Simple Prediction Model Research Articles

Precision psychiatry: thinking beyond simple prediction models - enhancing causal predictions.

Making informed clinical decisions based on individualised outcome predictions is the cornerstone of precision psychiatry. Prediction models currently employed in psychiatry rely on algorithms that map a statistical relationship between clinical features (predictors/risk factors) and subsequent clinical outcomes. They rely on associations that overlook the underlying causal structures within the data, including the presence of latent variables, and the evolution of predictors and outcomes over time. As a result, predictions from sparse associative models from routinely collected data are rarely actionable at an individual level. To be actionable, prediction models should address these shortcomings. We provide a brief overview of a general framework for the rationale for implementing causal and actionable predictions using counterfactual explanations to advance predictive modelling studies, which has translational implications. We have included an extensive glossary of terminology used in this paper and the literature (Supplementary Box 1) and provide a concrete example to demonstrate this conceptually, and a reading list for those interested in this field (Supplementary Box 2).

Machine learning validation of a simple prediction model for the correlation between advanced age and clinical outcomes in patients with aneurysmal subarachnoid hemorrhage.

Adverse effects of advanced age and poor initial neurological status on outcomes of patients with aneurysmal subarachnoid hemorrhage (SAH) have been documented. While a predictive model of the non-linear correlation between advanced age and clinical outcome has been reported, no previous model has been validated. Therefore, we created a prediction model of the non-linear correlation between advanced age and clinical outcome by machine learning and validated it using a separate cohort. Data of aneurysmal SAH patients treated by surgical clipping or endovascular coiling between 2003 and 2019 were obtained from the Japanese Stroke Databank (derivation cohort, n = 9,657) and "Predict for Outcome Study of Aneurysmal Subarachnoid Hemorrhage" (validation cohort, n = 5,085). Generalized additive models (GAMs) for poor outcome (modified Rankin Scale score ≥ 3 at discharge) were fitted with age transformation using spline curves for each World Federation of Neurological Societies grade. The discrimination property and calibration plot of unadjusted and adjusted models were assessed using the validation cohort. The derivation and validation cohorts included 3,610 and 3,251 patients, respectively. Regarding discrimination, areas under the receiver operating characteristic curve for the derivation and validation cohorts were 0.835 and 0.827, respectively, in the unadjusted model and 0.844 and 0.836, respectively, in the adjusted model. An unbiased correlation was confirmed between predicted and observed probabilities of poor outcomes. GAM could help visualize the correlation between age and clinical outcomes. Our prediction model can quantitatively aid in treatment decision-making and can be effective for most diseases and treatment settings. Trial Registration: UMIN Clinical Trials Registry (Date 2/22/2022/ ID, UMIN000046282 number, R000052809 URL, https//www.umin.ac.jp/ctr/index.htm) and the Japan Registry of Clinical Trials (Date 3/28/2022 /No. jRCT1060210092 URL, https//jrct.niph.go.jp/).

Validity and Utility of a Risk Prediction Model for Wound Infection After Lower Third Molar Surgery.

To externally validate a clinical prediction model for surgical site infection (SSI) after lower third molar (L3M) surgery and evaluate its clinical usefulness. We conducted a retrospective cohort study of patients who underwent L3M surgery at Hokkaido University Hospital. The study was designed to evaluate the historical and methodological transportability. Clinical usefulness was evaluated using decision curve analysis on the data of the non-antibiotic-treated patients. We obtained 2543 validation cohorts from April 2020 to March 2023, and 640 non-antibiotic cohorts from July 2010 to September 2023. The incidences of SSI after L3M surgery were 5.3% (135/2543) and 7.7% (49/640) in the validation and non-antibiotic cohorts, respectively. The discrimination ability of the prediction model was acceptable for the external validation cohort (c-statistic: 0.67; 95% CI: 0.62-0.71) and adequate for the non-antibiotic cohort (c-statistic: 0.72; 95% CI: 0.63-0.79). In both cohorts, the model showed excellent calibration between the observed and predicted probabilities. Decision curve analysis showed increased net benefit across a range of meaningful risk thresholds. A simple risk prediction model for SSI after L3M surgery demonstrated clinical transportability and usefulness. This model may help surgeons/clinicians determine the appropriateness of prophylactic antibiotics administration for patients in L3M surgery.

