Logistic Regression Method Research Articles

Predictive modeling of preoperative acute heart failure in older adults with hypertension: a dual perspective of SHAP values and interaction analysis.

In older adults with hypertension, hip fractures accompanied by preoperative acute heart failure significantly elevate surgical risks and adverse outcomes, necessitating timely identification and management to improve patient outcomes. This study aims to enhance the early recognition of acute heart failure in older hypertensive adults prior to hip fracture surgery by developing a predictive model using logistic regression (LR) and machine learning methods, optimizing preoperative assessment and management. Employing a retrospective study design, we analyzed hypertensive older adults who underwent hip fracture surgery at Hebei Medical University Third Hospital from January 2018 to December 2022. Predictive models were constructed using LASSO regression and multivariable logistic regression, evaluated via nomogram charts. Five additional machine learning methods were utilized, with variable importance assessed using SHAP values and the impact of key variables evaluated through multivariate correlation analysis and interaction effects. The study included 1,370 patients. LASSO regression selected 18 key variables, including sex, age, coronary heart disease, pulmonary infection, ventricular arrhythmias, acute myocardial infarction, and anemia. The logistic regression model demonstrated robust performance with an AUC of 0.753. Although other models outperformed it in sensitivity and F1 score, logistic regression's discriminative ability was significant for clinical decision-making. The Gradient Boosting Machine model, notable for a sensitivity of 95.2%, indicated substantial capability in identifying patients at risk, crucial for reducing missed diagnoses. We developed and compared efficacy of predictive models using logistic regression and machine learning, interpreting them with SHAP values and analyzing key variable interactions. This offers a scientific basis for assessing preoperative heart failure risk in older adults with hypertension and hip fractures, providing significant guidance for individualized treatment strategies and underscoring the value of applying machine learning in clinical settings.

The relative data hungriness of unpenalized and penalized logistic regression and ensemble-based machine learning methods: the case of calibration.

Machine learning methods are increasingly being used to predict clinical outcomes. Optimism is the difference in model performance between derivation and validation samples. The term "data hungriness" refers to the sample size needed for a modelling technique to generate a prediction model with minimal optimism. Our objective was to compare the relative data hungriness of different statistical and machine learning methods when assessed using calibration. We used Monte Carlo simulations to assess the effect of number of events per variable (EPV) on the optimism of six learning methods when assessing model calibration: unpenalized logistic regression, ridge regression, lasso regression, bagged classification trees, random forests, and stochastic gradient boosting machines using trees as the base learners. We performed simulations in two large cardiovascular datasets each of which comprised an independent derivation and validation sample: patients hospitalized with acute myocardial infarction and patients hospitalized with heart failure. We used six data-generating processes, each based on one of the six learning methods. We allowed the sample sizes to be such that the number of EPV ranged from 10 to 200 in increments of 10. We applied six prediction methods in each of the simulated derivation samples and evaluated calibration in the simulated validation samples using the integrated calibration index, the calibration intercept, and the calibration slope. We also examined Nagelkerke's R2, the scaled Brier score, and the c-statistic. Across all 12 scenarios (2 diseases × 6 data-generating processes), penalized logistic regression displayed very low optimism even when the number of EPV was very low. Random forests and bagged trees tended to be the most data hungry and displayed the greatest optimism. When assessed using calibration, penalized logistic regression was substantially less data hungry than methods from the machine learning literature.

Association between dietary iron intake and hyperuricemia in U.S. adults: a cross-sectional study

