AUTHOR=Rueangket Ploywarong , Rittiluechai Kristsanamon 

TITLE=Predictive Analytic Model for Diagnosis of Ectopic Pregnancy

JOURNAL=Frontiers in Medicine

VOLUME=Volume 8 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2021.646258

DOI=10.3389/fmed.2021.646258

ISSN=2296-858X

ABSTRACT=Objective: Ectopic pregnancy is a serious condition. Delayed in diagnosis could lead to life-threatening outcomes. The aimed of this study was to develop the diagnostic predictive model for ectopic pregnancy by using statistical method. 
Material and methods: A cross-sectional study of 377 pregnant women presented with first trimester complications (abdominal pain or vaginal bleeding), who diagnosed with pregnancy of unknown location in Phramongkutklao Hospital between October 2010 to April 2020. The data including clinical (risk factor, sign and symptoms), initial serum human chorionic gonadotropin (hCG) level and ultrasound findings were analyzed. The statistical predictive score was developed by performing logistic regression analysis using 9:1 divided ratio (n = 347:30) for training and testing data model. The testing data of 30 patients were performed to test the validation of predictive scoring.
Results: From total of 22 factors, logistic regression method derived scoring model based on five factors (history of sexual transmitted disease, current use of emergency pills, cervical motion tenderness, serum hCG ≥ 1,000 mIU/ml and ultrasound finding of adnexal mass) using cut off score ≥ 3. The predictive score was able to predict ectopic pregnancy with an accuracy of 77.81% (95% CI = 73.07-82.07), specificity of 91.0% (95% CI = 62.07-71.96), sensitivity of 67.02% (95% CI = 88.02-94.03) and area under the curve of 0.906. In validation group, no patient with negative result of this score had an ectopic pregnancy.
Conclusions: Statistical predictive score derived a well performance for ectopic pregnancy diagnosis with suitable interpretability function of risk factor association, which could lead to a good evidence-based decision making for healthcare physicians.