Development of a simple prediction model for tracheostomy requirement after surgical resection of medulloblastoma in children

ObjectivePostoperative tracheostomy is a significant complication following medulloblastoma (MB) resection. This study aimed to develop a predictive model for postoperative tracheostomy requirement in children undergoing MB surgical resection. This model was derived as a side product of a larger research project analyzing surgical outcomes in pediatric MB patients.ResultsForty-five patients (26%) required tracheostomy postoperatively. Using multivariable logistic regression, five models were developed, and the final model was selected based on performance and simplicity. The simplified version included two predictors: preoperative brainstem invasion and postoperative brainstem contusion, each contributing equally to the score. The model demonstrated an AUC of 0.845. Predicted risks of requiring a tracheostomy were 5.8%, 57.7%, and 75% for scores of 0, 1, and 2, respectively. This tool provides clinicians with a quantitative approach to assess tracheostomy risk, improving decision-making and patient management.

Validation of the 5-SNP score for the prediction of venous thromboembolism in a Danish fast-track cohort of 6,789 total hip and total knee arthroplasty patients

BackgroundVenous thromboembolism (VTE) is a serious complication following total hip arthroplasty (THA) and total knee arthroplasty (TKA). Despite improvements with fast-track treatment protocols, still 0.5% of patients develop a VTE within 90-days postoperatively. Previously, the 5-SNP genetic risk scores (weighted and simplified), were developed to identify people at a high risk of VTE within the general population. ObjectivesWe aimed to assess whether the 5-SNP scores could be used to identify high-risk patients in a cohort of fast-track THA/TKA patients. Patients/MethodsA subset of patients from the Lundbeck Centre for Fast-track Hip and Knee Replacement Database, based on the availability of genetic information, was included. Within these patients, the 5-SNP scores were calculated and their discriminatory performance was determined by c-statistic. Furthermore, the 5-SNP scores were added to a simple logistic prediction model containing clinical predictors to assess the added predictive value. Results7,753 THA and TKA procedures (6,798 patients) were included in this study. The c-statistics for the weighted and simple 5-SNP score were 0.50 (95% confidence interval [CI] 0.39 – 0.61) and 0.48 (95%CI 0.38 – 0.58), respectively. For the model with clinical predictors, the c-statistic was 0.67 (95%CI 0.56 – 0.77). Addition of either of the 5-SNP scores did not improve discrimination in this model. ConclusionsThese findings do not support genetic risk profiling in fast-track THA/TKA patients to predict VTE. Hence, efforts should be directed at optimizing prediction models with clinical predictors.

The eSAH score: a simple practical predictive model for SAH mortality and outcomes

We developed a simple quantifiable scoring system that predicts aneurysmal subarachnoid hemorrhage (aSAH) mortality, delayed cerebral ischemia (DCI), and modified Rankin scale (mRS) outcomes using readily available SAH admission data with SAH volume (SAHV) measured on computed tomography (CT). We retrospectively analyzed a cohort of 277 patients with aSAH admitted at our Comprehensive Stroke Center at Mayo Clinic in Jacksonville, Florida, between January 5, 2012, and February 24, 2022. We developed a mathematical radiographic model SAHV that measures basal cisternal SAH blood volume using a derivation of the ABC/2 ellipsoid formula (A = width/thickness, B = length, C = vertical extension) on noncontrast CT, which we previously demonstrated is comparable to pixel-based manual segmentation on noncontrast CT. Data were analyzed using t test, χ2 test, receiver operator characteristics curve, and area under the curve (AUC) analysis. Multivariate logistic regression analysis with stepwise elimination of variables not contributing to the model (0.05 significance level for entry into the model) was used to develop an enhanced SAH (eSAH) scoring system. Using multivariate logistic regression, we found that age, Glasgow Coma Scale score, and SAHV were significantly associated with mRS outcomes at discharge, in-hospital DCI, and in-hospital mortality. Using these factors, we developed a weighted eSAH score, ranging from 0 to 5, that was strongly predictive of mRS outcomes (AUC = 0.89), DCI (AUC = 0.75), and in-hospital mortality (AUC = 0.88). Our proposed eSAH score, a simple quantitative model based on SAHV, Glasgow Coma Scale score, and age, appears to predict mortality and outcomes in patients with aSAH. A larger cohort validation study is planned.