BackgroundElevating blood iron levels may increase the risk of hyperuricemia. However, the association between dietary iron intake and hyperuricemia remains unclear. This study examines the association between dietary iron and hyperuricemia.MethodsThe data for this cross-sectional study came from the 2011–2020 National Health and Nutrition Examination Survey (NHANES), which included adults aged 20 and above in the United States. Detailed information on their demographics, dietary iron intake, serum uric acid, and other variables were gathered. This study employs a multifactorial logistic regression method to assess the relationship between dietary iron intake and hyperuricemia. Additionally, it utilizes Restricted Cubic Spline analysis to determine whether there is a non-linear relationship between dietary iron and hyperuricemia. Furthermore, sub-group analyses are conducted to further validate the robustness of this correlation.ResultsThis study included 11,441 patients, with a prevalence of hyperuricemia at 14.96%. Regardless of whether confounding factors were adjusted for, there was a positive correlation between dietary iron intake and hyperuricemia. In Model III, after adjusting for age, gender, race, educational background, body mass index, hypertension, carbohydrate intake, dietary fiber intake, total saturated fatty acid intake, vitamin K intake, vitamin C intake, marital status, poverty income ratio, smoking status, drinking status, work activity, diabetes, sodium intake, energy intake, protein intake, zinc intake, copper intake, selenium intake and total sugars intake, we found that compared to the lowest quintile of dietary iron intake (Q1, ≤ 8.03 mg/day), the odds ratios (ORs) for hyperuricemia in the second (Q2, 8.04–11.07 mg/day), third (Q3, 11.08–14.27 mg/day), fourth (Q4, 14.28–19.33 mg/day), and fifth quintiles (Q5, ≥ 19.34 mg/day) were 1.05 (95% confidence interval: 0.79–1.40), 1.30 (95% confidence interval: 0.99–1.73), 1.39 (95% confidence interval: 1.02–1.89), and 1.36 (95% confidence interval: 0.95–1.97), respectively. The study results indicate that a higher dietary iron intake is significantly positively associated with the likelihood of hyperuricemia.ConclusionIn the adult population of the United States, there is a positive correlation between dietary iron intake and hyperuricemia, with higher dietary iron intake associated with an increased likelihood of hyperuricemia.

RELATIONSHIP BETWEEN BENIGN PROSTATIC HYPERPLASIA AND METABOLIC SYNDROME

Introduction: Factors that may increase the risk of Benign Prostatic Hyperplasia (BPH), and one theory suggests a possible connection between metabolic syndrome and the development of BPH. This study aims to investigate the potential relationship between metabolic syndrome and the development of BPH. Methods: This study analyzed medical records from Saiful Anwar Hospital in Malang from 2015 to 2020, using a case-control research design. It applied Chi-Square and Logistic Regression methods for data analysis. Results: The study, based on data from 90 patients, found that the highest number of BPH cases (75.56%) occurred in those with hypertension. Factors like body mass index and blood pressure were significantly associated with BPH (p-values = 0.046; OR = 2.473). Logistic regression confirmed the impact of fasting blood glucose, blood pressure, and body mass index categories on BPH. Conclusion: This research illustrates that there is an association between type 2 diabetes, hypertension, and obesity with the occurrence of BPH. However, there is no statistically significant relationship between the components of metabolic syndrome and the occurrence of BPH.

A goodness-of-fit measure for logistic regression under separation

Logistic regression models have a severe problem called separation. The maximum likelihood estimator does not exist in logistic regression models for data structures under separation. Under separation, the forcibly estimated maximum likelihood estimate may have an extremely large value. Separation often occurs when the size of dataset is small. Consequently, goodness-of-fit measures based on the likelihood ratio and those based on covariance functions using the maximum likelihood estimate indicate that the model is excessively good regardless of the cause of the separation. The Firth and exact logistic regression methods are valid estimation methods for separation problems. Therefore, we propose methods to reasonably evaluate the goodness-of-fit measures of statistical models under separation with dataset of a small sample size with the abovementioned methods. The goodness-of-fit measures based on covariance functions which are a generalization of the multiple correlation coefficient, referred to as the regression correlation coefficient and the entropy coefficient of determination are then used combined with the abovementioned methods for the separation data. In addition, we conducted a data analysis using the definition of the non separation ratio based on the regression depth.

Digitalization of cotton farming in the municipality of Banikoara in Northwestern Benin

The widespread use of mobile telephony and the internet constitutes an asset for digitalization in the agricultural sectors in Africa. Cotton being the leading export crop in Benin, this study aims to understand the determinants of the adoption of digitalization (use of digital machines, membership of a social network for information exchange, use of a mobile money account) by cotton producers in the municipality of Banikoara. In this context, a socio-economic survey was carried out with a sample of 314 producers, obtained by purposive sampling. The binary logistic regression method made it possible to identify the factors affecting the adoption of digitalization. Thus, the level of banking, the use of labor, the area of cotton sown and knowledge of agricultural information and exchange platforms had a significant impact on the use of digital machines in agriculture. Membership of a social network for information exchange between producers was influenced by the level of banking, the type of cotton grown, the strengthening of relationships with others, the risk linked to the use of digitalization and the assessment of the level of security of digitization by the producers. Finally, the level of banking, the exercise of a secondary activity and the use of labor were the significant variables in the adoption of mobile money by cotton farmers in Banikoara. However, the use of more advanced technologies such as drones and sensors was not yet a reality for these producers. This information is very useful for any project to popularize and promote these new technologies in the modernization of agriculture.

