Search

BACKGROUND: Pre-diagnostic disturbances in the microbiome-derived metabolome have been associated with an increased risk of diabetes in non-pregnant populations. However, the roles of microbiome-derived metabolites, the end-products of microbial metabolism, in gestational diabetes (GDM) remain understudied. We examined the prospective association of microbiome-derived metabolites in early to mid-pregnancy with GDM risk in a diverse population. METHODS: We conducted a prospective discovery and validation study, including a case-control sample of 91 GDM and 180 non-GDM individuals within the multi-racial/ethnic The Pregnancy Environment and Lifestyle Study (PETALS) as the discovery set, a random sample from the PETALS (42 GDM, 372 non-GDM) as validation set 1, and a case-control sample (35 GDM, 70 non-GDM) from the Gestational Weight Gain and Optimal Wellness randomized controlled trial as validation set 2. We measured untargeted fasting serum metabolomics at gestational weeks (GW) 10-13 and 16-19 by gas chromatography/time-of-flight mass spectrometry (TOF-MS), liquid chromatography (LC)/quadrupole TOF-MS, and hydrophilic interaction LC/quadrupole TOF-MS. GDM was diagnosed using the 3-h, 100-g oral glucose tolerance test according to the Carpenter-Coustan criteria around GW 24-28. RESULTS: Among 1362 annotated compounds, we identified 140 of gut microbiome metabolism origin. Multivariate enrichment analysis illustrated that carbocyclic acids and branched-chain amino acid clusters at GW 10-13 and the unsaturated fatty acids cluster at GW 16-19 were positively associated with GDM risk (FDR

Gun-related crime continues to be an urgent public health and safety problem in cities across the US. A key question is: how are firearms diverted from the legal retail market into the hands of gun offenders? With close to 8 million legal firearm transaction records in California (2010-2020) linked to over 380,000 records of recovered crime guns (2010-2021), we employ supervised machine learning to predict which firearms are used in crimes shortly after purchase. Specifically, using random forest (RF) with stratified under-sampling, we predict any crime gun recovery within a year (0.2% of transactions) and violent crime gun recovery within a year (0.03% of transactions). We also identify the purchaser, firearm, and dealer characteristics most predictive of this short time-to-crime gun recovery using SHapley Additive exPlanations and mean decrease in accuracy variable importance measures. Overall, our models show good discrimination, and we are able to identify firearms at extreme risk for diversion into criminal hands. The test set AUC is 0.85 for both models. For the model predicting any recovery, a default threshold of 0.50 results in a sensitivity of 0.63 and a specificity of 0.88. Among transactions identified as extremely risky, e.g., transactions with a score of 0.98 and above, 74% (35/47 in the test data) are recovered within a year. The most important predictive features include purchaser age and caliber size. This study suggests the potential utility of transaction records combined with machine learning to identify firearms at the highest risk for diversion and criminal use soon after purchase.

Existing risk prediction models for hospitalized heart failure patients are limited. We identified patients hospitalized with a diagnosis of heart failure between 7 May 2013 and 26 April 2022 from a large academic, quaternary care medical centre (training cohort). Demographics, medical comorbidities, vitals, and labs were collected and were used to construct random forest machine learning models to predict in-hospital mortality. Models were compared with logistic regression, and to commonly used heart failure risk scores. The models were subsequently validated in patients hospitalized with a diagnosis of heart failure from a second academic, community medical centre (validation cohort). The entire cohort comprised 21 802 patients, of which 14 539 were in the training cohort and 7263 were in the validation cohort. The median age (25th-75th percentile) was 70 (58-82) for the entire cohort, 43.2% were female, and 6.7% experienced inpatient mortality. In the overall cohort, 7621 (35.0%) patients had heart failure with reduced ejection fraction (EF ≤ 40%), 1271 (5.8%) had heart failure with mildly reduced EF (EF 41-49%), and 12 910 (59.2%) had heart failure with preserved EF (EF ≥ 50%). Random forest models in the validation cohort demonstrated a c-statistic (95% confidence interval) of 0.96 (0.95-0.97), sensitivity (SN) of 87.3%, and specificity (SP) of 90.6% for the prediction of in-hospital mortality. Models for those with HFrEF demonstrated a c-statistic of 0.96 (0.94-0.98), SN 88.2%, and SP 91.0%, and those for patients with HFpEF showed a c-statistic of 0.95 (0.93-0.97), SN 87.4%, and SP 89.5% for predicting in-hospital mortality. The random forest model significantly outperformed logistic regression (c-statistic 0.87, SN 75.9%, and SP 86.9%), and current existing risk scores including the Acute Decompensated Heart Failure National Registry risk score (c-statistic of 0.70, SN 69%, and SP 62%), and the Get With the Guidelines-Heart Failure risk score (c-statistic 0.69, SN 67%, and SP 63%); P

IntroductionEarly risk stratification for clinical dementia could lead to preventive therapies. We identified and validated a magnetic resonance imaging (MRI) signature for Alzheimer's disease (AD) and related dementias (ARDR).MethodsAn MRI ADRD signature was derived from cortical thickness maps in Framingham Heart Study (FHS) participants with AD dementia and matched controls. The signature was related to the risk of ADRD and cognitive function in FHS. Results were replicated in the University of California Davis Alzheimer's Disease Research Center (UCD-ADRC) cohort.ResultsParticipants in the bottom quartile of the signature had more than three times increased risk for ADRD compared to those in the upper three quartiles (P

Objective: To validate a serum biomarker developed in the USA for preterm birth (PTB) risk stratification in Viet Nam. Methods: Women with singleton pregnancies (n = 5000) were recruited between 19+0-23+6 weeks' gestation at Tu Du Hospital, Ho Chi Minh City. Maternal serum was collected from 19+0-22+6 weeks' gestation and participants followed to neonatal discharge. Relative insulin-like growth factor binding protein 4 (IGFBP4) and sex hormone binding globulin (SHBG) abundances were measured by mass spectrometry and their ratio compared between PTB cases and term controls. Discrimination (area under the receiver operating characteristic curve, AUC) and calibration for PTB <37 and <34 weeks' gestation were tested, with model tuning using clinical factors. Measured outcomes included all PTBs (any birth ≤37 weeks' gestation) and spontaneous PTBs (birth ≤37 weeks' gestation with clinical signs of initiation of parturition). Results: Complete data were available for 4984 (99.7%) individuals. The cohort PTB rate was 6.7% (n = 335). We observed an inverse association between the IGFBP4/SHBG ratio and gestational age at birth (p = 0.017; AUC 0.60 [95% CI, 0.53-0.68]). Including previous PTB (for multiparous women) or prior miscarriage (for primiparous women) improved performance (AUC 0.65 and 0.70, respectively, for PTB <37 and <34 weeks' gestation). Optimal performance (AUC 0.74) was seen within 19-20 weeks' gestation, for BMI >21 kg/m2 and age 20-35 years. Conclusion: We have validated a novel serum biomarker for PTB risk stratification in a very different setting to the original study. Further research is required to determine appropriate ratio thresholds based on the prevalence of risk factors and the availability of resources and preventative therapies. [ABSTRACT FROM AUTHOR]

Risk assessment in pulmonary arterial hypertension (PAH) is fundamental to guiding treatment and improved outcomes. Clinical models are excellent at identifying high-risk patients, but leave uncertainty amongst moderate-risk patients. Can a multiple blood biomarker model of PAH, using previously described biomarkers, improve risk discrimination over current models? Using a multiplex enzyme-linked immunosorbent assay, we measured N-terminal pro-B-type natriuretic peptide (NT-proBNP), soluble suppressor of tumorigenicity, IL-6, endostatin, galectin 3, hepatoma derived growth factor, and insulin-like growth factor binding proteins (IGFBP1-7) in training (n = 1,623), test (n = 696), and validation (n = 237) cohorts. Clinical variables and biomarkers were evaluated by principal component analysis. NT-proBNP was not included to develop a model independent of NT-proBNP. Unsupervised k-means clustering classified participants into clusters. Transplant-free survival by cluster was examined using Kaplan-Meier and Cox proportional hazard regressions. Hazard by cluster was compared with NT-proBNP, Registry to Evaluate Early and Long-Term PAH Disease Management (REVEAL), and European Society of Cardiology and European Respiratory Society risk models alone and combined clinical and biomarker models. The algorithm generated five clusters with good risk discrimination using six biomarkers, weight, height, and age at PAH diagnosis. In the test and validation cohorts, the biomarker model alone performed equivalent to REVEAL (area under the receiver operating characteristic curve, 0.74). Adding the biomarker model to the European Society of Cardiology and European Respiratory Society score and REVEAL score improved the European Society of Cardiology and European Respiratory Society score and REVEAL score. The best overall model was the biomarker model adjusted for NT-proBNP with the best C statistic, Akaike information criterion, and calibration for the adjusted model compared with either the biomarker or NT-proBNP model alone. In this study, a multibiomarker model alone was equivalent to current PAH clinical mortality risk prediction models and improved performance when combined and added to NT-proBNP. Clinical risk scores offer excellent predictive models, but require multiple tests; adding blood biomarkers to models can improve prediction or can enable more frequent, noninvasive monitoring of risk in PAH to support therapeutic decision-making. [ABSTRACT FROM AUTHOR]

There have been at least 7 separate randomised controlled trials published between 2011 and 2023 that have examined primary care for nonspecific low back pain informed by the STarT Back approach to stratified care based on risk prediction, compared with care not informed by this approach. The results, across 4 countries, have been contrasting—some demonstrating effectiveness and/or efficiency of this approach, others finding no benefits over comparison interventions. This review considers possible explanations for the differences, particularly whether this is related to poor predictive performance of the STarT Back risk-prediction tool or to variable degrees of success in implementing the whole STarT Back approach (subgrouping and matching treatments to predicted risk of poor outcomes) in different healthcare systems. The review concludes that although there is room for improving and expanding the predictive value of the STarT Back tool, its performance in allocating individuals to their appropriate risk categories cannot alone explain the variation in results of the trials to date. Rather, the learning thus far suggests that challenges in implementing stratified care in clinical practice and in changing professional practice largely explain the contrasting trial results. The review makes recommendations for future research, including greater focus on studying facilitators of implementation of stratified care and developing better treatments for patients with nonspecific low back pain at high risk of poor outcomes. [ABSTRACT FROM AUTHOR]

