Related Experiment Video
Updated: Feb 18, 2026
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
Comparison of three algorithms for prediction preeclampsia in the first trimester of pregnancy
Rebeca Silveira Rocha1, Júlio Augusto Gurgel Alves2, Sammya Bezerra Maia E Holanda Moura3
1Department of Nursing, Federal University of Ceará (UFC), Fortaleza, State of Ceará, Brazil.
Insights
A new algorithm using maternal factors and mean arterial pressure (MAP) shows better prediction for preterm preeclampsia (PE) in Brazilian women. Existing international guidelines from NICE and ACOG demonstrated lower accuracy for preterm PE prediction in this population.
Area of Science:
- Obstetrics and Gynecology
- Maternal-Fetal Medicine
- Clinical Prediction Algorithms
Background:
- Preeclampsia (PE) prediction is crucial for maternal and fetal outcomes.
- International guidelines like NICE and ACOG are used for PE risk assessment.
- The effectiveness of these guidelines in diverse populations, such as Brazil, requires evaluation.
Purpose of the Study:
- To compare the predictive accuracy of a novel, simple algorithm for preeclampsia (PE) with the NICE and ACOG guidelines.
- To assess the performance of these prediction models in a Brazilian cohort.
- To identify the most effective model for predicting preterm PE in Brazilian women.
Main Methods:
- Secondary analysis of two prospective cohort studies (August 2009 - January 2014).
- Prediction of total, early, preterm, and term PE using data from 11 to 13+6 weeks of gestation.
- Assessment of predictive accuracy using Area Under the Receiver Operator Characteristic Curve (AUC-ROC), sensitivity, and specificity.
Main Results:
- The NICE (AUC-ROC: 0.657) and ACOG (AUC-ROC: 0.562) algorithms showed limited accuracy for preterm PE prediction in Brazilian women.
- A simple algorithm incorporating maternal factors (MF) and mean arterial pressure (MAP) achieved a higher AUC-ROC of 0.842 for preterm PE.
- This MF + MAP model demonstrated statistically significant improvement over ACOG (p<0.0001) and NICE (p=0.0002) for preterm PE prediction.
Conclusions:
- The predictive performance of NICE and ACOG guidelines for preterm PE is suboptimal in the Brazilian population.
- A straightforward algorithm combining maternal factors and MAP offers superior predictive accuracy for preterm PE in Brazilian women.
- This finding suggests the need for population-specific adjustments in PE prediction strategies.
Objective:
To compare a new simple algorithm for preeclampsia (PE) prediction among Brazilian women with two international guidelines - National Institute for Clinical Excellence (NICE) and American College of Obstetricians and Gynecologists (ACOG).
Methods:
We performed a secondary analysis of two prospective cohort studies to predict PE between 11 and 13+6weeks of gestation, developed between August 2009 and January 2014. Outcomes measured were total PE, early PE (<34weeks), preterm PE (<37weeks), and term PE (≥37weeks). The predictive accuracy of the models was assessed using the area under the receiver operator characteristic curve (AUC-ROC) and via calculation of sensitivity and specificity for each outcome.
Results:
Of a total of 733 patients, 55 patients developed PE, 12 at early, 21 at preterm and 34 at term. The AUC-ROC values were low, which compromised the accuracy of NICE (AUC-ROC: 0.657) and ACOG (AUC-ROC: 0.562) algorithms for preterm PE prediction in the Brazilian population. The best predictive model for preterm PE included maternal factors (MF) and mean arterial pressure (MAP) (AUC-ROC: 0.842), with a statistically significant difference compared with ACOG (p<0.0001) and NICE (p=0.0002) guidelines.
Conclusion:
The predictive accuracies of NICE and ACOG guidelines to predict preterm PE were low and a simple algorithm involving maternal factors and MAP performed better for the Brazilian population.