Does inflammation matter in cardiogenic shock?

Abstract Background The current Society for Cardiovascular Angiography and Interventions (SCAI) shock stage classification does not include the systemic inflammatory response to risk stratify patients with cardiogenic shock. Purpose To evaluate the prognostic impact of the inflammatory response, on top of the SCAI shock classification, during cardiogenic shock using C-Reactive protein (CRP). Methods We analyzed CRP levels and SCAI shock classification in consecutive patients admitted for cardiogenic shock between January 2010 and December 2020, included in the ACTION-SHOCK registry of our tertiary center. Patients with ongoing sepsis or other extracardiac causes for elevated CRP were excluded. Using logistic regression method, we assessed both the association between CRP levels at admission and all-cause mortality at 90 days, and the value of the addition of baseline CRP level to SCAI stage of shock for the prediction of 90-day mortality. Results Among 755 patients admitted with cardiogenic shock (median age 61 years, interquartile range [IQR] 50-72, 27% of female), only 128 (17%) had normal values of CRP (< 5mg/L) at admission. Median CRP at admission was 19mg/l (IQR 6-55). There was an association between 90-day mortality and baseline CRP (Odds Ratio [OR] = 1.02, 95% Confidence Interval [95%CI] 1.01-1.03 per 5mg/L of CRP, p=0.004), still significant after adjustment on clinical and biological variables associated with 90-day mortality in univariate analysis, including the SCAI shock stage (adjusted OR 1.02, 95%CI 1.01 - 1.03 per 5 mg/L, p=0.042). The addition of CRP increased the accuracy of the SCAI shock stages for the prediction of 90-day mortality (Area under the Curve increased from 0.59, 95% CI 0.50 - 0.66 to 0.65, 95% CI 0.61 - 0.72). Interestingly, the association of CRP with mortality was driven by patients in SCAI shock classes A and B (figure). Conclusions CRP elevation was frequently observed in this population of patients with any cause of cardiogenic shock and was associated with 90-day mortality. This was particularly evident for patients in lower SCAI shock stages on admission.Figure 1

Toothbrushing habits are associated with self perceived stress

Abstract Background Personal self-care is one of the most important factors in maintaining and improving oral health. Public health guidelines are recommending tooth brushing at least twice daily, once before bed time and at least on one occasion. On the other hand, our regular routines and caring for our own health is influenced by different factors including stress. We wanted to research how much does daily stress influencing tooth brushing habits. Methods We analyzed data from 2020 CINDI nationwide cross-sectional survey in Slovenia. Tooth brushing once a day or less was assessed as inadequate. Stress was assessed with question >>How often in the last 14 days have you felt tense, stressed or under a lot of pressure? << Participants who chose often or always as an answer were assessed as persons under stress. Method of binary logistic regression was used for statistical analysis of association between tooth brushing frequency and self-perceived stress, modified for some socioeconomic factors. Results Prevalence of inadequate frequency of tooth brushing was present in 31,6 % of sample, 31,5% in participants who were not feeling tense and 31,8% in participants who often felt tense or stressed. At first it seemed stress is not associated with tooth brushing habits. We adjusted this association for gender, education level and age and this association became clearer. Participants who felt stress and/or under pressure had higher odds for poor tooth brushing habits, their odds ratio was 1,283 (p < 0,001; 95% CI 1,124; 1,464) when adjusted for aforementioned factors. Conclusions Oral health is integral part of general health. Association between oral and mental health is often overlooked even though psycho-social aspects are part of modern oral health definition. As personal self-care is cornerstone of maintaining and improving oral health, implementation of mental health aspects should be incorporated in public health advises and programmes for promoting oral health. Key messages • Self-perceived stress is associated with toothbrushing habits. • People who are under stress less regularly brush their teeth.