Optimisation and modelling of eddy viscosity in the resolvent analysis of turbulent channel flows

The eddy-viscosity model, as initially used to model the mean Reynolds stress, has been widely used in the linear analysis of turbulence recently by direct extension. In this study, the mechanism of the eddy viscosity in improving the prediction of fluctuation structures with linear analysis is clarified by investigating the statistical properties of forcing, eddy-viscosity term and their correlations. From the direct numerical simulation (DNS) results of turbulent channel flows with $Re_{\tau }=186$ – $2003$ , the spatial correlation of forcing is partially cancelled due to its interaction with eddy-viscosity terms. The stochastic forcing after excluding the eddy-viscosity term is nearly uncorrelated spatially, which better matches the condition where the resolvent modes are consistent with the spectral proper orthogonal decomposition (SPOD) modes from DNS. With the above findings, an optimisation framework is developed by minimising the spatial correlations of the stochastic forcing. The optimised eddy-viscosity profiles nearly overlap with the mean-quantity-based model in the near-wall region, but have different maximum values. Compared with the mean-quantity-based model, the optimised results enhance the consistency between the resolvent and DNS results significantly. Based on the optimised results, a new modelling framework is further abstracted, leaving only one to-be-modelled parameter of the maximum value of the eddy-viscosity profile. This parameter follows distinctive rules with spanwise flow scales, based on which a simplified predictive model is constructed. The resolvent modes predicted by the new model exhibit high consistency with SPOD modes, which are essentially comparable to the optimised results for wide ranges of streamwise and spanwise scales.

Impact assessment of wastewater on water quality using new quality indices and multiple linear regression

The Wastewater Quality Index (WWQI) is a unique measure that synthesizes the overall quality of wastewater. The main objective of the study was to evaluate whether the quality of wastewater reaching the treatment plant is suitable for optimal disposal using the WWQI approach. To understand the correlation between the determined indicators, Pearson correlation coefficients were calculated, so that the results of the analyses reflect the overall quality of the evaluated wastewater. Additionally, statistical analysis was employed to develop a simple prediction model through multiple linear regression (MLR) to predict WWQI based on various wastewater quality parameters. The results indicate good and regular quality of treated wastewater for safe disposal, according to the classification of WWQI levels, and the prediction model yielded values similar to those estimated.

Impact assessment of wastewater on water quality using new quality indices and multiple linear regression

The Wastewater Quality Index (WWQI) is a unique measure that synthesizes the overall quality of wastewater. The main objective of the study was to evaluate whether the quality of wastewater reaching the treatment plant is suitable for optimal disposal using the WWQI approach. To understand the correlation between the determined indicators, Pearson correlation coefficients were calculated, so that the results of the analyses reflect the overall quality of the evaluated wastewater. Additionally, statistical analysis was employed to develop a simple prediction model through multiple linear regression (MLR) to predict WWQI based on various wastewater quality parameters. The results indicate good and regular quality of treated wastewater for safe disposal, according to the classification of WWQI levels, and the prediction model yielded values similar to those estimated.

A simple predictive model for puerperal infections: emphasizing risk factors and pathogen analysis.

