Challenges exist when translating artificial intelligence (AI)-driven clinical decision support systems (CDSSs) from research into health-care settings, particularly in infectious diseases, an area in which behaviour, culture, uncertainty, and frequent absence of a ground truth enhance the complexity of medical decision making. We aimed to evaluate clinicians' perceptions of an AI CDSS for intravenous-to-oral antibiotic switching and how the system influences their decision making. This randomised, multimethod study enrolled health-care professionals in the UK who were regularly involved in antibiotic prescribing. Participants were recruited through personal networks and the general email list of the British Infection Association. The first part of the study involved a semistructured interview about participants' experience of antibiotic prescribing and their perception of AI. The second part used a custom web app to run a clinical vignette experiment: each of the 12 case vignettes consisted of a patient currently receiving intravenous antibiotics, and participants were asked to decide whether or not the patient was suitable for switching to oral antibiotics. Participants were assigned to receive either standard of care (SOC) information, or SOC alongside our previously developed AI-driven CDSS and its explanations, for each vignette across two groups. We assessed differences in participant choices according to the intervention they were assigned, both for each vignette and overall; evaluated the aggregate effect of the CDSS across all switching decisions; and characterised the decision diversity across participants. In the third part of the study, participants completed the system usability scale (SUS) and technology acceptance model (TAM) questionnaires to enable their opinions of the AI CDSS to be assessed. 59 clinicians were directly contacted or responded to recruitment emails, 42 of whom from 23 hospitals in the UK completed the study between April 23, 2024, and Aug 16, 2024. The median age of participants was 39 years (IQR 37-47), 19 (45%) were female and 23 (55%) were male, 26 (62%) were consultants and 16 (38%) were training-grade doctors, and 14 (33%) specialised in infectious diseases. Interviews revealed mixed individualisation of prescribing and uneven use of technology, alongside enthusiasm for AI, which was conditional on evidence and usability but constrained by behavioural inertia and infrastructure limitations. Case vignette completion times and many decisions were equivalent between SOC and CDSS interventions, with clinicians able to identify and ignore incorrect advice. When a statistical difference was observed, the CDSS influenced participants towards not switching (χ2 7·73, p=0·0054; logistic regression odds ratio 0·13 [95% CI 0·03-0·50]; p=0·0031). AI explanations were used only 9% of the time when available. Our software and AI CDSS obtained a good SUS score of 72·3 out of 100 (SD 8·79) and, for the TAM questionnaire, scores of 3·6 out of 5 (0·31) for perceived usefulness, 3·8 out of 5 (0·20) for perceived ease of use, and 4·1 out of 5 (0·05) for self-efficacy. This AI CDSS was positively received and has the potential to support antimicrobial prescribing, with the greatest influence on clinicians when it recommended not switching from intravenous to oral treatment. Further prospective research is needed to gather safety and benefit data and to understand behavioural changes as AI CDSSs enter clinical practice. Our research suggests that AI explanations are likely to have a minor role at the point of care, and that AI CDSS adoption and utilisation depends on systems being easy to use and trusted, primarily through clinical evidence. The UK Research and Innovation Centre for Doctoral Training in AI for Healthcare, and the National Institute for Health and Care Research Health Protection Research Unit in Healthcare Associated Infections and Antimicrobial Resistance at Imperial College London.
The 2023 iteration of the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) estimated prevalence, incidence, and health burden for 375 diseases and injuries, including 12 mental disorders. We assess past, current, and emerging trends in the prevalence and burden of mental disorders across sexes and age groups, for 21 regions, 204 countries and territories, and by Socio-demographic Index (SDI) quintile, from 1990 to 2023. Mental disorders included in GBD 2023 were anxiety disorders, major depressive disorder, dysthymia, bipolar disorder, schizophrenia, autism spectrum disorders, conduct disorder, attention-deficit hyperactivity disorder, anorexia nervosa, bulimia nervosa, idiopathic developmental intellectual disability, and a residual category of other mental disorders. A literature review identified epidemiological data for each disorder. These were analysed via a Bayesian meta-regression to estimate prevalence by disorder, sex, age, location, and year. Disorder-specific prevalence was multiplied by disability weights representing the severity of health loss associated with each disorder to estimate years lived with disability (YLDs). Deaths due to anorexia nervosa were assessed with a Cause of Death Ensemble modelling strategy to estimate deaths by sex, age, location, and year, and then multiplied by the standard life expectancy at age of death to estimate years of life lost (YLLs). YLDs equalled disability-adjusted life-years (DALYs) for all mental disorders except anorexia nervosa (the only mental disorder considered as an underlying cause of death in GBD), for which DALYs represented the sum of YLDs and YLLs. We presented prevalence, deaths, YLDs, YLLs, and DALYs as counts, age-specific rates per 100 000 population, and age-standardised rates per 100 000 population. We estimated 1·17 billion (95% uncertainty interval 1·06-1·31) prevalent cases of mental disorders globally in 2023, equivalent to an age-standardised prevalence rate of 14 210·7 cases (12 849·5-15 940·1) per 100 000 population. These estimates represented a 95·5% (75·0-121·2) increase in prevalent cases and 24·2% (11·4-41·4) increase in age-standardised prevalence rate between 1990 and 2023. All mental disorders showed increases in prevalent cases between 1990 and 2023, while notable increases were seen in age-standardised prevalence rates for anxiety disorders, major depressive disorder, dysthymia, anorexia nervosa, bulimia nervosa, schizophrenia, and conduct disorder. There were an estimated 171 million (127-228) DALYs due to mental disorders globally across sex and age in 2023, equivalent to an age-standardised DALY rate of 2070·5 DALYs (1519·1-2750·5) per 100 000 population. Mental disorders contributed to 6·1% (4·8-7·6) of all-cause DALYs in 2023, making them the fifth leading cause of global DALYs (up from 12th in 1990). DALYs were almost entirely composed of YLDs. Mental disorders were the leading cause of YLDs in 2023 (up from second in 1990), explaining 17·3% (14·8-20·6) of all-cause global YLDs. Leading causes of mental disorder DALYs were anxiety disorders (ranked 11th among the 304 diseases and injuries at Level 4 of the GBD cause hierarchy), major depressive disorder (15th), and schizophrenia (41st). Globally in 2023, mental disorder age-standardised DALY rates were higher among females (2239·6 [1643·7-3014·1] per 100 000) than among males (1900·2 [1399·8-2510·8] per 100 000), and peaked in the 15-19 years age group (2617·3 [1850·6-3696·8] per 100 000). All locations showed increased mental disorder DALY rates in 2023 compared with 1990, ranging across countries and territories from 1302·4 (952·7-1683·7) per 100 000 in Viet Nam to 3555·8 (2661·9-4715·0) per 100 000 in the Netherlands. Across SDI quintiles, DALY rates ranged from 1853·0 (1352·1-2469·3) per 100 000 for middle SDI to 2184·1 (1606·1-2890·3) per 100 000 for high SDI. A significant health burden was imposed by mental disorders in all countries and territories in 2023, irrespective of the health resources available. In some instances, this burden has increased over time and is unevenly distributed across populations. Stronger surveillance systems, particularly in low-income and middle-income countries, are required. Additionally, we need more coordinated and inclusive policies to reduce the burden through early treatment and prevention, tailored to sex and age differences across locations. Responding to the mental health needs of our global population, especially those most vulnerable, is an obligation, not a choice. Gates Foundation, Queensland Health, and University of Queensland.
Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease with a life expectancy of only 3-5 years and few approved treatments. To identify drug repurposing candidates for the treatment of ALS, we analysed the electronic health records (EHRs) of a large cohort of military veterans with ALS. We analysed the EHRs of individuals in the US Veterans Health Administration (VHA) database who were diagnosed with ALS between Jan 1, 2009 and Dec 31, 2019 to assess medication effects. Individuals without recorded prescriptions after the date of diagnosis were excluded. Two sets of criteria were applied to ascertain exposure. Exposure criteria A were met if the dispense date or the end date of the medication was within 12 months of ALS diagnosis and the end date was at least 6 months after the dispense date. Exposure criteria B were met if there were at least two dispenses within 6 months before diagnosis and 12 months after diagnosis. Propensity score-matched control groups were generated on the basis of confounders included in the EHR, with methodology of potential outcomes used to infer treatment effects. The primary outcome was death. A standard Cox proportional hazards analysis was done to assess association with survival. Survival was defined as the time from diagnosis date recorded in the EHR to death reported in the Department for Veterans Affairs Vital Status File. Follow-up survival time was censored on Dec 31, 2020, for those alive on this date. Downstream protein targets of drugs with clinically significant effects were analysed using the protein-protein interaction networks-based algorithm PathFX. The EHRs of 11 003 individuals with ALS in the VHA database were appropriate for analysis. 162 medications with treatment groups of 30 or more individuals were identified. Among these 162 medications, 27 were associated with statistically significant changes (≥0·1) in the hazard ratio (HR) for death. 18 of the medications were associated with a reduced HR for death (prolonged survival), and nine were associated with an increased HR for death (reduced survival). Drugs associated with reduced HR included HMG-CoA reductase inhibitors (simvastatin, pravastatin, lovastatin, and atorvastatin), PDE5 inhibitors (vardenafil and sildenafil), and α-adrenergic antagonists (tamsulosin and terazosin). The medications associated with an increased HR were drugs used either in the management of clinical features of ALS associated with poor outcomes or in end-of-life care. PathFx analysis identified a complex of proteins interacting with several of the identified drugs. To our knowledge, this analysis is the largest EHR-based study for identifying drug repurposing candidates for ALS. We identified several drugs that warrant further assessment as therapeutic options in ALS, as well as a protein network complex that might serve as a therapeutic target for ALS. Congressionally Directed Medical Research Programs, US Department of Defense.
Lower respiratory infections (LRIs) remain the world's leading infectious cause of death. This analysis from the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2023 provides global, regional, and national estimates of LRI incidence, mortality, and disability-adjusted life-years (DALYs), with attribution to 26 pathogens, including 11 newly modelled pathogens, across 204 countries and territories from 1990 to 2023. With new data and revised modelling techniques, these estimates serve as an update and expansion to GBD 2021. Through these estimates, we also aimed to assess progress towards the 2025 Global Action Plan for the Prevention and Control of Pneumonia and Diarrhoea (GAPPD) target for pneumonia mortality in children younger than 5 years. Mortality from LRIs, defined as physician-diagnosed pneumonia or bronchiolitis, was estimated using the Cause of Death Ensemble model with data from vital registration, verbal autopsy, surveillance, and minimally invasive tissue sampling. The Bayesian meta-regression tool DisMod-MR 2.1 was used to model overall morbidity due to LRIs. DALYs were calculated as the sum of years of life lost (YLLs) and years lived with disability (YLDs) for all locations, years, age groups, and sexes. We modelled pathogen-specific case-fatality ratios (CFRs) for each age group and location using splined binomial regression to create internally consistent estimates of incidence and mortality proportions attributable to viral, fungal, parasitic, and bacterial pathogens. Progress was assessed towards the GAPPD target of less than three deaths from pneumonia per 1000 livebirths, which is roughly equivalent to a mortality rate of less than 60 deaths per 100 000 children younger than 5 years. In 2023, LRIs were responsible for 2·50 million (95% uncertainty interval [UI] 2·24-2·81) deaths and 98·7 million (87·7-112) DALYs, with children younger than 5 years and adults aged 70 years and older carrying the highest burden. LRI mortality in children younger than 5 years fell by 33·4% (10·4-47·4) since 2010, with a global mortality rate of 94·8 (75·6-116·4) per 100 000 person-years in 2023. Among adults aged 70 years and older, the burden remained substantial with only marginal declines since 2010. A mortality rate of less than 60 deaths per 100 000 for children younger than 5 years was met by 129 of the 204 modelled countries in 2023. At a super-regional level, sub-Saharan Africa had an aggregate mortality rate in children younger than 5 years (hereafter referred to as under-5 mortality rate) furthest from the GAPPD target. Streptococcus pneumoniae continued to account for the largest number of LRI deaths globally (634 000 [95% UI 565 000-721 000] deaths or 25·3% [24·5-26·1] of all LRI deaths), followed by Staphylococcus aureus (271 000 [243 000-298 000] deaths or 10·9% [10·3-11·3]), and Klebsiella pneumoniae (228 000 [204 000-261 000] deaths or 9·1% [8·8-9·5]). Among pathogens newly modelled in this study, non-tuberculous mycobacteria (responsible for 177 000 [95% UI 155 000-201 000] deaths) and Aspergillus spp (responsible for 67 800 [59 900-75 900] deaths) emerged as important contributors. Altogether, the 11 newly modelled pathogens accounted for approximately 22% of LRI deaths. This comprehensive analysis underscores both the gains achieved through vaccination and the challenges that remain in controlling the LRI burden globally. Furthermore, it demonstrates persistent disparities in disease burden, with the highest mortality rates concentrated in countries in sub-Saharan Africa. Globally, as well as in these high-burden locations, the under-5 LRI mortality rate remains well above the GAPPD target. Progress towards this target requires equitable access to vaccines and preventive therapies-including newer interventions such as respiratory syncytial virus monoclonal antibodies-and health systems capable of early diagnosis and treatment. Expanding surveillance of emerging pathogens, strengthening adult immunisation programmes, and combating vaccine hesitancy are also crucial. As the global population ages, the dual challenge of sustaining gains in child survival while addressing the rising vulnerability in older adults will shape future pneumonia control strategies. Gates Foundation.
