Discriminative Ability Research Articles

Subtyping strokes using blood-based protein biomarkers: A high-throughput proteomics and machine learning approach.

Rapid diagnosis of stroke and its subtypes is critical in early stages. We aimed to discover and validate blood-based protein biomarkers to differentiate ischemic stroke (IS) from intracerebral haemorrhage (ICH) using high-throughput proteomics. We collected serum samples within 24 h from acute stroke (IS & ICH) and mimics patients. In the discovery phase, SWATH-MS proteomics identified differentially expressed proteins, which were validated using targeted proteomics in the validation phase. We conducted interaction network and pathway analyses using Cytoscape 3.10.0. We determined cut-off points using the Youden Index. We developed three prediction models using multivariable logistic regression analyses. We assessed the model performance using statistical tests. We included 20 IS and 20 ICH in the discovery phase and 150 IS, 150 ICH, and six stroke mimics in the validation phase. We quantified 375 proteins using SWATH-MS. Between IS and ICH, we discovered 20 differentially expressed proteins. In the validation phase, the combined prediction model including three biomarkers: GFAP (aOR 0.04; 95%CI .02-.11), MMP-9 (aOR .09; .03-.28), APO-C1 (aOR 5.76; 2.66-12.47) and clinical variables independently differentiated IS from ICH (accuracy: 92%, negative predictive value: 94%). Adding biomarkers to clinical variables improved discrimination by 26% (p < .001). Additionally, nine biomarkers differentiated IS from ICH within 6 h, while three biomarkers differentiated IS from mimics. Our study demonstrated that GFAP, MMP-9 and APO-C1 biomarkers independently differentiated IS from ICH within 24 h and significantly improved the discrimination ability of prediction models. Temporal profiling of these biomarkers in the acute phase of stroke is warranted.

Development of a model predicting falls in older emergency department patients using smartphone-based mobility measures.

While emergency departments (EDs) are crucial for identifying patients at risk for falls, existing fall risk measures are often inaccurate. This study aimed to assess whether iPhone sensor-based mobility measures collected after ED discharge can improve fall prediction compared with traditional ED-based screening measures. This single-center, observational cohort study recruited ED patients aged 60 or older who owned an iPhone. Participants completed baseline assessments, downloaded a custom app to track mobility measures from the iPhone, and were followed for 90 days post-discharge. Fall outcomes were self-reported via the app or follow-up phone calls. Logistic regression and the LASSO technique were employed to identify significant predictors. The discriminative ability of the models was assessed by comparing the C-statistics. Of the 149 participants enrolled, 76.5% (N = 114) provided at least 7 days of post-discharge iPhone sensor-based mobility data. The cohort had a mean age of 73 years, with 16.7% (N = 19) experiencing a fall. Participants who fell showed a significantly greater increase in daily steps over time compared with those who did not (p = 0.002). The extended logistic regression model, by incorporating mean gait asymmetry and change in step count, demonstrated a higher but nonsignificant improvement in discriminative ability (C-statistic = 0.84) compared with the base model (C-statistic = 0.79). This study demonstrates that iPhone mobility measures collected after ED discharge can enhance fall prediction relative to self-reported fall risk screening questions in older adults. The strongest mobility predictors were gait asymmetry and changes in step count. While the findings suggest that post-discharge mobility monitoring could improve fall prevention strategies, further validation in diverse populations is necessary.

Predicting Glaucoma Progression with Artificial Intelligence: A Meta-Analysis of Machine Learning Models

Introduction: Glaucoma, a leading cause of irreversible blindness, requires early detection and prediction of progression to preserve vision. Artificial intelligence (AI) offers promising tools for analyzing complex ophthalmic data and identifying high-risk individuals. This meta-analysis evaluates the performance of machine learning (ML) models in predicting glaucoma progression. Methods: A systematic search of PubMed, Scopus, and Web of Science databases was conducted for studies published between 2013 and 2024 that investigated the use of ML models to predict glaucoma progression. Studies reporting performance metrics like sensitivity, specificity, area under the receiver operating characteristic curve (AUC), and accuracy were included. Results: Six studies met the inclusion criteria, encompassing 1,250 participants. The pooled sensitivity and specificity of ML models for predicting glaucoma progression were 0.81 (95% CI: 0.78-0.84) and 0.77 (95% CI: 0.73-0.81), respectively. The pooled AUC was 0.88 (95% CI: 0.86-0.90), indicating excellent discriminatory ability. Conclusion: ML models hold significant potential for predicting glaucoma progression with high accuracy. Further research with larger, more diverse datasets is needed to validate these findings and develop clinically applicable tools.

