ABSTRACT Context: Bone stress injuries (BSI) have been recognized as one of the most common and potentially serious overuse injuries in military training and result in negative impacts on service member health and force readiness. Several studies have purported to develop a prediction model that could successfully identify individuals in military training at high risk for BSI, but none are currently acceptable for implementation for one or more reasons. Objective: To develop an accurate, parsimonious prediction model for BSI risk in a military training population using easily obtained and interpreted predictor variables. Design: Prospective cohort study. Setting: US Military Academy at XXX. Participants: 3,227 (749 females, 23.2%) incoming cadets. Main Outcome Measures: A multivariable prediction model for BSI risk during the first year of cadet training was created using potential predictor variables related to demographics, anthropometrics, exercise and injury histories, and lower extremity movement quality. A scree plot of change in model log likelihood value was used to guide selection of variables in the final model. Performance of this model was assessed for calibration (i.e. goodness-of-fit) and discrimination (i.e. prognostic accuracy). Minimum acceptable criteria for each were determined a priori . Results: A total of 63 BSI occurred in the study period. The final model consisted of sex and running frequency prior to entry. Performance of this model was sufficient on some measures analyses revealed that an expanded model consisting of all predictor variables also did not reach minimum acceptable calibration or discrimination criteria. Conclusions: Despite use of a large dataset and several predictor variables with well-established associations with BSI risk, we were unable to develop a prediction model for BSI risk with adequate prognostic accuracy properties using a set of easily obtained predictor variables.

Prediction models for maltreatment risk: TRIPOD/PROBAST compliance, calibration, and fairness-A systematic review.

Research on multivariate time series prediction method for upper motion intention perception

A novel structured discrete grey Gompertz multivariable prediction model and its application

Assessing adherence to TRIPOD+AI guidelines in machine learning models for predicting small for gestational age and fetal growth restriction: a systematic review.

Multiparametric integration of cardiac markers in differentiating myocardial infarction with non-obstructive coronary arteries: LASSO regression.

MAformer: A multivariate prediction framework with adaptive multi-scale decomposition and phase correction for water quality in aquaculture environments

Multivariate time series analysis and prediction are of great significance in traffic management, weather forecasting and other practical applications. However, most of the existing research focuses on using the traditional transformer model as the framework to predict short series or predict with time domain features, and the effect of removing noise interference with irregular frequency is not good. At the same time, because the model based on the transformer framework needs to adjust too many hyperparameters, improper parameter design in practical use will lead to model performance degradation, so we propose the WOA-WTConv-KANformer model. The model is optimized based on the itransformer time series prediction model. Before embedding the time nodes of each series into the variable token and input into the encoder layer, the WTConv2d model is used to process the data with wavelet frequency to extract the frequency domain and time domain features. So that the model can solve the non-stationarity problem of time series data caused by the frequency domain problem. In order to realize the effective training of the model, we also use the whale optimization algorithm to make the model reasonably adjust the hyperparameters before formal training. At the same time, the KAN module is used as the linear layer instead of MLP during the training and use of the model, so that the model can improve the performance of different prediction lengths. The number of training parameters is also reduced. Through five public prediction datasets, it is shown that our model can achieve performance improvement on different prediction lengths, and the training efficiency is also improved, which proves the potential of the model in the field of real-world time series prediction.