Puerperal infection (PI) accounting for approximately 11% of maternal deaths globally is an important preventable cause of maternal morbidity and mortality. This study aims to analyze the high-risk factors and pathogenic bacteria of PI, design a nomogram to predict the risk of PI occurrence, and provide clinical guidance for prevention and treatment to improve maternal outcomes. A total of 525 pregnant women were included in the study. The mothers were randomly divided into a training cohort (n=367) and a test cohort (n=158). The performance of our model was assessed using the area under the receiver operating characteristic (ROC) curve, calibration curve, and decision curve analyses. All the women in the group of PI underwent blood culture tests, if the bacteria were detected, drug sensitivity tests were performed. The drug sensitivity spectrum was recorded and analyzed. Univariate analysis showed that 12 indicators were significantly different (P< 0.05). Logistic regression analysis showed 6 factors, such as parity, number of vaginal examinations, amount of postpartum bleeding, antibiotics administered in one week before admission, induced labor, and indwelling catheter were significantly different between the PI group and control group (P < 0.05). The area under the ROC curve was 0.904 (95% CI: 0.871-0.936) in the training set and 0.890 (95% CI: 0.837-0.942) in the test set. The calibration curve of the nomogram showed good agreement between prediction and observation. The analysis of the clinical decision curve showed that the nomogram is of practical significance. There were 100 patients with positive blood cultures in the PI group, and Escherichia.coli was the main pathogenic bacteria, accounting for 89%. The sensitivity to Meropenem and Imipenem was 100%, to Piperacillin tazobactam 97.75%, to Ceftazidime 95.51%, and to Amoxicillin/Clavulanat (AMC) was 93.26%. The risk of PI will be significantly reduced by controlling the number of vaginal examinations less than 4 times, postpartum hemorrhage less than 414ml, and reducing the time of urethral catheter indwelling. If PI was clinically diagnosed or highly suspected, it was recommended to use antibiotics that were sensitive to Escherichia. coli, such as Piperacillin tazobactam, Ceftazidime, and AMC.

Creating simple predictive models in ecology, conservation and environmental policy based on Bayesian belief networks.

Predictive models are often complex to produce and interpret, yet can offer valuable insights for management, conservation and policy-making. Here we introduce a new modelling tool (the R package 'BBNet'), which is simple to use, and requires little mathematical or computer programming background. By using straightforward concepts to describe interactions between model components, predictive models can be effectively constructed using basic spreadsheet tools and loaded into the R package. These models can be analysed, visualised, and sensitivity tested to assess how information flows through the system's components and provide predictions for future outcomes of the systems. This paper provides a theoretical background to the models, which are modified Bayesian belief networks (BBNs), and an overview of how the package can be used. The models are not fully quantitative, but outcomes between different modelled scenarios can be considered ordinally (i.e. ranked from 'best' to 'worse'). Parameterisation of models can also be through data, literature, expert opinion, questionnaires and/or surveys of opinion, which are expressed as a simple 'weak' to 'very strong' or 1-4 integer value for interactions between model components. While we have focussed on the use of the models in environmental and ecological problems (including with links to management and social outcomes), their application does not need to be restricted to these disciplines, and use in financial systems, molecular biology, political sciences and many other disciplines are possible.

THE IMPACT OF ARTIFICIAL INTELLIGENCE (AI) ON CONSUMER ELECTRONICS

Artificial intelligence (AI) significantly impacts consumer electronics by making devices more intuitive, personalized, and efficient, enabling them to learn from user behavior, optimize performance, and automate routine tasks, ultimately enhancing the overall user experience across various applications like smart homes, wearables, and communication tools; however, concerns around data privacy and security remain crucial considerations when integrating AI into consumer electronics. AI improves the computational capabilities of consumer electronics by incorporating complex algorithms that can learn and adapt based on data. These algorithms range from simple predictive models used in personalizing user interfaces to sophisticated neural networks that enable advanced image and speech recognition.

