Abstract Objective Using data from the PREECARDIA study, we aimed to evaluate the performance of the miniPIERS model for adverse maternal outcomes and the prognostic model for perinatal death in Bissau. Methods We conducted an external validation of the two predictor models using prospectively collected data from women with preeclampsia recruited at Hospital Nacional Simão Mendes, in Guinea‐Bissau, between June 2023 and January 2025. Model performance was assessed by evaluating discrimination, calibration, and risk stratification across different probability thresholds. Results During this period, 140 preeclamptic women were recruited, of whom 50 were excluded from the analysis because they had incomplete predictor or outcome information. The miniPIERS model demonstrated poor discrimination (area under the receiver operating characteristic curve AUROC: 0.664, 95% confidence interval CI: 0.50–0.78) and suboptimal calibration, with Calibration‐in‐the large (CITL) = 0.27, slope: 0.53. The prognostic model for perinatal death showed good discrimination (AUROC: 0.854, 95% CI: 0.77–0.93) but marked miscalibration (CITL = 1.43, slope = 2.11). The risk stratification table indicated low sensitivity and positive predictive value across different probability thresholds for both models. Conclusion The miniPIERS model showed limited utility for risk prediction in this setting, where healthcare infrastructure and population characteristics differ from the model's development context. While the prognostic model for perinatal death showed better discrimination, its poor calibration and high baseline neonatal mortality in this setting suggest that the model's utility must be interpreted cautiously. Future research should focus on larger multicenter validation studies in similar low‐resource settings to optimize and adapt risk prediction tools for maternal conditions such as preeclampsia.
Turé et al. (Tue,) studied this question.