Temporal Validation Research Articles

Potential and Limitations of the New European Ground Motion Service in Landslides at a Local Scale

Mass movements represent one of the most significant geohazards worldwide. The aim of this research is to highlight the potential and limitations of the European Ground Motion Service (EGMS) in detecting and monitoring mass movements at a local scale, especially in cases where data from in situ instrumental devices are unavailable. The study area corresponds to the La Miera landslide, located in Asturias (NW Spain). The multidisciplinary methodology applied involved the following steps: (1) downloading, acquiring, and analyzing Sentinel-1 A-DInSAR datasets (2015–2021) through the EGMS; (2) conducting a detailed geomorphological map and identifying evidence of movement; (3) classifying building damage by means of a damage inventory; (4) compiling and analyzing daily rainfall records with respect to deformation time series. Sentinel-1 A-DInSAR results revealed maximum LOS and East–West velocities of −11.6 and −7.9 mm/yr related to the landslide activity. Geomorphological mapping allowed for the updating of the landslide boundaries and its characterization as an active, complex movement. Registered building damage, which ranged from moderate to serious, was correlated with LOS and East–West velocities. The displacement recorded by the EGMS closely corresponds with rainfall periods, while periods of reduced rainfall coincide with the stabilization and recovery phases of displacement. This emphasizes a noteworthy quantitative correlation between rainfall events and EGMS data, evident both spatially and temporally. This work highlights that areas in which the EGMS data indicate deformation but lack in situ instrumental records, geomorphological techniques, and building damage surveys can provide spatial validation of the EGMS displacement, while rainfall records can provide temporal validation.

Integrating StEP-COMPAC definition and enhanced recovery after surgery status in a machine-learning-based model for postoperative pulmonary complications in laparoscopic hepatectomy

BackgroundPostoperative pulmonary complications (PPCs) contribute to high mortality rates and impose significant financial burdens. In this study, a machine learning-based prediction model was developed to identify patients at high risk of developing PPCs following laparoscopic hepatectomy. MethodsData were collected from 1022 adult patients who underwent laparoscopic hepatectomy at two centres between January 2015 and February 2021. The dataset was divided into a development set and a temporal external validation set based on the year of surgery. A total of 42 factors were extracted for pre-modelling, including the implementation status of Enhanced Recovery after Surgery (ERAS). Feature selection was performed using the least absolute shrinkage and selection operator (LASSO) method. Model performance was assessed using the area under the receiver operating characteristic curve (AUC). The model with the best performance was externally validated using temporal data. ResultsThe incidence of PPCs was 8.7%. Lambda.1se was selected as the optimal lambda for LASSO feature selection. For implementation of ERAS, serum gamma-glutamyl transferase levels, malignant tumour presence, total bilirubin levels, and age-adjusted Charleston Comorbidities Index were the selected factors. Seven models were developed. Among them, logistic regression demonstrated the best performance, with an AUC of 0.745 in the internal validation set and 0.680 in the temporal external validation set. ConclusionsBased on the most recent definition, a machine learning model was employed to predict the risk of PPCs following laparoscopic hepatectomy. Logistic regression was identified as the best-performing model. ERAS implementation was associated with a reduction in the number of PPCs.

Spleen stiffness measurement by vibration-controlled transient elastography at 100 Hz for non-invasive predicted diagnosis of clinically significant portal hypertension in patients with compensated advanced chronic liver disease: a modelling study