Development and validation of a predictive model (CHASE-OSA) for preoperative assessment of moderate-to-severe pediatric obstructive sleep apnea

PurposeUndetected obstructive sleep apnea (OSA) in children increases the likelihood of perioperative respiratory complications. Current screening tools for OSA often lack sensitivity or are overly complex. This study aimed to develop and validate a simplified preoperative predictive model for moderate-to-severe pediatric OSA.MethodsThe study included children aged 1 to 18 years who underwent either polysomnography or nocturnal pulse oximetry from January 2013 to December 2020. OSA severity was categorized using these tests, and potential predictors were identified using multivariable logistic regression. The outcomes of the tests were used to create a risk-based scoring system. Internal validation was performed using bootstrapping procedures.ResultsOut of the 1,327 participants, 882 individuals (66.5%) were diagnosed with moderate-to-severe OSA. Predictors considered for developing the scoring system included Craniofacial abnormalities, adenotonsillar Hypertrophy, Age 1–5 years, Snoring > 5 nights/week, Excessive daytime sleepiness, Obesity, Stopping breathing, and Awakening during sleep (CHASE-OSA). The scoring system developed demonstrated an area under the receiver operating characteristic curve of 0.85 (95% CI: 0.83–0.88). The CHASE-OSA score, ranging from 0 to 14, classified scores < 6 as low-risk and ≥ 6 as high-risk for moderate-to-severe pediatric OSA. This cutoff demonstrated a sensitivity of 86%, specificity of 70%, and positive and negative predictive values of 85% and 71%, respectively.ConclusionThe CHASE-OSA predictive model provides a concise and user-friendly preoperative screening tool for identifying moderate-to-severe pediatric OSA. It facilitates risk assessment, enhances perioperative care optimization, and informs postoperative management planning. Further research is needed to comprehensively validate its clinical utility.

A Simple Credit Rating Prediction Model for FinTech Companies Using SMOTE and MRMR Techniques

FinTech companies have made the financial industry more efficient and have increased financial inclusion. However, it has also brought new risks to the financial system. Regulators, investors, and researchers are concerned that their financial difficulties could affect the financial system. Our study aims to delve deeper into the effectiveness of machine learning techniques in identifying early warnings of FinTech companies’ credit risk impairment. Using commonly employed accounting and market measures in the literature, we created various classifiers to predict FinTech credit ratings. Classification algorithms face a challenge when the number of observations between classes is not equivalent, affecting their performance. Due to the limited size of publicly traded FinTech stocks with an issuer-level credit rating, our database has few observations and is highly imbalanced. The results of our study show that the SMOTE oversampling technique improves the predictive power of machine learning algorithms and that feature selection algorithms such as MRMR allow the generation of less complex and easierto-understand models. Our results suggest that the KNN classification algorithm has higher accuracy in predicting FinTech’s credit ratings.

The optimal duration of continuous respiratory rate monitoring to predict in-hospital mortality within seven days of admission – A pilot study in a low resource setting

BackgroundCurrently there are no established benefits from the continuous monitoring of vital signs, and the optimal time period for respiratory rate measurement is unknown. SettingLow resource Ugandan hospital, MethodsProspective observational study. Respiratory rates of acutely ill patients were continuously measured by a piezoelectric device for up to seven hours after admission to hospital. Results22 (5.5%) out of 402 patients died within 7 days of hospital admission. The highest c-statistic of discrimination for 7-day mortality (0.737 SE 0.078) was obtained after four hours of continuously measured respiratory rates transformed into a weighted respiratory rate score (wRRS). After seven hours of measurement the c-statistic of the wRRS fell to 0.535 SE 0.078. 20% the patients who died within seven days did not have an elevated National Early Warning Score (NEWS) on admission but were identified by the 4-hour wRRS. None of the 88 patients whose average respiratory rate remained between 12 and 20 bpm throughout four hours of observation died within 7 days of admission. A simple predictive model that included the four-hour wRRS, Shock Index and altered mental status had a c-statistic for 7-day in-hospital mortality of 0.843 SE. 0.057. ConclusionFour hours of continuously measured respiratory rates was the observation period that best predicted 7-day in-hospital mortality. After four hours the discrimination of a weighted respiratory rate score deteriorated rapidly.