To investigate the value of baseline CT imaging for the prediction of functional outcome and benefit of endovascular thrombectomy (EVT) for anterior large vessel occlusion (LVO). We used individual patient data from seven randomized EVT trials and included patients with available baseline CT imaging and outcome data. We developed a model to predict functional outcome and benefit of EVT, including baseline stroke-related and brain frailty CT imaging features alone. We compared the discriminative performance of our model for predicting good functional outcome (modified Rankin Scale [mRS] 0-2) and treatment benefit (difference between the probability of mRS 0-2 with vs without EVT) with MR PREDICTS by calculating the difference in C-statistics (delta C and delta C-for-benefit). We included 1391 patients (median age, 67 years, interquartile range 59-76; 53% male). Discrimination of the model based on CT imaging alone was substantial for the prediction of good functional outcome (C-statistic 0.700, 95% CI: 0.666-0.731) and treatment benefit (C-for-benefit 0.640, 95% CI: 0.586-0.690). After adding the known strongest clinical predictors namely age and National Institutes of Health Stroke Scale score, discrimination improved to slightly lower than MR PREDICTS for prediction of good functional outcome (C-statistic 0.733 vs 0.750; delta C, -0.017 [95% CI: -0.037 to 0.003]) and treatment benefit (C-for-benefit 0.675 vs 0.692; delta C-for-benefit -0.017 [95% CI: -0.084 to 0.050]). Baseline CT imaging holds considerable predictive value with regard to functional outcome and treatment benefit, but a combination of clinical and imaging features offers the best predictive performance. Question The predictive value of baseline CT imaging for the prediction of functional outcome and benefit of EVT for anterior LVO stroke is uncertain. Findings Discrimination of a model based on CT imaging alone is substantial, but can further be improved by the addition of limited clinical characteristics. Clinical relevance Baseline CT imaging holds considerable predictive value with regard to functional outcome and treatment benefit. The addition of limited clinical information is needed to achieve predictive performance similar to an established prediction model.
Quantitative structure-activity/property relationship (QSAR/QSPR) is a well-established methodology widely used to model molecular properties based on structure and is applied in fields such as drug design and environmental protection. The knowledge and procedures developed and used in QSPR modelling will be applied to the validation of protein folding rate models. Understanding the protein folding process is considered one of the most important scientific topics, and identifying the fundamental factors responsible for protein folding has been the subject of intensive research over the past 30 years. Among the structural descriptors determining the protein folding rate, the length of the protein sequence, the content of regular secondary structures, and the average contact row distance between amino acids in the 3D structure are the most important. Comparative studies of different methods for predicting protein folding rates are occasionally published, and we conducted one such study. We found that the experimental data in literature databases and the data available online are inconsistent and scattered. This is partly due to differences in experimental data and protein sequence lengths, but more so due to the questionable quality of the models themselves. We observed very large deviations in the predictions of ln(kf) by some of the analysed models implemented as web servers. The root mean square errors (RMSEs) of some of the analysed models in predicting ln(kf) for a new external set of proteins are much larger than the RMSEs obtained for the same models on the training sets. External validation demonstrates that protein folding rate models available on web servers have accuracy for external protein sets comparable to that of a simple model based solely on the logarithm of protein chain length. This finding, which highlights the importance of external model validation as recommended by the OECD guidelines for QSAR validation, is fundamental and offers a new perspective for improving protein folding rate models by applying the knowledge and procedures used in the QSPR methodology.
