"Synthetic Indian Clinical Notes for Natural Language Processing Research"

"Abstract\u2014The application of Natural Language Processing (NLP) in healthcare holds immense potential, but its progress is often hindered by the scarcity of publicly available clinical data, especially from non-English-centric regions like India. Indian clinical notes present unique challenges due to their linguistic diversity (including code-mixing with languages like Hindi, i.e., \u2018Hinglish\u2019), distinct documentation structures, and context-specific terminology. To address this gap, we present a new, large-scale synthetic dataset of 10,000 Indian clinical notes. This paper details the methodology used to generate the dataset, including the analysis of real-world templates and the programmatic incorporation of India-specific semantic features. The dataset spans six medical specialties (cardiology, neurology, pediatrics, general medicine, orthopaedics, and oncology) and is designed to facilitate research in semantic induction, information extraction, and the development of language models tailored to the Indian healthcare ecosystem. The dataset is structured in an accessible JSON format and is made available to the research community to foster innovation and development in this critical domain."