A Prostate Imaging-Reporting and Data System version 2.1-based predictive model for clinically significant prostate cancer diagnosis.

To develop and validate a Prostate Imaging-Reporting and Data System (PI-RADS) version 2.1 (v2.1)-based predictive model for diagnosis of clinically significant prostate cancer (csPCa), integrating clinical and multiparametric magnetic resonance imaging (mpMRI) data, and compare its performance with existing models. We retrospectively analysed data from patients who underwent prospective mpMRI assessment using the PI-RADS v2.1 scoring system and biopsy at our institution between April 2019 and December 2023. A 'Clinical Baseline' model using patient demographics and laboratory results and an 'MRI Added' model additionally incorporating PI-RADS v2.1 scores and prostate volumes were created and validated on internal and external patients. Both models were compared against two previously published MRI-based algorithms for csPCa using area under the receiver operating characteristic curve (AUC) and decision curve analysis. A total of 1319 patients across internal and external cohorts were included. Our 'MRI Added' model demonstrated significantly improved discriminative ability (AUCinternal 0.88, AUCexternal 0.79) compared to our 'Clinical Baseline' model (AUCinternal 0.75, AUCexternal 0.68) (P < 0.001). The 'MRI Added' model also showed higher net benefits across various clinical threshold probabilities and compared to a 'biopsy all' approach, it reduced unnecessary biopsies (defined as biopsies without Gleason Grade Group ≥2 csPCa) by 27% in the internal cohort and 10% in the external cohort at a risk threshold of 25%. However, there was no significant difference in predictive ability and reduction in unnecessary biopsies between our model and comparative ones developed for PI-RADS v2 and v1. Our PI-RADS v2.1-based mpMRI model significantly enhances csPCa prediction, outperforming the traditional clinical model in accuracy and reduction of unnecessary biopsies. It proves promising across diverse patient populations, establishing an updated, integrated approach for detection and management of prostate cancer.

Predictive model of the surgical difficulty of robot-assisted total mesorectal excision for rectal cancer: a multicenter, retrospective study

Rectal cancer robotic surgery is becoming more and more common, but evidence for predicting surgical difficulty is scarce. Our goal was to look at the elements that influence the complexity of robot-assisted total mesorectal excision (R-TME) in the medical care of middle and low rectal cancer as well as to establish and validate a predictive model on the basis of these factors. Within this multicenter retrospective investigation, 166 consecutive patients receiving R-TME between January 2021 and December 2022 with middle and low rectal cancer were included and categorized according to the median operation time. A nomogram was created to forecast the procedure’s complexity after variables that could affect its difficulty were found using logistic regression analysis. Using R software, a total of 166 patients were randomly split into two groups: a test group (48 patients) and a training group (118 patients) at a ratio of 7 to 3. The median operation time of all patients was 207.5 min; patients whose operation time was ≥ 207.5 min were allocated to the difficult surgery group (83 patients), and patients whose operation time was < 207.5 min were allocated to the nondifficult surgery group. Multivariate analysis revealed that body mass index (BMI), the gap between the tumor and the anal verge and the posterior rectal mesenteric thickness were independent predictors of surgical duration. A clinical predictive model was created and assessed employing the above independent predictors. The results of the receiver operating characteristic (ROC) analysis revealed the adequate discriminative ability of the predictive model. Our study revealed that it is feasible to predict surgical difficulty by obtaining clinical and magnetic resonance parameters for imaging (the gap between the anal verge and the tumour, and posterior mesorectal thickness), and these predictions could be useful in making clinical decisions.

