AI Driven Pregnancy Risk Prediction in Women with Thalassemia Using CBC Data: A Proof-of-Concept Study

Abstract

Thalassemia is a common blood disorder in India, and women who carry this condition in their child-bearing years. Even though thalassemia trait is usually considered mild, pregnancy can put extra pressure on their body. Because of this, affected women may face problems like low haemoglobin(anaemia) and high blood pressure during pregnancy. In many hospitals, especially those with limited resources, advanced genetic tests are not easily available. However, basic blood tests like Complete Blood Count (CBC) and haemoglobin analysis are routinely done. This study suggests a simple AI-based approach to predict pregnancy-related risk in women with thalassemia traits using only regular blood test data. An openly available thalassemia dataset was used, which included the CBC values and haemoglobin components. Since actual pregnancy data were not available, a simulated risk label was created based on known medical risk patterns. A machine learning model (Random Forest) was then trained using these blood parameters and carrier status. The result showed that factors such as small red blood cells, haemoglobin levels, and thalassemia carrier status played an important role in predicting risk. Although the study used simulated data, it shows that low-cost and easily available blood tests can be useful for identifying risk in thalassemia carriers. This approach can be improved in the future by using real pregnancy data.