Individual-level prediction models of societal costs and health-related quality of life in pediatric cerebral palsy: a population-based study from Spain

Individual-level economic models more effectively capture the heterogeneity of clinical severity, caregiving needs, and social context in pediatric cerebral palsy (CP). We developed and validated predictive models to estimate how clinical, sociodemographic, socioeconomic, and health-related quality-of-life (HRQoL) factors affect disaggregated categories of societal costs. We then examined how these factors and cost components influence HRQoL outcomes for children with CP and their caregivers in Spain. Using data from 148 children with CP and their caregivers in a Spanish population-based registry, we estimated annual costs from a societal perspective. Cost components included healthcare, informal care, public subsidies, and out-of-pocket expenses (adjusted to €2023). We applied generalized linear models (GLMs) to model costs: one-part GLMs with gamma distribution and log link for most categories, and two-part models for intensive rehabilitation therapies (logistic regression for the probability of any cost, followed by a GLM for the positive costs). Child and caregiver HRQoL were modelled using ordinary least squares (OLS) and Tobit regression, respectively. Predictive performance was assessed via mean absolute error (MAE), root mean squared error (RMSE), and mean error (ME). Greater functional impairment in children was the primary driver of higher costs and lower HRQoL. Socioeconomic factors, including caregivers’ job loss due to caregiving, low household income, and lower social class, predicted higher out-of-pocket spending and worse caregiver HRQoL. Latin American family origin was associated with higher total illness costs but lower informal care costs. Use of non-standard therapies (reported by 64% of families) increased overall societal costs and was linked to marginally better child HRQoL. Individual-level, disaggregated models can inform value-based healthcare assessments by capturing the complexity of pediatric disability. Our findings underscore the importance of caregiver burden and informal care in determining costs and well-being. We provide a web-based Shiny calculator to enable practical application of these predictive models.