In patients with compensated advanced chronic liver disease (cACLD), risk of clinically significant portal hypertension (CSPH) can be estimated by applying non-invasive tests such as liver stiffness measurement (LSM), platelet count, and, in some cases, BMI. We aimed to assess the diagnostic utility of spleen stiffness measurement (SSM) at 100 Hz as a standalone non-invasive test for CSPH and to evaluate its incremental value compared with the ANTICIPATE±NASH model in patients with cACLD. For this modelling study, patients were recruited from 16 expert centres in Europe. Patients who underwent characterisation by hepatic venous pressure gradient (HVPG) and non-invasive tests (ie, LSM, platelet count, and SSM at 100 Hz) at one of the participating centres between Jan 1, 2020, and Dec 31, 2023, were considered for inclusion. Only patients aged 18 years or older with Child-Pugh class A cACLD, shown by LSM 10 kPa or more or F3 or F4 fibrosis on liver histology, were included. The overall cohort was split into the derivation cohort (patients recruited between Jan 1, 2020, and Dec 31, 2022) and the temporal validation cohort (patients recruited between Jan 1, 2023, and Dec 31, 2023). The ANTICIPATE±NASH model was applied to assess individual CSPH probability and SSM was investigated as a standalone non-invasive test for CPSH; in combination with platelet count and BMI; and in a full model of SSM, LSM, platelet count, and BMI (ie, the Non-Invasive CSPH Estimated Risk [NICER] model). All models were binary logistic regression models. The primary outcome was CSPH. We evaluated the discriminative utility of the models by calculating the area under the receiver operating characteristics curve (AUC) and creating calibration plots and calibration of intercept, slope, and integrated calibration index. 407 patients with cACLD were included, 202 (50%) in the derivation cohort and 205 (50%) in the validation cohort. Median age was 60·0 years (IQR 55·0-66·8); 275 (68%) of 407 patients were male and 132 (32%) were female. 164 (40%) of 407 patients had metabolic dysfunction-associated steatotic liver disease (MASLD), 133 (33%) had MASLD with increased alcohol intake or alcohol-related liver disease, 75 (18%) had viral hepatitis (61 [81%] of whom had sustained virologic response of hepatitis C virus or suppression of hepatitis B virus DNA), and 35 (9%) had other chronic liver diseases. 241 (59%) patients had CSPH. Median SSM was 45·0 kPa (32·1-65·4) and LSM was 21·4 kPa (14·1-31·6). SSM and LSM had similar AUCs for prediction of CSPH in the derivation cohort (0·779 [95% CI 0·717-0·842] vs 0·781 [0·718-0·844]; p=0·97) and in the validation cohort (0·830 [0·772-0·887] vs 0·804 [0·743-0·864]; p=0·50). The SSM-based model comprising platelet count and BMI had a similar AUC as the ANTICIPATE±NASH model in both the derivation cohort (0·849 [0·794-0·903] vs 0·849 [0·794-0·903]; p=0·999) and in the validation cohort (0·873 [0·819-0·922] vs 0·863 [0·810-0·916]; p=0·75). The NICER model had a significantly higher AUC for prediction of CSPH than the ANTICIPATE±NASH model in the derivation cohort (0·889 [0·843-0·934] vs 0·849 [0·794-0·903]; p=0·022) and in the validation cohort (0·906 [0·864-0·946] vs 0·863 [0·810-0·916]; p=0·012). The addition of SSM to LSM, BMI, and platelet count outperformed the ANTICIPATE±NASH model for CSPH risk stratification in our cohort of contemporary patients with cACLD. SSM improves the non-invasive diagnosis of CSPH, supporting its implementation into clinical practice. Echosens.

Development and Validation of a Deep Learning Model for Prediction of Adult Physiological Deterioration.

Prediction-based strategies for physiologic deterioration offer the potential for earlier clinical interventions that improve patient outcomes. Current strategies are limited because they operate on inconsistent definitions of deterioration, attempt to dichotomize a dynamic and progressive phenomenon, and offer poor performance. Can a deep learning deterioration prediction model (Deep Learning Enhanced Triage and Emergency Response for Inpatient Optimization [DETERIO]) based on a consensus definition of deterioration (the Adult Inpatient Decompensation Event [AIDE] criteria) and that approaches deterioration as a state "value-estimation" problem outperform a commercially available deterioration score? The derivation cohort contained retrospective patient data collected from both inpatient services (inpatient) and emergency departments (EDs) of two hospitals within the University of California San Diego Health System. There were 330,729 total patients; 71,735 were inpatient and 258,994 were ED. Of these data, 20% were randomly sampled as a retrospective "testing set." The validation cohort contained temporal patient data. There were 65,898 total patients; 13,750 were inpatient and 52,148 were ED. DETERIO was developed and validated on these data, using the AIDE criteria to generate a composite score. DETERIO's architecture builds upon previous work. DETERIO's prediction performance up to 12 hours before T0 was compared against Epic Deterioration Index (EDI). In the retrospective testing set, DETERIO's area under the receiver operating characteristic curve (AUC) was 0.797 and 0.874 for inpatient and ED subsets, respectively. In the temporal validation cohort, the corresponding AUC were 0.775 and 0.856, respectively. DETERIO outperformed EDI in the inpatient validation cohort (AUC, 0.775 vs. 0.721; p < 0.01) while maintaining superior sensitivity and a comparable rate of false alarms (sensitivity, 45.50% vs. 30.00%; positive predictive value, 20.50% vs. 16.11%). DETERIO demonstrates promise in the viability of a state value-estimation approach for predicting adult physiologic deterioration. It may outperform EDI while offering additional clinical utility in triage and clinician interaction with prediction confidence and explanations. Additional studies are needed to assess generalizability and real-world clinical impact.