Earlier heart failure (HF) diagnosis in the community could allow timely treatment initiation and prevent unnecessary hospitalisation, but identifying those at risk remains challenging. We aimed to summarise the performance of risk prediction models for a new diagnosis of HF. Systematic review of multivariable incident HF risk prediction models in the community setting. MEDLINE and Embase were searched from inception to 9 November 2023. Observational, community-based studies reporting prediction model performance for incident HF within a 5-year time horizon. Two reviewers independently screened and extracted data. Where possible, C-statistics (or area under the receiver operating characteristic curve) with 95% CIs were extracted. Risk of bias was assessed using the Prediction model Risk Of Bias ASsessment Tool and certainty of evidence by the Grading of Recommendations, Assessment, Development and Evaluation. Eighteen studies described 45 prediction models, 27 used traditional statistical methods and 18 applied machine learning. Most (39/45) demonstrated acceptable discrimination (C-statistic >0.70). Overall, C-statistics ranged from 0.675 to 0.954, typically with narrow 95% CIs. External validation was performed for 31 models, but only two-the modified PCP-HF models for white men and women-were validated in three cohorts, the highest among all the models. Exploratory random-effects meta-analysis of these models showed pooled C-statistics of 0.82 (95% CI 0.82 to 0.82) for men and 0.85 (95% CI 0.82 to 0.88) for women, indicating excellent discrimination but more heterogenous performance among women. Model performance was at high risk of bias due to unreported or inappropriate handling of missing data, and the certainty of evidence was very low. Risk prediction models for a new diagnosis of HF in the community performed well, but were at high risk of bias and lacked external validation. Future model development requires appropriate data sources, robust handling of missing data, external validation and clinical testing to assess their impact on earlier HF diagnosis and outcomes. CRD42022347120.

Genomic selection (GS) is a promising strategy for accelerating genetic gains of complex traits in breeding programs. Despite the recent advancements in high-throughput genotyping technologies, the selection of the type of marker systems needed for GS remains challenging in breeding programs. In this study, we explored 3K array single nucleotide polymorphisms (SNPs) and genotyping by sequencing (GBS) SNP markers for genomic prediction of oat biomass yield using different statistical and machine learning approaches. An oat panel consisting of 420 lines was phenotyped for biomass-related traits for 3 years and genotyped using two different marker platforms (3K array and GBS). Our results showed similar performance of both the 3K array and GBS-based SNPs in terms of training population optimization, forward prediction, and univariate and multivariate genomic prediction of forage yield. The genomic best linear unbiased prediction (GBLUP), Bayes-B, and random forest models gave similar predictive ability for dry matter yield (DMY) in different harvest-year combinations and for both marker platforms. The multivariate models involving various combinations of secondary traits (simple breeders' field notes and data) resulted in more than twofold increases in predictive abilities compared to the univariate models. Comparison of the 25% top-performing observed and predicted genotypes showed a higher overlap percentage (30.10%-66.99%) for multivariate GBLUP models compared to the univariate models (27.18%-51.46%). This further elucidates the great potential of multivariate GS models incorporating the more robust and easily reproducible 3K array SNP markers for improving the genetic gains of DMY in breeding programs.

This study employed a combined experimental and theoretical approach to investigate the influence of foreign object damage (FOD) on the fatigue limit of surface-strengthened EA4T axles. FOD was introduced on surface-strengthened axle specimens to generate surface defects, and finite element analysis was subsequently performed to evaluate the stress fields in the damaged regions. Fatigue tests were conducted on prefabricated defective specimens to characterize their fatigue behavior. Based on test results, an improved backward statistical inference method was used to fit the fatigue P-S-N curves for each specimen group and derive corresponding fatigue limits. The fatigue limit of full-size damaged axles was estimated by extrapolating from small-scale test results, with due consideration of geometric scale effects on mechanical performance. Considering the stochastic distribution of impact defect depths, an exponential fitting was performed to establish the quantitative relationship between defect depth and full-size axle fatigue limit. Finally, a multivariate fatigue limit prediction model was developed for surface-strengthened full-size EA4T axles based on the El-Haddad formula framework. This model enables comprehensive assessment of fatigue performance under multi-parameter coupling conditions, providing a robust theoretical basis for safety evaluation and maintenance strategies of high-speed train axles subjected to foreign object impacts.