Metabolic dysfunction-associated steatotic liver disease (MASLD), previously known as non-alcoholic fatty liver disease, is one of the most prevalent liver diseases globally, contributing to both economic and health-related challenges. We aimed to evaluate the global, regional, and national burden of MASLD from 1990 to 2023, quantify the contribution of identified modifiable risk factors, and project future prevalence up to the year 2050. Estimates of MASLD prevalence and disability-adjusted life-years (DALYs) were produced by age, sex, region, Socio-demographic Index (SDI), and Healthcare Access and Quality (HAQ) index across 204 countries and territories from 1990 to 2023 as part of the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2023. The MASLD burden attributable to three risk factors (smoking, high BMI, and high fasting plasma glucose) was assessed as part of the GBD comparative risk assessment. As a secondary analysis, we used these estimates to forecast MASLD prevalence up to 2050 using fasting plasma glucose and mean BMI as predictors. Furthermore, to examine the relative contributions of population ageing, population growth, and changes in MASLD prevalence rate to the forecasted changes in case counts from 2023 to 2050, we conducted a decomposition analysis. In 2023, approximately 1·3 billion (95% uncertainty interval [UI] 1·2 to 1·4) individuals were estimated to be living with MASLD (ie, 16·1% of the global population), with an age-standardised prevalence rate of 14 429·3 (95% UI 13 268·3 to 15 990·6) per 100 000 population, representing a percentage increase of 142·7% (95% UI 139·2 to 146·7) in crude numbers from 1990 (0·5 billion [0·5 to 0·6]) and of 28·6% (27·8 to 29·5) in the rate (11 217·2 [10 276·8 to 12 467·0] per 100 000 in 1990). An estimated 3·6 million (2·8 to 4·5) total DALYs were attributable to MASLD worldwide in 2023, corresponding to an age-standardised DALY rate of 39·6 (31·2 to 49·9) per 100 000 population. Despite a 116·3% (93·3 to 139·4) increase in crude DALYs (from 1·7 million [1·3 to 2·1] in 1990), its age-standardised estimate remained consistent (1·8% [-8·6 to 12·8]) from 1990 (38·9 [30·1 to 49·8] per 100 000) to 2023. There was substantial variation in age-standardised estimates across regions. North Africa and the Middle East had the highest prevalence rate (29 246·1 [26 848·3 to 32 048·7] per 100 000) and Andean Latin America showed the highest DALY rate (152·3 [114·1 to 194·7] per 100 000). By contrast, the high-income Asia Pacific region had the lowest prevalence rate (8653·5 [7923·7 to 9592·8] per 100 000) and east Asia had the lowest DALY rate (16·3 [13·5 to 19·9] per 100 000) among all GBD regions. North Africa and the Middle East showed disproportionately higher prevalence rates relative to other regions with similar SDIs. Lower SDIs and HAQs were associated with higher age-standardised DALY rates. The age-standardised prevalence rate was consistently higher in males (15 616·4 [14 349·2 to 17 263·3] per 100 000 people in 2023) than in females (13 245·2 [12 132·0 to 14 692·6] per 100 000 people), and peaked at age 80-84 years in both sexes. The number of MASLD prevalent cases was the highest in younger adults, peaking at age 35-39 years for males and age 55-59 years for females. Among the risk factors for MASLD, high fasting plasma glucose presented the largest contribution to the age-standardised DALY rate of total MASLD in 2023 (2·2 [95% UI 1·6 to 3·1] per 100 000 people), followed by high BMI (1·4 [0·6 to 2·4] per 100 000 people) and smoking (1·0 [0·3 to 1·8] per 100 000 people). Our forecasting model estimates that 1·8 billion (95% UI 1·6 to 2·0) individuals are likely to have MASLD by 2050, representing a 42·0% increase from 2023. The age-standardised prevalence rate is expected to increase to 15 774·9 (95% UI 14 613·9 to 17 336·2) per 100 000 people in 2050, representing an average annual percentage change of 0·3% (95% UI 0·3-0·3). According to our decomposition analysis, this change will be primarily due to population growth, particularly in sub-Saharan Africa and North Africa and Middle East, and less by population ageing or epidemiological change. With a global prevalence of 16·1% and approximately 1·3 billion people already living with MASLD in 2023, the condition has and will continue to have substantial health and economic impacts worldwide. An inverse association between the HAQ Index and age-standardised DALY rates suggests that countries with lower health-care access and quality might be less well positioned to manage the growing MASLD burden, underscoring the need for strengthened health-system capacity in these settings. Gates Foundation.
To determine the accuracy for progressing records to full-text screening using one vs two reviewers to screen title and abstracts in 3 reviews of the effectiveness of interventions for chronic primary low back pain. Secondary objectives include computing inter-rater reliability, describing misclassified records and reviewer performance across reviews, and conducting sensitivity analysis limited to English records and falsely excluded records. One reviewer screened title and abstracts using standardized eligibility criteria and results were compared to consensus screening from two reviewers. We computed sensitivity, specificity, positive (PPV) and negative predictive values (NPV) with 95% confidence intervals using the two reviewers as the comparison. We calculated the inter-rater reliability, proportion of misclassified citations, and the reasons for misclassification. We conducted sensitivity analyses by restricting the analysis to English records. The sensitivity of one reviewer ranged from 48.8% to 66.3% and the specificity ranged from 88.0% to 93.3%. The PPV ranged from 40.6% to 51.8% and NPV 93.6% to 95%. The inter-rater reliability ranged from 0.39 to 0.50. Between 5.0% and 6.3% of records were misclassified as false negative by a single reviewer. Reasons for misclassification were primarily related to the assessment of relevant interventions and comparators, such as whether the intervention could be isolated. Our sensitivity analysis showed that screening English records only compared to all languages improved sensitivity and PPV, with no change in specificity and NPV. Using a single reviewer to screen titles and abstracts may lead to the exclusion of eligible records during title and abstract screening in rapid reviews of the literature. We caution against using Kappa alone as an indicator of the quality of screening, as it is influenced by classification imbalances and suggest including accuracy measures to describe the potential for differences between reviewer screening classifications. This study investigated whether one reviewer can accurately screen research articles for inclusion in a systematic review, compared to the usual approach of having 2 people do the screening. This was tested in three reviews of common treatments for chronic primary low back pain. The single reviewer who screened titles and abstracts was likely to miss relevant articles that were identified as relevant by 2 reviewers. However, the single reviewer was good at correctly excluding irrelevant articles. Between 5% and 6% of eligible articles were incorrectly excluded by the single reviewer. Most mistakes happened when the single reviewer was uncertain about a treatment's eligibility. Limiting screening to English language articles slightly improved the accuracy of the screening but it did not eliminate the risk of missing relevant research. Since artificial intelligence was used to translate Chinese studies to English, further research on the usefulness for this approach is warranted. In summary, restricting screening of articles to one reviewer may save time, but it increases the probability that important evidence will be overlooked. Researchers should be cautious about relying on a single reviewer and should use additional quality assurance to limit bias.
