel movements (OR 0.97, 95%CI 0.30-3.1, P=0.96). An intention-to-treat comparison of overall KESS, GIQLI, Wexner, USP and SF36 scores did not reveal significant differences between the two arms 10 years postoperatively. Longitudinal Generalized Estimating Equations analysis revealed no differences between functional outcome trajectories over time for conservative and radical rectal surgery. The 10-year recurrence rates of rectal endometriosis in the excision vs. the segmental resection arms were 7.4 % vs. 3.6% (OR 0.46, 95%CI 0.04-5.4, P=0.54). Among patients with pregnancy intention after surgery, the pregnancy rate was 85.3%, with most conceiving naturally (64.4%). 45 children were born to 27 women. Second surgeries related to endometriosis were recorded in 32.7% of cases, with no difference between groups. Ten-year follow-up data show no statistically significant differences between conservative and radical rectal surgery for long-term functional digestive outcomes, rectal recurrence rate and risk of reoperation in this population of women with large involvement of the rectum. Most patients considered their bowel movements normal. Our study suggests that in patients undergoing surgical management of severe deep endometriosis of the rectum, the surgery-related benefits persist over 10 years postoperatively, with a low risk of rectal recurrence.

Vaccination and screening services are essential components of preventive care, yet access to these services remains unequal in many low- and middle-income countries (LMICs). This narrative review examines health equity barriers affecting access to vaccination and screening services in LMICs, with screening evidence drawn predominantly from cervical cancer/HPV-related programs and selected other preventive screening contexts, and identifies strategies to improve equity in preventive care. Relevant English-language literature was identified through PubMed and Scopus and synthesized thematically. The review examined cross-cutting equity barriers across vaccination and the screening contexts represented in this evidence base. Barriers operated across individual and household, community and sociocultural, health system, and policy and structural domains. Common barriers included poverty, transportation costs, low health literacy, stigma, weak primary care, workforce shortages, and inadequate financing. Vaccination-specific challenges included weak cold-chain systems, reliance on campaign-based delivery, vaccine hesitancy, and missed immunization schedules, whereas screening-specific challenges included repeated visits, weak referral and follow-up systems, stigma related to cancer and reproductive health screening, and delayed diagnosis. These barriers disproportionately affected rural populations, low-income households, women and girls, migrants, ethnic minorities, people with disabilities, and other marginalized groups, contributing to persistent inequities by income, gender, education, and rural-urban residence. Advancing equity in preventive care requires strengthening primary health care, reducing financial and geographic barriers, improving culturally responsive communication, using disaggregated data for equity-focused planning, and investing in workforce and service delivery improvements.

Background/Objectives: This study aimed to develop and externally validate machine learning (ML) models for low-density lipoprotein cholesterol (LDL-C) estimation and compare their analytical and clinical performance with conventional formulas, particularly in individuals with elevated triglyceride (TG) levels. Methods: This retrospective study included 11,681 adults whose lipid profiles were retrieved using a laboratory information system. ML models (linear regression, random forest, support vector regression, and XGBoost) were developed using routine lipid parameters and evaluated using 10-fold cross-validation. Performance was assessed using the mean absolute error (MAE), root mean squared error (RMSE), bias, correlation, Bland-Altman agreement, and clinical classification according to LDL-C categories. Subgroup analyses were conducted across TG strata, with an emphasis on TG ≥ 400 mg/dL. Results: ML models generally demonstrated lower error and higher agreement with directly measured LDL-C levels than conventional formulas. XGBoost showed the best overall performance (MAE: 14.7 mg/dL; RMSE: 20.22 mg/dL; R2 = 0.780; r = 0.88) and the lowest deviation. The ML models also showed a higher clinical classification accuracy (up to 66%). Performance declined with increasing TG levels, particularly for conventional formulas, whereas ML models remained more stable, including patients with TG ≥ 400 mg/dL. External validation across independent cohorts and analytical platforms demonstrated stable performance of the XGBoost model and generally higher classification accuracy than conventional LDL-C estimation formulas. Conclusions: ML-based LDL-C estimation may represent a complementary alternative to conventional formulas, particularly in hypertriglyceridemic populations.