With implementation of newborn screening (NBS) for Duchenne muscular dystrophy (DMD), creatine kinase-muscle (CK) values will be reported on newborns. Maternal, labor, delivery, and newborn factors may elevate CK levels, raising concern for DMD. Predictive modeling could aid hyperCKemia interpretation while awaiting diagnostic confirmation. In this single-center, prospective cohort study, parents of 8365 newborns were offered DMD-NBS. Electronic health records provided data on candidate predictors of hyperCKemia defined by values > 97th or 99th percentiles, or 2000 ng/mL in babies with normal DMD sequences. Exposures included maternal, newborn and perinatal factors. Associations between predictors and hyperCKemia were evaluated using univariate logistic regression. A multivariable prediction model for the 97th percentile was derived using backward stepwise logistic regression and externally validated in a cohort of 2672 newborns. HyperCKemia > 97th percentile was the main outcome. Univariate analyses revealed associations between hyperCKemia and maternal ethnicity, primiparity, labor and delivery complications, oxytocin induction or augmentation, duration of ruptured membranes, forceps or vacuum-assisted delivery, neonatal resuscitation, sex, gestational age, birth weight, and Apgar scores. Lower odds of hyperCKemia were associated with later hour-of-life at sample collection and birth by cesarean section. The final model included parity, shoulder dystocia, forceps or vacuum-assisted delivery, gestational age, neonatal resuscitation, Apgar score (1 min), and time of sample collection. Elevated CK levels may be used for DMD-NBS. Multiple perinatal factors are associated with transient non-DMD hyperCKemia. Our model considers the potential combined impact of such factors and generates a non-disease likelihood for preliminary hyperCKemia interpretation.

Background/Objectives: To construct and compare multivariable prediction models for the early prediction of large-for-gestational-age (LGA) neonates, using ultrasound biometry and maternal characteristics. Methods: This retrospective cohort study analyzed data from singleton pregnancies that underwent routine ultrasound examinations at 30+0–34+0 weeks of gestation. Ultrasound parameters included fetal abdominal circumference (AC), head circumference (HC), femur length (FL), HC-to-AC ratio, mean uterine artery pulsatility index (mUtA-PI), and presence of polyhydramnios. LGA neonates were defined as those having a birthweight > 90th percentile. Logistic regression was used to evaluate associations between ultrasound markers and LGA after adjusting for the following maternal and pregnancy-related covariates: maternal age, body mass index, parity, gestational diabetes mellitus (GDM), pre-existing diabetes, previous cesarean section (PCS), assisted reproductive technology (ART) use, smoking, hypothyroidism, and chronic hypertension. Associations were expressed as adjusted odds ratios (aORs) with 95% confidence intervals (CIs). Three prognostic models were developed utilizing the following predictors: (i) biometric ultrasound measurements including AC, HC-to-AC ratio, FL, UtA-PI, and polyhydramnios (Model 1), (ii) a combination of biometric ultrasound measurements and clinical–maternal data (Model 2), and (iii) only the estimated fetal weight (EFW) (Model 3). Results: In total, 3808 singleton pregnancies were included in the analyses. The multivariable analysis revealed that AC (aOR 1.07, 95% CI [1.06, 1.08]), HC to AC (aOR 1.01, 95% CI [1.006, 1.01]), FL (aOR 1.01, 95% CI [1.009, 1.01]), and the presence of polyhydramnios (aOR 4.97, 95% CI [0.7, 58.8]) were associated with an increased risk of LGA, while a higher mUtA-PI was associated with a reduced risk (aOR 0.98, 95% CI [0.98, 0.99]). Maternal parameters, such as GDM, pre-existing diabetes, elevated pre-pregnancy BMI, absence of uterine artery notching, mUtA-PI, and multiparity, were significantly higher in the LGA group. Both models 1 and 2 showed similar performance (AUCs: 84.7% and 85.3%, respectively) and outperformed model 3 (AUC: 77.5%). Bootstrap and temporal validation indicated minimal overfitting and stable model performance, while decision curve analysis supported potential clinical utility. Conclusions: Models using biometric and Doppler ultrasound at 30–34 weeks demonstrated good discriminative ability for predicting LGA neonates, with an AUC up to 84.7%. Adding maternal characteristics did not significantly improve performance, while the biometric model performed better than EFW alone. Sensitivity at conventional thresholds was low but increased substantially when lower probability cut-offs were applied, illustrating the model’s threshold-dependent flexibility for early risk stratification in different clinical screening needs. Although decision curve analysis was performed to explore potential clinical utility, external validation and prospective assessment in clinical settings are still needed to confirm generalizability and to determine optimal decision thresholds for clinical application.