Digital health (DH) has emerged as the dominant terminology for technology-enabled healthcare, reshaping institutional discourse, educational programmes and policy agendas. DH is often portrayed as a novel discipline that could supersede Biomedical and Health Informatics (BMHI), creating conceptual ambiguity and the risk of weakening scientific rigour in the design, evaluation and governance of digital interventions. This paper clarifies BMHI's disciplinary identity, characterises DH as a practice-oriented and implementation-oriented domain, and establishes the epistemic relationship between the two in order to inform academic strategy, curricula and professional identity. A comparative conceptual analysis examined scope, theoretical foundations, methodological approaches, educational frameworks, professional structures and stakeholder ecosystems in both domains, informed by recent debates on degree programmes and workforce training. We found that BMHI is a mature scientific discipline with explicit theoretical foundations encompassing data representation, knowledge modelling, clinical reasoning, decision support, interoperability, sociotechnical analysis, evaluation science, safety, ethics and governance. DH, by contrast, represents an implementation-oriented translational domain that operationalises BMHI through products, services and organisational transformation (eg, telehealth, mobile applications, artificial intelligence-enabled clinical services). While DH has amplified visibility and accelerated adoption, it lacks autonomous theoretical foundations and depends on BMHI for scientific validity, safety and sustainability. Three evolutionary scenarios are examined: DH absorbing BMHI (epistemically incoherent), convergence as parallel disciplines (conceptually flawed), and reintegration of DH within BMHI as its applied layer (scientifically coherent and strategically sustainable) but requiring disciplinary communication, curricular alignment and institutional leadership explicitly positioning DH as a translational layer of BMHI.
Breast cancer is a leading cause of mortality and morbidity among females worldwide. As part of the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2023, we provided an updated comprehensive assessment of the epidemiological trends, disease burden, and risk factors associated with breast cancer globally, regionally, and nationally from 1990 to 2023. Breast cancer incidence, mortality, prevalence, years lived with disability (YLDs), years of life lost (YLLs), and disability-adjusted life-years (DALYs) were estimated by age and sex for 204 countries and territories from 1990 to 2023. Mortality estimates were generated using GBD Cause of Death Ensemble models, leveraging data from population-based cancer registration systems, vital registration systems, and verbal autopsies. Mortality-to-incidence ratios were calculated to derive both mortality and incidence estimates. Prevalence was calculated by combining incidence and modelled survival estimates. YLLs were established by multiplying age-specific deaths with the GBD standard life expectancy at the age of death. YLDs were estimated by applying disability weights to prevalence estimates. The sum of YLLs and YLDs equalled the number of DALYs. Breast cancer burden attributable to seven risk factors was examined through the comparative risk assessment framework. The GBD forecasting framework was used to forecast breast cancer incidence and mortality from 2024 to 2050. Age-standardised rates were calculated for each metric using the GBD 2023 world standard population. In 2023, there were an estimated 2·30 million (95% uncertainty interval [UI] 2·01 to 2·61) breast cancer incident cases, 764 000 deaths (672 000 to 854 000), and 24·1 million (21·3 to 27·5) DALYs among females globally. In the World Bank low-income group, where a low age-standardised incidence rate (ASIR) was estimated (44·2 per 100 000 person-years [31·2 to 58·4]), the age-standardised mortality rate (ASMR) was the highest (24·1 per 100 000 [16·8 to 31·9]). The highest ASIR was in the high-income group (75·7 per 100 000 [67·1 to 84·0]), and the lowest ASMR was in the upper-middle-income group (11·2 per 100 000 [10·2 to 12·3]). Between 1990 and 2023, the ASIR in the low-income group increased by 147·2% (38·1 to 271·7), compared with a 1·2% (-11·5 to 17·2) change in the high-income group. The ASMR decreased in the high-income group, changing by -29·9% (-33·6 to -25·9), but increased by 99·3% (12·5 to 202·9) in the low-income group. The increase in age-standardised DALY rates followed that of ASMRs. Risk factors such as dietary risks, tobacco use, and high fasting plasma glucose contributed to 28·3% (16·6 to 38·9) of breast cancer DALYs in 2023. The risk factors with a decrease in attributable DALYs between 1990 and 2023 were high alcohol use and tobacco. By 2050, the global incident cases of breast cancer among females were forecast to reach 3·56 million (2·29 to 4·83), with 1·37 million (0·841 to 2·02) deaths. The stable incidence and declining mortality rates of female breast cancer in high-income nations reflect success in screening, diagnosis, and treatment. In contrast, the concurrent rise in incidence and mortality in other regions signals health system deficits. Without effective interventions, many countries will fall short of the WHO Global Breast Cancer Initiative's ambitious target of achieving an annual reduction of 2·5% in age-standardised mortality rates by 2040. The mounting breast cancer burden, disproportionately affecting some of the world's most vulnerable populations, will further exacerbate health inequalities across the globe without decisive immediate action. Gates Foundation, St Jude Children's Research Hospital.
The ongoing digitalization of medicine, increased computing power and low-cost storage capacities enable the use of AI-based algorithms for epidemiological big data analysis of electronic patient records. The aim of this study was to evaluate the representativeness of a data network with infrastructure for federated machine learning (ML) across numerous German hospitals. This was done by comparing basic data from the ML data network with publicly available data from the Federal Statistical Office (DESTATIS) to test the scientific validity for future epidemiological analyses. In a retrospective epidemiological secondary analysis, 8,106,105 case files from the ML network were examined and compared to DESTATIS data regarding age, gender, length of hospital stay, ICD-10 diagnoses, and OPS codes. In addition, ICD-10 codes for substance abuse and the regional distribution were compared to examine socioeconomic confounders. The variables age, gender and length of stay, as well as the most common general ICD-10 and OPS codes and ENT-specific OPS codes, showed a high level of concordance based on clinical relevance. For the ENT-specific ICD-10 codes, 2 out of 11 of the most frequent codes showed a maximum deviation of 3.71%. The analysis of socioeconomic factors and regional distribution showed no evidence for deviations. The high level of agreement for the variables examined indicates the representativeness of the ML dataset in comparison to the DESTATIS data. This finding paves the way for future epidemiological studies based on big data, which were previously unavailable in research.
