Modeling the determinants of attrition in a two-stage epilepsy prevalence survey in Nairobi using machine learning - Part 1

Aug 5, 2026

Attrition undermines parameter estimation in both longitudinal studies and multi-stage cross-sectional surveys. This work applies machine learning to predict attrition and uncover associated factors in a two-stage, population-based epilepsy prevalence study in Nairobi. Participants screened in the Nairobi Urban Health and Demographic Surveillance System were assessed for probable epilepsy at stage I and neurologist-confirmed at stage II, with attrition defined as non-attendance at stage II.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages13

Reading Time33 min

AccessFree to download

Abstract

Background

Global epilepsy burden and service coverage targets

Prevalence estimation in urban settings

Methods

Study design and attrition definition

Feature processing and class imbalance handling

Model training, tuning, and evaluation

Results

Best-performing base learners

Key predictors of attrition

Conclusion

Implications for targeted mobilization and risk prediction

Global Epidemiology 9 (2025) 100183 Contents lists available at ScienceDirect Global Epidemiology journal [homepage:](homepage: www.sciencedirect.com/journal/global-epidemiology)[ www.sciencedirect.com/journal/global-epidemiology](homepage: www.sciencedirect.com/journal/global-epidemiology) | Modeling the determinants of attrition in a two-stage epilepsy prevalence survey in Nairobi using machine learning | | | | | --- | --- | --- | --- | | Daniel M. Mwangaa,b,* , Isaac C. Kipchirchir a , George O. Muhua a , Charles R. Newton c,d , Damazo T. Kadengye b , for EPInA Study Team
a Department of Mathematics, University of Nairobi, Kenya b African Population and Health Research Center, Nairobi, Kenya c Department of Psychiatry, University of Oxford, United Kingdom
d Kenya Medical Research Institute, Wellcome Trust Research Programme, Kilifi, Kenya | | | | | A R T I C L E I N F O | | A B S T R A C T | | | Keywords: Machine learning Attrition
Loss to follow-up
Urban settlements Epilepsy Prevalence Surveillance | | Background: Attrition is a challenge in parameter estimation in both longitudinal and multi-stage cross-sectional studies. Here, we examine utility of machine learning to predict attrition and identify associated factors in a twostage population-based epilepsy prevalence study in Nairobi.
Methods: All individuals in the Nairobi Urban Health and Demographic Surveillance System (NUHDSS) (Korogocho and Viwandani) were screened for epilepsy in two stages. Attrition was defined as probable epilepsy cases identified at stage-I but who did not attend stage-II (neurologist assessment). Categorical variables were one-hot encoded, class imbalance was addressed using synthetic minority over-sampling technique (SMOTE) and numeric variables were scaled and centered. The dataset was split into training and testing sets (7:3 ratio), and seven machine learning models, including the ensemble Super Learner, were trained. Hyperparameters were tuned using 10-fold cross-validation, and model performance evaluated using metrics like Area under the curve (AUC), accuracy, Brier score and F1 score over 500 bootstrap samples of the test data.
Results: Random forest (AUC = 0.98, accuracy = 0.95, Brier score = 0.06, and F1 = 0.94), extreme gradient boost (XGB) (AUC = 0.96, accuracy = 0.91, Brier score = 0.08, F1 = 0.90) and support vector machine (SVM)(AUC = 0.93, accuracy = 0.93, Brier score = 0.07, F1 = 0.92) were the best performing models (base learners). Ensemble Super Learner had similarly high performance. Important predictors of attrition included proximity to industrial areas, male gender, employment, education, smaller households, and a history of complex partial seizures.
Conclusion: These findings can aid researchers plan targeted mobilization for scheduled clinical appointments to improve follow-up rates. These findings will inform development of a web-based algorithm to predict attrition risk and aid in targeted follow-up efforts in similar studies. | | Background Epilepsy is among the most common neurological disorders, affecting over 50 million people worldwide and over 80 % of the cases are in low- and middle-income countries [1]. The World Health Organization in 2022 published the Intersectoral Global Action Plan (IGAP) on epilepsy and other neurological disorders which outlines five strategic objectives including to strengthen public health approach to epilepsy (strategic objective 5). One of the key global targets under strategic objective (SO) 5 of IGAP is to increase epilepsy service coverage by 50 % by 2031. The denominator to compute service coverage is the number of people with epilepsy (prevalence). Accurate estimation of prevalence for epilepsy is therefore important to contribute effective measurement of progress towards IGAP goals. It is also important for evidence-based decisions on policy and national planning of epilepsy services. Estimation of prevalence of epilepsy in urban

FAQs

What does attrition mean in this two-stage epilepsy prevalence survey?

Attrition refers to probable epilepsy cases identified at stage I who did not attend stage II for neurologist assessment.

Which machine learning models performed best at predicting attrition?

Random forest, extreme gradient boosting (XGB), and support vector machine showed the strongest performance across AUC, accuracy, Brier score, and F1 score.

What factors were identified as important predictors of attrition?

Proximity to industrial areas, male gender, employment, education, smaller households, and a history of complex partial seizures were among the key predictors.

Aug 5, 2026

Upload by Clementine