Cien Saude Colet. 2026 May;31(5):e23482025. doi: 10.1590/1413-81232026315.23482025. Epub 2025 Dec 12.
ABSTRACT
This study aims to evaluate the viability of the NPS as a parsimonious predictor of the Primary Care Assessment Tool (PCAT). The data used in this study were obtained from a cross-sectional survey conducted in Rio de Janeiro in 2024, including adults and children in separate samples. Both datasets were divided into training (70%) and test (30%) sets. The main predictor was the NPS. Additional predictors consisted of 44 variables for adults and 39 for pediatric populations. Six machine learning algorithms were assessed. After evaluating all variables, six variables were selected for adult sample without showing a decrease in the model performance: 1) NPS, 2) age, 3) years of education; 4) the same doctor was seen consistently, 5) the number of times the service was used, and 6) the number of children. Four similar variables were selected for the children sample: 1) NPS, 2) age (respondent), 3) years of education, 4) the same doctor was seen consistently. We observed a superior performance of predictive models in the children population (AUC=0.82) compared to adults (AUC=0.76). The transfer learning approach achieved robust performance in adapting the adult-derived model to children's data. This study demonstrated that a predictive model based on NPS and a minimal set of additional variables can effectively estimate PCAT scores.
PMID:42385074 | DOI:10.1590/1413-81232026315.23482025