The PROgnostic ModEl for chronic lung disease (PRO-MEL): development and temporal validation

BackgroundPatients with chronic lung diseases (CLDs), defined as progressive and life-limiting respiratory conditions, experience a heavy symptom burden as the conditions become more advanced, but palliative referral rates are low and late. Prognostic tools can help clinicians identify CLD patients at high risk of deterioration for needs assessments and referral to palliative care. As current prognostic tools may not generalize well across all CLD conditions, we aim to develop and validate a general model to predict one-year mortality in patients presenting with any CLD.MethodsA retrospective cohort study of patients with a CLD diagnosis at a public hospital from July 2016 to October 2017 was conducted. The outcome of interest was all-cause mortality within one-year of diagnosis. Potential prognostic factors were identified from reviews of prognostic studies in CLD, and data was extracted from electronic medical records. Missing data was imputed using multiple imputation by chained equations. Logistic regression models were developed using variable selection methods and validated in patients seen from January 2018 to December 2019. Discriminative ability, calibration and clinical usefulness of the model was assessed. Model coefficients and performance were pooled across all imputed datasets and reported.ResultsOf the 1000 patients, 122 (12.2%) died within one year. Patients had chronic obstructive pulmonary disease or emphysema (55%), bronchiectasis (38%), interstitial lung diseases (12%), or multiple diagnoses (6%). The model selected through forward stepwise variable selection had the highest AUC (0.77 (0.72–0.82)) and consisted of ten prognostic factors. The model AUC for the validation cohort was 0.75 (0.70, 0.81), and the calibration intercept and slope were − 0.14 (-0.54, 0.26) and 0.74 (0.53, 0.95) respectively. Classifying patients with a predicted risk of death exceeding 0.30 as high risk, the model would correctly identify 3 out 10 decedents and 9 of 10 survivors.ConclusionsWe developed and validated a prognostic model for one-year mortality in patients with CLD using routinely available administrative data. The model will support clinicians in identifying patients across various CLD etiologies who are at risk of deterioration for a basic palliative care assessment to identify unmet needs and trigger an early referral to palliative medicine.Trial registrationNot applicable (retrospective study).

Identification of variables and development of a prediction model for DIBH eligibility in left-sided breast cancer radiotherapy: a prospective cohort study with temporal validation

ObjectiveTo identify variables associated with a patients’ ability to reproducibly hold their breath for deep-inspiration breath-hold (DIBH) radiotherapy (RT) and to develop a predictive model for DIBH eligibility.MethodsThis prospective, single-institution, IRB-approved observational study included women with left-sided breast cancer treated between January 2023 and March 2024. Patients underwent multiple breath-hold sessions over 2–3 consecutive days. DIBH waveform metrics and clinical factors were recorded and analysed. Logistic mixed modelling was used to predict DIBH eligibility, and a temporal validation cohort was used to assess model performance.ResultsIn total, 253 patients were included, with 206 in the model development cohort and 47 in the temporal validation cohort. The final logistic mixed model identified increasing average breath-hold duration (OR, 95% CI: 0.308, 0.104–0.910. p = 0.033) and lower amplitude (OR, 95% CI: 0.737, 0.641–0.848. p < 0.001) as significant predictors of DIBH eligibility. Increasing age was associated with higher odds of being ineligible for DIBH (OR, 95% CI: 1.040, 1.001–1.081. p = 0.044). The model demonstrated good discriminative performance in the validation cohort with an AUC of 80.9% (95% CI: 73.0-88.8).ConclusionThe identification of variables associated with DIBH eligibility and development of a predictive model has the potential to serve as a decision-support tool. Further external validation is required before its integration into routine clinical practice.

Prediction for Perioperative Stroke Using Intraoperative Parameters.

