UNASSIGNED: Recently, a method was developed to predict the motor Functional Independence Measure (FIM) score at discharge in patients with stroke by stratifying the effects of factors such as age and cognitive function and multiplying those by the influence coefficients of these factors. However, an evaluation of the predictive performance of the method is required for clinical application. The present study aimed to evaluate the predictive performance of this prediction method.
UNASSIGNED: Patients with stroke discharged from a rehabilitation ward between April 2021 and September 2022 were included. Predicted values of the motor FIM score at discharge were calculated after data collection from the hospital\'s patient database. The concordance between predicted and actual values was evaluated using the interclass correlation coefficient; moreover, the residual values were calculated.
UNASSIGNED: In total, 207 patients were included in the analysis. The median age was 79 (69-85) years, and 112 (54.1%) patients were male. The interclass correlation coefficient between predicted and actual values was 0.84 (95% confidence interval 0.75-0.89) for the motor FIM score at discharge. Meanwhile, the median residual value was 5.3 (-2.0-10.3) for the motor FIM score at discharge.
UNASSIGNED: The prediction method was validated with good performance. However, the residual values indicated that some cases deviated from the prediction. In future studies, it will be necessary to improve the predictive performance of the method by clarifying the characteristics of cases that deviate from the prediction.

Background and Objectives: The predicted serum concentrations of vancomycin are determined using population pharmacokinetic parameters. However, the accuracy of predicting vancomycin serum concentrations in the older population remains unclear. Therefore, this study aimed to investigate the accuracy of predicting vancomycin serum concentrations and identifying elements that diminish the prediction accuracy in older people. Materials and Methods: A total of 144 patients aged 75 years or older were included. The serum vancomycin concentrations in the patients were predicted based on population pharmacokinetic parameters common in Japan. We examined the accuracy of serum vancomycin concentration prediction in elderly individuals by comparing the predicted and measured serum vancomycin concentrations in each patient. The prediction accuracy was evaluated using the mean prediction error (ME) and mean absolute error of prediction (MAE) calculated from the measured and predicted serum vancomycin concentrations in each patient. Results: The ME for all patients was 0.27, and the 95% CI included 0, indicating that the predicted values were not significantly biased compared to the measured values. However, the predicted serum concentrations in the <50 kg body weight and serum creatinine (Scr) < 0.6 mg/dL groups were significantly biased compared to the measured values. The group with a history of intensive care unit (ICU) admission showed the largest values for the ME and MAE. Conclusions: Our prediction accuracy was satisfactory but tended to be lower in underweight patients, those with low creatinine levels, and patients admitted to the ICU. Patients with multiple of these factors may experience a greater degree of decreased predictive accuracy.

BACKGROUND: The development of emergency department (ED) triage systems remains challenging in accurately differentiating patients with acute abdominal pain (AAP) who are critical and urgent for surgery due to subjectivity and limitations. We use machine learning models to predict emergency surgical abdominal pain patients in triage, and then compare their performance with conventional Logistic regression models.
METHODS: Using 38 214 patients presenting with acute abdominal pain at Zhongnan Hospital of Wuhan University between March 1, 2014, and March 1, 2022, we identified all adult patients (aged ≥18 years). We utilized routinely available triage data in electronic medical records as predictors, including structured data (eg, triage vital signs, gender, and age) and unstructured data (chief complaints and physical examinations in free-text format). The primary outcome measure was whether emergency surgery was performed. The dataset was randomly sampled, with 80% assigned to the training set and 20% to the test set. We developed 5 machine learning models: Light Gradient Boosting Machine (Light GBM), eXtreme Gradient Boosting (XGBoost), Deep Neural Network (DNN), and Random Forest (RF). Logistic regression (LR) served as the reference model. Model performance was calculated for each model, including the area under the receiver-work characteristic curve (AUC) and net benefit (decision curve), as well as the confusion matrix.
RESULTS: Of all the 38 214 acute abdominal pain patients, 4208 underwent emergency abdominal surgery while 34 006 received non-surgical treatment. In the surgery outcome prediction, all 4 machine learning models outperformed the reference model (eg, AUC, 0.899 [95%CI 0.891-0.903] in the Light GBM vs. 0.885 [95%CI 0.876-0.891] in the reference model), Similarly, most machine learning models exhibited significant improvements in net reclassification compared to the reference model (eg, NRIs of 0.0812[95%CI, 0.055-0.1105] in the XGBoost), with the exception of the RF model. Decision curve analysis shows that across the entire range of thresholds, the net benefits of the XGBoost and the Light GBM models were higher than the reference model. In particular, the Light GBM model performed well in predicting the need for emergency abdominal surgery with higher sensitivity, specificity, and accuracy.
CONCLUSIONS: Machine learning models have demonstrated superior performance in predicting emergency abdominal pain surgery compared to traditional models. Modern machine learning improves clinical triage decisions and ensures that critically needy patients receive priority for emergency resources and timely, effective treatment.