IntroductionBioelectrical impedance analysis (BIA) is a common technique for assessing body composition in clinical and epidemiological settings. However, its accuracy is limited compared to reference methods such as dual-energy X-ray absorptiometry (DXA).PurposeThis study aimed to evaluate the agreement between fat-free mass (FFM) and fat mass (FM) measured using BIA (Tanita BC-418) and DXA and to develop a calibration model to correct BIA estimates in a heterogeneous sample of Brazilian adults and older adults.MethodsWe analyzed data from 945 participants (aged≥18 years; 611 female participants) who underwent both BIA and DXA assessments across multiple cross-sectional research projects. Agreement between the BIA and DXA measures of FFM (BIAFFM and DXAFFM) and fat mass (FM) was assessed using Pearson correlation coefficients (r) to evaluate precision and Lin’s concordance correlation coefficients (CCCs) to evaluate accuracy. Mean absolute and relative differences were evaluated using paired t-tests or analysis of variance (ANOVA) by sex, age, and nutritional status based on body mass index (BMI). Linear regression was employed to calibrate BIAFFM against DXAFFM. A multivariate prediction model for DXAFFM was developed using BIA-derived resistance, stature, body mass (BM), and age in a randomly selected subsample comprising 70% of the participants (n = 659) and was validated in the remaining 30% (n = 286).ResultsBIA and DXA measures were highly correlated for both FFM and FM (r = 0.97) and demonstrated moderate to high accuracy (CCC ≥ 0.93). For the entire sample, BIA overestimated FFM by 3.1 kg (SD = 2.4; +7.2%) and underestimated FM by 2.9 kg (2.3; −13.0%) compared to DXA (both p < 0.0001). The resulting calibration equation for FFM was DXAFFM = 0.94420 × BIAFFM–(0.01128 x Age) + 0.20516. The multivariate prediction equation derived from the development group was as follows: FFM (kg) = (Sex × 4.1797) + [Stature (cm) × 0.1062] + [Resistance Index (cm2/Ω) × 0.5289] + [Body Mass (kg) × 0.1797] – [Age (yrs) × 0.0705] – 5.4286 (female participants = 0, male participants = 1). In the validation group, the mean FFM values obtained by the calibrated regression and by the new multivariate equation showed no statistically significant difference from the actual DXAFFM measurement.ConclusionSignificant discrepancies existed between BIA- and DXA-derived body composition measures in this heterogeneous sample of Brazilian individuals. The developed prediction equations effectively calibrated BIAFFM estimates to align with DXA values, providing a practical method to enhance the accuracy of BIA for body composition assessment in this population.

Artificial intelligence (AI) applied to screening mammography may non-invasively predict breast cancer molecular subtype and receptor status. We conducted a PRISMA-DTA systematic review and bivariate random-effects meta-analysis (PROSPERO CRD420251032810) on this subject. Methods: We conducted a thorough search in MEDLINE, Embase, Scopus, Web of Science, and IEEE Xplore up to May 2025. Eligible studies compared mammogram AI predictions with histopathologic findings. Risk of bias was assessed with the PROBAST tool, and quality assessment was done using transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD). Twenty-five studies met the inclusion criteria. On internal test sets, pooled AUC/sensitivity/specificity were 0.86/84%/80% for luminal subtype, 0.80/70%/78% for HER2-enriched tumors, and 0.76/75%/83% for triple-negative breast cancer. Multi-class receptor-status tasks yielded AUCs: estrogen receptor 0.71, progesterone receptor 0.59, HER2 0.64, and Ki-67 0.60. Binary receptor-status tasks provided AUCs: HER2 0.80 and hormone receptor positive 0.71. Heterogeneity was substantial (I2 often > 75%). AI from mammograms shows moderate-to-high discrimination, strongest for luminal and triple-negative disease, but evidence is insufficient for clinical deployment. Priorities include larger multicenter cohorts, standardized pipelines, preregistered external validation, uncertainty quantification, and multimodal fusion.