Intelligibility Sound Therapy Enhances the Ability of Speech-in-Noise Perception and Pre-Perceptual Neurophysiological Response

Aural rehabilitation with hearing aids can decrease the attentional requirements of cognitive resources by amplifying deteriorated-frequency sound in hearing loss patients and improving auditory discrimination ability like speech-in-noise perception. As aural rehabilitation with an intelligible-hearing sound also can be hopeful, the aim of this study was to evaluate the effectiveness of aural rehabilitation with intelligible-hearing sound for hearing loss patients. Adult native Japanese speakers (17 males and 23 females, 68.43 ± 9.23 years) with hearing thresholds exceeding 30 dB at any of the following frequencies: 125, 250, 500, 1000, 2000, 3000, 4000, 8000, 10,000, and 12,000 Hz in either ear, were recruited. on any side were recruited and underwent the Mini-Mental State Examination Japanese. We conducted a self-evaluation questionnaire for hearing problems of voice, a gap detection test, a fast speech test, a speech-in-noise test, a pure tone audiogram, and a speech perception test using a Japanese 67-S, cortical auditory-evoked fields, and magnetic mismatch negativity before and after the non-intelligible-hearing (N = 20) and intelligible-hearing (N = 20) sound therapy, which involved listening to music for one hour a day for 35 days. The better hearing ear was defined using a four-frequency pure-tone average at the thresholds of 500, 1000, 2000, and 4000 Hz. After the sound therapy, the speech-in-noise test with a signal-to-noise ratio +10 in the better hearing ear showed significant improvement (p &lt; 0.05), and N1m-P2m amplitudes showed a significant increase in the Lt superior temporal gyrus in response to the stimulus from the better hearing ear (p &lt; 0.05). A significant enhancement of the magnetic mismatch negativity amplitude at the Lt superior temporal gyrus was exhibited after the sound therapy (p &lt; 0.01). Intelligible-hearing sound therapy can improve the ability of speech-in-noise perception in the better hearing ear and enhancement of central cortex response, which reflects the ability of working memory, was proved by cortical auditory-evoked fields and magnetic mismatch negativity. Intelligible-hearing sound therapy can be a valuable aural rehabilitation method for sensory neural hearing loss, the same as hearing aids.

Articulatory precision from connected speech as a marker of cognitive decline in Alzheimer's disease risk-enriched cohorts.

Mild cognitive impairment (MCI) has been recognized as a possible precursor to Alzheimer's disease (AD). Recent research focusing on connected speech has uncovered various features strongly correlated with MCI due to AD and related dementias. Despite these advancements, the impact of early cognitive decline on articulatory precision has not been thoroughly investigated. This study introduced the phoneme log-likelihood ratio (PLLR) to assess the articulatory precision of speakers across different cognitive status levels and compared its effectiveness to existing well-studied acoustic features. The study utilized speech recordings from a picture description task, which included data from 324 cognitively unimpaired participants with low amyloid-β burden (CU, Aβ()), 47 cognitively unimpaired participants with high amyloid-β burden (CU, Aβ()), 69 participants with MCI, and 20 participants with dementia. Nine acoustic features were extracted from the speech recordings, covering three categories: speech fluency, speech pace, and articulatory precision. Welch's -test and Hedge's were adopted to assess their discriminative ability. A reduction in speech fluency and pace was observed among participants in the MCI and dementia groups. The PLLR shows large effect sizes in distinguishing between cognitively unimpaired participants with low Aβ burden and those diagnosed with MCI or dementia. Additionally, the test-retest reliability experiment indicated moderate repeatability of the features under study. The study reveals PLLR as a sensitive indicator capable of detecting subtle articulatory variations across groups, while also providing further support for the association between reduced articulatory precision and early cognitive decline.

EXPRESS: Alexithymia does not explain facial expression recognition difficulties across the dark triad spectrum.

Using an individual differences approach (N=236) we assessed whether trait alexithymia was able to predict unique variance in facial expression discrimination ability and facial expression labelling ability above and beyond an individual's level of dark triad traits. Alexithymia has been shown to be associated with the Dark Triad, and recent evidence has suggested that co-occurring alexithymia may explain facial expression recognition difficulties in autism spectrum disorder. Here we tested this alexithymia hypothesis for individuals on the dark triad spectrum. Results showed that autistic traits, alexithymic traits and dark triad traits all correlated with expression recognition ability. However, linear regression models showed that an individual's level of dark triad traits, their level of autistic traits, and a brief measure of general cognitive ability each predicted unique variance in facial expression discrimination and facial expression labelling ability, but an individual's level of alexithymic traits predicted no additional unique variance. We suggest that Dark Triad and Autistic traits each uniquely contribute to expression recognition ability alongside general cognitive ability.