BACKGROUND: The English PUMA questionnaire emerges as an effective COPD case-finding tool. We aimed to use the PUMA questionnaire in combination with peak expiratory flow rate (PEFR) to improve case-finding efficacy in Chinese population.
METHODS: This cross-sectional, observational study included two stages: translating English to Chinese PUMA (C-PUMA) questionnaire with linguistic validation and psychometric evaluation, followed by clinical validation. Eligible participants (with age ≥40 years, respiratory symptoms, smoking history ≥10 pack-years) were enrolled and subjected to three questionnaires (C-PUMA, COPD assessment test [CAT], and generic health survey [SF-12V2]), PEFR measurement, and confirmatory spirometry. The C-PUMA score and PEFR were incorporated into a PUMA-PEFR prediction model applying binary logistic regression coefficients to estimate the probability of COPD (PCOPD).
RESULTS: C-PUMA was finalized through standard forward-backward translation processes and confirmation of good readability, comprehensibility, and reliability. In clinical validation, 240 participants completed the study. 78/240 (32.5%) were diagnosed with COPD. C-PUMA exhibited significant validity (correlated with CAT or physical component scores of SF-12V2, both P<0.05, respectively). PUMA-PEFR model had higher diagnostic accuracy than C-PUMA alone (area under ROC curve, 0.893 vs. 0.749, P<0.05). The best cutoff values of C-PUMA and PUMA-PEFR model (PCOPD) were ≥6 and ≥0.39, accounting for a sensitivity/specificity/numbers needed to screen of 77%/64%/3 and 79%/88%/2, respectively. C-PUMA ≥5 detected more underdiagnosed patients, up to 11.5% (vs. C-PUMA ≥6).
CONCLUSIONS: C-PUMA is well-validated. The PUMA-PEFR model provides more accurate and cost-effective case-finding efficacy than C-PUMA alone in at-risk, undiagnosed COPD patients. These tools can be useful to detect COPD early.

Item response theory (IRT) models are often compared with respect to predictive performance to determine the dimensionality of rating scale data. However, such model comparisons could be biased toward nested-dimensionality IRT models (e.g., the bifactor model) when comparing those models with non-nested-dimensionality IRT models (e.g., a unidimensional or a between-item-dimensionality model). The reason is that, compared with non-nested-dimensionality models, nested-dimensionality models could have a greater propensity to fit data that do not represent a specific dimensional structure. However, it is unclear as to what degree model comparison results are biased toward nested-dimensionality IRT models when the data represent specific dimensional structures and when Bayesian estimation and model comparison indices are used. We conducted a simulation study to add clarity to this issue. We examined the accuracy of four Bayesian predictive performance indices at differentiating among non-nested- and nested-dimensionality IRT models. The deviance information criterion (DIC), a commonly used index to compare Bayesian models, was extremely biased toward nested-dimensionality IRT models, favoring them even when non-nested-dimensionality models were the correct models. The Pareto-smoothed importance sampling approximation of the leave-one-out cross-validation was the least biased, with the Watanabe information criterion and the log-predicted marginal likelihood closely following. The findings demonstrate that nested-dimensionality IRT models are not automatically favored when the data represent specific dimensional structures as long as an appropriate predictive performance index is used.

