ABSTRACTObjectiveThe aim of the research was to determine effectiveness of the model for prenatal control in reducing new cases of severe thalassemia.MethodsPregnant women at six tertiary centers were recruited to follow the model, consisting of (1) carrier screening using mean corpuscular volume (for alpha‐thal‐1 and beta‐thal) and CMU‐E screen (for HbE trait), (2) carrier diagnosis, (3) the couples at risk were counseled and offered prenatal diagnosis, and (4) termination of affected pregnancy. All neonates were evaluated for thalassemia.ResultsOf the 12 874 recruited pregnancies, 7008 were valid for analysis. Of them, 281 couples were identified to be at risk, Of the 281, 58 affected fetuses were identified and 55 pregnancies were terminated, whereas three did not accept pregnancy termination. All 6727 neonates at no risk were proven to be unaffected. The model had sensitivity and positive predictive value of 100% and 20%, respectively. The model could detect all of affected fetuses.ConclusionThe model could prenatally identify affected fetuses with a detection rate and negative predictive value of 100%. The model was highly effective to prenatally detect affected fetuses with an acceptable false positive rate. © 2013 John Wiley & Sons, Ltd.