Purpose: To develop a machine learning algorithm, using patient-reported data from early pregnancy, to predict later onset of first time moderate-to-severe depression. Methods: A sample of 944 U.S. patient participants from a larger longitudinal observational cohortused a prenatal support mobile app from September 2019 to April 2022. Participants self-reported clinical and social risk factors during first trimester initiation of app use and completed voluntary depression screenings in each trimester. Several machine learning algorithms were applied to self-reported data, including a novel algorithm for causal discovery. Training and test datasets were built from a randomized 80/20 data split. Models were evaluated on their predictive accuracy and their simplicity (i.e., fewest variables required for prediction). Results: Among participants, 78% identified as white with an average age of 30 [IQR 26–34]; 61% had income ≥ $50,000; 70% had a college degree or higher; and 49% were nulliparous. All models accurately predicted first time moderate-severe depression using first trimester baseline data (AUC 0.74–0.89, sensitivity 0.35–0.81, specificity 0.78–0.95). Several predictors were common across models, including anxiety history, partnered status, psychosocial factors, and pregnancy-specific stressors. The optimal model used only 14 (26%) of the possible variables and had excellent accuracy (AUC = 0.89, sensitivity = 0.81, specificity = 0.83). When food insecurity reports were included among a subset of participants, demographics, including race and income, dropped out and the model became more accurate (AUC = 0.93) and simpler (9 variables). Conclusion: A relatively small amount of self-report data produced a highly predictive model of first time depression among pregnant individuals. [ABSTRACT FROM AUTHOR]

Background: The clinical relevance of cardiac troponin (cTn) elevation in takotsubo syndrome (TTS) remains uncertain. The present study sought to investigate the role of cardiac troponin (cTn) elevations in mortality prediction of patients with Takotsubo syndrome (TTS). Methods: Patients enrolled in the International Takotsubo (InterTAK) Registry from January 2011 to February 2020 with available data on peak cTn levels were included in the analysis. Peak cTn levels during the index hospitalization were used to define clinically relevant myocardial injury. The threshold at which clinically relevant myocardial injury drives mortality at 1 year was identified using restricted cubic spline analysis. Results: Out of 2′938 patients, 222 (7.6%) patients died during 1‐year follow‐up. A more than 28.8‐fold increase of cTn above the upper reference limit was identified as threshold for clinically relevant myocardial injury. The presence of clinically relevant myocardial injury was significantly associated with an increased risk of mortality at 5 years (adjusted HR 1.58, 95% CI 1.18–2.12, p =.002). Clinically relevant myocardial injury was related to an increased 5‐year mortality risk in patients with apical TTS (adjusted HR 1.57, 95% CI 1.21–2.03, p =.001), in presence of physical stressors (adjusted HR 1.60, 95% CI 1.22–2.11, p =.001), and in absence of emotional stressors (adjusted HR 1.49, 95% CI, 1.17–1.89, p =.001). Conclusion: This study for the first time determined a troponin threshold for the identification of TTS patients at excess risk of mortality. These findings advance risk stratification in TTS and assist in identifying patients in need for close monitoring and follow‐up. [ABSTRACT FROM AUTHOR]

Risk prediction scores are used to guide clinical decision-making. Our primary objective was to externally validate two patient-specific risk scores for 30-day in-hospital mortality using the Multicenter Perioperative Outcomes Group (MPOG) registry: the Pediatric Risk Assessment (PRAm) score and the intrinsic surgical risk score. The secondary objective was to recalibrate these scores. Data from 56 US and Dutch hospitals with paediatric caseloads were included. The primary outcome was 30-day mortality. To assess model discrimination, the area under the receiver operating characteristic curve (AUROC) and area under the precision–recall curve (AUC-PR) were calculated. Model calibration was assessed by plotting the observed and predicted probabilities. Decision analytic curves were fit. The 30-day mortality was 0.14% (822/606 488). The AUROC for the PRAm upon external validation was 0.856 (95% confidence interval 0.844–0.869), and the AUC-PR was 0.008. Upon recalibration, the AUROC was 0.873 (0.861–0.886), and the AUC-PR was 0.031. The AUROC for the external validation of the intrinsic surgical risk score was 0.925 (0.914–0.936) and AUC-PR was 0.085. Upon recalibration, the AUROC was 0.925 (0.915–0.936), and the AUC-PR was 0.094. Calibration metrics for both scores were favourable because of the large cluster of cases with low probabilities of mortality. Decision curve analyses showed limited benefit to using either score. The intrinsic surgical risk score performed better than the PRAm, but both resulted in large numbers of false positives. Both scores exhibited decreased performance compared with the original studies. ASA physical status scores in sicker patients drove the superior performance of the intrinsic surgical risk score, suggesting the use of a risk score does not improve prediction. [ABSTRACT FROM AUTHOR]

To identify longitudinal metabolomic fingerprints of diabetic retinopathy (DR) and to evaluate their usefulness in predicting DR development and progression. Multicenter, multiethnic cohort study. This study included 17 675 participants from the UK Biobank (UKB) who had baseline prediabetes or diabetes, identified in accordance with the 2021 American Diabetes Association guidelines, and were free of baseline DR and an additional 638 participants with type 2 diabetes mellitus from the Guangzhou Diabetic Eye Study (GDES) for external validation. Diabetic retinopathy was determined by ICD-10 codes in the UKB cohort and revised ETDRS grading criteria in the GDES cohort. Longitudinal DR metabolomic fingerprints were identified through nuclear magnetic resonance (NMR) assay in UKB participants. The predictive value of these fingerprints for predicting DR development were assessed in a fully withheld test set. External validation and extrapolation analyses of DR progression and microvascular damage were conducted in the GDES cohort using NMR technology. Model assessments included the concordance (C) statistic, net classification improvement (NRI), integrated discrimination improvement (IDI), calibration, and clinical usefulness in both cohorts. DR development and progression and retinal microvascular damage. Of 168 metabolites, 118 were identified as candidate metabolomic fingerprints for future DR development. These fingerprints significantly improved the predictability for DR development beyond traditional indicators (C statistic, 0.802 [95% confidence interval (CI), 0.760–0.843] vs. 0.751 [95% CI, 0.706–0.796]; P = 5.56 × 10−4). Glucose, lactate, and citrate were among the fingerprints validated in the GDES cohort. Using these parsimonious and replicable fingerprints yielded similar improvements for predicting DR development (C statistic, 0.807 [95% CI, 0.711–0.903] vs. 0.617 [95% CI, 0.494–0.740]; P = 1.68 × 10−4) and progression (C statistic, 0.797 [95% CI, 0.712–0.882] vs. 0.665 [95% CI, 0.545–0.784]; P = 0.003) in the external GDES cohort. Improvements in NRIs, IDIs, and clinical usefulness also were evident in both cohorts (all P < 0.05). In addition, lactate and citrate were associated with microvascular damage across macular and optic nerve head regions among Chinese GDES (all P < 0.05). Metabolomic profiling may be effective in identifying robust fingerprints for predicting future DR development and progression, providing novel insights into the early and advanced stages of DR pathophysiology. The author(s) have no proprietary or commercial interest in any materials discussed in this article. [ABSTRACT FROM AUTHOR]

Following a comprehensive analysis of the initial three generations of prisoner risk assessment tools, the field has observed a notable prominence in the integration of fourth-generation tools and machine learning techniques. However, limited efforts have been made to address the explainability of data-driven prediction models and their connection with treatment recommendations. Our primary objective was to develop predictive models for assessing the likelihood of recidivism among prisoners released from their index incarceration within 1-year, 2-year, and 5-year timeframes. We aimed to enhance interpretability using SHapley Additive exPlanations (SHAP). We collected data from 20,457 in-prison records from February 10, 2005, to August 25, 2021, sourced from a Southwestern China prison's data management system. Recidivism records were officially determined through data mining from an official website and combined identification data from neighboring prisons. We employed five machine learning algorithms, considering sociodemographic, physical health, psychological assessments, criminological characteristics, crime history, social support, and in-prison behaviors as factors. For interpretability, SHAP was applied to reveal feature contributions. Findings indicated that young prisoners accused of larceny, previous convictions, lower fines, and limited family support faced higher reoffending risk. Conversely, middle-aged and senior prisoners with no prior convictions, lower monthly supermarket expenses, and positive psychological test results had lower reoffending risk. We also explored interactions between significant predictive features, such as prisoner age at incarceration initiation and primary accusation, and the duration of current incarceration and cumulative prior incarcerations. Notably, our models consistently exhibited high performance, as shown by AUC on the test dataset across time windows. Interpretability results provided insights into evolving risk factors over time, valuable for intervention with high-risk individuals. These insights, with additional validation, could offer dynamic prisoner information for stakeholders. Moreover, interpretability results can be seamlessly integrated into prison and court management systems as a valuable risk assessment tool. [ABSTRACT FROM AUTHOR]

Neonatal AKI (acute kidney injury) is an underreported entity in sick neonates associated with significant morbidity and mortality. This was a prospective cohort study, to study the incidence, risk factors, and outcomes of AKI among sick neonates. We included sick inborn neonates admitted at a level III neonatal intensive care unit. Neonates with congenital anomalies and who died within 72 h of life were excluded. AKI was defined and categorized as per KDIGO guidelines. Two hundred and seventy-six neonates were enrolled over 1 year, of which 115 (42%) had AKI. The incidence of AKI was highest n = 27/38 (71%) among extremely preterm (< 28 weeks) infants. On Cox regression analysis, sepsis, invasive ventilation, acidosis, and perinatal asphyxia were significantly associated with AKI with a hazard ratio (95% CI) of 4 (1.21–13.42), 2.3 (1.32–4.03), 1.9 (1.13–3.36), and 1.5 (1.04–2.31), respectively. The risk prediction model, using the 4 predictors mentioned above, had good diagnostic accuracy (area under the curve, 83.6%) with a sensitivity and specificity of 77% and 80%, respectively. Infants with AKI have significantly higher mortality, compared to those who did not have AKI n = 45/115 (39%) vs. n = 5/161 (3%), p < 0.01. Conclusions: Nearly half of sick neonates admitted to NICU have AKI, which is maximum in extremely preterm infants. Sepsis, invasive ventilation, acidosis, and perinatal asphyxia have good diagnostic accuracy in identifying neonates likely to develop AKI. What is Known: • Asphyxia, prematurity, sepsis, shock, hypotension, drugs, congenital heart diseases contribute to neonatal AKI. What is New: • Our simple risk prediction model can be used in sick neonates to identify infants who are at risk for developing AKI. [ABSTRACT FROM AUTHOR]