UNASSIGNED: The mortality rate of acute coronary syndrome (ACS) remains high. Therefore, patients with ACS should undergo early risk stratification, for which various risk calculation tools are available. However, it remains uncertain whether the predictive performance varies over time between risk calculation tools for different target periods. This study aimed to compare the predictive performance of risk calculation tools in estimating short- and long-term mortality risks in patients with ACS, while considering different observation periods using time-dependent receiver operating characteristic (ROC) analysis.
UNASSIGNED: This study included 404 consecutive patients with ACS who underwent coronary angiography at our hospital from March 2017 to January 2021. The ACTION and GRACE scores for short-term risk stratification purposes and CRUSADE scores for long-term risk stratification purposes were calculated for all participants. The participants were followed up for 36 months to assess mortality. Using time-dependent ROC analysis, we evaluated the area under the curve (AUC) of the ACTION, CRUSADE, and GRACE scores at 1, 6, 12, 24, and 36 months.
UNASSIGNED: Sixty-six patients died during the observation periods. The AUCs at 1, 6, 12, 24, and 36 months of the ACTION score were 0.942, 0.925, 0.889, 0.856, and 0.832; those of the CRUSADE score were 0.881, 0.883, 0.862, 0.876, and 0.862; and those of the GRACE score 0.949, 0.928, 0.888, 0.875, and 0.860, respectively.
UNASSIGNED: The ACTION and GRACE scores were excellent risk stratification tools for mortality in the short term. The prognostic performance of each risk score was almost similar in the long term, but the CRUSADE score might be a superior risk stratification tool in the longer term than 3 years.

Hierarchical models can express ecological dynamics using a combination of fixed and random effects, and measurement of their complexity (effective degrees of freedom, EDF) requires estimating how much random effects are shrunk toward a shared mean. Estimating EDF is helpful to (1) penalize complexity during model selection and (2) to improve understanding of model behavior. I applied the conditional Akaike Information Criterion (cAIC) to estimate EDF from the finite-difference approximation to the gradient of model predictions with respect to each datum. I confirmed that this has similar behavior to widely used Bayesian criteria, and I illustrated ecological applications using three case studies. The first compared model parsimony with or without time-varying parameters when predicting density-dependent survival, where cAIC favors time-varying demographic parameters more than conventional Akaike Information Criterion. The second estimates EDF in a phylogenetic structural equation model, and identifies a larger EDF when predicting longevity than mortality rates in fishes. The third compares EDF for a species distribution model fitted for 20 bird species and identifies those species requiring more model complexity. These highlight the ecological and statistical insight from comparing EDF among experimental units, models, and data partitions, using an approach that can be broadly adopted for nonlinear ecological models.

BACKGROUND: Recent studies have suggested that insulin resistance (IR) contributes to the development of cardiovascular diseases (CVD), and the estimated glucose disposal rate (eGDR) is considered to be a reliable surrogate marker of IR. However, most existing evidence stems from studies involving diabetic patients, potentially overstating the effects of eGDR on CVD. Therefore, the primary objective of this study is to examine the relationship of eGDR with incidence of CVD in non-diabetic participants.
METHODS: The current analysis included individuals from the China Health and Retirement Longitudinal Study (CHARLS) who were free of CVD and diabetes mellitus but had complete data on eGDR at baseline. The formula for calculating eGDR was as follows: eGDR (mg/kg/min) = 21.158 - (0.09 × WC) - (3.407 × hypertension) - (0.551 × HbA1c) [WC (cm), hypertension (yes = 1/no = 0), and HbA1c (%)]. The individuals were categorized into four subgroups according to the quartiles (Q) of eGDR. Crude incidence rate and hazard ratios (HRs) with 95% confidence intervals (CIs) were computed to investigate the association between eGDR and incident CVD, with the lowest quartile of eGDR (indicating the highest grade of insulin resistance) serving as the reference. Additionally, the multivariate adjusted restricted cubic spine (RCS) was employed to examine the dose-response relationship.
RESULTS: We included 5512 participants in this study, with a mean age of 58.2 ± 8.8 years, and 54.1% were female. Over a median follow-up duration of 79.4 months, 1213 incident CVD cases, including 927 heart disease and 391 stroke, were recorded. The RCS curves demonstrated a significant and linear relationship between eGDR and all outcomes (all P for non-linearity > 0.05). After multivariate adjustment, the lower eGDR levels were founded to be significantly associated with a higher risk of CVD. Compared with participants with Q1 of eGDR, the HRs (95% CIs) for those with Q2 - 4 were 0.88 (0.76 - 1.02), 0.69 (0.58 - 0.82), and 0.66 (0.56 - 0.79). When assessed as a continuous variable, per 1.0-SD increase in eGDR was associated a 17% (HR: 0.83, 95% CI: 0.78 - 0.89) lower risk of CVD, with the subgroup analyses indicating that smoking status modified the association (P for interaction = 0.012). Moreover, the mediation analysis revealed that obesity partly mediated the association. Additionally, incorporating eGDR into the basic model considerably improve the predictive ability for CVD.
CONCLUSIONS: A lower level of eGDR was found to be associated with increased risk of incident CVD among non-diabetic participants. This suggests that eGDR may serve as a promising and preferable predictor and intervention target for CVD.