Perioperative stroke is a severe complication following surgery. To identify patients at risk for perioperative stroke, several prediction models based on the preoperative factors were suggested. Prediction models often focus on preoperative patient characteristics to assess stroke risk. However, most existing models primarily base their predictions on the patient's baseline characteristics before surgery. We aimed to develop a machine-learning model incorporating both pre- and intraoperative variables to predict perioperative stroke. This study included patients who underwent noncardiac surgery at 2hospitals with the data of 15 752 patients from Seoul National University Hospital used for development and temporal internal validation, and the data of 449 patients from Boramae Medical Center used for external validation. Perioperative stroke was defined as a newly developed ischemic lesion on diffusion-weighted imaging within 30 days of surgery. We developed a prediction model composed of pre- and intraoperative factors (integrated model) and compared it with a model consisting of preoperative features alone (preoperative model). Perioperative stroke developed in 109 (0.69%) patients in the Seoul National University Hospital group and 11 patients (2.45%) in the Boramae Medical Center group. The integrated model demonstrated superior predictive performance with area under the curve values of 0.824 (95% CI, 0.762-0.880) versus 0.584 (95% CI, 0.499-0.667; P<0.001) in the internal validation; and 0.716 (95% CI, 0.560-0.859) versus 0.505 (95% CI, 0.343-0.654; P=0.018) in the external validation, compared to the preoperative model. We suggest that incorporating intraoperative factors into perioperative stroke prediction models can improve their accuracy.

Development and validation of early prediction models for new-onset functional impairment of patients with trauma at hospital discharge.

Early identification of individuals at risk of functional impairment after trauma is crucial for the timely clinical decision-making and intervention to improve reintegration into the society. This study aimed to develop and validate models for predicting new-onset functional impairment after trauma using predictors that are routinely collected within 2 days of hospital admission. In this multicenter retrospective cohort study of acute care hospitals in Japan, we identified adult patients with trauma with independence in carrying out activities of daily living before hospitalization, treated in the intensive or high-dependency care unit, and survived for at least 2 days between April 2008 and September 2023. The primary outcome was functional impairment defined as Barthel Index ≤60 at hospital discharge. In the internal validation data set (between April 2008 and August 2022), using the routinely collected 129 candidate predictors within 2 days of admission, we trained and tuned the four conventional and machine learning models with repeated random subsampling cross-validation. We measured the performance of these models in the temporal validation data set (between September 2022 and September 2023). We also computed the importance of each predictor variable in our model. We identified 8,529 eligible patients. Functional impairment at discharge was observed in 41% of the patients (n = 3,506/8,529). In the temporal validation data set, all four models showed moderate discrimination ability, with areas under the curve above 0.79, and extreme gradient boosting showing the best performance (0.83). In the variable importance analyses, age was the most important predictor, followed by consciousness, severity score, cervical spinal cord injury, mild dementia, and serum albumin level at admission. We successfully developed early prediction models for patients with trauma with new-onset functional impairment at discharge that achieved high predictive performance using routinely collected data within 2 days of hospital admission. Prognostic/Epidemiological; Level II.

Development and validation of a risk scoring tool for predicting incident reversible cognitive frailty among community-dwelling older adults: A prospective cohort study.

Reversible cognitive frailty (RCF) is an ideal target to prevent asymptomatic cognitive impairment and dependency. This study aimed to develop and validate prediction models for incident RCF. A total of 1230 older adults aged ≥60 years from China Health and Retirement Longitudinal Study 2011-2013 survey were included as the training set. The modified Poisson regression and three machine learning algorithms including eXtreme Gradient Boosting, support vector machine and random forest were used to develop prediction models. All models were evaluated internally with fivefold cross-validation, and evaluated externally using a temporal validation method through the China Health and Retirement Longitudinal Study 2013-2015 survey. The incidence of RCF was 27.4% in the training set and 27.5% in the external validation set. A total of 13 important predictors were selected to develop the model, including age, education, contact with their children, medical insurance, vision impairment, heart diseases, medication types, self-rated health, pain locations, loneliness, self-medication, night-time sleep and having running water. All models showed acceptable or approximately acceptable discrimination (AUC 0.683-0.809) for the training set, but fair discrimination (AUC 0.568-0.666) for the internal and external validation. For calibration, only modified Poisson regression and eXtreme Gradient Boosting were acceptable in the training set. All models had acceptable overall prediction performance and clinical usefulness. Older adults were divided into three groups by the risk scoring tool constructed based on modified Poisson regression: low risk (≤24), median risk (24-29) and high risk (>29). This risk tool could assist healthcare providers to predict incident RCF among older adults in the next 2 years, facilitating early identification of a high-risk population of RCF. Geriatr Gerontol Int 2024; ••: ••-••.