Machine learning (ML) models for early identification of patients at risk of hospital-acquired urinary tract infection (HA-UTI) may enable timely and targeted preventive and therapeutic strategies. However, clinicians are often challenged in the interpretation of the predictive outcomes provided by the ML models, which often reach different performances. To train ML models for predicting patients at risk of HA-UTI using available data from electronic health records at the time of hospital admission. This study focused on the performance of different ML models and clinical explainability. This retrospective study investigated patient data representing 138,560 hospital admissions in the North Denmark Region from 1st January 2017 to 31st December 2018. Fifty-one health sociodemographic and clinical features were extracted as the full dataset, and χ2 test and expert knowledge were used for feature selection, resulting in two reduced datasets. Seven different ML models were trained and compared between the three datasets. The SHapley Additive exPlanation (SHAP) method was used to support population- and patient-level explainability. The best-performing ML model was the neural network model based on the full dataset, with an area under the curve (AUC) of 0.758. The neural network model was also the best-performing ML model based on the reduced datasets, with an AUC of 0.746. Clinical explainability was demonstrated with a SHAP summary and forceplot. Within 24 h of hospital admission, the ML models were able to identify patients at risk of developing HA-UTI, providing new opportunities to develop efficient strategies for the prevention of HA-UTI. SHAP was used to demonstrate how risk predictions can be explained at individual patient level and for the patient population in general. [ABSTRACT FROM AUTHOR]

Background: The accurate identification of patients with high cardiovascular risk in suspected myocardial infarction (MI) is an unmet clinical need. Therefore, we sought to investigate the prognostic utility of a multi-biomarker panel with 29 different biomarkers in in 748 consecutive patients with symptoms indicative of MI using a machine learning-based approach. Methods: Incident major cardiovascular events (MACE) were documented within 1 year after the index admission. The selection of the best multi-biomarker model was performed using the least absolute shrinkage and selection operator (LASSO). The independent and additive utility of selected biomarkers was compared to a clinical reference model and the Global Registry of Acute Coronary Events (GRACE) Score, respectively. Findings were validated using internal cross-validation. Results: Median age of the study population was 64 years. At 1 year of follow-up, 160 cases of incident MACE were documented. 16 of the investigated 29 biomarkers were significantly associated with 1-year MACE. Three biomarkers including NT-proBNP (HR per SD 1.24), Apolipoprotein A-I (Apo A-I; HR per SD 0.98) and kidney injury molecule-1 (KIM-1; HR per SD 1.06) were identified as independent predictors of 1-year MACE. Although the discriminative ability of the selected multi-biomarker model was rather moderate, the addition of these biomarkers to the clinical reference model and the GRACE score improved model performances markedly (∆C-index 0.047 and 0.04, respectively). Conclusion: NT-proBNP, Apo A-I and KIM-1 emerged as strongest independent predictors of 1-year MACE in patients with suspected MI. Their integration into clinical risk prediction models may improve personalized risk stratification. Prognostic utility of a multi-biomarker approach in suspected myocardial infarction. In a cohort of 748 patients with symptoms indicative of myocardial infarction (MI) to the emergency department, we measured a 29-biomarker panel and performed regressions, machine learning (ML)-based variable selection and discriminative/reclassification analyses. We identified three biomarkers as top predictors for 1-year major adverse cardiovascular events (MACE). Their integration into a clinical risk prediction model and the Global Registry of Acute Coronary Events (GRACE) Score allowed for marked improvement in discrimination and reclassification for 1-year MACE. Apo apolipoprotein; CRP C-reactive protein; CRS clinical risk score; ECG electrocardiogram; EN-RAGE extracellular newly identified receptor for advanced glycation end-products binding protein; FABP fatty acid–binding protein; GS Grace Score; hs-cTnI high-sensitivity cardiac troponin I; KIM-1 kidney injury molecule–1; LASSO least absolute shrinkage and selection operator; MACE major adverse cardiovascular events; MI myocardial infarction; NRI net reclassification improvement; NT-proBNP N-terminal prohormone of brain natriuretic peptide. [ABSTRACT FROM AUTHOR]

Purpose: Head and neck cancer surgery often requires postoperative monitoring in an intensive care unit (ICU) or intermediate care unit (IMC). With a variety of different risk scores, it is incumbent upon the investigator to plan a risk-adapted allocation of resources. Tumor surgery in the head and neck region itself offers a wide range of procedures in terms of resection extent and reconstruction methods, which can be stratified only vaguely by a cross-disciplinary score. Facing a variety of different risk scores we aimed to develop a new Tumor Risk Score (TRS) enabling anterograde preoperative risk evaluation, resource allocation and optimization of cost and outcome measurements in tumor surgery of the head and neck. Methods: A collective of 547 patients (2010–2021) with intraoral tumors was studied to develop the TRS by grading the preoperative tumor size and location as well as the invasiveness of the planned surgery by means of statistical modeling. Two postoperative complications were defined: (1) prolonged postoperative stay in IMC/ICU and (2) prolonged total length of stay (LOS). Each parameter was analyzed using TRS and all preoperative patient parameters (age, sex, preoperative hemoglobin, body-mass-index, preexisting medical conditions) using predictive modeling design. Established risk scores (Charlson Comorbidity Index (CCI), American Society of Anesthesiologists risk classification (ASA), Functional Comorbidity Index (FCI)) and Patient Clinical Complexity Level (PCCL) were used as benchmarks for model performance of the TRS. Results: The TRS is significantly correlated with surgery duration (p < 0.001) and LOS (p = 0.001). With every increase in TRS, LOS rises by 9.3% (95%CI 4.7–13.9; p < 0.001) or 1.9 days (95%CI 1.0-2.8; p < 0.001), respectively. For each increase in TRS, the LOS in IMC/ICU wards increases by 0.33 days (95%CI 0.12–0.54; p = 0.002), and the probability of an overall prolonged IMC/ICU stay increased by 32.3% per TRS class (p < 0.001). Exceeding the planned IMC/ICU LOS, overall LOS increased by 7.7 days (95%CI 5.35–10.08; p < 0.001) and increases the likelihood of also exceeding the upper limit LOS by 70.1% (95%CI 1.02–2.85; p = 0.041). In terms of predictive power of a prolonged IMC/ICU stay, the TRS performs better than previously established risk scores such as ASA or CCI (p = 0.031). Conclusion: The lack of a standardized needs assessment can lead to both under- and overutilization of the IMC/ICU and therefore increased costs and losses in total revenue. Our index helps to stratify the risk of a prolonged IMC/ICU stay preoperatively and to adjust resource allocation in major head and neck tumor surgery. [ABSTRACT FROM AUTHOR]

Purpose of Review: The objective of this narrative review is to summarize data from recently published prospective observational studies that analyze the association between circulating interleukin-6 (IL-6) levels and cardiovascular clinical or imaging endpoints. Recent Findings: Higher levels of IL-6 are associated with a higher risk of cardiovascular death, major adverse cardiovascular events, myocardial infarction, stroke, peripheral artery disease, and heart failure. Imaging studies have also shown an association between IL-6 and carotid intima-media thickness progression, carotid plaque progression, severity, and vulnerability. These observations have been consistent across a wide range of study populations and after adjusting for traditional and emerging risk factors including high-sensitivity C-reactive protein. Summary: Robust epidemiologic evidence supports IL-6 as a central mediator of cardiovascular risk along with human genetic studies and mechanistic experiments. Ongoing clinical studies are testing the therapeutic hypothesis of IL-6 inhibition in patients with atherosclerotic cardiovascular disease or heart failure. [ABSTRACT FROM AUTHOR]

Background: Elevated mammographic density is associated with increased breast cancer risk. However, the contribution of longitudinal changes in mammographic density to breast cancer risk beyond initial mammographic density levels, considering familial breast cancer risk and menopausal status, remains uncertain but holds important clinical implications. Methods: In a nested case–control study within the Sister Study (323 cases, 899 controls; 12,095 mammograms), a cohort enriched for family history of breast cancer, we examined case–control status in relation to the largest annual change in percent density and dense area using mammograms available spanning 5.4 years, on average, using multivariable logistic regression and to the rate of mammographic density change using linear mixed-effects models. We considered effect modification by: mammographic density level of the earlier mammogram, the extent of family history, Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation (BOADICEA) risk strata, and menopausal status. Results: Cases (diagnosed < 60 years) had greater initial percent density and dense area levels and a slower rate of decline in dense area than controls. Women with stable mammographic density (≤ 10% annual change) had an increased breast cancer risk as compared with women whose largest mammographic density change was > 10% annual decline (e.g., Odds Ratio (OR) 2.34, 95% Confidence Interval (CI) 1.63–3.37 for dense area). Increasing vs. decreasing dense area was also associated with elevated risk, especially in women with the highest dense area levels at the earlier mammogram (OR: 2.56, 95%CI 1.50–4.36). Although generally similar across menopausal and familial risk categories, the associations of MD change with risk appeared stronger in pre-menopausal and lower-risk women. Conclusions: Women who maintain higher levels of mammographic density (i.e. do not decrease over time) or have increasing mammographic density over time have a higher risk of subsequent breast cancer than women with high mammographic density that decreases over time. These findings suggest potential for incorporating mammographic density trajectories in clinical risk assessment, and the importance of additional breast cancer monitoring in women not experiencing declines in mammographic density over time. [ABSTRACT FROM AUTHOR]

