Análise dos factores associados a não adesão ás consultas pre-natais: um estudo de caso no Hospital Geral José Macamo

Abstract

Non-adherence to antenatal care (ANC) appointments represents a significant public health chal- lenge, highlighting persistent difficulties in ensuring adequate and continuous pregnancy monito- ring. This study aimed to identify the factors associated with non-adherence to ANC at the Jos´e Macamo General Hospital and to develop a predictive model to estimate the likelihood of non- adherence. A quantitative, cross-sectional study was conducted with a sample of 1,026 pregnant women. Initially, a descriptive analysis was performed, followed by the chi-square test of inde- pendence to identify the independent variables significantly associated with the outcome. The predictive model was developed using binary logistic regression with stepwise selection based on the Akaike Information Criterion (AIC). To address class imbalance, the SMOTE technique was applied to the training set (80% of the sample), and model performance was evaluated using stra- tified 10-fold cross-validation. The main predictors of non-adherence were low educational level, multiparity, absence of a partner, younger maternal age, and late initiation of ANC. The model demonstrated satisfactory performance, with a sensitivity of 88.8%, specificity of 43.1%, positive predictive value (PPV) of 61.6%, and negative predictive value (NPV) of 79.2%. Overall accuracy reached 66.1%, and the area under the ROC curve (AUC) was 76.6%, reflecting good discrimina- tory ability to identify pregnant women at higher risk of non-adherence to ANC