METHODS: Using a subset of survey data from the National Health and Morbidity Survey (NHMS) 2019, a secondary data analysis was performed. Trained research assistants collected data through face-to-face method using a mobile tablet questionnaire system application. Logistic regression analysis was performed to examine the relationship between sociodemographic factors, physical activity, and cervical cancer screening. The analyses were conducted using STATA version 14 (Stata Corp, College Station, Texas, USA), accounting for sample weighs and complex sampling design.

RESULTS: The analysis included 5,650 female respondents, representing an estimated 10.3 million Malaysian female adults aged 18 and above. Overall, 35.2% (95%CI 33.2, 37.4) respondents had a Pap smear test within the past three years. Respondents who were physically active were 1.41 times more likely to have a Pap smear test. Similarly, respondents aged 35-59 (OR 1.84; 95%CI 1.46, 2.34) and those living in rural localities (OR 1.38; 95%CI 1.13, 1.70) had higher odds of receiving a Pap smear test. Compared to married respondents, single respondents (OR 0.04; 95%CI 0.02, 0.07) and widowed/divorcee respondents (OR 0.72; 95%CI 0.56, 0.82) were less likely to receive a Pap smear test. Educated respondents were more likely to have had a Pap smear test.

PATIENTS AND METHODS: The dataset encompassed patient data from a tertiary cardiothoracic center in Malaysia between 2011 and 2015, sourced from electronic health records. Extensive preprocessing and feature selection ensured data quality and relevance. Four machine learning algorithms were applied: Logistic Regression, Gradient Boosted Trees, Support Vector Machine, and Random Forest. The dataset was split into training and validation sets and the hyperparameters were tuned. Accuracy, Area Under the ROC Curve (AUC), precision, F-measure, sensitivity, and specificity were some of the evaluation criteria. Ethical guidelines for data use and patient privacy were rigorously followed throughout the study.

RESULTS: With the highest accuracy (88.66%), AUC (94.61%), and sensitivity (91.30%), Gradient Boosted Trees emerged as the top performance. Random Forest displayed strong AUC (94.78%) and accuracy (87.39%). In contrast, the Support Vector Machine showed higher sensitivity (98.57%) with lower specificity (59.55%), but lower accuracy (79.02%) and precision (70.81%). Sensitivity (87.70%) and specificity (87.05%) were maintained in balance via Logistic Regression.

MATERIALS AND METHODS: A retrospective cohort study was conducted involving 592 prediabetic adults from 28 health clinics in Terengganu between January 2019 and June 2023. Participants were selected based on oral glucose tolerance test (OGTT) results indicating prediabetes. Sociodemographic, medical background, and clinical data, including body mass index (BMI), blood pressure, fasting blood sugar (FBS), and lipid profiles, were extracted from medical records. Glycaemic outcomes were classified into three categories: reversion to normoglyacemia, persistent prediabetes, or progression to DM, based on glycated haemoglobin (HbA1c) levels taken within two years of follow-up. Ordinal logistic regression analysis was used to identify the significant predictors influencing these outcomes.

RESULTS: Analysis showed age, BMI, underlying dyslipidaemia, FBS, and triglyceride levels as significant predictors of glycaemic progression. Specifically, each additional year of age and each one-unit increase in BMI raised the likelihood of progression to DM by 3% and 6%, respectively. Participants with dyslipidaemia were noted to have a 67% higher risk of worsening glycaemic status, while increases in FBS and triglyceride levels were associated with 65% and 34% greater odds of diabetic progression, respectively.

METHODS: This study analyzed death records from January 2017 to June 2022, sourced from Malaysia's Health Informatics Centre, coded into ICD-10. Data anonymization adhered to ethical standards, with 387,650 death registrations included after quality checks. The dataset, limited to three-digit ICD-10 codes, underwent cleaning and an 80:20 training-testing split. Preprocessing involved HTML tag removal and tokenization. ML approaches, including BERT (Bidirectional Encoder Representations from Transformers), Gzip+KNN (K-Nearest Neighbors), XGBoost (Extreme Gradient Boosting), TensorFlow, SVM (Support Vector Machine), and Naive Bayes, were evaluated for automated ICD-10 coding. Models were fine-tuned and assessed across accuracy, F1-score, precision, recall, specificity, and precision-recall curves using Amazon SageMaker (Amazon Web Services, Seattle, WA). Sensitivity analysis addressed unbalanced data scenarios, enhancing model robustness.

RESULTS: In assessing ICD-10 coding with ML, Gzip+KNN had the longest training time at 10 hours, with BERT leading in memory use. BERT performed best for the F1-score (0.71) and accuracy (0.82), closely followed by Gzip+KNN. TensorFlow excelled in recall, whereas SVM had the highest specificity but lower overall performance. XGBoost was notably less effective across metrics. Precision-recall analysis showed Gzip+KNN's superiority. On an unbalanced dataset, BERT and Gzip+KNN demonstrated consistent accuracy.