Purpose: The availability of population datasets and machine learning techniques heralded a new era of sophisticated prediction models involving a large number of routinely collected variables. However, severe class imbalance in clinical datasets is a major challenge. The aim of this study is to investigate the impact of commonly-used resampling techniques in combination with commonly-used machine learning algorithms in a clinical dataset, to determine whether combination(s) of these approaches improve upon the original multivariable logistic regression with no resampling. Methods: We previously developed and internally validated a multivariable logistic regression 30-day mortality prediction model in 30,619 patients using preoperative and intraoperative features. Using the same dataset, we systematically evaluated and compared model performances after application of resampling techniques [random under-sampling, near miss under-sampling, random oversampling, and synthetic minority oversampling (SMOTE)] in combination with machine learning algorithms (logistic regression, elastic net, decision trees, random forest, and extreme gradient boosting). Results: We found that in the setting of severe class imbalance, the impact of resampling techniques on model performance varied by the machine learning algorithm and the evaluation metric. Existing resampling techniques did not meaningfully improve area under receiving operating curve (AUROC). The area under the precision recall curve (AUPRC) was only increased by random under-sampling and SMOTE for decision trees, and oversampling and SMOTE for extreme gradient boosting. Importantly, some combinations of algorithm and resampling technique decreased AUROC and AUPRC compared to no resampling. Conclusion: Existing resampling techniques had a variable impact on models, depending on the algorithms and the evaluation metrics. Future research is needed to improve predictive performances in the setting of severe class imbalance. [ABSTRACT FROM AUTHOR]

Objectives: In adult hypertensive patients, blood pressure variability is considered a risk factor for heart failure. The relationship between pulse pressure variability and the risk of heart failure remains unclear. This study aims to explore the impact of pulse pressure variability (PPV) on heart failure through a secondary analysis of the SPRINT randomized controlled trial. Methods: The data were derived from the SPRINT (Systolic Blood Pressure Intervention Trial) study. The trial recruited participants 50 years or older, with SBP ≥ 130 mm Hg and at least one additional CVD risk factor. We calculated pulse pressure based on the systolic and diastolic blood pressure obtained during follow-up, and used the coefficient of variation to represent pulse pressure variability (PPV) for statistical analysis. We considered the incidence of acute decompensated heart failure as the outcome measure. We employed multivariable Cox regression analysis to examine the relationship between PPV and the risk of heart failure occurrence. Additionally, we used a restricted cubic spline model to analyze the dose–response relationship between PPV and the risk of heart failure occurrence. Results: In this study, a total of 9429 participants were included. During a median follow-up time of 3.87 years, 188 new cases of heart failure were observed. The mean age of the study population was 67.9 ± 9.4 years and 3382 participants (35.5%) were females. The average PPCV was 13.85 ± 5.37%. The results from the multivariable Cox regression analysis indicated that the risk of heart failure increased by 3% for every 1% increase in PPCV (HR = 1.030 [95% CI 1.016–1.044]; P < 0.001). Conclusions: The study found that PPV is an independent risk factor for the occurrence of heart failure. This underscores the importance of maintaining long-term stability in pulse pressure, in preventing the development of heart failure. [ABSTRACT FROM AUTHOR]

Background and purpose Methods Results Conclusions Neuropeptide Y (NPY) has been reported to be involved in the pathophysiology of several cardiovascular disease processes and might contribute to the incidence of stroke, but the prognostic utility of circulating NPY after acute ischaemic stroke is unclear. This study aimed to prospectively examine the association between plasma NPY levels and adverse clinical outcomes after acute ischaemic stroke.Plasma NPY levels were measured in 3250 patients (2066 men and 1184 women) from the China Antihypertensive Trial in Acute Ischaemic Stroke. The primary outcome was the combination of death and major disability (modified Rankin Scale score ≥3) at 12 months after stroke onset, and secondary outcomes included major disability, death and cardiovascular events.During the 12‐month follow‐up, 702 participants (21.6%) experienced major disability or died. After multivariable adjustment, odds ratio (95% confidence interval) for the highest quartile of NPY was 1.56 (1.19–2.04) for the primary outcome, compared to the lowest quartile. Each standard deviation (0.27 ng/mL) higher log‐transformed NPY was associated with an odds ratio (95% confidence interval) of 1.18 (1.07–1.30) for the primary outcome, 1.28 (1.15–1.42) for major disability. The addition of NPY to a conventional risk factors model improved risk prediction of the composite outcome of death and major disability (category‐free net reclassification index 8.82%, p = 0.040; integrated discrimination improvement 0.38%, p = 0.011).Elevated plasma NPY levels in the acute phase of ischaemic stroke were associated with increased risk of poor clinical outcomes after ischaemic stroke, suggesting that plasma NPY may be a potential prognostic biomarker for ischaemic stroke. [ABSTRACT FROM AUTHOR]

This study examined the relationship between the systemic inflammation response index (SIRI) and kidney stone occurrence in adults in the United States. It also evaluated its potential as a predictor of kidney stones. A total of 24,833 adult participants were included in the study using cross-sectional data from the 2007–2018 National Health and Nutrition Examination Survey (NHANES) database. A history of kidney stones was ascertained through the administration of a questionnaire, and SIRI values (calculated based on neutrophil, monocyte, and lymphocyte counts) were determined for each participant. Logistic regression models were employed to examine the relationship between SIRI and kidney stones while accounting for potential confounding variables such as gender, age, race, lifestyle, and history of chronic disease. Subgroup analyses were also conducted. A significant positive correlation was observed between SIRI and kidney stones. In the unadjusted model, elevated SIRI was significantly and positively associated with an increased risk of kidney stones (OR = 1.17). Analysis of SIRI quartiles demonstrated a gradual increase in the risk ratio of kidney stones with increasing SIRI levels, indicating a clear dose-response relationship. In particular, in the model adjusted for multiple confounding variables, the risk of developing kidney stones in the highest SIRI quartile was increased by 20% compared to the lowest SIRI quartile (OR = 1.20, P = 0.007). There is a significant positive correlation between SIRI and kidney stones. SIRI may predict kidney stone risk and highlight the systemic inflammatory state's substantial contribution to kidney stones' pathogenesis. [ABSTRACT FROM AUTHOR]

Introduction: Polygenic Risk Scores (PRS) are an emerging tool for predicting an individual's genetic risk to a complex trait. Several methods have been proposed to construct and calculate these scores. Here, we develop a biologically driven PRS using the UK BioBank cohort through validated protein interactions (PPI) and network construction for psoriasis, incorporating variants mapped to the interacting genes of 14 psoriasis susceptibility (PSORS) loci, as identified from previous genetic linkage studies. Methods: We constructed the PPI network via the implementation of two major meta-databases of protein interactions, and identified variants mapped to the identified PSORS-interacting genes. We selected only European unrelated participants including individuals with psoriasis and randomly selected healthy controls using an at least 1:4 ratio to maximize statistical power. We next compared our PPI-PRS model to (i) clinical risk models and (ii) conventional PRS calculations through p -value thresholding. Results: Our PPI-PRS model provides comparable results to both clinical risk models and conventional approaches, despite the incorporation of a limited number of variants which have not necessarily reached genome-wide significance (GWS). Exclusion of variants mapped to the HLA-C locus, an established risk locus for psoriasis resulted in highly similar associations compared to our primary model, indicating the contribution of the genetic variability mapped to non-GWS variants in PRS computations. Discussion: Our findings support the implementation of biologically driven approaches in PRS calculations in psoriasis, highlighting their potential clinical utility in risk assessment and treatment management. [ABSTRACT FROM AUTHOR]

Simple Summary: This study looked at different ways to predict the risk of colorectal cancer using blood tests. Sixteen studies from 2015 to 2024 were reviewed, focusing on people with an average risk of cancer who were undergoing a colonoscopy for screening. The most common test was a complete blood count, which showed good accuracy. Other tests included tests of proteins, metabolites, enzymes, and markers related to insulin and anemia. The best results came from a test using a group of plasma metabolites. While these tests could make cancer screening more accurate, many studies did not show how these models could be used in real-life settings. More research is needed to understand how these models can be effectively used in current screening programs. Background/Objectives: Risk prediction models (RPMs) for colorectal cancer (CRC) could facilitate risk-based screening. Models incorporating biomarkers may improve the utility of current RPMs. We performed a systematic review of studies reporting RPMs for CRC that evaluated the impact of blood-based biomarkers on clinical outcome prediction at the time of screening colonoscopy in average-risk populations. Methods: We conducted a search of MEDLINE, Web of Science, and PubMed databases from inception through April 2024. Studies that developed or validated a model to predict risk of CRC or its precursors were included. Studies were limited to those including patients undergoing average-risk CRC screening. Results: Sixteen studies published between 2015 and 2024 were included. Outcomes included CRC (16 studies) and high-risk adenomas (1 study). Using a complete blood count was the most common biomarker and was able to achieve an AUC of 0.82 and a specificity of 0.88. Other blood-based biomarkers included were various serum proteins/metabolites/enzymes, plasma metabolites, insulin-related factors, and anemia markers. The highest-performing model, with an AUC of 0.99, involved the use of a plasma metabolite panel. Conclusions: The evidence base of RPMs for CRC screening is expanding and incorporating biomarkers, which remain a prominent aspect of model discovery. Most RPMs included a lack of internal/external validation or discussion as to how the model could be implemented clinically. As biomarkers improve the discriminatory potential of RPMs, more research is needed for the evaluation and implementation of RPMs within existing CRC screening frameworks. [ABSTRACT FROM AUTHOR]