BACKGROUND: Vessels Encapsulating Tumor Clusters (VETC) are now recognized as independent indicators of recurrence and overall survival in hepatocellular carcinoma (HCC) patients. However, there has been limited investigation into predicting the VETC pattern using hepatobiliary phase (HBP) features from preoperative gadobenate-enhanced MRI.
METHODS: This study involved 252 HCC patients with confirmed VETC status from three different hospitals (Hospital 1: training set with 142 patients; Hospital 2: test set with 64 patients; Hospital 3: validation set with 46 patients). Independent predictive factors for VETC status were determined through univariate and multivariate logistic analyses. Subsequently, these factors were used to construct two distinct VETC prediction models. Model 1 included all independent predictive factors, while Model 2 excluded HBP features. The performance of both models was assessed using the Area Under the Curve (AUC), Decision Curve Analysis, and Calibration Curve. Prediction accuracy between the two models was compared using Net Reclassification Improvement (NRI) and Integrated Discriminant Improvement (IDI).
RESULTS: CA199, IBIL, shape, peritumoral hyperintensity on HBP, and arterial peritumoral enhancement were independent predictors of VETC. Model 1 showed robust predictive performance, with AUCs of 0.836 (training), 0.811 (test), and 0.802 (validation). Model 2 exhibited moderate performance, with AUCs of 0.813, 0.773, and 0.783 in the respective sets. Calibration and decision curves for both models indicated consistent predictions between predicted and actual VETC, benefiting HCC patients. NRI showed Model 1 increased by 0.326, 0.389, and 0.478 in the training, test, and validation sets compared to Model 2. IDI indicated Model 1 increased by 0.036, 0.028, and 0.025 in the training, test, and validation sets compared to Model 2.
CONCLUSIONS: HBP features from preoperative gadobenate-enhanced MRI can enhance the predictive performance of VETC in HCC.

OBJECTIVE: To clarify the predictive performance of different measures of frailty, including Clinical Frailty Scale (CFS), 11-factor modified Frailty Index (mFI-11), and 5-factor modified Frailty Index (mFI-5), on adverse outcomes.
METHODS: PubMed, Embase, Web of Science, and other databases were retrieved from the inception of each database to June 2023. The pooled sensitivity, specificity, and the area under the summary receiver operating curve (SROC) values were analyzed to determine the predictive power of CFS, mFI-11, and mFI-5 for adverse outcomes.
RESULTS: A total of 25 studies were included in quantitative synthesis. The pooled sensitivity values of CFS for predicting anastomotic leakage, total complications, and major complications were 0.39, 0.57, 0.45; pooled specificity values were 0.70, 0.58, 0.73; the area under SROC values were 0.58, 0.6, 0.66. The pooled sensitivity values of mFI-11 for predicting total complications and delirium were 0.38 and 0.64; pooled specificity values were 0.83 and 0.72; the area under SROC values were 0.64 and 0.74. The pooled sensitivity values of mFI-5 for predicting total complications, 30-day mortality, and major complications were 0.27, 0.54, 0.25; pooled specificity values were 0.82, 0.84, 0.81; the area under SROC values were 0.63, 0.82, 0.5.
CONCLUSIONS: The results showed that CFS could predict anastomotic leakage, total complications, and major complications; mFI-11 could predict total complications and delirium; mFI-5 could predict total complications and 30-day mortality. More high-quality research is needed to support the conclusions of this study further.