Predicting the progression/regression of coronary plaque burden is challenging. We aimed to develop a deep learning model to forecast changes in percent atheroma volume (ΔPAV) using intravascular ultrasound (IVUS). We analysed data from IBIS-4 and PACMAN-AMI. Core lab measurements of plaque burden were available from IVUS pullbacks. Each model consists of a bidirectional Long Short-Term Memory (biLSTM) layer followed by two fully connected layers with one neuron each, resulting in both a classification for input progression/regression and an estimation of the ΔPAV. For the derivation and validation, a total of 1,960 regions of interest (ROIs) from the IBIS-4 dataset were used. The mean±standard deviation of the model accuracy was 0.85±0.02, the Matthews correlation coefficient was 0.70±0.04, and the F1 score was 0.85±0.02 for both progression and regression classes. In the testing (external validation) process with the PACMAN-AMI dataset, 5,283 ROIs were utilised. The mean ΔPAV was -0.31±5.63, for which 2,665 featured regression with a mean ΔPAV of -4.57±3.73, and 2,618 presented progression with a mean ΔPAV of 4.02±3.55, representing 49.6% of plaque progression prevalence. The predictive performance across the 100 trained models in the testing dataset showed an accuracy of 0.84, a Matthews correlation coefficient of 0.68, and an F1 score for the progression and regression classes of 0.84. This is the first deep learning model capable of detecting changes in plaque progression by analysing the rate of plaque burden change between adjacent frames.
The global burden of sepsis, a life-threatening dysregulated host response to infection leading to organ dysfunction, remains challenging to quantify. We aimed to comprehensively estimate the global, regional, and national burden of sepsis, including the impact of the COVID-19 pandemic and underlying causes of sepsis-related deaths with co-occurring infectious syndromes. We used multiple cause-of-death, hospital, minimally invasive tissue sampling, and linked death certificate and hospital record data representing 149 million deaths, covering 4290 location-years with mortality estimates from the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2021 to capture explicit and implicit sepsis cases and deaths. We estimated age-location-sex-specific fractions of sepsis-related deaths from 195 underlying causes of death and 22 infectious syndromes from 1990 to 2021 using binomial logistic regression models, and estimated sepsis-related deaths using GBD cause-specific mortality estimates. Using 250 million hospital admissions and 7·82 million deaths from hospital data, representing 1310 location-years, we modelled case fatality rates by use of binomial logistic regression, applied to sepsis death estimates to estimate sepsis incidence by age, location, and year. In 2021, we estimated 166 million (95% uncertainty interval 135-201) sepsis cases and 21·4 million (20·3-22·5) all-cause sepsis-related deaths globally, representing 31·5% of total global deaths. Sepsis-related deaths decreased between 1990 and 2019, followed by a surge in 2020 and 2021. As of 2021, individuals aged 15 years and older experienced increases across incidence (230%) and mortality (26·3%) since 1990. Those aged 70 years and older had the highest sepsis-related mortality in 2021 (9·28 million [8·74-9·86] deaths). Sepsis-related deaths from infectious underlying causes decreased from 11·8 million (11·1-12·5) in 1990 to 8·34 million (7·72-9·01) in 2019, then increased by 86·4% to 15·5 million (14·7-16·4) in 2021. Sepsis-related mortality due to non-infectious underlying causes of death increased from 4·69 million (4·35-5·05) in 1990 to 5·81 million (5·40-6·25) in 2021; the leading non-infectious underlying causes of death with sepsis were stroke, chronic obstructive pulmonary disease, and cirrhosis. In 2021, bloodstream infections inclusive of HIV and malaria (3·08 million [2·83-3·35]) and lower respiratory infections inclusive of COVID-19 (11·33 million [1·20-1·47]) were the most prominent infectious syndromes complicating sepsis-related deaths from non-infectious underlying causes, representing a consistent trend since 1990. The global burden of sepsis increased in 2020 and 2021, reversing progress from 1990. Sepsis incidence and mortality increased in people aged 15 years and older, especially those aged 70 years and older, and as a complication of non-infectious underlying causes of death such as stroke, primarily through bloodstream infections and lower respiratory infections. The global burden of sepsis is substantial, and sepsis is increasingly a complication of non-infectious causes of death. Gates Foundation, Wellcome Trust, and Department of Health and Social Care using UK aid funding managed by the Fleming Fund.
