Regional
Artificial intelligence in Transfusion Medicine: initiatives in India
The healthcare landscape, including transfusion medicine, stands at the threshold of a new era driven by Artificial Intelligence (AI). We may soon witness a watershed moment that separates the pre and post-AI chapters in modern medicine. While AI adoption in transfusion medicine remains in its infancy, as reflected by the limited body of literature and its constrained clinical use, momentum is building.
Countries such as the United States, China, and India, with a large pool of software developers and data engineers, are poised to propel AI across various healthcare domains 1. In the following sections, we will briefly highlight recent contributions of AI to transfusion medicine in India and outline the developments that may shape its future trajectory.
One of the earliest studies on using AI in transfusion medicine was presented by Dr. Anand Deshpande (PD Hinduja Hospital, Mumbai) at the 2020 International Society of Blood Transfusion (ISBT) virtual congress. The study, “Using artificial intelligence (AI) to evaluate emotional aspects of the donors – a way forward for donor reaction (ADR) prevention/management” (Abstract no. 5C-S36-05), retrospectively analysed the facial expressions of over 2,000 donors using Microsoft’s cognitive intelligence tools to identify emotional parameters 2. A webcast of this presentation can be accessed by ISBT members under the “Brains and Hearts - Dealing with Blood Donors” section (13 Dec 2020) 3.
Subsequent work has further demonstrated the potential of AI in clinical decision support. Ansari et al. from Max Super Speciality Hospital, New Delhi, developed machine-learning models, including Random Forest and XGBoost, to predict blood component requirements in hospitalized dengue patients 4. This study, published in 2023, exemplifies how predictive modelling can be helpful in implementing transfusion strategies in clinical practice.
A futuristic blood centre
The advent of accessible, customizable Large Language Models (LLMs) has attracted substantial interest from both clinicians and data scientists. A recent study from the All India Institute of Medical Sciences (AIIMS), Nagpur, presented at the Annual ISBT Congress in Barcelona, explored Retrieval Augmented Generation (RAG) techniques for refining and training AI models for transfusion medicine(Abstract no. PA03-L04) 5. Another similar RAG-based AI model from the same institute, presented at HAEMATOCON 2024 (Abstract no. OP-BHC-58), explored AI-driven clinical decision support in hematology 6. These initiatives are evolving toward integrated AI platforms that leverage domain-specific APIs (Application Programming Interfaces).
Recognizing the need for multidisciplinary collaboration, a workshop focusing on Data Informatics and Artificial Intelligence was conducted during TRANSMEDCON 2024 (Indian Society of Haematology and Transfusion Medicine), held in Mahabalipuram/Chennai. National and international experts addressed fundamental aspects of data management, storage, processing, and integration for AI-driven applications. Additionally, the Asian Association of Transfusion Medicine (AATM) has established a working group dedicated to exploring the role of AI in transfusion medicine.
Several projects are currently in progress and are expected to yield tangible outcomes by 2025. Although India’s large population (>1.4 billion) suggests significant data availability for AI and machine learning model development, progress has been hindered by the lack of clean, structured healthcare data. The Indian Government’s Ayushman Bharat Digital Mission (ABDM) aims to address these challenges by digitizing healthcare data, and further initiatives are underway to leverage this data for the development of public AI tools in healthcare 7,8. Such efforts may ultimately facilitate more robust and clinically meaningful AI implementations within transfusion medicine.
Every breakthrough in healthcare once began as a curiosity met with scepticism. Today, AI in transfusion medicine stands at that inflection point, where apprehensions are slowly making way for adoption. As India’s digital infrastructure matures and data becomes increasingly robust, it will provide the essential momentum needed to advance AI in transfusion medicine.
References
1. Business Standard. India among critical tech leaders, behind only US and China in AI. [Internet]. Available from: https://www.business-standard.com/technology/tech-news/india-among-critical-tech-leaders-behind-only-us-and-china-in-ai-124082900976_1.html.
2. The 36th International ISBT Congress, Virtual meeting, 12–16 December 2020. Vox Sang. 2020;115 Suppl 1:5–396. doi: 10.1111/vox.13031. PMID: 33302324.
3. ISBT. Brains and Hearts - Dealing with Blood Donors. [Internet]. Available from: https://www.isbtweb.org/resource/brainsand-hearts-dealing-with-blood-donors.html.
4. Ansari MS, Jain D, Budhiraja S. Machine-learning prediction models for any blood component transfusion in hospitalized dengue patients. Hematology, Transfusion and Cell Therapy. 2023 Nov 17.
5. Abstracts of the 38th International Congress of the ISBT, Barcelona, Spain, 23–27 June 2024. Vox Sang. 2024;119:7–596.
6. Abstracts of HAEMATOCON 2024. Indian J Hematol Blood Transfus. 2024;40(Suppl 1):1–168. doi: 10.1007/s12288-024-01926-4.
7. Sharma RS, Rohatgi A, Jain S, Singh D. The Ayushman Bharat Digital Mission (ABDM): making of India’s Digital Health Story. CSIT. 2023;11(1):3–9. doi: 10.1007/s40012-023-00375-0. Epub 2023 Mar 31. PMCID: PMC10064942.
8. Healthcare IT News. India harnesses ABDM data to develop public AI. [Internet]. Available from: https://www.healthcareitnews.com/news/asia/india-harness-abdm-data-develop-public-ai.