A nomogram for predicting nutritional risk before gastric cancer surgery.

Gastric cancer (GC) is the fourth leading cause of cancer death worldwide. Patients with GC have higher nutritional risk. This study aimed to construct a nomogram model for predicting preoperative nutritional risk in patients with GC in order to assess preoperative nutritional risk in patients more precisely. Patients diagnosed with GC and undergoing surgical treatment were included in this study. Data was collected through clinical information, laboratory testing, and radiomics-derived characteristics. Least absolute shrinkage selection operator (LASSO) regression analysis and multi-variable logistic regression were employed to construct a clinical prediction model, which takes the form of a logistic nomogram. The effectiveness of the nomogram model was evaluated using receiver operat-ing characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA). A total of three predictors, namely body mass index (BMI), hemoglobin (Hb) and radiomics characteristic score (Radscore) were identified by LASSO regression analysis from a total of 21 variables studied. The model constructed using these three predictors displayed medium prediction ability. The area under the ROC curve was 0.895 (95% CI 0.844-0.945) in the training set, with a cutoff value of 0.651, precision of 0.957, and sensitivity of 0.718. In the validation set, it was 0.880 (95% CI 0.806-0.954), with a cutoff value of 0.655, precision of 0.930, and sensitivity of 0.698. DCA also confirmed the clinical benefit of the combined model. This simple and dependable nomogram model for clinical prediction can assist physicians in assessing preoperative nutritional risk in GC patients in a time-efficient and accurate manner to facilitate early identification and diagnosis.

Simple prediction model for vitamin D deficiency in women with osteoporosis or risk factors for osteoporosis in Thailand

IntroductionIn Thailand, the assessment of vitamin D status by measuring 25-hydroxyvitamin D[25(OH)D] levels in individuals at risk for osteoporosis is constrained by limited facilities and high costs. This study aimed to create a clinical model for predicting vitamin D deficiency in women with osteoporosis or risk factors for osteoporosis. Materials and MethodsThis was a cross-sectional study of 490 women. All participants had 25(OH)D levels measured. A questionnaire was used to assess factors related to vitamin D status. Vitamin D deficiency was defined as 25(OH)D levels < 30 ng/mL. Logistic regression analyses were conducted to investigate predictors of vitamin D deficiency. In the model, odds ratios (ORs) were converted into simple scores. The optimal cutoff for women at a high risk of vitamin D deficiency was established. Internal validation was assessed using a Bootstrap. ResultsSixty percent had vitamin D deficiency. The final model for predicting vitamin D deficiency consisted of a body mass index ≥ 25 kg/m2 (OR:1.15), lack of exercise (OR:1.59), exercise 1–2 times/week (OR:1.40), sunlight exposure < 15 min/day (OR:1.70), no vitamin D supplementation (OR:8.76), and vitamin D supplementation of 1–20,000 IU/week (OR:2.31). The area under the curve was 0.747. At a cutoff of 6.6 in total risk score (range 4–13.6), the model predicted vitamin D deficiency with a sensitivity of 71.9 % and a specificity of 65.3 %. The internal validation by Bootstrap revealed a ROC of 0.737. ConclusionsIn women at risk of osteoporosis, a simple risk score can identify individuals with a high risk of vitamin D deficiency. These women could benefit from vitamin D supplementation without requiring 25(OH)D measurements.

Development of a predictive model for gastrointestinal side effects of metformin treatment in Chinese individuals with type 2 diabetes based on four randomised clinical trials.