Optimization of high-throughput marker systems for genomic prediction in alfalfa family bulks.

Alfalfa (Medicago sativa L.) is a perennial forage legume esteemed for its exceptional quality and dry matter yield (DMY); however, alfalfa has historically exhibited low genetic gain for DMY. Advances in genotyping platforms paved the way for a cost-effective application of genomic prediction in alfalfa family bulks. In this context, the optimization of marker density holds potential to reallocate resources within genomic prediction pipelines. This study aimed to (i) test two genotyping platforms for population structure discrimination and predictive ability (PA) of genomic prediction models (G-BLUP) for DMY, and (ii) explore optimal levels of marker density to predict DMY in family bulks. For this, 160 nondormant alfalfa families were phenotyped for DMY across 11 harvests and genotyped via targeted sequencing using Capture-seq with 17K probes and the DArTag 3K panel. Both platforms discriminated similarly against the population structure and resulted in comparable PA for DMY. For genotyping optimization, different levels of marker density were randomly extracted from each platform. In both cases, a plateau was achieved around 500 markers, yielding similar PA as the full set of markers. For phenotyping optimization, models with 500 markers built with data from five harvests resulted in similar PA compared to the full set of 11 harvests and full set of markers. Altogether, genotyping and phenotyping efforts were optimized in terms of number of markers and harvests. Capture-seq and DArTag yielded similar results and have the flexibility to adjust their panels to meet breeders' needs in terms of marker density.

Systematic review of prognostic models for predicting recurrence and survival in patients with treated oropharyngeal cancer

ObjectivesThis systematic review aims to evaluate externally validated models for individualised prediction of recurrence or survival in adults treated with curative intent for oropharyngeal cancer.DesignSystematic review.SettingHospital care.MethodsSystematic searches were conducted...

Evaluation of indices for the assessment and classification of keratoconus based on optical coherence tomography and Scheimpflug technology.

To compare the parameters and indices of a novel swept-source optical coherence tomography device (SS-OCT, ANTERION) with those of a rotating Scheimpflug camera (RSC)-based tomograph (Pentacam) in normal and keratoconic (KC) eyes. This prospective, monocentric, cross-sectional study included individuals with unoperated normal and KC eyes, selecting one eye per subject. Ectasia-specific parameters analysed with the SS-OCT were difference in mean keratometry (Kmean) in the inferior and superior meridians, maximum keratometry value (Kmax), elevation of the posterior surface at the thinnest point, screening corneal objective risk of ectasia (SCORE) and thinnest point thickness. With the RSC, parameters determined were Belin/Ambrosio total deviation value (BAD-D), index of height decentration and index of vertical asymmetry. KC classification with the SS-OCT was based on the anterior and posterior radii of curvature and thinnest point thickness according to the ABCD classification of the RSC system. This study included 117 individuals with healthy eyes and 335 eyes with KC. The indices with the highest diagnostic discriminatory ability between the two cohorts were SCORE, difference of Kmean in the inferior and superior meridians and posterior elevation of the thinnest point (SS-OCT), as well as the index of height decentration, index of vertical asymmetry and BAD-D (RSC). The classifications using SS-OCT defined mild-stage KC as Kmax, posterior elevation of the thinnest point and thinnest point thickness as ≤50.9 D, ≤30 and ≥472 μm, respectively. Moderate stage values were 51-55.9 D, 31-69 and 471-438 μm, respectively, while respective advanced stage were ≥56 D, ≥70 and ≤437 μm. The diagnostic capabilities for both devices were found to be comparable. KC classification using SS-OCT can be independently based on the anterior surface, posterior surface and corneal thickness.