Background: Breast cancer patients often experience cognitive impairment as a complication during treatment, which seriously affects their quality of life. This study aimed to assess the risk factors associated with cognitive impairment in breast cancer patients and to construct and validate a nomogram model to predict cognitive impairment in this population. Methods: In this study, we used a convenience sampling method to select 423 breast cancer patients who attended the Department of Breast Surgery at the First Hospital of Jinzhou Medical University from September 2023 to March 2024. We analyzed these patients' cognitive impairment risk factors through LASSO regression and logistic regression analysis to develop a predictive model. The model was evaluated using the area under the curve (AUC) from the receiver operating characteristic (ROC) curve and the calibration curve and decision curve analysis. Results: This study found a prevalence of cognitive impairment of 19.62% among breast cancer patients. A nomogram model was developed based on six influencing factors: age, educational level, pathological type, treatment program, emotional state, and fatigue. The area under the curve (AUC) for the model's training and validation groups was 0.944 and 0.931, respectively. The model calibration curves showed a high degree of consistency, and the decision curve analysis (DCA) indicated good clinical applicability of the model. Conclusions: This nomogram demonstrates good discrimination, calibration, and clinical applicability, making it a more intuitive predictor of the risk of cognitive impairment in breast cancer patients. [ABSTRACT FROM AUTHOR]

Over 50% of children with a parent with severe mental illness will develop mental illness by early adulthood. However, intergenerational transmission of risk for mental illness in one's children is insufficiently considered in clinical practice, nor is it sufficiently utilised into diagnostics and care for children of ill parents. This leads to delays in diagnosing young offspring and missed opportunities for protective actions and resilience strengthening. Prior twin, family, and adoption studies suggest that the aetiology of mental illness is governed by a complex interplay of genetic and environmental factors, potentially mediated by changes in epigenetic programming and brain development. However, how these factors ultimately materialise into mental disorders remains unclear. Here, we present the FAMILY consortium, an interdisciplinary, multimodal (e.g., (epi)genetics, neuroimaging, environment, behaviour), multilevel (e.g., individual-level, family-level), and multisite study funded by a European Union Horizon-Staying-Healthy-2021 grant. FAMILY focuses on understanding and prediction of intergenerational transmission of mental illness, using genetically informed causal inference, multimodal normative prediction, and animal modelling. Moreover, FAMILY applies methods from social sciences to map social and ethical consequences of risk prediction to prepare clinical practice for future implementation. FAMILY aims to deliver: (i) new discoveries clarifying the aetiology of mental illness and the process of resilience, thereby providing new targets for prevention and intervention studies; (ii) a risk prediction model within a normative modelling framework to predict who is at risk for developing mental illness; and (iii) insight into social and ethical issues related to risk prediction to inform clinical guidelines. [ABSTRACT FROM AUTHOR]

Simple Summary: Cardiovascular diseases are among the most frequent, although rare, long-term sequalae in adolescents and young adult survivors of breast cancer. However, no dedicated tool exists to help clinicians with planning personalized follow-up strategies for these patients. To make up for this lack, in this work, we developed a Bayesian network, an artificial intelligence model, to predict the 5-year risk for cardiovascular diseases in these patients, leveraging real-world data from two different cohorts. The model showed a very good ability to identify patients at risk and select those that should be prioritized because they are at higher risk, making it useful for guiding clinicians in everyday practice. Finally, the methodological approach proposed in this work is particularly interesting for all researchers who aim at developing causally interpretable tools, also dealing with real-world data and their biases. Background: In the last decades, the increasing number of adolescent and young adult (AYA) survivors of breast cancer (BC) has highlighted the cardiotoxic role of cancer therapies, making cardiovascular diseases (CVDs) among the most frequent, although rare, long-term sequalae. Leveraging innovative artificial intelligence (AI) tools and real-world data (RWD), we aimed to develop a causally interpretable model to identify young BC survivors at risk of developing CVDs. Methods: We designed and trained a Bayesian network (BN), an AI model, making use of expert knowledge and data from population-based (1036 patients) and clinical (339 patient) cohorts of female AYA (i.e., aged 18 to 39 years) 1-year survivors of BC, diagnosed in 2009–2019. The performance achieved by the BN model was validated against standard classification metrics, and two clinical applications were proposed. Results: The model showed a very good classification performance and a clear causal semantic. According to the predictions made by the model, focusing on the 25% of AYA BC survivors at higher risk of developing CVDs, we could identify 81% of the patients who would actually develop it. Moreover, a desktop-based app was implemented to calculate the individual patient's risk. Conclusions: In this study, we developed the first causal model for predicting the CVD risk in AYA survivors of BC, also proposing an innovative AI approach that could be useful for all researchers dealing with RWD. The model could be pivotal for clinicians who aim to plan personalized follow-up strategies for AYA BC survivors. [ABSTRACT FROM AUTHOR]

• Skeletal muscle is displayed and easily quantifiable on standard CMR T1-maps. • High native skeletal muscle T1 times are linked to a more unfavorable prognosis. • Native skeletal muscle T1 time may serve as a risk stratification marker in HFpEF. Heart failure with preserved ejection fraction (HFpEF) is associated with heart failure (HF) hospitalizations and death. Previous studies have shown that altered muscle composition is associated with higher risk of adverse outcome in HFpEF patients. The purpose of our study was to investigate the association between skeletal muscle composition, as measured by skeletal muscle T1-times on cardiac magnetic resonance (CMR) imaging, and adverse outcome. We measured skeletal muscle T1-times of the back muscles on standard CMR images in a prospective cohort of HFpEF patients. Cox regression models were used to test the association of skeletal muscle T1-times and adverse outcome defined as hospitalization for HF and/or cardiovascular death. We included 101 patients (mean age 72±7 years, 71 % female) in our study. The median skeletal muscle T1-times were 842 ms (IQR 806–881 ms). In univariate analysis high muscle T1-time was associated with adverse outcome (HR=1.96 [95 % CI, 1.31–2.94] per every 100 ms increase; p =.001). After adjustment for age, sex, body mass index, left- and right ventricular ejection fraction, N-terminal pro-brain natriuretic peptide and myocardial native T1-times, native skeletal muscle T1-time remained an independent predictor for adverse outcome (HR=1.94 [95 % CI, 1.24–3.03] per every 100 ms increase; p =.004). In patients with HFpEF, high skeletal muscle T1-times on standard CMR scans are associated with higher rates of HF hospitalizations and cardiovascular death. Condensed abstract: Skeletal muscle abnormalities are common in patients with heart failure with preserved ejection fraction (HFpEF). The present study evaluates skeletal muscle composition, as quantified by native skeletal muscle T1-times of the back muscles on standard cardiac magnetic resonance imaging, and assessed the association with adverse outcome, defined as hospitalization for heart failure and/or cardiovascular death. In a prospective cohort of 101 patients with HFpEF, we found that high native skeletal muscle T1-times are associated with an increased risk for adverse outcome. These findings suggest that native skeletal muscle T1-time may serve as marker for improved risk prediction. High skeletal muscle T1-time is associated with highest risk of adverse cardiovascular events including heart failure hospitalizations and cardiovascular death 101 patients with HFpEF underwent CMR for measurement of skeletal muscle T1-time of the back muscles (green arrows on panel 3) using T1-mapping to assess muscle tissue. The right panel shows time-to-event curves of the study cohort for the combined endpoint of HF hospitalizations and/or cardiovascular death, stratified by skeletal muscle T1-time median. Patients with high skeletal muscle T1-time (purple line), suggesting that inflammation and fibrosis are present, had a higher risk for adverse outcome compared to patients with a low skeletal muscle T1-time (blue line). CMR, cardiac magnetic resonance; HF, heart failure; HFpEF, heart failure with preserved ejection fraction. [Display omitted] [ABSTRACT FROM AUTHOR]

[Display omitted] • A panel of seven lipids increased c-statistic by an increment of 0.077, the greatest added value for prediction of incident CVD ever achieved by prospective lipidomics studies to the best of our knowledge. • Considerable mediations were observed through SM 41:1b, presumably SM(d18:1/23:0), in the associations between established cardiovascular risk factors and CHD risk, diabetes in particular. • CHD risk associated with increased levels of PC 36:0a, SM 41:1b, LPC 18:0 and LPC 20:3, and decreased levels of CE 18:2 were aggravated by long-term exposure to ambient air pollution. Previous lipidomics studies have identified various lipid predictors for cardiovascular risk, however, with limited predictive increment, sometimes using too many predictor variables at the expense of practical efficiency. To search for lipid predictors of future coronary heart disease (CHD) with stronger predictive power and efficiency to guide primary intervention. We conducted a prospective nested case-control study involving 1,621 incident CHD cases and 1:1 matched controls. Lipid profiling of 161 lipid species for baseline fasting plasma was performed by liquid chromatography-mass spectrometry. In search of CHD predictors, seven lipids were selected by elastic-net regression during over 90% of 1000 cross-validation repetitions, and the derived composite lipid score showed an adjusted odds ratio of 3.75 (95% confidence interval: 3.15, 4.46) per standard deviation increase. Addition of the lipid score into traditional risk model increased c-statistic to 0.736 by an increment of 0.077 (0.063, 0.092). From the seven lipids, we found mediation of CHD risk from baseline diabetes through sphingomyelin (SM) 41:1b with a considerable mediation proportion of 36.97% (P < 0.05). We further found that the positive associations of phosphatidylcholine (PC) 36:0a, SM 41:1b, lysophosphatidylcholine (LPC) 18:0 and LPC 20:3 were more pronounced among participants with higher exposure to fine particulate matter or its certain components, also to ozone for LPC 18:0 and LPC 20:3, while the negative association of cholesteryl ester (CE) 18:2 was attenuated with higher black carbon exposure (P < 0.05). We identified seven lipid species with greatest predictive increment so-far achieved for incident CHD, and also found novel biomarkers for CHD risk stratification among individuals with diabetes or heavy air pollution exposure. [ABSTRACT FROM AUTHOR]

Preoperative risk prediction is an important component of perioperative medicine. Machine learning is a powerful tool that could lead to increasingly complex risk prediction models with improved predictive performance. Careful consideration is required to guide the machine learning approach to ensure appropriate decisions are made with regard to what we are trying to predict, when we are trying to predict it, and what we seek to do with the results. [ABSTRACT FROM AUTHOR]

