Journal of Biostatistics and Epidemiology

Impact of the Power of Adaptive Weight on Penalized Logistic Regression: Application to Adipocytic Tumors Classification

Narumol Sudjai, Monthira Duangsaphon, Chandhanarat Chandhanayingyong — Sun, 23 Feb 2025 06:57:07 +0000

Introduction: MRI-based texture features in adipocytic tumors to serve as non-invasive predictive biomarkersthat can provide precise outcomes for decision-making. Power of adaptive weight and the initial weight forthe adaptive Lasso is one of the important parameters. This study aimed to compare the impact of the initialweight together with the power of adaptive weight for this adaptive Lasso under high-dimensional sparse datawith multicollinearity.

Methods: All independent variables in the Monte Carlo simulation were generated using the Toeplitzcorrelation structure. Performance of the initial weight together with the power of adaptive weight on penalizedapproaches was evaluated using the mean of the predicted mean squared error (MPMSE) for simulationstudy and the area under the receiver operator characteristic curve (AUC), precision, recall, F1-score, and theclassification accuracy of models for real-data applications.

Results: The simulation study showed that the smallest MPMSE value was obtained from the square rootof the adaptive Lasso together with the initial weight using Lasso. Additionally, the results of this approachon the real-data application achieved high performance to distinguish the intramuscular lipomas from well-differentiated liposarcomas: the values of AUC, accuracy, precision, recall, and F1-score for the model basedon penalized logistic regression classifier were 0.935, 0.928, 0.919, 0.921, and 0.925 respectively, and 0.946,0.935, 0.932, 0.934, and 0.930 respectively for the model based on support vector machine classifier. Both thesimulation study and the real-data application presented that the square root of the adaptive Lasso together withthe initial weight using Lasso was the best option under high-dimensional sparse data with multicollinearity.

Conclusion: Our finding showed that the power of adaptive weight on penalty function and the initialweight can affect certain the classification accuracy of machine-learning model. In practice, if choosing theseparameters are appropriate, it produces models that have good performance.

Risk Factors Related to Congenital Hypothyroidism: A Systematic Review and Meta-Analysis

Sun, 23 Feb 2025 00:00:00 +0000

Introduction: Congenital endocrine disorders have a global impact on the morbidity and mortality of children and are a public health problem that heavily affects society and the daily lives of affected children and their families. The severity and consequences of congenital hypothyroidism (CH) on physical and especially cerebral maturation combined with lifetime mental retardation make CH neonatal screening one of the costliest preventive health programs. Thus, early diagnosis can improve the prognosis of the disease. The objective of the study is to examine CH’s risk factors reported in previous studies.

Methods: Systematic review was performed according to the PRISMA checklist. PUBMED, Google Scholar, Scopus, MEDLINE, Web of Science, and Springer were analyzed using R version 4.0.3. For further review, we assessed eligibility analysis to identify influential studies.

Results: Of 63 studies, 21 studies were suitable for synthesis. Based on this review, risk factors related to CH were birth weight, age of pregnancy, female sex, home environment, notion of inbreeding, seasonality, multiple pregnancy, gestational diabetes, parity, advanced maternal age, parental thyroid disease, gestational diabetes, ethnicity, maternal body mass index (BMI), and socio-economic status.

Conclusion: This systematic review indicates that the risk factors related to CH vary by country and even by inter-region according to geographical, genetic, and socioeconomic specificities.

The Association between Oral Contraceptive Pills and Subtypes of Ovarian Cancer: A Sys- tematic Review and Meta-Analysis

Sun, 23 Feb 2025 07:40:03 +0000

Introduction: Limited studies have been conducted on the effect of oral contraceptive pills on the subgroupsof ovarian cancer, so we decided that conduct a systematic review and meta-analysis to investigate the effectof preventive pills on ovarian cancer subgroups.

Methods: Scopus, PubMed, Web of Science and EMBASE were searched to identify studies on the associationbetween OCPs and subtypes of ovarian cancer from January 1, 2000, through February 5, 2023. The pooledrelative risk (RR) and odds ratio (OR) were used to measure this association.

Results: A total of 48 studies were included. In the association between ever-use compared with never-use ofOCPs and ovarian cancer risk, the pooled RR in cohort studies was 0.69 [95% CI: 0.61, 0.78], and the pooledOR of the case-control studies was 0.64 [95% CI: 0.59, 0.69]. For the association between OCPs and subtypesof ovarian cancer, there is a significant inverse association between OCPs and serous 0.72 [95% CI: 0.23, 0.82]and endometrioid 0.74 [95% CI: 0.64, 0.86], but no association between OCPs and clear cell 0.84 [95% CI:0.60, 1.16] and mucinous 0.80 [95% CI: 0.63, 1.01].