Evaluation, Calibration, and Application of Probabilistic 100-Meter Wind Speed Forecasts Produced by the WRF Ensemble Prediction System in Taiwan

Abstract The use of renewable energy is on the rise; however, it brings more challenges for grid operations and unit scheduling due to the intermittent nature and limitations in predicting renewable power generation. High-quality meteorological forecasts play a critical role in the effective application and management of renewable energy resources. This study evaluates the quality and economic benefits of probabilistic forecasts for 100-m wind speeds from the Weather Research and Forecasting model ensemble prediction system (WEPS). A multiple linear regression (MLR) method for calibration is also used to improve forecast quality. Additionally, this study explores the application of probabilistic hub-height wind speed forecasts in conjunction with an economic value analysis to determine whether wind turbines should be shut down during typhoon periods. The findings reveal that the forecast in the central offshore region of Taiwan exhibits the highest reliability and discrimination ability compared to those in the northern and southern regions. Additionally, employing the MLR calibration method significantly improved the reliability and discrimination ability of the probabilistic forecasts compared to the raw forecasts. Furthermore, during the typhoon seasons, almost all users, regardless of their cost-loss ratio, can benefit from basing their decisions on WEPS forecasts compared to using a single deterministic forecast. Importantly, decision-makers can benefit more from probabilistic than ensemble mean forecasts when both are derived from the same ensemble prediction system, since the former incorporates forecast uncertainty.

An integrated machine learning framework for developing and validating a prognostic risk model of gastric cancer based on endoplasmic reticulum stress-associated genes

BackgroundGastric cancer (GC), a prevalent and deadly malignancy, demonstrates poor survival outcomes. Evidence has emerged indicating that disruptions in endoplasmic reticulum homeostasis are significantly implicated in the onset and progression of various oncological conditions. This study was designed to construct a prognostic model based on genes related to endoplasmic reticulum stress(ERS) to predict survival outcomes in patients with GC. MethodsExpression profiling data for GC samples were extracted and analyzed from TCGA-STAD, revealing 214 genes related to endoplasmic reticulum stress that show differential expression when compared with normal gastric tissue. Building on these insights, a prognostic model was formulated using data from TCGA-STAD and validated through subsequent analyses of GEO datasets. The tumor immune dysfunction and exclusion(TIDE) algorithm was applied to determine the susceptibility of individuals in high- and low-risk categories to immunotherapy. The presence of immune and stromal cells within the tumor microenvironment was assessed with the aid of the ESTIMATE algorithm. Sensitivity variations to prevalent anticancer drugs between the risk groups were evaluated using the Genomics of Drug Sensitivity in Cancer(GDSC) database, and prospective therapeutic agents were confirmed through molecular docking techniques. ResultsThirty-one endoplasmic reticulum stress (ERS)-related differentially expressed genes (DEGs) crucial for prognosis in GC were pinpointed. These DEGs were then used to construct a prognostic model and were considered as independent prognostic factors for GC patients. This risk model proved to have a good predictive performance for estimating the overall survival of these patients. The patients placed into the high-risk group showed worse results and lower sensitivity to immunotherapy. Moreover, five specific targeted therapy drugs, namely BMS-754807, Dasatinib, JQ1, AZD8055 and SB505124, produced better results in the treatment of the high-risk group of patients. ConclusionsA new molecular prognostic model associated with ERS was established and validated for GC and showed relatively good discriminative and predictive ability. This model greatly expands the collection of weapons in the armoury of prognostic analysis in GC.

High expression of RHOF is an effective diagnostic marker and a potential prognostic indicator for primary mediastinal large B-cell lymphoma.