What factors preventing the older adults in China from living longer: a machine learning study

BackgroundThe fact that most older people do not live long means that they do not have more time to pursue self-actualization and contribute value to society. Although there are many studies on the longevity of the elderly, the limitations of traditional statistics lack the good ability to study together the important influencing factors and build a simple and effective prediction model.MethodsBased on the the data of Chinese Longitudinal Healthy Longevity Survey (CLHLS), 2008–2018 cohort and 2014–2018 cohort were selected and 16 features were filtered and integrated. Five machine learning algorithms, Elastic-Net Regression (ENR), Decision Tree (DT), Random Forest (RF), K-Nearest Neighbor (KNN), and eXtreme Gradient Boosting (XGBoost), were used to develop models and assessed by internal validation with CLHLS 2008–2018 cohort and temporal validation with CLHLS 2014–2018 cohort. Besides, the best performing model was explained and according to the variable importance results, simpler models would be developed.ResultsThe results showed that the model developed by XGBoost algorithm had the best performance with AUC of 0.788 in internal validation and 0.806 in temporal validation. Instrumental activity of daily living (IADL), leisure activity, marital status, sex, activity of daily living (ADL), cognitive function, overall plant-based diet index (PDI) and psychological resilience, 8 features were more important in the model. Finally, with these 8 features simpler models were developed, it was found that the model performance did not decrease in both internal and temporal validation.ConclusionsThe study indicated that the importance of these 8 factors for predicting the death of elderly people in China and built a simple machine learning model with good predictive performance. It can inspire future key research directions to promote longevity of the elderly, as well as in practical life to make the elderly healthy longevity, or timely end-of-life care for the elderly, and can use predictive model to aid decision-making.

Measuring childbirth-related posttraumatic stress disorder: Psychometric properties of the Italian version of the City Birth Trauma Scale (City BiTS).

The City Birth Trauma Scale (City BiTS) assesses postpartum posttraumatic stress disorder (PTSD) based on the Diagnostic and Statistical Manual of Mental Disorders (5th ed.) criteria. Although it has been validated worldwide, predictive validity has not been previously examined. Moreover, no Italian version of the scale exists. This study aimed to test the bifactor latent structure and alternative models, internal consistency, test-retest reliability, convergent validity, divergent validity, and predictive validity of the City BiTS. Women (N = 629) who had given birth within the past 3 months completed an online survey including sociodemographic and obstetric characteristics, the City BiTS, the Impact of Event Scale-Revised, and the Edinburgh Postnatal Depression Scale. After 3 months, women completed the City BiTS again and reported their intention to breastfeed during the 1-year postpartum. Exploratory factor analysis confirmed the two-factorial structure. In confirmatory factor analysis, the two-factorial solution showed the best model fit. Internal consistency was good to excellent for the subscales and the total scale. Correlation analyses showed strong convergent validity with the Impact of Event Scale-Revised, high divergent validity with the Edinburgh Postnatal Depression Scale, high test-retest reliability, and good predictive validity with the intention to exclusively breastfeed. Moreover, the Birth-Related Symptoms subscale distinguished between different types of delivery. The City BiTS-Italian is the first measure evaluating and diagnosing childbirth-related PTSD symptoms based on Diagnostic and Statistical Manual of Mental Disorders (5th ed.) in Italy. The factorial structure and validity reported in other cultural contexts were confirmed; moreover, findings add evidence to the scale's temporal stability and predictive validity. Besides contributing to clinical purposes, the City BiTS-Italian will facilitate international comparability regarding the prevalence of PTSD following childbirth. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Temporal and external validation of the algorithm predicting first trimester outcome of a viable pregnancy.

