Work

A Comparative Assessment of Parametric and Semi- parametric Regression Models for Functional Difficulties among Children Aged 5–17 Years in Bangladesh

Year: 2026
Type: preprint
Abstract: Background Functional difficulties among children are becoming a major public health challenge in Bangladesh. It is affecting children’s development, education, and daily life. Understanding the determinants of these difficulties is required to develop effective strategies to reduce this burden. Objective The present study aimed to identify the key socio-demographic determinants of child functional difficulties among children aged 5–17 years in Bangladesh and further compare the performance of the semi-parametric transition model with the parametric count regression models. Methods Analyzing data from the 2019 MICS (37,579 children aged 5–17), this study examined child functional difficulties derived from 13 domains. Key variables were identified using one-way ANOVA. After fitting multiple classical count regression models (Poisson, Negative Binomial, discrete Lindley, and discrete Weibull), the discrete Weibull model was selected based on AIC and dispersion criteria. Its performance was then evaluated against the semi-parametric transition model. Finally, transition models were used to analyze the data and determine critical risk factors of child functional difficulties. Results Functional difficulties were less common among children aged 5–9 years. Maternal functional difficulty was significantly associated with functional difficulties in children (incidence rate ratio [IRR]: 2.31, 95% confidence interval [CI]: 2.09–2.57). In addition, administrative division, household wealth index, children’s educational status, and the sex of the household head were identified as significant determinants of functional difficulties among children. Conclusion The semi-parametric transition model provides robust estimates compared to the best-performing parametric model without requiring distributional assumptions, making it a reliable alternative to parametric models. The findings of this study can help policymakers to develop evidence-based policies to reduce the burden of child functional difficulties in Bangladesh.
Institution: University of Dhaka
Language: English
Open Access status: green
Funder: UNICEF
OpenAlex ID: W7236775771
Cites: 23
Cited by:
Related to: 10