Machine learning model for prediction of obstructive hydrocephalus in Aminu Kano Teaching Hospital (AKTH) Kano State

Abstract

Hydrocephalus is a neurological condition marked by an abnormal accumulation of cerebrospinal fluid (CSF) within the brain’s ventricles, leading to increased intracranial pressure and potential developmental impairments in children. Obstructive hydrocephalus remains a significant neurosurgical challenge, especially in resource-limited settings. This study aimed to classify the types and causes of hydrocephalus among paediatric patients in AKTH and to develop a machine learning model for the prediction of obstructive hydrocephalus using clinical variables and anthropometric parameters. A hospital-based cross-sectional study was conducted using real-time data from paediatric patients aged 0–7 years. Variables collected included age, sex, head circumference, mid-upper arm circumference (MUAC), maternal demographics, and type and cause of hydrocephalus. Statistical
analysis was performed using SPSS version 27 to assess the relationship between type of hydrocephalus and selected variables. A machine learning model was developed and validated using WEKA version 3.8.6 to predict the type of hydrocephalus based on input features. The statistical analysis revealed no significant associations between the type of hydrocephalus and variables such as gender and head circumference. The machine learning model demonstrated high predictive accuracy in classifying
hydrocephalus type. In conclusion, the integration of statistical analysis with machine learning methods improved classification and prediction of hydrocephalus, offering a supportive tool for early diagnosis and clinical decision-making in resource-limited neurosurgical settings.