The intra-tumoral microbiome as a potential biomarker of response to external beam radiation therapy in cervical cancer

BackgroundWe aimed to determine the potential predictive value of the intra-tumoral microbiome as a marker of the response to external beam radiation therapy (EBRT) in cervical cancer (CC).MethodsA prospective longitudinal trial of 36 CC patients receiving pelvic radiotherapy was designed to investigate microbial characteristic signatures and diversity (alpha and beta) of multiple sites (tumor, vaginal, gut, urethral, and oral) in the superior response (SR) and inferior response (IR) groups of CC patients by 16S rRNA sequencing. Utilized the least absolute shrinkage and selection operator (LASSO) logistic regression method to analyze clinicopathological factors that potentially influenced the efficacy of EBRT. LEfSe analysis highlighted the microbiome features that best distinguished the categorized patient samples. Selected parameters were validated with Spearman correlation analysis, receiver operating characteristic (ROC) area under the curve (AUC) analysis and Kaplan-Meier survival analysis.ResultsFirstly, in our cohort, LASSO logistic regression analysis revealed no association between clinicopathological factors and EBRT efficacy. Subsequently, we employed 16S rRNA sequencing to compare microbiome differences across multiple sites and their correlations with major clinicopathological factors. We discovered that the intra-tumoral microbiome was independent of clinicopathologic features and represented the most direct and reliable reflection of the microbial differences between the SR and IR groups. We found lower alpha diversity in the tumor microbiome of SR group and identified the most relevant microbiome taxa (Bifidobacteriaceae, Beijerinckiaceae, and Orbaceae) associated with the efficacy of the response to EBRT in CC patients. We then conducted ROC analysis, finding that specific microbial taxa had an AUC of 0.831 (95% CI, 0.667–0.995), indicating the potential of these taxa as biomarkers for predicting EBRT efficacy. Kaplan-Meier survival analysis showed a better prognosis for patients with lower alpha diversity and higher relative abundance of Bifidobacteriaceae.ConclusionsOur data suggested that intra-tumoral specific microbiome taxa and lower alpha diversity may play an important role in the CC patient sensitivity to EBRT and offer novel potential biomarkers for predicting the response to EBRT efficacy.

Relationship Between Gross Motor Function and Nutritional Status of Children with Cerebral Palsy-A cross-sectional study in a Tertiary Care Hospital of Bangladesh

Background: Cerebral Palsy (CP) is one of the most common causes of chronic disability in children. Childhood malnutrition is a major public health issue in low- and middle-income countries. The gross motor function of children with CP can easily be assessed by the tool Gross Motor Function Classification System- Expanded &Revised (GMFCS-E&R). This study aimed to determine the relationship between GMFCS and nutritional status in children with cerebral palsy. Method: This cross-sectional observational study was conducted in the Department of Pediatric Neurology, IPNA, BSMMU from October 2019 to August 2020. All the patients aged 2-12 years were diagnosed as CP by clinical criteria and classified by GMFCS. Nutritional status (weight for age, height for age, weight for height, BMI) is assessed and classified according to the World Health Organization growth charts. Then linear and multinomial logistic regression methods to determine the associations between GMFCS and nutritional status. The correlation between motor function and nutritional status was assessed. Results: Among the 50 children 68% were male and 32% were female. The mean age of the CP children was 4.90±3.05 years. Most patients (66.0%) came from poor families followed by 24% from the middle class. Functionally majority of patients were in GMFCS level IV (28.0% ) and level V (20.0%). Maximum patients in spastic quadriplegia 42%, spastic hemiplegia 16%. The level of GMFCS was associated with nutritional status, as determined by anthropometry. Moderate and severe underweight significantly related with GMFC level IV and V. Moderate and severe stunted was significantly related with IV and V GMFC level. Moderate wasted was significantly related to GMFC levels III and IV. Mild malnutrition was significantly with GMFC level II and III. Severe malnutrition is significantly related to GMFC level IV and V. Conclusion: This study showed majority of cerebral palsy cases had severe functional gross motor disability and also severe grade of malnutrition. A significant association between severe grade of malnutrition and severe grade of motor disability was found. J Bangladesh Coll Phys Surg 2024; 42: 342-349

Analysis of perinatal outcomes in pregnant women with the resolution of placenta previa in the second trimester