Due to the overlap of histological and molecular features of primary mediastinal large B-cell lymphoma (PMBL) with other lymphomas and its low incidence, the diagnosis and prognostic assessment of PMBL pose certain challenges. This study included 51 PMBL and 375 diffuse large B-cell lymphoma, not otherwise specified (DLBCL-NOS) from the GEO database and 65 PMBL and 117 DLBCL-NOS from a single center. At the transcriptional and protein levels, RHOF expression in PMBL was significantly higher than in DLBCL-NOS (P < 0.001). ROC curve analysis suggested that high RHOF expression had a high discriminative diagnostic ability for PMBL and DLBCL-NOS (transcriptional level, AUC = 0.913-0.940; protein level, AUC = 0.878). Comparing RHOF with commonly used PMBL diagnostic markers CD23, CD30, and PD-L1, it was found that high RHOF expression is a more useful diagnostic marker for PMBL (AUC (RHOF) = 0.878 > AUC (CD23) = 0.818 > AUC (CD30) = 0.756 > AUC (PD-L1) = 0.590). The diagnostic model based on CD23, CD30, and RHOF exhibits higher sensitivity for the diagnosis of PMBL than any of the individual markers mentioned above. Concerning prognosis, high RHOF transcriptional expression was significantly associated with poorer overall survival (OS) in PMBL (P = 0.00037), while high RHOF protein expression was significantly associated with poorer OS and progression-free survival (PFS) in PMBL (P = 0.034; P = 0.034). This study indicates that high RHOF expression in PMBL is closely associated with the diagnosis, prognosis, and clinical pathological features of the disease. High RHOF expression demonstrates a superior diagnostic ability in distinguishing PMBL from DLBCL-NOS compared to existing markers. Furthermore, research on RHOF expression in PMBL holds promise for providing new targets and insights for prognosis assessment and treatment of PMBL.

A Feature Extraction Method of Ship Underwater Noise Using Enhanced Peak Cross-Correlation Empirical Mode Decomposition Method and Multi-Scale Permutation Entropy

A feature extraction method based on the combination of improved empirical modal decomposition (IEMD) and multi-scale permutation entropy (MPE) is proposed to address the problem of inaccurate recognition and classification of ship noise signals under complex environmental conditions. In order to eliminate the end effects, this paper proposes an extended model based on the principle of peak cross-correlation for improved empirical modal decomposition (EMD). In this paper, the IEMD method is used to decompose three ship underwater noise signals to extract the MPE features of the highest order intrinsic modal function (IMF) of energy. The results show that the IEMD-MPE method performs well in extracting the feature information of the signals and has a strong discriminative ability. Compared with the IEMD-aligned entropy (IEMD-PE) method, which describes the signals only at a single scale, the IEMD-MPE method achieves an improvement in the minimum difference distance ranging from 101.36% to 212.98%. In addition, two sets of highly similar ship propulsion noise signals were applied to validate the IEMD-MPE method, and the minimum differences of the experimental results were 0.0814 and 0.0057 entropy units, which verified the validity and generality of the method. This study provides theoretical support for the development of ship target recognition technology for propulsion.

Linguistic and clinical validation of the Chinese version of the Acute Cystitis Symptom Score (ACSS) questionnaire in Taiwanese women with uncomplicated acute cystitis

Purpose: The Acute Cystitis Symptom Score (ACSS) questionnaire is designed to diagnose acute cystitis in women. By quantifying the severity of symptoms, ACSS provides objective diagnostic criteria. The aim of this study is to translate the ACSS questionnaire into a Chinese version and validate its clinical use for diagnosing acute cystitis in Taiwanese women. Materials and methods: After rigorously translating the ACSS questionnaire into traditional Chinese used by the Mandarin-speaking Taiwanese people, it was clinically validated. Women aged 20 and above with suspected acute cystitis were recruited as the patients, and healthy women undergoing health check-ups as controls. Discriminative ability was assessed by comparing ACSS scores between the 2 groups, and the optimal diagnostic cutoff was determined using receiver operating characteristics analysis. In the patient group, treatment response was evaluated as patient-reported outcome by comparing ACSS scores pretreatment and posttreatment. Results: A total of 89 and 43 participants were recruited for the patient and control groups, respectively. The total score of typical symptoms between the patient and the control groups was significantly different (P &lt; 0.001). After antibiotic treatment, the total score of typical symptoms in the patient group significantly decreased (P &lt; 0.001). Using receiver operating characteristics curve analysis, the best cutoff score for diagnosing acute cystitis was 4, with a sensitivity and a specificity of 75.3% and 95.3%, respectively. Conclusion: After clinical validation, the Chinese version of the ACSS questionnaire can now be used as a symptom-oriented diagnostic and patient-reported outcome tool for acute cystitis in the Mandarin-speaking female population in Taiwan.

Development of a Brief, Positively Framed Care Team Experience Measure.

