A decision support scheme for beta thalassemia and HbE carrier screening

Area of Science:

• Medical Genetics
• Computational Biology
• Hematology

Background:

• β-thalassemias pose a significant public health challenge, necessitating effective carrier screening.
• Current screening methods can be costly and time-consuming, impacting widespread implementation.
• Accurate identification of β-thalassemia traits (BTT) and hemoglobin E (HbE) carriers is crucial for prevention programs.

Purpose of the Study:

• To develop and validate cost-effective scoring mechanisms for identifying BTT carriers.
• To create a joint scoring system for simultaneous detection of BTT and HbE traits.
• To reduce reliance on expensive diagnostic techniques like HPLC in population screening.

Main Methods:

• Utilized five hematological parameters to formulate scoring systems.
• Employed machine learning algorithms including decision trees, Naïve Bayes, and Artificial Neural Networks.
• Validated the developed scores on independent datasets from a tertiary care hospital in India.

Main Results:

• Achieved 100% sensitivity for both BTT and joint HbE/BTT screening scores.
• Reported specificities of 79.25% and 91.74% for BTT, and 58.62% and 78.03% for the joint score.
• Observed lower Youden's index compared to some existing indices, indicating potential for optimization.

Conclusions:

• The proposed scoring mechanisms offer a sensitive and potentially more economical alternative for BTT and HbE carrier screening.
• These scores can significantly reduce the number of individuals requiring expensive HPLC analysis.
• Implementation of these methods could lead to substantial cost savings in thalassemia prevention programs.