AI-Augmented Hematological Signatures for Equitable Detection of Hereditary Hemolytic Anemia Carriers: A Global Systematic Review and Meta-Analysis
Abstract
Artificial intelligence (AI) augmentation of routine hematological tests offers a promising strategy to improve hereditary hemolytic anemia (HHA) carrier detection in premarital screening, especially in resource‐limited settings. HHA in this review specifically encompasses β‐thalassemia, α‐thalassemia, sickle cell disease (including HbS and HbC variants), and other hemoglobinopathies with autosomal recessive inheritance patterns requiring carrier detection for prevention. These conditions share hemolytic phenotypes but differ in hematological signatures, necessitating separate subgroup analyses. This global systematic review and meta‐analysis evaluated the diagnostic accuracy, equity implications, and implementation challenges of AI‐augmented complete blood count (CBC), blood smear, and erythrocyte sedimentation rate (ESR) for HHA carrier identification. We systematically searched seven databases and included 85 studies (n = 133,498 participants, 23 countries). AI‐augmented screening achieved a pooled sensitivity of 92.8% (95% CI: 91.3%–94.1%) and specificity of 91.5% (89.7%–93.0%), representing a 12.3% sensitivity improvement over conventional interpretation (p < 0.001). However, significant geographic disparities were observed: sensitivity in Sub‐Saharan Africa was 86.5% compared with 94.8% in the Middle East (p < 0.001), partly due to algorithmic bias against African HbS/HbC variants and infrastructural barriers. Deep learning models achieved the highest sensitivity…
How strong is the evidence?
Evidence Hub reads this as a meta-analysis, done in people. Its evidence level is 1 of 5 (Highest · pooled evidence). Pooled evidence from several studies sits at the top of the evidence pyramid. Reported size: 133,498 participants or records.
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.