Symptoms like vaginal bleeding or abdominal pain in early pregnancy can create anxiety about potential miscarriage. Previous studies have demonstrated ultrasonographic variables at the first trimester transvaginal scan (TVS) which can assist in predicting outcomes by 12 weeks gestation. To validate the miscarriage risk prediction model (MRP) in women who present with a viable intrauterine pregnancy (IUP) at the primary ultrasound. A multi-centre diagnostic study of 1490 patients was performed between 2011 and 2019 for retrospective external and 2017-2019 for prospective temporal validation. The reference standard was a viable pregnancy at 12 + 6 weeks. The MRP model is a multinomial logistic regression model based on maternal age, embryonic heart rate, logarithm (gestational sac volume/crown-rump length (CRL)) ratio, CRL and presence or absence of clots. Temporal validation data from 290 viable IUPs were collected: 225 were viable at the end of the first trimester, 31 had miscarried and 34 were lost to follow-up. External validation data from 1203 viable IUPs were collected at two other ultrasound units: 1062 were viable, 69 had miscarried and 72 were lost to follow-up. Temporal validation with a cut-off of 0.1 demonstrated: area under the curve (AUC) of 0.8 (0.7-0.9), sensitivity 66.7%, specificity 83.9%, positive predictive value (PPV) 35.7%, negative predictive value (NPV) 94.9%, positive likelihood ration (LR+) 4.1 and negative LR (LR-) 0.4. External validation demonstrated: AUC 0.7 (0.7-0.8), sensitivity 44.9%, specificity 90.4%, PPV 23.3%, NPV 96.2%, LR+ 4.6 and LR- 0.6 (0.4-0.7). The MRP model is not able to be used in real time for counselling, and management should be individualised.

Forty-Year Fire History Reconstruction from Landsat Data in Mediterranean Ecosystems of Algeria following International Standards

Algeria, the main fire hotspot on the southern rim of the Mediterranean Basin, lacks a complete fire dataset with official fire perimeters, and the existing one contains inconsistencies. Preprocessed global and regional burned area (BA) products provide valuable insights into fire patterns, characteristics, and dynamics over time and space, and into their impact on climate change. Nevertheless, they exhibit certain limitations linked with their inherent spatio-temporal resolutions as well as temporal and geographical coverage. To address the need for reliable BA information in Algeria, we systematically reconstructed, validated, and analyzed a 40-year (1984–2023) BA product (NEALGEBA; North Eastern ALGeria Burned Area) at 30 m spatial resolution in the typical Mediterranean ecosystems of this region, following international standards. We used Landsat data and the BA Mapping Tools (BAMTs) in the Google Earth Engine (GEE) to map BAs. The spatial validation of NEALGEBA, performed for 2017 and 2021 using independent 10 m spatial resolution Sentinel-2 reference data, showed overall accuracies &gt; 98.10%; commission and omission errors &lt; 8.20%; Dice coefficients &gt; 91.90%; and relative biases &lt; 3.44%. The temporal validation, however, using MODIS and VIIRS active fire hotspots, emphasized the limitation of Landsat-based BA products in temporal fire reporting accuracy terms. The intercomparison with five readily available BA products for 2017, by using the same validation process, demonstrated the overall outperformance of NEALGEBA. Furthermore, our BA product exhibited the highest correspondence with the ground-based BA estimates. NEALGEBA currently represents the most continuous and reliable time series of BA history at fine spatial resolution for NE Algeria, offering a significant contribution to further national and international fire hazard and impact assessments and acts as a reference dataset for contextualizing future weather extremes, such as the 2023 exceptional heat wave, which we show not to have led to the most extreme fire year over the last four decades.

Development and validation of machine learning models to predict MDRO colonization or infection on ICU admission by using electronic health record data

BackgroundMultidrug-resistant organisms (MDRO) pose a significant threat to public health. Intensive Care Units (ICU), characterized by the extensive use of antimicrobial agents and a high prevalence of bacterial resistance, are hotspots for MDRO proliferation. Timely identification of patients at high risk for MDRO can aid in curbing transmission, enhancing patient outcomes, and maintaining the cleanliness of the ICU environment. This study focused on developing a machine learning (ML) model to identify patients at risk of MDRO during the initial phase of their ICU stay.MethodsUtilizing patient data from the First Medical Center of the People’s Liberation Army General Hospital (PLAGH-ICU) and the Medical Information Mart for Intensive Care (MIMIC-IV), the study analyzed variables within 24 h of ICU admission. Machine learning algorithms were applied to these datasets, emphasizing the early detection of MDRO colonization or infection. Model efficacy was evaluated by the area under the receiver operating characteristics curve (AUROC), alongside internal and external validation sets.ResultsThe study evaluated 3,536 patients in PLAGH-ICU and 34,923 in MIMIC-IV, revealing MDRO prevalence of 11.96% and 8.81%, respectively. Significant differences in ICU and hospital stays, along with mortality rates, were observed between MDRO positive and negative patients. In the temporal validation, the PLAGH-ICU model achieved an AUROC of 0.786 [0.748, 0.825], while the MIMIC-IV model reached 0.744 [0.723, 0.766]. External validation demonstrated reduced model performance across different datasets. Key predictors included biochemical markers and the duration of pre-ICU hospital stay.ConclusionsThe ML models developed in this study demonstrated their capability in early identification of MDRO risks in ICU patients. Continuous refinement and validation in varied clinical contexts remain essential for future applications.