Triglyceride-glucose (TyG) index, a surrogate measure of insulin resistance, is associated with hypertension mediated organ damage (HMOD) and cardiovascular disease. This study investigated the association between TyG index and major adverse cardiovascular events (MACE) and its interaction with traditional risk factors and HMOD. Healthy subjects recruited from the general population were thoroughly examined and followed for MACE using nation-wide registries. Cox proportional hazard models were used to calculate the association between TyG index and MACE occurrence. Models were adjusted for Systematic Coronary Risk Evaluation (SCORE) risk factors, pulse wave velocity, left ventricular mass index, carotid atherosclerotic plaque status, and microalbuminuria. Continuous net reclassification and Harrell's Concordance index (C-index) were used to assess the added prognostic value of TyG index. During a follow-up period of mean 15.4 ± 4.7 years, MACE were observed in 332 (17%) of 1970 included participants. TyG index was associated with MACE; HR = 1.44 [95%CI:1.30–1.59] per standard deviation. After adjustment for traditional cardiovascular (CV) risk factors, HR was 1.16 [95%CI:1.03–1.31]. The association between TyG index and MACE remained significant after further adjustment for each HMOD component. However, this finding was evident only in subjects aged 41 or 51 years (HR = 1.39; 95%CI:1.15–1.69). Including TyG index in a risk model based on traditional CV risk factors improved C-index with 0.005 (P = 0.042). In this population-based study of healthy middle-aged subjects, TyG index was associated with MACE independently of traditional CV risk factors and HMOD. TyG index may have a potential role in future risk prediction systems. • Triglyceride-glucose index (TyG index) is a surrogate marker of insulin resistance. • TyG index can predict major adverse cardiovascular events independently of both traditional risk factors and different types of HMOD. • TyG index can improve individual cardiovascular risk prediction in apparently healthy individuals. [ABSTRACT FROM AUTHOR]

Aims This study aimed to compare the association between measures of left atrial (LA) structure and function, derived from cardiovascular magnetic resonance (CMR), with cardiovascular death or non-fatal heart failure events in patients with non-ischaemic dilated cardiomyopathy (DCM). Methods and results CMR studies of 580 prospectively recruited patients with DCM in sinus rhythm [median age 54 (interquartile range 44–64) years, 61% men, median left ventricular ejection fraction 42% (30–51%)] were analysed for measures of LA structure [LA maximum volume index (LAVImax) and LA minimum volume index (LAVImin)] and function (LA emptying fraction, LA reservoir strain, LA conduit strain (LACS), and LA booster strain]. Over a median follow-up of 7.4 years, 103 patients (18%) met the primary endpoint. Apart from LACS, each measure of LA structure and function was associated with the primary endpoint after adjusting for other important prognostic variables. The addition of each LA metric to a baseline model containing the same important prognostic covariates improved model discrimination, with LAVImin providing the greatest improvement [ C -statistic improvement: 0.702–0.738; χ 2 test comparing likelihood ratio P < 0.0001; categorical net reclassification index: 0.210 (95% CI 0.023–0.392)]. Patients in the highest tercile of LAVImin had similar event rates to those with persistent atrial fibrillation. Measures of LA strain did not enhance model discrimination above LA volumetric measures. Conclusion Measures of LA structure and function offer important prognostic information in patients with DCM and enhance the prediction of adverse outcomes. LA strain was not incremental to volumetric analysis for risk prediction. [ABSTRACT FROM AUTHOR]

The Enriching New-Onset Diabetes for Pancreatic Cancer (ENDPAC) model relies primarily on fasting glucose values. Health systems have increasingly shifted practice towards use of glycated hemoglobin (HbA1c) measurement. We modified the ENDPAC model using patients with new onset hyperglycemia. Four cohorts of patients 50–84 years of age with HbA1c results ≥6.2–6.5 % in 2011–2018 were identified. A combine cohort was formed. A widened eligibility criterion was applied to form additional four individual cohorts and one combined cohort. The primary outcome was the diagnosis of pancreatic cancer within 3 years after the first elevated HbA1c testing. The performance of the modified ENDPAC model was evaluated by AUC, sensitivity, positive predictive value, cases detected, and total number of patients screened. The individual and combined cohorts consisted of 39,001–79,060 and 69,334–92,818 patients, respectively (mean age 63.5–65.0 years). The three-year PC incidence rates were 0.47%–0.54 %. The AUC measures were in the range of 0.75–0.77 for the individual cohorts and 0.75 for the combined cohorts. When the four individual cohorts were combined, more PC cases can be identified (149 by the combined vs. 113–116 by individual cohorts when risk score was 5+). Performance measures were compromised in nonwhites. Asian and Pacific islanders had lower sensitivity compared to other racial and ethnic groups (29 % vs. 50–60 %) when risk score was 5+. The modified ENDPAC model targets a broader population and thus identifies more high-risk patients for cancer screening. The differential performance needs to be considered when the model is applied to non-white population. [ABSTRACT FROM AUTHOR]

The acceptance of genetically modified (GM) crops is still a controversial topic that presents major obstacles to their general use. Few studies, nevertheless, have emphasized the use of artificial intelligence (AI) in forecasting the dangers of GM crops. This review delves into the emerging field of applying AI to forecast the hazards linked to GM crops and examines how it could increase public acceptance of GM products. AI algorithms, and predictive modeling approaches examine enormous datasets that include genetic, environmental, and agronomic factors. Utilizing AI, researchers may accelerate risk assessment procedures on the safety and effectiveness of GM crops. In addressing public concerns and skepticism, AI-generated risk assessments foster transparency and confidence among consumers, regulators, and stakeholders, thereby might fostering greater acceptance of GM products. Although the lack of available data on genetic modifications or developing crop varieties, the amount of training and validation needed for AI algorithms before they can be trusted, the complexity of AI models, and ethical issues about AI, like data privacy and algorithm bias, may present difficulties for precise AI-driven risk assessment. This review outlines recent developments and future directions in utilizing AI as a promising strategy to enhance the acceptability of GM products. [ABSTRACT FROM AUTHOR]

Several risk prediction models exist to predict atherosclerotic cardiovascular disease in asymptomatic individuals, but systematic reviews have generally found these models to be of limited utility. The coronary artery calcium score (CACS) offers an improvement in risk prediction, yet its role remains contentious. Notably, its negative predictive value has a high ability to rule out clinically relevant atherosclerotic cardiovascular disease. Nonetheless, CACS 0 does not permanently reclassify to a lower cardiovascular risk and periodic reassessment every 5 to 10 years remains necessary. Conversely, elevated CACS (> 100 or > 75th percentile adjusted for age, sex and ethnicity) can reclassify intermediate-risk individuals to a high risk, benefiting from preventive medication. The forthcoming update to the Dutch cardiovascular risk management guideline intends to re-position CACS for cardiovascular risk assessment as such in asymptomatic individuals. Beyond CACS as a single number, several guidelines recommend coronary CT angiography (CCTA), which provides additional information about luminal stenosis and (high-risk) plaque composition, as the first choice of test in symptomatic patients and high-risk patients. Ongoing randomised studies will have to determine the value of atherosclerosis evaluation with CCTA for primary prevention in asymptomatic individuals. [ABSTRACT FROM AUTHOR]

Background: We aimed to develop risk tools for dementia, stroke, myocardial infarction (MI), and diabetes, for adults aged ≥ 65 years using shared risk factors. Methods: Data were obtained from 10 population-based cohorts (N = 41,755) with median follow-up time (years) for dementia, stroke, MI, and diabetes of 6.2, 7.0, 6.8, and 7.4, respectively. Disease-free participants at baseline were included, and 22 risk factors (sociodemographic, medical, lifestyle, laboratory biomarkers) were evaluated. Two risk tools (DemNCD and DemNCD-LR based on Fine and Gray sub-distribution and logistic regression [LR], respectively) were developed and validated. Predictive accuracies of these risk tools were assessed using Harrel's C-statistics and area under the curve (AUC) and 95% confidence interval (CI). Model calibration was conducted using Hosmer–Lemeshow goodness of fit test along calibration plots. Results: Both the DemNCD and DemNCD-LR resulted in similar predictive accuracy for each outcome. The overall AUC (95% CI) for dementia, stroke, MI, and diabetes risk tool were 0·68 (0·65, 0·70), 0·58 (0·54, 0·61), 0·65 (0·61, 0·68), and 0·68 (0·64, 0·72), respectively, for males. For females, these figures were 0·65 (0·63, 0·67), 0·55 (0·52, 0·57), 0·65 (0·62, 0·68), and 0·61 (0·57, 0·65). Conclusions: The DemNCD is the first tool to predict both dementia and multiple cardio-metabolic diseases using comprehensive risk factors and provided similar predictive accuracy to existing risk tools. It has similar predictive accuracy as tools designed for single outcomes in this age-group. DemNCD has the potential to be used in community and clinical settings as it includes self-reported and routinely available clinical measures. [ABSTRACT FROM AUTHOR]

Background: Medical outcomes of interest to clinicians may have multiple categories. Researchers face several options for risk prediction of such outcomes, including dichotomized logistic regression and multinomial logit regression modeling. We aimed to compare these methods and provide guidance needed for practice. Methods: We described dichotomized logistic regression, multinomial continuation-ratio logit regression, which is an alternative to standard multinomial logit regression for ordinal outcomes, and logistic competing risks regression. We then applied these methods to develop prediction models of survival and neurodevelopmental outcomes based on the NICHD Extremely Preterm Birth Outcome Tool model. The statistical and practical advantages and flaws of these methods were examined. Both discrimination and calibration of the estimated logistic models of dichotomized outcomes and continuation-ratio logit model were assessed. Results: The dichotomized logistic models and multinomial continuation-ratio logit model had similar discrimination and calibration in predicting death and survival without neurodevelopmental impairment. But the continuation-ratio logit model had better discrimination and calibration in predicting neurodevelopmental impairment. The sum of predicted probabilities of outcome categories from the dichotomized logistic models could deviate from 100% substantially, ranging from 87.7 to 124.0%, and the dichotomized logistic model of neurodevelopmental impairment greatly overpredicted low risks and underpredicted high risks. Conclusions: Estimating multiple logistic regression models of dichotomized outcomes may result in poorly calibrated predictions for an outcome with multiple ordinal categories. Multinomial continuation-ratio logit regression produces better calibrated predictions, constrains the sum of predicted probabilities to 100%, and has the advantages of simplicity in model interpretation, flexibility to include outcome category-specific predictors and random-effect terms for patient heterogeneity by hospital. It also accounts for mutual dependence among multiple categories and accommodates competing risks. [ABSTRACT FROM AUTHOR]