Tuberculosis (TB) is the leading global cause of death from a single infectious agent. Recent reductions in global health funding have threatened TB control, making comprehensive assessment of TB, HIV-related TB, and drug-resistant TB burdens before these disruptions essential for shaping effective responses. The WHO End TB Strategy sets targets of a 95% reduction in TB deaths and a 90% reduction in TB incidence between 2015 and 2035. Using results from the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2023, this study aims to assess the burden of TB and multidrug-resistant TB (MDR-TB) across 204 countries and territories, and to evaluate progress towards the WHO End TB incidence and mortality targets. We quantified TB mortality using the Cause of Death Ensemble modelling platform with global vital registration, surveillance, verbal autopsy, and minimally invasive tissue sampling data. For TB morbidity estimation, we simultaneously modelled incidence, prevalence, and mortality by age and sex using DisMod-MR 2.1. A population attributable fraction (PAF) approach was applied to stratify morbidity and mortality estimates by HIV and drug-resistance status. We also calculated disability-adjusted life-years (DALYs) as the sum of years of life lost and years lived with disability. For the risk factor analysis, a comparative risk assessment framework was used and PAFs were derived for alcohol use, smoking, and high fasting plasma glucose to determine the proportion of TB burden associated with these risk factors. In 2023, there were an estimated 9·11 million (95% uncertainty interval 8·04-10·3) incident cases of all-form TB, 1·22 million (0·98-1·49) deaths, and 54·6 million (43·8-65·5) DALYs globally. HIV-related TB comprised 781 000 (690 000-879 000) incident cases and 210 000 (142 000-279 000) deaths, contributing 11·0 million (7·56-14·3) DALYs. MDR-TB accounted for 466 000 (198 000-1 080 000) incident cases, 102 000 (31 700-238 000) deaths, and 3·96 million (1·31-9·01) DALYs. From 2015 to 2023, global all-form TB incidence rates declined by 19·2% (17·8-20·5) and deaths declined by 22·6% (4·7-35·7); declines were larger for drug-susceptible TB than for MDR-TB. Sub-Saharan Africa and south Asia had the highest mortality burdens in 2023; reductions in all-form TB incidence and mortality were uneven between 2000 and 2023, with limited progress in both measures in Latin America and the Caribbean. Removing smoking, alcohol use, and high fasting plasma glucose would reduce global TB deaths to 768 000 (592 000-970 000) and DALYs to 34·9 million (27·8-43·8) in 2023; MDR-TB deaths would decrease to 77 200 (23 400-183 000) and DALYs to 3·12 million (1·03-7·29). Global progress towards WHO End TB targets is disparate and fragile. Although many regions achieved meaningful gains, others have stagnated in recent years. The complexity of TB prevention is amplified by divergent MDR-TB trends, the persistent burden of HIV, and growing exposure to modifiable risk factors. Recent volatility in global health financing threatens to further destabilise this vulnerable epidemiological landscape; concerted action is urgently needed to temper disruptions and preserve progress. Gates Foundation.

This study applied a spatial multilevel modelling approach to the latest Demographic and Health Survey data from the Democratic Republic of the Congo to examine the association between maternal empowerment and infant and under-five mortality. To improve temporal alignment between women's empowerment and child mortality outcomes, the analysis was restricted to births occurring within the five years preceding the survey. The econometric modelling accounted for the hierarchical structure of the data, with children nested within clusters and clusters within provinces, and decomposed maternal empowerment into within-cluster and contextual (cluster-level) components. The results indicate that, after controlling for child-, maternal-, household-, cluster-, and province-level factors, neither the within-cluster component nor the cluster-level mean of maternal empowerment shows a robust average association with infant or under-five mortality. In contrast, substantial heterogeneity in child mortality persists at both the cluster and province levels, with particularly strong variation at the cluster level. These findings suggest that child mortality in the Democratic Republic of the Congo is more strongly shaped by local contextual conditions and demographic factors than by a uniform average effect of maternal empowerment, underscoring the importance of multilevel policy responses that address both household-level disadvantage and place-based inequalities.