Measures that quantify clinician fulfillment and engagement with work (clinician experience measures) have the potential to improve the quality, effectiveness, and enjoyment of patient care. Many existing measures of clinician fulfillment and engagement with work are relatively long and potentially burdensome, negatively framed, or address personal well-being. A measure with a small number of items that address positive attributes to which a team aspires may be a more useful measure to guide organizational improvement efforts. In a series of cross-sectional studies we asked: (1) Can we develop a brief, positively framed, and team-oriented care team experience measure (CTEM)? (2) How does the CTEM perform relative to the Team Climate Inventory (TCI)? In a first step, the clinical leadership group of a nascent multispecialty group practice curated 21 items, some from existing engagement and burnout measures and others based on suggestions from clinical leaders. After pilot testing and interviews with specific clinicians, these items were then administered to all clinicians in the practice, and 44% (120 of 274) participated. Factor analysis identified three groups of items (factors). In a second step, 43% (493 of 1135) of patient-facing employees (75% [368] of whom were women and 45% [222] of whom were ages 42 to 61 years) of a statewide musculoskeletal specialty practice rated 12 items (two from each of the factors identified in the first stage and six new items recommended by administrative leaders) and completed the TCI. Participants were then randomly allocated to a learning cohort (70% [343 participants]), in which we performed factor analysis and item response theory to develop a new CTEM and validation cohort (30% [150 participants]) to test the consistency of the findings. The best-performing items in terms of factor loadings (the degree to which an individual item represents an identified factor or "theme"; we chose a minimum of 0.4 as items above this threshold are generally regarded as stable), item discrimination (ability to measure different experiences among participants), item difficulty (the ability of an item to contribute to variation in scores), and model fit (the degree to which the estimates of the statistical model align with the observed data) were selected to form a new three-item CTEM. The CTEM was subsequently evaluated for internal consistency (Cronbach alpha, which measures the extent to which different subsets of the included items would provide the same measurement), floor and ceiling effects (the percentage of the lowest and highest possible scores, with high percentages indicating that information is lost because of the inability of a tool to measure the extremes of variation), and correlation with the TCI. Factor analysis identified two groups of items representing "effectiveness" (nine items, including "I am proud of the work we do") and "collaboration" (three items, including "Our team encourages everyone's input before making changes"). The best-performing items of "effectiveness" (two items) and "collaboration" (one item) were selected to form the three-item CTEM, which had good internal consistency (Cronbach α = 0.77; > 0.7 generally considered acceptable); a notable ceiling (37%; higher than desired, common with experience measures, and-in our opinion-acceptable for routine use given the trade-off of limited participant burden) and limited floor effect (1.3%); strong correlation with the TCI (Spearman ρ 0.77 [95% confidence interval 0.70 to 0.85]), which supports convergent validity (part of construct validity, meaning correlation with measures of related concepts); and similar correlations with participant characteristics compared with the TCI, supporting construct and criterion validity (association with real-world outcomes). A small number of positively framed, team-oriented questions (the CTEM) can quantify clinician and care team experience and provide a practical and feasible method for health organizations to monitor and improve the experience of their clinicians. Routine measurement of organizational climate at regular intervals using a brief, user-friendly, team-oriented measure might help organizations improve their culture and help their clinicians feel appreciated and engaged.

Customer Churn Prediction Approach Based on LLM Embeddings and Logistic Regression

Nowadays, predicting customer churn is essential for the success of any company. Loyal customers generate continuous revenue streams, resulting in long-term success and growth. Moreover, companies are increasingly prioritizing the retention of existing customers due to the higher costs associated with attracting new ones. Consequently, there has been a growing demand for advanced methods aimed at enhancing customer loyalty and satisfaction, as well as predicting churners. In our work, we focused on building a robust churn prediction model for the telecommunications industry based on large embeddings from large language models and logistic regression to accurately identify churners. We conducted extensive experiments using a range of embedding techniques, including OpenAI Text-embedding, Google Gemini Text Embedding, bidirectional encoder representations from transformers (BERT), Sentence-Transformers, Sent2vec, and Doc2vec, to extract meaningful features. Additionally, we tested various classifiers, including logistic regression, support vector machine, random forest, K-nearest neighbors, multilayer perceptron, naive Bayes, decision tree, and zero-shot classification, to build a robust model capable of making accurate predictions. The best-performing model in our experiments is the logistic regression classifier, which we trained using the extracted feature from the OpenAI Text-embedding-ada-002 model, achieving an accuracy of 89%. The proposed model demonstrates a high discriminative ability between churning and loyal customers.