Objective: The long-term glucose monitoring is essential to the risk assessment of diabetic retinopathy (DR), the aim of this study was to investigate the predictive ability of visit-to-visit fasting blood glucose (FBG) indices on the risk of DR. Methods: This was a community-based, cohort study conducted from 2013 to 2021. DR was diagnosed by digital fundus photography. The FPG indices included FBG, var. Associations of each FBG indices and DR were estimated using multinomial logistic regression models adjusting for confounders, and discrimination was determined by area under the curve (AUC). Predictive utility of different models was compared by changes in AUC, integrated discrimination improvement (IDI), and net reclassification index (NRI). Results: This study analyzed 5054 participants, the mean age was 46.26 ± 11.44 years, and 2620 (51.84%) were women. After adjustment for confounders, the adjusted odds ratios (ORs) with 95% confidence intervals (CIs) for FBG, SD, CV, VIM, ARV, M-FBG, and cumulative FBG load were 1.62 (1.52--1.73), 2.74 (2.38--3.16), 1.78 (1.62--1.95), 1.11 (0.95--1.29), 1.72 (1.56--1.91), 2.15 (1.96--2.36), and 2.57 (2.31--2.85), respectively. The AUC of the model with separate cumulative FBG load and classical risk factors was 0.9135 (95%CI 0.8890--0.9380), and no substantive improvement in discrimination was achieved with the addition of other FBG indices once cumulative FBG load was in the model. Conclusions: Cumulative FBG load is adequate for capturing the glucose-related DR risk, and the predictive utility of cumulative FBG load is not significantly improved by adding or replacing other FBG indices in the assessment of DR risk. [ABSTRACT FROM AUTHOR]

Serum neurofilament light chain (sNFL) is a biomarker for axonal injury. Previous studies have linked sNFL levels to cardiovascular risk factors such as diabetes and hypertension, but its association with cardiovascular diseases (CVD) remains unclear. This study aims to explore the association between sNFL and CVD and evaluates its predictive value. Utilizing NHANES 2013–2014 data, this study included 2,035 participants aged ≥ 20 years with measured sNFL quantified using a Siemens immunoassay. CVD was self-reported and included myocardial infarction, stroke, heart failure, coronary heart disease, or angina. Logistic regression models assessed the association between sNFL levels and CVD. The predictive value of sNFL for CVD was evaluated using area under the curve (AUC) and DeLong test. Participants with higher sNFL levels were typically older, male, non-Hispanic white, smokers, and had lower socioeconomic status, higher CVD, hypertension, and diabetes prevalence. Higher sNFL levels were significantly associated with increased odds of CVD (adjusted OR = 1.41, 95% CI: 1.05–1.88). The association was significant in non-hypertensive individuals (OR = 2.72, 95% CI: 1.61–4.62) but not in hypertensive individuals (OR = 1.13, 95% CI: 0.81–1.56). sNFL addition to traditional risk models improved predictive accuracy, especially in non-hypertensive individuals (AUC from 0.827 to 0.856). sNFL levels are significantly associated with CVD in the general population, with a strong predictive value in non-hypertensive individuals. Future longitudinal studies should validate sNFL's efficacy in various populations and explore the underlying mechanisms of its relationship with hypertension and CVD. [ABSTRACT FROM AUTHOR]

Left atrial thrombus (LAT) impacts humans greatly and can result in ischemia and necrosis in severe cases. Therefore, health workers appeal to the social community to emphasize the importance of preventive treatment for LAT. This paper proposes an ensemble framework for risk prediction of LAT based on undersampling with replacement (EFRP-UR), addressing the problem of data imbalance. Firstly, in the feature selection process, we adopt the method of separately counting the essential features of data subsets. In view of the characteristics of class imbalance in medical data, we apply our improved undersamling method, "undersampling with replacement", to obtain a number of training subsets, train multiple base-classifiers, and use an iterative method to select the classifiers with better performance for subsequent integration, improving the prediction accuracy of the proposed EFRP-UR. To aim for disease risk prediction, we synthesize the results of different ensemble algorithms in the end to increase the recall rate. Applied to the LAT dataset obtained from the Regional Medical Center, our experimental results prove that the proposed EFRP-UR has improved in accuracy, recall rate and F1 value, compared with any single base-classifier. In addition, if comprehensive data on other diseases exist, EFRP-UR can also be transferred to predict other diseases. [ABSTRACT FROM AUTHOR]

Background: We aimed to develop multiple machine learning models to predict the risk of early intracranial aneurysms (IAs) rupture, evaluate and compare the performance of predictive models. Methods: Information related to patients diagnosed with IA by CT angiography and clinicians in Central hospital of Dalian University of Technology from January 2010 to June 2022 was collected, including clinical characteristics, blood indicators and IA morphological parameters. IA with rupture or maximum growth ≥ 0.5 mm within 1 month of first diagnosis was considered unstable. The relevant factors affecting IA stability were screened and predictive models were developed based on the above three levels, including random forest (RF), support vector machine (SVM), and artificial neural network (ANN). Sensitivity, specificity, accuracy and area under curve (AUC) value were used to evaluate the predictive models. Results: A total of 989 IA patients were included in the study, including 561 stable patients and 428 unstable patients. For RF models, the training set showed that sensitivity, specificity, accuracy and the AUC values were 72.8–83.7%, 76.9–86.9%, 75.1–84.1% and 0.748 (0.719–0.778)–0.839 (0.814–0.864), respectively; after test set validation, the results were 71.9–78.8%, 75.0–84.0%, 73.6–81.1% and 0.734 (0.688–0.781)–0.809 (0.768–0.850), respectively. For SVM models, the training set were 66.0–80.2%, 76.5–85.5%, 71.7–82.3%, 0.712 (0.682–0.743)–0.913 (0.884–0.924), respectively; the test set were 44.2–78.3%, 63.4–84.4%, 57.9–80.9%, 0.699 (0.651–0.747)–0.806 (0.765–0.848), respectively. For ANN models, the training set were 66.8–83.0%, 75.3–82.3%, 71.6–82.1%, 0.783 (0.757–0.808)–0.897 (0.879–0.914); the test set were 63.1–76.3%, 65.5–84.0%, 64.4–80.6%, 0.680 (0.593–0.694)–0.860 (0.821–0.899). The results of variable importance showed that age, white blood cell count (WBC) and uric acid (UA) played an important role in predicting the stability of IA. Conclusions: The predictive stability models of IA based on three artificial intelligence methods shows good clinical application. Age, WBC and UA played an important role in predicting the IA stability, and were potentially important predictors. [ABSTRACT FROM AUTHOR]

Background: The early detection of transthyretin cardiac amyloidosis (ATTR-CM) is essential, with Tc-99m pyrophosphate scintigraphy (PYP scan) being a key diagnostic tool. Although a previously validated score has shown promise in predicting PYP scan positivity among patients with HFpEF, further evaluation in diverse cohorts is necessary. Objectives: To assess the effectiveness of the ATTR-CM score in predicting PYP scan positivity within our patient population. Methods: We analyzed patients referred for PYP with SPECT at the Cleveland Clinic from January 2012 to January 2020, all of whom had undergone echocardiography within the previous year. The ATTR-CM score was determined using the following criteria: Age (60–69, +2; 70–79, +3; ≥80, +4), sex (male, +2), hypertension (present, −1), left ventricular ejection fraction (LVEF <60%, +1), posterior wall thickness (≥12 mm, +1), and relative wall thickness (>0.57, +2). A score of ≥6 indicated high risk. Results: Among the 540 patients (32% female, 33% black), 27% had an LVEF <40%. The score demonstrated good discrimination by AUC, with consistent performance across different racial groups, sexes, and LVEF categories. For scores ≥6, sensitivity was lower in women and black patients; however, lowering the cutoff to 5 markedly improved sensitivity. Conclusions: The ATTR-CM score displayed consistently good performance by AUC across our cohort, including patients with HFrEF. Nevertheless, its sensitivity was reduced in black patients and women. Efforts to scale ATTR-CM diagnosis tools should be mindful of demographic differences in risk prediction models. [ABSTRACT FROM AUTHOR]

Correlation between echocardiographic and pathoanatomic variables and their prognostic value in murine cardiomyopathy models remain unknown. Using echocardiography, morphometrics, and survival monitoring, we characterized transgenic (TG) mice with dilated cardiomyopathy due to cardiac overexpression of β2‐adrenoceptors focusing on predicting heart failure (HF) risk and HF mortality. In 12‐month‐old non‐TG and TG mice, echocardiography was performed to determine left ventricular (LV) dimensions (d), wall thickness (h), and fractional shortening (FS). Animals were monitored for 3 months for survival. Organ weights and pathological events indicating left HF were determined. TG mice (n = 76) had reduced FS and enlarged LV, and 79% died of HF or likely arrhythmias during the follow‐up period while all non‐TG mice (n = 26) survived. These mice with left HF also had pulmonary congestion and hypertrophy/dilatation of the right ventricle (RV). Weights of lungs, RV, and atria were intercorrelated (r = 0.79–0.83) and also negatively correlated with FS × (h/d) index (r = −0.502 to −0.609). By FS × (h/d) tertiles, TG mice of low tertiles were identified with the highest mortality (96%) largely due to HF (76%). In conclusion, in aged cardiomyopathy mice a good correlation existed between echocardiographic and pathoanatomic variables. Echocardiography‐derived LV function and remodeling were useful in identifying a subgroup of TG mice with a high risk of HF and HF fatality. [ABSTRACT FROM AUTHOR]