Child growth failure (CGF), which includes underweight, wasting, and stunting, is among the factors most strongly associated with mortality and morbidity in children younger than 5 years worldwide. Poor height and bodyweight gain arise from a variety of biological and sociodemographic factors and are associated with increased vulnerability to infectious diseases. We used data from the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2023 to estimate CGF prevalence, the risk of infectious diseases associated with CGF, and the disease mortality, morbidity, and overall burden associated with CGF. In this analysis we estimated the all-cause and cause-specific (diarrhoea, lower respiratory tract infections, malaria, and measles) disability-adjusted life-years (DALYs) lost and mortality associated with stunting, wasting, underweight, and CGF in aggregate. We combined the burden associated with mild, moderate, and severe forms of CGF: stunting was defined as height-for-age Z scores (HAZ) less than -1, underweight was defined as weight-for-age Z scores (WAZ) less than -1, and wasting was defined as weight-for-height Z scores (WHZ) less than -1, according to WHO Child Growth Standards. Population-level continuous distributions of HAZ, WAZ, and WHZ were estimated for 2000 to 2023 using data from surveys, literature, and individual-level study data. The risk of incidence of, and mortality due to, diarrhoea, lower respiratory infections, malaria, and measles was separately estimated in a meta-regression framework from longitudinal cohort data for Z scores less than -1. Finally, fatal outcomes associated with these diseases were estimated with vital registration, verbal autopsy, and case-fatality data, while non-fatal outcomes were estimated with surveys as well as health-care utilisation and case reporting data. The exposure prevalence and relative risk estimates were from continuous distributions, allowing for direct assessment of the attributable fractions for mild, moderate, and severe stunting, underweight, wasting, and the combined impact of child growth failure within populations. All estimates were age-specific, sex-specific, geography-specific, and year-specific. We estimated that, in children younger than 5 years in 2023, CGF was associated with 79·4 million (95% uncertainty interval [UI] 47·0-106) DALYs lost and 880 000 (517 000-1 170 000) deaths. This represented 17·9% (10·6-23·8) of 444 million (434-457) total under-5 DALYs and 18·8% (11·1-25·0) of all 4·67 million (4·59-4·75) under-5 deaths. Compared to stunting (33·0 million [24·1-42·2] DALYs, 373 000 [272 000-477 000] deaths) and wasting (39·2 million [23·8-53·0] DALYs, 428 000 [256 000-583 000] deaths), childhood underweight was associated with the largest share of CGF-related disease burden: 52·2 million (21·9-75·1) DALYs and 573 000 (236 000-824 000) deaths in children younger than 5 years in 2023. CGF remains a leading factor associated with death and disability in children younger than 5 years, despite global attention and focused interventions to reduce the prevalence of associated CGF indicators. Our findings underscore the need for policies, strategies, and interventions that focus on all indicators of CGF to reduce its associated health burden. Gates Foundation.
Chronic kidney disease (CKD) is common and ranks among the leading causes of mortality and morbidity. This analysis aimed to present global CKD estimates using the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2023 to inform evidence-based policies for CKD identification and treatment. This analysis focused on adults aged 20 years and older over the period 1990 to 2023, from 204 countries and territories. Data sources used were published literature, vital registration systems, kidney failure treatment registries, and household surveys. Estimates of CKD burden, including deaths, incidence, prevalence, and disability-adjusted life-years (DALYs), were produced using a Cause of Death Ensemble model and a Bayesian meta-regression analytical tool. A comparative risk assessment approach estimated the proportion of cardiovascular deaths attributable to impaired kidney function and estimated risk factors for CKD. Globally, in 2023, 788 million (95% uncertainty interval 743-843) people aged 20 years and older were estimated to have CKD, up from 378 million (354-407) in 1990. The global age-standardised prevalence of CKD in adults was 14·2% (13·4-15·2), a relative rise of 3·5% (2·7-4·1) from 1990. The region with the highest age-standardised prevalence was north Africa and the Middle East (18·0%; 16·9-19·4). Most people had stage 1-3 CKD, with a combined prevalence of 13·9% (13·1-15·0). In 2023, CKD was the ninth leading cause of death globally, accounting for 1·48 million (1·30-1·65) deaths, and the 12th leading cause of DALYs, with an age-standardised DALY rate of 769·2 (691·8-857·4) per 100 000. Impaired kidney function as a risk factor accounted for 11·5% (8·4-14·5) of cardiovascular deaths. High fasting plasma glucose, body-mass index, and systolic blood pressure were all leading risk factors for CKD DALYs. CKD is a major global health issue, with rising prevalence and increasing importance as a cause of death and as a risk factor for cardiovascular death. A better understating of aetiology, appropriate screening, and implementation programmes are needed to translate advances in CKD treatment into improved patient outcomes. Gates Foundation, Wellcome, US National Kidney Foundation, and US National Institute of Diabetes and Digestive and Kidney Diseases.
Body size is one of the most important traits governing individual-level demographic rates and modulating population-level processes. Multiple size-dependent demographic rates can simultaneously change population structure, so distinguishing their individual contributions to overall population dynamics remains a challenge. Disentangling size-dependent harvest rates from other demographic rates is critical for assessing the impact of removal on populations of invasive species. Inference about invasive populations can be difficult, however, as observations are often collected opportunistically as part of removal programs, rather than experimentally designed. Yet accurate inference is essential for understanding the feasibility of population suppression and optimising management decisions. We develop an integrated integral projection model (IPM2) that leverages the strengths of the integrated population model and integral projection model to enable inference about complex, size-structured demographic rates from imperfect observations. We apply the IPM2 in the context of invasive European green crab (Carcinus maenas), a species for which individual body size strongly regulates both the observation-generating process and latent, population dynamics. The IPM2 facilitates the distinct estimation of green crab size-structured harvest and natural mortality rates, parameters for which no explicit data is collected and that are unidentifiable in component datasets of the integrated population model. The model represents how the green crab population changes over time, providing the first estimates of size-structured abundance of this high-priority species. By forecasting the stable size distribution and equilibrium population size under varying removal efforts, we demonstrate that extremely high levels of removal effort can reduce the equilibrium green crab population size. Yet these high mortality rates also shift the stable size distribution and increase the equilibrium abundance of smaller crabs, since size-selective removal alters intraspecific interactions. The ecological outcome of this shift in size structure will be variable, as green crab size modulates only some of its interactions with other species. These results highlight the value of the IPM2 framework for inferring complex population dynamics with information needs that outpace information in individual observational datasets, providing a path forward for accurate assessment of conservation programs.
Flibanserin is the initial pharmaceutical treatment for hypoactive sexual desire disorder (HSDD). The analysis of urine samples plays a crucial role in the quantitation of flibanserin since a portion of flibanserin is excreted unchanged in the urine. An analytical method was proposed to quantify flibanserin in artificial urine samples (as model matrices). The integration of the spray-assisted fine droplet formation-liquid phase microextraction (SFDF-LPME) method and gas chromatography-mass spectrometry (GC-MS) system was performed for the first time to improve the sensitivity of the GC-MS system for flibanserin. Several parameters, including spraying cycle, extraction solvent type, mixing type and period, and sample volume, were systematically optimized to enhance the signal-to-noise ratio (S/N) of the analyte. After determining the optimal conditions, the analytical performance measurements of the system were figured out. The limit of detection (LOD), the limit of quantification (LOQ), and coefficient of determination (R2) values were 6.91, 23.05 μg kg-1, and 0.9989, respectively. Recovery experiments were performed in artificial urine samples within the specified linear working range of 33.15-505.66 μg kg-1. The SFDF-LPME-GC-MS method was efficiently applied to artificial urine samples by computing the matrix-matching calibration strategy, with percentage recovery values ranging from 90.0% to 105.9%.