Fair prediction of 2-year stroke risk in patients with atrial fibrillation.

This study aims to develop machine learning models that provide both accurate and equitable predictions of 2-year stroke risk for patients with atrial fibrillation across diverse racial groups. Our study utilized structured electronic health records (EHR) data from the All of Us Research Program. Machine learning models (LightGBM) were utilized to capture the relations between stroke risks and the predictors used by the widely recognized CHADS2 and CHA2DS2-VASc scores. We mitigated the racial disparity by creating a representative tuning set, customizing tuning criteria, and setting binary thresholds separately for subgroups. We constructed a hold-out test set that not only supports temporal validation but also includes a larger proportion of Black/African Americans for fairness validation. Compared to the original CHADS2 and CHA2DS2-VASc scores, significant improvements were achieved by modeling their predictors using machine learning models (Area Under the Receiver Operating Characteristic curve from near 0.70 to above 0.80). Furthermore, applying our disparity mitigation strategies can effectively enhance model fairness compared to the conventional cross-validation approach. Modeling CHADS2 and CHA2DS2-VASc risk factors with LightGBM and our disparity mitigation strategies achieved decent discriminative performance and excellent fairness performance. In addition, this approach can provide a complete interpretation of each predictor. These highlight its potential utility in clinical practice. Our research presents a practical example of addressing clinical challenges through the All of Us Research Program data. The disparity mitigation framework we proposed is adaptable across various models and data modalities, demonstrating broad potential in clinical informatics.

A pilot clinical risk model to predict polymyxin-induced nephrotoxicity: a real-world, retrospective cohort study.

Polymyxin-induced nephrotoxicity (PIN) is a major safety concern and challenge in clinical practice, which limits the clinical use of polymyxins. This study aims to investigate the risk factors and to develop a scoring tool for the early prediction of PIN. Data on critically ill patients who received intravenous polymyxin B or colistin sulfate for over 24 h were collected. Logistic regression with the least absolute shrinkage and selection operator (LASSO) was used to identify variables that are associated with outcomes. The eXtreme Gradient Boosting (XGB) classifier algorithm was used to further visualize factors with significant differences. A prediction model for PIN was developed through binary logistic regression analysis and the model was assessed by temporal validation and external validation. Finally, a risk-scoring system was developed based on the prediction model. Of 508 patients, 161 (31.6%) patients developed PIN. Polymyxin type, loading dose, septic shock, concomitant vasopressors and baseline blood urea nitrogen (BUN) level were identified as significant predictors of PIN. All validation exhibited great discrimination, with the AUC of 0.742 (95% CI: 0.696-0.787) for internal validation, of 0.708 (95% CI: 0.605-0.810) for temporal validation and of 0.874 (95% CI: 0.759-0.989) for external validation, respectively. A simple risk-scoring tool was developed with a total risk score ranging from -3 to 4, corresponding to a risk of PIN from 0.79% to 81.24%. This study established a prediction model for PIN. Before using polymyxins, the simple risk-scoring tool can effectively identify patients at risk of developing PIN within a range of 7% to 65%.

Impact of relative and absolute values on orienting attention in time.

Reward has been known to render the reward-associated stimulus more salient to block effective attentional orienting in space. However, whether and how reward influences goal-directed attention in time remains unclear. Here, we used a modified attentional cueing paradigm to explore the effect of reward on temporal attention, in which the valid targets were given a low monetary reward and invalid targets were given a high monetary reward. The results showed that the temporal cue validity effect was significantly smaller when the competitive reward structure was employed (Experiment 1), and we ruled out the possibility that the results were due to the practice effect (Experiment 2a) or a reward-promoting effect (Experiment 2b). When further strengthening the intensity of the reward from 1:10 to 1:100 (Experiment 3), we found a similar pattern of results to those in Experiment 1. These results suggest that reward information which was based on relative instead of absolute values can weaken, but not reverse, the orienting attention in time.

A nomogram for risk stratification of central cervical lymph node metastasis in patients with papillary thyroid carcinoma.