Objective: To investigate perinatal outcomes of pregnant women with the resolution of placenta previa in the second trimester. Methods: This study was a retrospective cohort study, which included singleton pregnant women who received prenatal care and delivered at Peking University Third Hospital from January 1st 2019 to December 31st 2020. A total of 403 pregnant women diagnosed with placenta previa by ultrasound at 20-24 weeks of gestation and the placental position returned to normal before delivery were included in the study group, and 403 pregnant women with normal placental position matched 1∶1 were the control group. The primary outcome was postpartum hemorrhage rate, and secondary outcomes included postpartum bleeding volume, severe hemorrhage complications, blood transfusion, drug application, the application of instrument or surgical hemostasis measures, gestational week of delivery, and neonatal outcomes. The perinatal outcomes were analyzed by univariate and multivariate logistic regression methods. Results: (1) Compared with the control group, the incidence of postpartum hemorrhage [10.4% (42/403) vs 17.6% (71/403)], postpartum hemorrhage volume (median: 375 vs 400 ml), the proportion of postpartum hemorrhage≥500 ml [18.6% (75/403) vs 30.5% (123/403)], and the proportion of application of instrument or surgical hemostasis measures [1.7% (7/403) vs 4.5% (18/403)] in the study group were increased, and the differences were statistically significant (all P<0.05). Neonatal outcomes, including birth weight, small for gestational age, hospitalization in neonatal intensive care unit and incidence of neonatal asphyxia, were compared between the two groups, and there were no statistically significant differences (all P>0.05). (2) In pregnant women with vaginal delivery, the postpartum hemorrhage rate [31.7% (66/208) vs 17.5% (39/223)], postpartum hemorrhage volume (median: 390 vs 380 ml), the proportion of instrument or surgical hemostasis measures [3.8% (8/208) vs 0.4% (1/223)] of the study group were higher than those of the control group, and the differences were statistically significant (all P<0.05). There was no significant difference in the rate of postpartum hemorrhage between the study group and the control group who gave birth by cesarean setion (P=0.545), but the proportion of postpartum hemorrhage≥500 ml in the study group and the control group were 29.2% (57/195) and 20.0% (36/180), and the difference was statistically significant (P=0.039). (3) The results of multivariate analysis showed that compared with the control group, the risk of postpartum hemorrhage (aOR=2.042, 95%CI: 1.313-3.175), the application of drugs (aOR=1.684, 95%CI: 1.142-2.484) and the application of instruments or surgical hemostasis measures (aOR=2.696, 95%CI: 1.089-6.675) were significantly increased in the study group (all P<0.05). Among women who delivered vaginally, the risk of postpartum hemorrhage in the study group was 2.021 times greater than that in the control group (95%CI: 1.269-3.220; P=0.003). Conclusion: In women with placental previa in the second trimester of pregnancy, even if the placental position returns to normal before delivery, it is still a high risk factor for postpartum hemorrhage, especially in vaginal delivery.

Nomogram predicts cervical lymph node metastasis of pathological subtypes of papillary thyroid carcinoma.

Pathological subtypes of papillary thyroid carcinoma (PTC) are important factors in thyroid cancer. Some rare subtypes exhibit extensive lymph node metastasis. These pathological subtypes should receive more attention in clinical practice. Patients with different pathological subtypes of PTC were selected from the SEER database. Logistic regression, random forest, and bootstrap aggregating (bagging) methods were employed to screen for risk factors associated with cervical lymph node metastasis in the training cohort. A nomogram was established based on the model with the largest area under the curve (AUC) and evaluated using calibration curves. Decision curve analysis (DCA) was used to evaluate the clinical benefit to patients. The nomogram was validated in depth by 200 iterations of tenfold cross-validation. A total of 7,882 patients were included in the analysis, with 5,516 patients in the training group and 2,366 patients in the testing group. The logistic regression model achieved the highest AUC of 0.7396. Sex, age, race, extension (extrathyroidal extension), pathological type, and primary tumour size were identified as independent risk factors for cervical lymph node metastasis (p < 0.05). The calibration curve indicated that the model was well calibrated. DCA indicated that the nomogram model had good clinical practicability. In clinical practice, it is important to consider the pathological subtypes of PTC. The established nomogram can serve as a predictive tool for assessing cervical lymph node metastasis.