Conclusion: This study shows a statistically significant inverse association between ever-use compared tonever-use of OCPs and ovarian cancer risk. Also shows a statistically significant inverse association betweenserous and endometrioid cancer and OCPs, but no association between OCPs and clear cell and mucinous.

Using Causal Attitude Network Model to Analyze the Factors Affecting Public Attitude and Acceptance of COVID-19 and Vaccination in Indonesia

Asti Meiza, Fithria Siti Hanifah, Han L.J. van der Maas, . Mahmudi — Sun, 23 Feb 2025 07:42:55 +0000

Introduction: Attitudes about COVID-19 relate to cognitions, feelings, and behaviors regarding the pandemic and vaccination, as well as other factors, such as demographic characteristics, and health- related information. This research uses the Causal Attitude Network (CAN) model to measure attitudes and acceptance of COVID-19 vaccination among 1385 Indonesian people from 15 cities.

Methods: Data was obtained from instruments that made in the Netherlands and adapted to the Indonesian language and culture. This research integrates psychometrics with network analysis, an advanced implementation of the field of Statistics to reveal the interaction between psychological factors that shape people's attitudes towards COVID-19 and vaccination in Indonesia. Data analysis used JASP, an open-source statistical analysis software.

Results: From this research, it was found that attitude elements regarding trust in vaccine development and awareness of the importance of vaccines in Indonesian society have a high influence on other attitude elements. Attitude elements regarding the habit of wearing masks and awareness about the importance of the COVID-19 vaccine are the attitude elements that have the highest impact on changing other attitude elements.

Conclusion: Two attitude elements, trust and awareness, most influence other attitude elements. Trust in the development of the COVID-19 vaccine is related to trust in the experts developing it. In other words, increasing public confidence in the development of a science-appropriate COVID-19 vaccine will be in line with increasing public trust in COVID-19 vaccine developers, and vice versa.

Analysis of Vertical Ground Reaction Force Data in Predicting Parkinson’s Disease

Varun Jain — Sun, 23 Feb 2025 07:45:29 +0000

Introduction: Parkinson’s disease is a complex, progressive neurodegenerative disorder known to negatively impair patient gait. Therefore, with gait and vertical ground reaction force data, an association can be made between the data and Parkinson’s disease.

Methods: Data from 146 participants; 93 with Parkinson’s disease and 73 without Parkinson’s disease was obtained from a PhysioNet database for use in this article. A Fourier Analysis and several support vector machine learning models were computed in MATLAB to classify whether an individual had Parkinson’s disease.

Results: From the Fourier analysis, it was determined that a statistically significant difference was present between the vertical ground reaction force data of individuals with and without Parkinson’s disease. Additionally, it was found that a Minimum Classification Error Optimized SVM machine learning model using Bayesian statistics was able to classify individuals with Parkinson’s disease using vertical ground reaction force data at an accuracy of 67.1%, and sensitivity of 80.43%.

Conclusion: Therefore, it can be determined that vertical ground reaction force can predict Parkinson’s Disease with considerable accuracy which could be improved with an increased number of participants.

Bayesian Approach in Modeling Prostate Cancer

Job Lusweti Sirengo, Drinold Aluda Mbete — Sun, 23 Feb 2025 07:48:00 +0000

Introduction: Prostate cancer is an emerging health problem in Sub-Saharan Africa. It is often diagnosed at an advanced stage due to lack of access to screening and diagnostic facilities.

Methods: This study therefore aimed to model the effects of risk factors on the outcome of prostate cancer screening using Generalized Bayesian ordinal logistic regression with random effects then compare the results obtained with the model without random effects. The study further used Mean Squared Errors to establish if the estimates for the two models were different.

Results: The findings in this study indicate that aged individuals have high chances of having prostate cancer at the early, late or advanced stage. The individual with traces of family history and hereditary breast & ovarian cancer syndrome are also most likely to be in late or advanced stage of prostate cancer.

Conclusion: From the findings aged individuals, having traces of family history and individuals with hereditary breast & ovarian cancer history, should make sure they understand all symptoms of prostate cancer so that incase of any signs they immediately seek for screening services. In addition, the Ministry of Health should create awareness training and increase screening facilities, this will also encourage for early screening and detection of prostate cancer. The models with presence of random effects were considered best since they had lowest Widely Applicable Information Criterion values in each category.