Whether to perform prophylactic central lymph node dissection for cN0 papillary thyroid carcinoma (PTC) patients is still controversial. This retrospective study aimed to develop and validate a nomogram based on ultrasound and dual-energy computed tomography (DECT) for the risk stratification of central lymph node metastasis (CLNM) in patients with PTC. A total of 525 patients from 2017 to 2019 [Tianjin First Central Hospital (Hospital A)] were retrospectively analyzed to form the training cohort and to conduct internal validation. Another group of 204 patients in 2020 (Hospital A) formed the temporal validation cohort. A total of 107 patients in 2020 [Binzhou Medical University Hospital (Hospital B)] formed the geographic validation cohort, which was a retrospective cohort study. The area under the curve (AUC), calibration curve, and decision curve were used to evaluate the performance of the nomogram. The locally weighted regression curve was used for risk stratification. Diameter, taller-than-wide, calcification, capsular invasion, and iodine concentration in the arterial and venous phases were independent risk predictors of CLNM. The AUC of the nomogram was 0.922 (95% confidence interval: 0.895-0.943) in the training cohort. Two external validation cohorts demonstrated the good performance of the nomogram in predicting CLNM, with AUCs of 0.912 and 0.861. The significantly improved net reclassification index and integrated discriminatory improvement index indicated that DECT was a powerful supplement to ultrasound for predicting CLNM. The risk stratification system divided all patients into low-risk (0-50 points), intermediate-risk (51-100 points), and high-risk groups (>100 points). The nomogram and risk stratification system estimated the utility of CLNM to guide individualized treatment of patients with PTC.

Prognostic Nomograms for Elderly Patients with Small Cell Lung Cancer Brain Metastasis: A Surveillance, Epidemiology, and End Results Population-Based Study with Temporal External Validation

This study aimed to pinpoint independent predictors influencing overall survival (OS) and cancer-specific survival (CSS) in elderly patients with small cell lung cancer (SCLC) brain metastasis (BM), and to create and validate nomograms for OS and CSS prediction. Data from elderly SCLC BM patients were extracted out of the Surveillance, Epidemiology, and End Results database, including 1200 patients identified from 2010 and 2015 who were randomly allocated into a training set and an internal validation set at a proportion of 7:3, and 666 patients diagnosed between 2018 and 2020 as a temporal external validation set. Independent predictors for OS and CSS were determined through univariate Cox analysis, least absolute shrinkage and selection operator analysis, and multivariate Cox analysis sequentially. Nomograms for OS and CSS were constructed, and validated by the internal and temporal external validation sets. Age, N stage, chemotherapy, and liver metastasis were determined as independent predictors of OS and CSS, while radiotherapy and surgery were not. Nomograms were constructed based on these independent predictors. The results of the receiver operator characteristic curves, the areas under the curve and calibration curve demonstrated that the nomograms exhibited commendable discriminative ability and calibration. Moreover, decision curve analysis, net reclassification improvement, and integrated discrimination improvement also suggested that the nomograms possessed superior clinical usefulness and predictive capability relative to the TNM system. Prognostic nomograms for elderly patients with SCLC BM have been developed, demonstrating good performance in terms of accuracy, reliability, and practicality.

Development and validation of a machine learning-based model for post-sepsis frailty.

The development of post-sepsis frailty is a common and significant problem, but it is a challenge to predict. Data for deep learning were extracted from a national multicentre prospective observational cohort of patients with sepsis in Korea between September 2019 and December 2021. The primary outcome was frailty at survival discharge, defined as a clinical frailty score on the Clinical Frailty Scale ≥5. We developed a deep learning model for predicting frailty after sepsis by 10 variables routinely collected at the recognition of sepsis. With cross-validation, we trained and tuned six machine learning models, including four conventional and two neural network models. Moreover, we computed the importance of each predictor variable in the model. We measured the performance of these models using a temporal validation data set. A total of 8518 patients were included in the analysis; 5463 (64.1%) were frail, and 3055 (35.9%) were non-frail at discharge. The Extreme Gradient Boosting (XGB) achieved the highest area under the receiver operating characteristic curve (AUC) (0.8175) and accuracy (0.7414). To confirm the generalisation performance of artificial intelligence in predicting frailty at discharge, we conducted external validation with the COVID-19 data set. The XGB still showed a good performance with an AUC of 0.7668. The machine learning model could predict frailty despite the disparity in data distribution. The machine learning-based model developed for predicting frailty after sepsis achieved high performance with limited baseline clinical parameters.