The Advent of Artificial Intelligence in Diabetes Diagnosis: Current Practices and Building Blocks for Future Prospects

Abstract

India has the highest proportion of diabetes patients, and it is estimated that there will be 134 Million diabetics in India by 2045 as per IDF. Also, the disease burden is increasing to the young population between ages 25-40 as more of them are diagnosed positive according to JAMA recently. Moreover, there are only 4.8 Doctors per 10,000 population, and in villages, the ratio is the lowest possible in this country, according to the Indian Journal of Public Health. Therefore, screening & predicting Diabetes at an early stage remains a priority for clinicians. It reduces the risk of major complications and improves patients' quality of life with diabetes, and builds resilience and well-being amongst other citizens. With the advancement of Computer Science & Artificial Intelligence, it is now possible to predict diabetes and other such diseases through applying deep learning algorithms in high-quality data sets. This helps in a more accurate and faster diagnosis of Pre-diabetes, Diabetes & diabetes-related progressive eye diseases. In this study, a systematic review of the Pubmed repository for current practices to diagnose Diabetes based on AI intervention in the Indian context is carried out. Also, a critical analysis was done on various pioneered companies currently offering AI-based Diabetes diagnostic services in India. The study represents different concepts of AI tools used to predict the diseases currently available in India. Although most of the studies were carried out on Diabetic Retinopathy screening, future opportunities can be in several other areas such as Clinical Decision Support, Predictive Population Risk Stratification and Patient Self-Management Tools.