Nomogram for the prediction of relapse factors in patients with neuromyelitis optica spectrum disorder during rituximab treatment.

Develop a nomogram to analyse the factors influencing the relapse of neuromyelitis optica spectrum disorder (NMOSD) during rituximab (RTX) treatment. A retrospective analysis of 214 NMOSD patients identified 181 with AQP4-IgG-seropositive. 32 patients who relapsed during RTX treatment were included, and 122 sets of lymphocyte subset monitoring data were collected. 110 sets of data were finally included and divided into relapse (n = 30) and nonrelapse (n = 80) groups depending on whether a relapse occurred between two adjacent RTX treatments. Logistic and LASSO regressions were used to identify the relevant factors influencing NMOSD relapse, and a nomogram was constructed. Receiver operating characteristic (ROC) curves were generated to evaluate the nomogram's ability to differentiate, and the bootstrap method was utilized for internal validation. Calibration curve and decision curve analysis were also conducted. Comparing baseline data revealed differences in the RTX administration interval, CD3-CD19+ B lymphocyte and CD3-CD56+ NK cell levels. The RTX administration interval and the level of CD3-CD19+ B lymphocytes were independent variables influencing relapse. The nomogram had an area under the curve (AUC) of 0.71 and a 95% confidence interval (CI) of 0.58-0.83. The Hosmer-Lemeshow (H-L) goodness-of-fit test yielded a χ2 = 11.80 (p = 0.16). Decision curve analysis revealed that the model provided greater net benefits within the threshold probability range of 0.18-0.98. The nomogram developed in this study showed that the RTX administration interval and CD3-CD19+ B lymphocyte levels independently influence NMOSD relapse, indicating good discriminative ability, consistency, and clinical benefits.

A novel nomogram prediction model for postoperative atrial fibrillation in patients undergoing laparotomy

BackgroundPostoperative atrial fibrillation (POAF) is an ordinary complication of surgery, particularly cardiac surgery. It significantly increases in-hospital mortality and costs. This study aimed to establish a nomogram prediction model for POAF in patients undergoing laparotomy. The model is expected to identify individuals at a high risk of POAF before surgery in clinical practice.MethodsA retrospective observational case–control study involving 230 adult patients (60 patients with POAF, 120 patients in the control group, and 50 patients in the validation group) who underwent laparotomy was retrieved from two hospitals. Independent risk variables for POAF were investigated using logistic regression and the least absolute shrinkage and selection operator (LASSO) regression analysis. Subsequently, a nomogram model for POAF was constructed by multivariate logistic regression equations. The prediction model was internally validated by bootstrap method and externally validated with the validation group data. To assess the discriminative ability of the nomogram model, a receiver operating characteristic (ROC) curve was generated and a calibration curve was employed to assess the concentricity between the model’s probability curve and the ideal curve. Subsequently, decision curve analysis (DCA) was performed to assess the clinical effectiveness of the model.ResultsC-reactive protein (CRP), lymphocyte-to-monocyte ratio(LMR), blood urea nitrogen (BUN), and Macruz index were independent risk variables for POAF in patients who underwent laparotomy. A user-friendly and efficient prediction nomogram was visualized using R software. This nomogram exhibited strong discrimination, as evidenced by an area under the ROC curve (AUC) of 0.90 (95% CI 0.8509–0.9488) for the training set, 0.86 (95% CI 0.7142–1) for the test set, and 0.9792 (95% CI 0.9293–1) for the validation group data. The C-index of the bootstrap nomogram model was 0.8998. Furthermore, DCA revealed that this model displayed excellent fit and calibration, as well as positive net benefits.ConclusionsA nomogram prediction model was constructed for POAF in patients who underwent abdominal surgery. The nomogram prediction model is expected to identify individuals at high risk of POAF in clinical practice for prophylactic therapeutic intervention prior to surgery.