Construction and Evaluation of a Model for the Prediction of Risk Factors Associated with Severe Pneumonia in Patients with Positive Hepatitis B Core Antibody and COVID-19 Infection in China

Background: Hepatitis B virus (HBV) infection is a high-risk factor for severe COVID-19 cases, leading clinicians to enhance the monitoring of patients with concurrent HBV and COVID-19 infections. However, this focus is typically on patients with active HBV infection. Patients who are HBV surface antigen (HBsAg) negative and HBV core antibody (anti-HBc) positive are often considered to have either a past infection that has naturally cleared or to be in a low-replication state, and thus receive less clinical attention. However, in immunosuppressed states, such as during immunosuppressive treatment following COVID-19 infection, the virus in these patients may reactivate. In such cases, this reactivation can increase the risk of COVID-19 progressing to a severe illness. Therefore, being HBsAg negative and anti-HBc positive could be a potential risk factor for severe COVID-19. Studying the combination of anti-HBc positivity and COVID-19 infection can help identify high-risk populations, allowing for the implementation of targeted prevention and management measures, thereby reducing the occurrence of severe COVID-19 cases. Objectives: To establish and evaluate a model for predicting severe pneumonia in patients with positive anti-HBc combined with COVID-19 infection. Methods: We retrospectively enrolled 380 patients who tested positive for anti-HBc, negative for HBsAg, and HBV e-antigen (HBeAg), combined with COVID-19 infection, in our hospital from December 2022 to May 2023. Based on the inclusion criteria, 163 patients were included in the study. We applied the Lasso binary logistic regression model to optimize feature selection, identifying eight non-zero coefficients using a minimum of one standard error. Using the multiple logistic regression method with backward selection, we screened six factors from the eight selected by the Lasso binary logistic regression model. These six factors were used to construct the predictive model and a nomogram. The validity of our nomogram was assessed using the area under the receiver operating characteristic curve (AUC), calibration curve, Hosmer-Lemeshow (HL) test, and decision-curve analysis (DCA). Results: Hypertension, diabetes, decreased absolute lymphocyte count, prolonged prothrombin time, elevated aspartate aminotransferase, and decreased albumin are high-risk factors for severe pneumonia in patients with positive anti-HBc combined with COVID-19 infection. The AUC of the predictive model constructed using these six factors is 0.785, with a 95% confidence interval of (0.709-0.862). The HL test, performed using the calibration curve, yielded a p-value of 0.868. The application of this diagnostic curve will increase the net benefit when the threshold probability is between 5% and 75%. Conclusions: The constructed nomogram can be used to predict the risk of patients with positive anti-HBc combined with COVID-19 infection progressing to severe pneumonia, based on routine blood parameters, liver function, coagulation, lactate dehydrogenase levels, and the patient's underlying disease. The predictive model demonstrates good discrimination, calibration, and clinical utility.

Machine learning and multisource data analysis approach towards traffic accident risk prediction

Traffic accidents are continually a concern across the world bringing with them huge economic costs and high rates of fatality and injuries each single year. Predictive modeling can lessen such risks if the accident hotspots and timing can easily be ascertained. This study collects various data including historical accident reports, water and meteorological conditions, road characteristics, and telematics data, to form traffic accident risk predictive models. To construct and analyze these predictive models, we use various machine learning techniques such as logistic regression, decision trees, and ensemble methods (Random Forest and Gradient Boosting). The results bear out as efficient and effective, ensemble models are streets ahead than regression-based techniques with Gradient Boosting achieving AUC-ROC scores of up to 0.89. The results of this research can also assist in applying effective safety measures in a more cost-effective manner as well as facilitate the development of efficient transport management systems and policies.

Predicting Clinical Outcomes in COVID-19 and Pneumonia Patients: A Machine Learning Approach.

In the clinical diagnosis of pneumonia, particularly during the COVID-19 pandemic, individuals who progress to a critical stage requiring mechanical ventilation are classified as mechanically ventilated critically ill patients. Accurately predicting the discharge outcomes for this specific cohort, especially those with COVID-19, is of paramount clinical importance. Missing data, a common issue in medical research, can significantly impact the validity of analyses. In this work, we address this challenge by employing two missing data imputation techniques: multiple imputation and missForest, to enhance data completeness. Additionally, we utilize the smoothly clipped absolute deviation (SCAD) penalized logistic regression method to select significant features. Our real data analysis compares the predictive performances of extreme learning machines, random forests, support vector machines, and XGBoost using 10-fold cross-validation. The results consistently show that XGBoost outperforms the other methods in predicting discharge outcomes, making it a reliable tool for clinical decision-making in the treatment of severe pneumonia, including COVID-19 cases. Within this context, the random forest imputation method generally enhances performance, underscoring its effectiveness in managing missing data compared to multiple imputation.