Epidemiology, Clustering and Spatial Distribution of Animal Bites in Algeria

Latifa Bouguerra, Mohamed L’Hadj, Schehrazad Selmane — Sun, 23 Feb 2025 07:51:49 +0000

Introduction: Animals can transmit many viral and bacterial diseases through bites and saliva that canbe potentially fatal to human. Rabies, one of these diseases, is rife in two-thirds of the world’s countries.Algeria is not spared. This study was scoped to provide insight into the demography and epidemiology, spatialdistribution and clustering patterns of animal bites in Algeria.

Methods: The global and local Moran's I were used to investigate geographic clustering patterns of animalbites in Algeria. The animal bites data provided by North West Health Region (NWHR) Observatory wasanalyzed to glean useful information.

Results: Over the past five decades, 1201 human rabies fatalities have been recorded in Algeria with a yearlyaverage of 20 cases and a male predominance. As for 2017, a total of 116403 animal attacks were recorded.Dog bites accounted for 64.1%, followed by cat bites for 30.5%. The rabies vaccine was practiced in 74% ofcases and vaccine with rabies immune globulin in 26% cases. The incidence was estimated at 279 per 100000inhabitants. The incidence of animal bites, dog and cat bites exhibited spatial autocorrelation globally; theMoran index values were 0.41, 0.43 and 0.60 respectively. Significant hot spots were located in Tell, andsignificant cold spots were located on Sahara and High-Plateaus. The analysis of the 21314 animal attacksreported in NWHR in 2019, showed that young children and men are the most-at-risk. Indeed, 71.3% weremale and 58.7% occurred outdoors. Among the 8275 bites that occurred in children under 15 years, 66.8%were boys and 29.3% were children under 5 years. Most of the bites were Category II(45.7%) followed byCategory III(38.6%).

Conclusion: The current strategy needs to be reviewed, reformed and strengthened while promoting cross-sectoral work with a collaborative approach of all relevant sectors for a One Health initiative.

Socioeconomic Inequality in Chronic Complications of Type 2 Diabetes Mellitus in Iran: Concentration Index and Decomposition Approach

Sun, 23 Feb 2025 07:56:26 +0000

Introduction: In Iran, evidence regarding the impact of socioeconomic status (SES) on the progression and complications of type 2 diabetes mellitus (T2DM) are sparse and needs growing body of research. Socioeconomic status is a complex construct and its impact on the health outcomes should be evaluated in an efficient and flexible way. The aim of this study is to investigate socioeconomic inequality in chronic complications among patients with T2DM using the concentration index and, also determine the contribution of various variables on inequality through the decomposition analysis.

Methods: This cross-sectional study included patients with T2DM who received care at the diabetes clinic in Hamadan from April to September 2023. Demographic information, household assets, and diabetes-related factors were obtained from medical records and face-to-face interviews. In this study, a healthy lifestyle was evaluated based on four characteristics of healthy behavior (smoking, dietary pattern, weight control, and physical activity) and the score obtained for each individual. The asset Index was considered as a measure of SES based on household assets and was created using principal component analysis. To examine the relationship between diabetes complications and independent variables, univariate logistic regression models were employed, and the concentration index (CI) was used to assess inequality. The decomposition approach was utilized to determine the contribution of each factor to the inequality.

Results: A total of 530 patients (60% females and 54.9% less than 60 years) were included. In the study population, 22.3%, 9.5%, and 4.7% had retinopathy, kidney failure, and diabetic foot ulcers, respectively. The CI for retinopathy, kidney failure, and foot ulcers were [(CI: -0.248, p<0.001), (CI: -0.085, p<0.001), (CI: -0.125, p<0.001), respectively]. Factors with the greatest contribution to socioeconomic inequality for retinopathy were economic status (57.25%), duration of T2DM (21.77%), and adherence to prescribed medication (10.89%), for kidney failure were economic status (38.83%), hypertension (24.71%), and education level (14.11%), and for foot ulcers were economic status (24%), duration of T2DM (24%), education level (20.80%), and HbA1c level (18.40%).

Conclusion: This study demonstrated that socioeconomic inequality in chronic complications of T2DM with greatest contribution for economic status. It is recommended that policymakers and health professionals consider the main causes of socioeconomic inequality in the chronic complications of T2DM when developing health strategies.