This study aimed to build a model-based predictive approach to evaluate the gastrointestinal side effects following an initial metformin medication. The model was developed from data from four randomised clinical cohorts. A prediction model was established using integrated or simplified indicators. Ten machine learning models were used for the construction of predictive models. The Shapley values were used to report the features' contribution. Four randomised clinical trial cohorts, including 1736 patients with type 2 diabetes, were first included in the analysis. Seventy percent of participants (1216) were allocated to the training set, 15% (260) were assigned to the internal validation set and 15% (260) were assigned to the test set. The Extra Tree model had the highest area under curve (AUC) (0.87) in the validation and test set. The top five crucial indicators were blood urea nitrogen (BUN), sex, triglyceride (TG), high-density lipoprotein-cholesterol (HDL-C) and total cholesterol (TC), and these five indicators were selected for constructing a simplified predictive model (AUC = 0.76). An online web-based tool was established based on the predictive model with integrated 17 features and top five indicators. To predict gastrointestinal side effects in diabetic patients for initial use of metformin, a few easily obtained features are needed to establish the model. The model can be applied to the Chinese population in clinical practice.

Risk prediction model for psoriatic arthritis: NHANES data and multi-algorithm approach.

To develop a simplified predictive model for identifying psoriatic arthritis (PsA) in psoriasis patients. Data from the National Health and Nutrition Examination Survey (NHANES) database were analyzed, including patients with psoriasis without arthritis (PsC) or PsA. The least absolute shrinkage and selection operator, Boruta algorithm, random forest, and stepwise regression were employed to select key variables from 38 potential predictors. Logistic regression models were constructed for each combination of selected variables and evaluated using receiver operating characteristic (ROC) curves, precision-recall (PR) curves, calibration plots, Brier scores, and decision curve analysis (DCA). The study included 587 patients with psoriasis, 238 of whom had PsA. The variable combinations proposed by the Boruta algorithm exhibited the best overall performance. Key predictors in the Borutamodel included age, fasting glucose, education level, thyroid disease, hypertension, and chronic bronchitis. This model achieved area under the curve (AUC) of 0.781 (95% CI, 0.737-0.826) for the training set and 0.780 (95% CI, 0.712-0.848) for the testing set in the ROC curve analyses. The AUC values in the PR curves were 0.687 (95% CI, 0.611-0.757) and 0.653 (95% CI, 0.535-0.770), respectively. The Brier scores of 0.186 and 0.191 for the testing and training sets indicated a good fit, further supported by the calibration curves. DCA showed a net clinical benefit for decision thresholds ranging from 0.2 to 0.8 in both datasets. The Borutamodel represents a promising tool for early risk assessment of PsA. Key Points • National Database Utilization: This study leverages the NHANES database to predict psoriatic arthritis risk, addressing previous limitations tied to regional or ethnic constraints. • Comprehensive Variable Analyses: The research examines 38 variables, including demographics, health conditions, laboratory results, and lifestyle factors, using four distinct screening methods and thorough evaluations of model performance. • Innovative Risk Model: The study introduces a novel risk assessment model that integrates age, fasting glucose, education, and comorbidities including hypertension, thyroid disease, and chronic bronchitis, thus moving beyond traditional focus on skin lesions and joint symptoms.

A Lifetime Prediction Model for the Anode Catalyst Layer in PEM Electroyzers

Electrolyzer stacks (ELY stacks) are anticipated to form a key component in the hydrogen economy by providing a clean hydrogen source which may be deployed modularly. Currently, proton-exchange-membrane (PEM) electolyzers are a promising candidate to meet this demand on account of their high performance relative other electrolysis methods. Amongst other considerations, economic barriers preventing the uptake of PEM ELY stacks are the durability and cost of the precious catalyst materials, especially iridium-loaded catalysts. Additionally, due to the ever-growing uptake of intermittent power sources, dynamic operation of the ELY stack is an attractive economic option, yet may also accelerate degradation. Taken together, a key question in the design and deployment of ELY stacks is predicting the lifetime of the anode catalyst layer under dynamic power loads.In this presentation we present a simple model for lifetime prediction of the anode catalyst layer of a PEM electrolyzer with iridium dissolution as the principal degradation mechanism. The model is parameterized from data in the literature and wind and solar profiles are used as validation. Implications in the design of accelerated stress tests are also discussed.