Prediction and feature selection of low birth weight using machine learning algorithms

Background and aimsThe birth weight of a newborn is a crucial factor that affects their overall health and future well-being. Low birth weight (LBW) is a widespread global issue, which the World Health Organization defines as weighing less than 2,500 g. LBW can have severe negative consequences on an individual’s health, including neonatal mortality and various health concerns throughout their life. To address this problem, this study has been conducted using BDHS 2017–2018 data to uncover important aspects of LBW using a variety of machine learning (ML) approaches and to determine the best feature selection technique and best predictive ML model.MethodsTo pick out the key features, the Boruta algorithm and wrapper method were used. Logistic Regression (LR) used as traditional method and several machine learning classifiers were then used, including, DT (Decision Tree), SVM (Support Vector Machine), NB (Naïve Bayes), RF (Random Forest), XGBoost (eXtreme Gradient Boosting), and AdaBoost (Adaptive Boosting), to determine the best model for predicting LBW. The model’s performance was evaluated based on the specificity, sensitivity, accuracy, F1 score and AUC value.ResultsResult shows, Boruta algorithm identifies eleven significant features including respondent’s age, highest education level, educational attainment, wealth index, age at first birth, weight, height, BMI, age at first sexual intercourse, birth order number, and whether the child is a twin. Incorporating Boruta algorithm’s significant features, the performance of traditional LR and ML methods including DT, SVM, NB, RF, XGBoost, and AB were evaluated where LR, had a specificity, sensitivity, accuracy and F1 score of 0.85, 0.5, 85.15% and 0.915. While the ML methods DT, SVM, NB, RF, XGBoost, and AB model’s respective accuracy values were 85.35%, 85.15%, 84.54%, 81.18%, and 84.41%. Based on the specificity, sensitivity, accuracy, F1 score and AUC, RF (specificity = 0.99, sensitivity = 0.58, accuracy = 85.86%, F1 score = 0.9243, AUC = 0.549) outperformed the other methods. Both the classical (LR) and machine learning (ML) models’ performance has improved dramatically when important characteristics are extracted using the wrapper method. The LR method identified five significant features with a specificity, sensitivity, accuracy and F1 score of 0.87, 0.33, 87.12% and 0.9309. The region, whether the infant is a twin, and cesarean delivery were the three key features discovered by the DT and RF models, which were implemented using the wrapper technique. All three models had the identical F1 score of 0.9318. However, “child is twin” was recognized as a significant feature by the SVM, NB, and AB models, with an F1 score of 0.9315. Ultimately, with an F1 score of 0.9315, the XGBoost model recognized “child is twin” and “age at first sex” as relevant features. Random Forest again beat the other approaches in this instance.ConclusionsThe study reveals Wrapper method as the optimal feature selection technique. The ML method outperforms traditional methods, with Random Forest (RF) being the most effective predictive model for Low-Birth-Weight prediction. The study suggests that policymakers in Bangladesh can mitigate low birth weight newborns by considering identified risk factors.

Role of Artificial intelligence model in prediction of low back pain using T2 weighted MRI of Lumbar spine.

Low back pain (LBP), the primary cause of disability, is the most common musculoskeletal disorder globally and the primary cause of disability. Magnetic resonance imaging (MRI) studies are inconclusive and less sensitive for identifying and classifying patients with LBP. Hence, this study aimed to investigate the role of artificial intelligence (AI) models in the prediction of LBP using T2 weighted MRI image of the lumbar spine. This was a prospective case-control study. A total of 200 MRI patients (100 cases and controls each) referred for lumbar spine and whole spine screening were included. The scans were performed using 3.0 Tesla MRI (United Imaging Healthcare). T2 weighted images of the lumbar spine were segmented to extract radiomic features. Machine learning (ML) models, such as random forest, decision tree, logistic regression, K-nearest neighbors, adaboost, and deep learning methods (DL), such as ResNet and GoogleNet, were used, and performance measures were calculated. Our study showed that Random forest and AdaBoost are the most reliable ML models for predicting LBP. Random forest showed high performance with area under curve (AUC) values from 0.83 to 0.88 across all lumbar vertebrae and L2-L3, L3-L4, and L4-L5 intervertebral discs (IVDs), with AUCs of 0.88 the highest at L5-S1 IVD (0.92). Adaboost demonstrated high performance at the L2-L5 vertebrae with AUC values of 0.82 to 0.90, with the highest AUC (0.97) at the L5-S1 IVD. Among the DL models, GoogleNet outperformed the other models at 30 epochs with an accuracy of 0.85, followed by ResNet 18 (30 epochs) with an accuracy of 0.84. The study demonstrated that ML and DL models can effectively predict LBP from MRI T2 weighted image of the lumbar spine. ML and DL models could also enhance the diagnostic accuracy of LBP, potentially leading to better patient management and outcomes.