BackgroundThe incidence of respiratory infections in children has been increasing in recent years, and co-infections can lead to additional complications. This study aimed to investigate predictors of Mycoplasma pneumoniae pneumonia (MPP) co-infected with influenza virus through a retrospective analysis of clinical data in pediatric patients.MethodsWe retrospectively reviewed the medical records of 195 children diagnosed with MPP at the Pediatric Internal Medicine Department of Gansu Provincial Hospital between November 2023 and November 2024. Patients were categorized into two groups: single-infection (n = 128, MPP alone) and mixed-infection (n = 67, MPP co-infected with influenza). Predictors of mixed infection were identified using a multivariate logistic regression-based prediction model. The model’s discrimination, accuracy, clinical utility, and generalizability were evaluated using receiver operating characteristic (ROC) curve analysis and decision curve analysis (DCA).ResultsMultivariate analysis showed that influenza season, fibrinogen (Fib) level, fever duration, and C-reactive protein (CRP) were significantly associated with MPP co-infection (p < 0.05). The prediction model demonstrated good discrimination, with an area under the curve (AUC) of 0.820 (95% CI: 0.760–0.879) for the ROC analysis. DCA confirmed the model’s strong clinical utility.ConclusionsA prediction model based on influenza season, Fib level, fever duration, and CRP provide accurate identification of children at risk for MPP co-infected with influenza, demonstrating strong discrimination and clinical applicability.Clinical trialNot applicable.

A Multi-Task End-to-End Multivariate Long-Sequence Time Series Prediction Model for Load Forecasting

Forecasting total electricity consumption using a novel sinusoidal driving seasonal multivariable grey prediction model

Prompt detection and intervention are crucial for improving outcomes in acute invasive fungal rhinosinusitis (AIFS). Diagnostic prediction models assist in risk-stratification, but their accuracy requires testing through external validation. This study aims to validate a previously published diagnostic prediction model for AIFS in an independent cohort. A retrospective chart review was conducted at a tertiary care center (2008-2023) to identify patients with an otolaryngology consult for suspected AIFS. Of 65 patients identified, 11 (16.9%) were diagnosed with AIFS based on histopathology. Risk was calculated using Yin etal.'s predictive model. Predictive performance was assessed by calibration and discrimination. Patients had significantly higher rates of diabetes (46.2%vs. 26.1%, p = 0.002), long-term steroid use (60%vs. 28.2%, p < 0.0001), and solid organ transplantation (38.5%vs. 8.5%, p < 0.001), compared with the development cohort, with conversely lower rates of hematologic malignancy (29.2%vs. 58.7, p < 0.001) and neutropenia (19.4%vs. 41%, p = 0.001). Despite these differences, both the three-variable (C-index: 0.844; 95% CI, 0.736-0.952) and four-variable models (C-index: 0.963; 95% CI, 0.919-1) showed adequate discrimination. Both models exhibited slight overprediction of risk, with a calibration-in-the-large predicted risk of 24.1% (95% CI, 13.68-34.46) for the three-variable model and 24.2% (95% CI, 13.76-34.57) for the four-variable model. Calibration plots confirmed overprediction. The AIFS diagnostic model demonstrates acceptable discrimination and calibration on external validation, with generalizability to patients with different comorbidities. Larger studies are recommended to further test the model's predictive performance and clinical applicability.

Multivariable Analysis-Based Risk Prediction Model for Intracranial Hematoma Expansion in Traumatic Brain Injury Patients.