Detecting glioma recurrence is fundamental for clinical patient outcomes; however, conventional MRI (cMRI) techniques may be limited, leading to diagnostic uncertainty relevant for therapeutic intervention. This study aimed to evaluate whether a microvascular perfusion (µPerf) imaging technique based on spin-echo DSC perfusion MRI could support the early detection of glioma recurrence compared with cMRI by characterizing subtle vascular changes preceding macroscopic tumor growth. A total of 351 patients with gliomas who underwent 2003 follow-up MRI examinations were retrospectively evaluated, with 422 of these examinations subjected to detailed quantitative analysis. The standard cMRI protocol was extended by applying the µPerf approach, with an additional 2 minutes for data acquisition. Custom-made Matlab software was used to generate imaging biomarker maps for microvascular CBV and microvascular type indicator. The clinical utility of µPerf was assessed by comparing its findings with radiologic interpretations of cMRI data, which were reviewed in consensus by at least 2 board-certified radiologists. Statistical analyses included the calculation of diagnostic performance metrics and the area under the receiver operating characteristic curve (AUROC) to evaluate glioma recurrence detection. The µPerf technique exhibited superior diagnostic performance, achieving an accuracy of 97.4% and AUROC values of 0.987 (95% CI, 0.976-0.999) for microvascular type indicator and 0.982 (95% CI, 0.965-0.998) for microvascular CBV, significantly surpassing cMRI (accuracy: 85.1%; AUROC: 0.941; 95% CI, 0.918-0.965 for CBV). µPerf identified glioma recurrence earlier than cMRI in 13.5% of cases, with the time interval ranging from 41 to 353 days (mean, 137 days). During this time, tumor volume increased by 38% to as much as 155-fold (mean, 9.1-fold). Notably, early recurrences of high-grade malignant gliomas were predominantly characterized by microvascular changes compared with later-stage recurrences. µPerf improves early detection of glioma recurrence and shows a higher sensitivity of microvascular changes compared with cMRI. µPerf has significant potential to promote more timely and personalized treatment strategies, which, in turn, could improve patient outcomes. Notably, µPerf works with standard clinical follow-up protocols, but its integration into clinical practice requires further validation through multicenter studies and long-term outcome analyses.
Meningitis remains the leading infectious cause of neurological disabilities globally, disproportionately affecting children younger than 5 years and populations in the African meningitis belt. Whereas previous global estimates focused on ten pathogen categories, this study presents the most comprehensive analysis to date, assessing the meningitis burden attributable to 17 causative pathogens based on the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2023 framework. GBD is a systematic, scientific effort aimed at quantifying the comparative magnitude of health loss caused by diseases, injuries, and risk factors across age groups, sexes, and geographical locations over time. We estimated meningitis mortality using the Cause of Death Ensemble model (CODEm) and morbidity using DisMod-MR 2.1, incorporating data from vital registration, verbal autopsy, surveillance, hospital data, and systematic reviews. Aetiology-specific estimates were generated with pathogen-linked case-fatality ratios and splined binomial regression models. Risk factor attribution was based on established risk-outcome pairs and population attributable fractions. In 2023, there were 259 000 (95% uncertainty interval 202 000-335 000) global deaths and 2·54 million (2·20-2·93) incident cases of meningitis. Children younger than 5 years accounted for more than a third of deaths (86 600 [53 300-149 000]). Streptococcus pneumoniae, Neisseria meningitidis, non-polio enteroviruses, and other viruses were the leading causes of death, while non-polio enteroviruses caused the most cases. The four WHO-defined preventable meningitis pathogens of interest (S pneumoniae, N meningitidis, Haemophilus influenzae, and Group B streptococcus) contributed to 98 700 deaths (77 000-127 000) and 594 000 cases (514 000-686 000). Low birthweight, short gestation, and household air pollution were the top risk factors for meningitis-related mortality. Although mortality and incidence have declined significantly since 1990, progress is insufficient to meet WHO 2030 targets. Despite marked progress in reducing bacterial meningitis via global vaccination campaigns, a substantial meningitis burden persists, attributable both to common pathogens such as S pneumoniae and N meningitidis and to emerging non-bacterial pathogens such as Candida spp and drug-resistant fungi. Achieving WHO goals will require sustained investment in surveillance, vaccination, maternal screening, and health-system strengthening, especially in high-burden settings. Gates Foundation, Wellcome Trust, and UK Department of Health and Social Care.
Guidelines for managing scientific data have been established under the FAIR principles, requiring that data be Findable, Accessible, Interoperable, and Reusable. In many scientific disciplines, especially computational biology, both data and models are key to progress. For this reason, and recognizing that such models are a very special type of "data", we argue that computational models, especially mechanistic models prevalent in medicine, physiology and systems biology, deserve a complementary set of guidelines. We propose the CURE principles, emphasizing that models should be Credible, Understandable, Reproducible, and Extensible. We delve into each principle, discussing verification, validation, and uncertainty quantification for model credibility; the clarity of model descriptions and annotations for understandability; adherence to standards and open science practices for reproducibility; and the use of open standards and modular code for extensibility and reuse. We outline recommended and baseline requirements for each aspect of CURE, aiming to enhance the impact and trustworthiness of computational models, particularly in biomedical applications where credibility is paramount. Our perspective underscores the need for a more disciplined approach to modeling, aligning with emerging trends such as Digital Twins and emphasizing the importance of data and modeling standards for interoperability and reuse. Finally, we emphasize that given the non-trivial effort required to implement the guidelines, the community should strive to automate as many of the guidelines as possible.