Features selection in a predictive model for cardiac surgery-associated acute kidney injury.

Cardiac surgery-associated acute kidney injury (CSA-AKI) is related to increased morbidity and mortality. However, limited studies have explored the influence of different feature selection (FS) methods on the predictive performance of CSA-AKI. Therefore, we aimed to compare the impact of different FS methods for CSA-AKI. CSA-AKI is defined according to the kidney disease: Improving Global Outcomes (KDIGO) criteria. Both traditional logistic regression and machine learning methods were used to select the potential risk factors for CSA-AKI. The area under the receiver operating characteristic curve (AUC) was used to evaluate the performance of the models. In addition, the importance matrix plot by random forest was used to rank the features' importance. A total of 1977 patients undergoing cardiac surgery at Fuwai hospital from December 2018 to April 2021 were enrolled. The incidence of CSA-AKI during the first postoperative week was 27.8%. We concluded that different enrolled numbers of features impact the final selected feature number. The more you input, the more likely its output with all FS methods. In terms of performance, all selected features by various FS methods demonstrated excellent AUCs. Meanwhile, the embedded method demonstrated the highest accuracy compared with the LR method, while the filter method showed the lowest accuracy. Furthermore, NT-proBNP was found to be strongly associated with AKI. Our results confirmed some features that previous studies have reported and found some novel clinical parameters. In our study, FS was as suitable as LR for predicting CSA-AKI. For FS, the embedded method demonstrated better efficacy than the other methods. Furthermore, NT-proBNP was confirmed to be strongly associated with AKI.

Investigation of Factors Associated with Gut Microbiota in Demodex-associated Skin Conditions.

This study describes the relationships of factors related to gut microbiota and skin conditions associated with Demodex, including demodicosis, rosacea, and perioral dermatitis. A total of 113 patients from Dokuz Eylül University Hospital Dermatology Department answered a cross-sectional questionnaire. They consisted of 42 cases of Demodex-related skin diseases and 71 healthy controls. Demographic data and medical history, dietary and lifestyle habits, and gastrointestinal symptoms were recorded. Statistical analysis included descriptive statistics, chi-square tests, Fisher's Exact tests, independent samples t-tests, and logistic regression methods. Our findings identified alcohol consumption [odds ratio (OR)=11.13, 95% confidence interval (CI): 4.11-17.22, p<0.01] and smoking (OR=10.32, 95% CI: 2.47-21.57, p<0.01) as strong risk factors for Demodex-related conditions. Low water intake (0-1 liter per day) (OR=3.39, 95% CI: 2.08-5.57, p=0.03) and infrequent exercise (less than 1 hour per week) (OR=4.87, 95% CI: 2.70-12.54, p=0.02) were also significant risk factors. Additional factors associated with increased Demodex risk included reduced bowel movements (OR=2.71, 95% CI: 1.45-4.06, p=0.01) and higher pet ownership (OR=2.85, 95% CI: 2.13-4.27, p=0.03). Although vegetarian and high-fat diets showed some associations, they were not independently significant. This study demonstrates key environmental and lifestyle factors, such as low water intake, infrequent exercise, reduced bowel movements, higher pet ownership, alcohol consumption, and smoking, that are significantly associated with Demodex-related skin conditions. These factors, related to gut microbiota, may provide valuable insights for managing these skin conditions and suggest promising directions for future research.

Micro-loans, macro-impacts: Examining the reverberating gains for Habru Woreda’s small-scale agrarian households.

Background Most people in third-world countries are impoverished and rely on small-holder farming as a source of income. Due to a lack of working capital to diversify their sources of income and acquire new sources, farmers are highly dependent on financial institutions to access microcredit. Method This study utilized logistic regression and propensity score matching methods to analyze the primary data collected from a sample of 385 household heads. Results The Estimation results of the study shows that Gender, age, family size, and education, access to irrigation, extension services, dependency ratio, and distance to credit sources were among the significant determinants of access to microcredit service. Conclusion Although microcredit led to an increase in household spending, real asset holdings did not improve. In addition to the need for supplementary interventions like financial literacy training and asset-building programs to encourage asset accumulation among these small-scale agrarian households, high interest rates and repayment burdens are blamed for this uneven result.