Metabolic syndrome increases the risks of cardiovascular diseases, type 2 diabetes, and certain cancers. The early detection of metabolic syndrome is clinically relevant, as it enables timely and targeted interventions. In the current study, we aimed to investigate the association between metabolic syndrome and heart rate measured using wearable devices in a real-world setting and compare this association with that for clinical resting heart rate. Data from 564 middle-aged adults who wore wearable devices for at least 7 days with a minimum daily wear time of 20 h were analyzed. The results showed significantly elevated all-day, sleeping, minimum, and inactive heart rates in men with pre-metabolic or metabolic syndrome compared with those in normal individuals, whereas sleeping heart rate and heart rate dips were significantly increased and decreased, respectively, in women with metabolic syndrome. After adjusting for confounders, every 10-beats-per-minute increment in all-day, sleeping, minimum, and inactive heart rates in men corresponded to odds ratios of 2.80 (95% confidence interval 1.53–5.44), 3.06 (1.57–6.40), 4.21 (1.87–10.47), and 3.09 (1.64–6.29), respectively, for the presence of pre-metabolic or metabolic syndrome. In women, the association was significant only for heart rate dips (odds ratio = 0.49 [95% confidence interval 0.25–0.96] for every 10% increment). Models incorporating inactive or minimum heart rate in men and heart rate dip in women demonstrated better fits, as indicated by lower Akaike information criterion values (170.3 in men and 364.9 in women), compared with models that included clinical resting heart rate (173.4 in men and 369.1 in women). These findings suggest that the heart rate indices obtained from wearable devices may facilitate early identification of metabolic syndrome. [ABSTRACT FROM AUTHOR]

Background: Colorectal cancer (CRC) is the third most frequently diagnosed cancer worldwide. Diabetes and CRC share many overlapping lifestyle risk factors such as obesity, heavy alcohol use, and diet. This study aims to develop a risk scoring system for CRC prediction among diabetes patients using routine medical records. Methods: A retrospective cohort study was conducted using electronic health records of Hong Kong. Patients who received diabetes care in public general outpatient clinics between 2010 and 2019 and had no cancer history were identified, and followed up until December 2019. The outcome was diagnosis of CRC during follow-up. For model building, predictors were first selected using random survival forest, and weights were subsequently assigned to selected predictors using Cox regression. Results: Of the 386,325 patients identified, 4,199 patients developed CRC during a median follow-up of 6.2 years. The overall incidence rate of CRC was 1.93 per 1000 person-years. In the final scoring system, age, waist-to-hip ratio, and serum creatinine were included as predictors. The C-index on test set was 0.651 (95%CI: 0.631-0.669). Elevated serum creatinine (=127 µmol/L) could be a potential important predictor of increased CRC risk. Conclusion: While obesity is a well-known risk factor for CRC, renal dysfunction could be potentially linked to an elevated risk of CRC among diabetes patients. Further studies are warranted to explore whether renal function could be a potential parameter to guide screening recommendation for diabetes patients. [ABSTRACT FROM AUTHOR]

Background: Traditional cardiovascular disease risk prediction models generate a combined risk assessment for myocardial infarction (MI) and ischemic stroke (IS), which may inadequately reflect the distinct etiologies and disparate risk factors of MI and IS. We aim to develop prediction models that separately estimate the risks of MI and IS. Methods: Our analysis included 6,242,404 individuals over 40 years old who participated in a cardiovascular health screening examination in 2009. Potential predictors were selected based on a literature review and the available data. Cox proportional hazards models were used to construct 5-year risk prediction models for MI, and IS. Model performance was assessed through discrimination and calibration. Results: During a follow-up of 39,322,434.39 person-years, 89,140 individuals were diagnosed with MI and 116,259 with IS. Both models included age, sex, body mass index, smoking, alcohol consumption, physical activity, diabetes, hypertension, dyslipidemia, chronic kidney disease, and family history. Statin use was factored into the classification of dyslipidemia. The c-indices for the prediction models were 0.709 (0.707–0.712) for MI, and 0.770 (0.768–0.772) for IS. Age and hypertension exhibited a more pronounced effect on IS risk prediction than MI, whereas smoking, body mass index, dyslipidemia, and chronic kidney disease showed the opposite effect. The models calibrated well for low-risk individuals. Conclusions: Our findings underscore the necessity of tailored risk assessments for MI and IS to facilitate the early detection and accurate identification of heterogeneous at-risk populations for atherosclerotic cardiovascular disease. [ABSTRACT FROM AUTHOR]

Background: Ensemble tree-based models such as Xgboost are highly prognostic in cardiovascular medicine, as measured by the Clinical Effectiveness Metric (CEM). However, their ability to handle correlated data, such as hospital-level effects, is limited. Objectives: The aim of this work is to develop a binary-outcome mixed-effects Xgboost (BME) model that integrates random effects at the hospital level. To ascertain how well the model handles correlated data in cardiovascular outcomes, we aim to assess its performance and compare it to fixed-effects Xgboost and traditional logistic regression models. Methods: A total of 227,087 patients over 17 years of age, undergoing cardiac surgery from 42 UK hospitals between 1 January 2012 and 31 March 2019, were included. The dataset was split into two cohorts: training/validation (n = 157,196; 2012–2016) and holdout (n = 69,891; 2017–2019). The outcome variable was 30-day mortality with hospitals considered as the clustering variable. The logistic regression, mixed-effects logistic regression, Xgboost and binary-outcome mixed-effects Xgboost (BME) were fitted to both standardized and unstandardized datasets across a range of sample sizes and the estimated prediction power metrics were compared to identify the best approach. Results: The exploratory study found high variability in hospital-related mortality across datasets, which supported the adoption of the mixed-effects models. Unstandardized Xgboost BME demonstrated marked improvements in prediction power over the Xgboost model at small sample size ranges, but performance differences decreased as dataset sizes increased. Generalized linear models (glms) and generalized linear mixed-effects models (glmers) followed similar results, with the Xgboost models also excelling at greater sample sizes. Conclusions: These findings suggest that integrating mixed effects into machine learning models can enhance their performance on datasets where the sample size is small. [ABSTRACT FROM AUTHOR]

Dementia is a leading cause of death and disability with over 60% of cases residing in low- and middle-income countries (LMICs). Therefore, new strategies to mitigate risk are urgently needed. However, despite the high burden of disease associated with dementia in LMICs, research into dementia risk profiling and risk prediction modelling is limited. Further, dementia risk prediction models developed in high income countries generally do not transport well to LMICs suggesting that context-specific models are instead needed. New prediction models have been developed, in China and Mexico only, with varying predictive accuracy. However, none has been externally validated or incorporated variables that may be important for predicting dementia risk in LMIC settings such as socio-economic status, literacy, healthcare access, nutrition, stress, pollutants, and occupational hazards. Since there is not yet any curative treatment for dementia, developing a context-specific dementia prediction model is urgently needed for planning early interventions for vulnerable groups, particularly for resource constrained LMIC settings. [ABSTRACT FROM AUTHOR]

Background: Secondary prevention interventions to reduce post-stroke cognitive impairment (PSCI) can be aided by the early identification of high-risk individuals who would benefit from risk factor modification. Aims: To develop and evaluate a predictive model to identify patients at increased risk of PSCI over 5 years using data easily accessible from electronic health records. Methods: Cohort study that included primary care patients from two academic medical centers. Patients were aged 45 years or older, without prior stroke or prevalent cognitive impairment, with primary care visits and an incident ischemic stroke between 2003 and 2016 (development/internal validation cohort) or 2010 and 2022 (external validation cohort). Predictors of PSCI were ascertained from the electronic health record. The outcome was incident dementia/cognitive impairment within 5 years and beginning 3 months following stroke, ascertained using International Classification of Diseases, Ninth/Tenth Revision (ICD-9/10) codes. For model variable selection, we considered potential predictors of PSCI and constructed 400 bootstrap samples with two-thirds of the model derivation sample. We ran 10-fold cross-validated Cox proportional hazards models using a least absolute shrinkage and selection operator (LASSO) penalty. Variables selected in >25% of samples were included. Results: The analysis included 332 incident diagnoses of PSCI in the development cohort (n = 3741), and 161 and 128 incident diagnoses in the internal (n = 1925) and external (n = 2237) validation cohorts, respectively. The C-statistic for predicting PSCI was 0.731 (95% confidence interval (CI): 0.694–0.768) in the internal validation cohort, and 0.724 (95% CI: 0.681–0.766) in the external validation cohort. A risk score based on the beta coefficients of predictors from the development cohort stratified patients into low (0–7 points), intermediate (8–11 points), and high (12–23 points) risk groups. The hazard ratios (HRs) for incident PSCI were significantly different by risk categories in internal (high, HR: 6.2, 95% CI: 4.1–9.3; Intermediate, HR: 2.7, 95% CI: 1.8–4.1) and external (high, HR: 6.1, 95% CI: 3.9–9.6; Intermediate, HR: 2.8, 95% CI: 1.9–4.3) validation cohorts. Conclusion: Five-year risk of PSCI can be accurately predicted using routinely collected data. Model output can be used to risk stratify and identify individuals at increased risk for PSCI for preventive efforts. Data access statement: Mass General Brigham data contain protected health information and cannot be shared publicly. The data processing scripts used to perform analyses will be made available to interested researchers upon reasonable request to the corresponding author. [ABSTRACT FROM AUTHOR]

Many individuals with serious mental illness are at high risk for hospitalization or death due to inadequate treatment of medical conditions or unhealthy behaviors. The authors describe demographic and clinical characteristics associated with increased risk in this population. Electronic data were obtained for individuals in treatment at a large Veterans' healthcare system who were at high risk according to a validated model. A random sample of these individuals was assessed in person. Multivariable regressions estimated the effect of numerous demographic, health, and clinical characteristics on risk. Emergency visits and hospitalizations were common. Greater risk was associated with being male, not married, and having more diagnoses. While risk varied by race, this effect was no longer significant after controlling for other factors. Health-related quality of life worsened with increasing risk. Routine data identify a large population of high-risk individuals who may benefit from outreach to provide healthcare services. [ABSTRACT FROM AUTHOR]