Clinical and Sociodemographic Predictors of Tuberculosis and Malnutrition Among Under-Five Migrant Children in Urban Pakistan: A Community-Based Cross-Sectional Study

Tuberculosis (TB) and malnutrition remain major health challenges among migrant children living in low-resource urban settings. Identifying clinically relevant patterns associated with these conditions may support standardized clinical assessment and public health planning. The present study aimed to identify clinical, anthropometric and sociodemographic factors associated with TB and malnutrition among migrant children under five years of age in urban Pakistan. A community-based cross-sectional study was conducted in migrant settlements of Rachna Town, Lahore, Pakistan. Clinical, demographic and anthropometric data were collected from 300 children under five years of age. A Genetic Algorithm (GA) combined with Random Forest analysis was applied to identify the most informative variables associated with TB and malnutrition. Model performance was assessed using internal cross-validation. The analytical approach identified a concise subset of variables showing stable internal performance. Key factors associated with malnutrition included Bacillus Calmette-Guérin vaccination status, formula feeding, chest X-ray findings and gastrointestinal symptoms. The most relevant variables associated with TB were Bacillus Calmette-Guérin vaccination score, erythrocyte sedimentation rate, gastrointestinal indicators and parental occupation. The final models demonstrated good internal discrimination and substantial agreement. These findings suggest that routinely collected clinical, anthropometric, laboratory, radiological, and sociodemographic variables may support the identification of clinically relevant patterns associated with TB and malnutrition during standardized community-based assessments. The proposed data-driven approach should be interpreted as an analytical and decision-support framework for prioritizing informative variables and supporting clinical assessment, rather than as a pre-diagnostic or early risk prediction tool.