Skip to main content
. 2025 Jun 4;13:e68138. doi: 10.2196/68138

Table 1.

Summary statistics of the Medical Information Mart for Intensive Care IV (MIMIC-IV) dataset.


Pretraining dataset Fine-tuning dataset Test dataset Total dataset
Patients, n 121,000 36,000 16,000 173,000
Visits, n 285,000 86,000 37,000 408,000
ICD-9a codes, n 1,940,000 579,000 248,000 2,767,000
ATCb codes, n 5,511,000 1,655,000 688,000 7,854,000
All codes, n 7,451,000 2,234,000 937,000 10,622,000

aICD-9: International Classification of Diseases, Ninth Revision.

bATC: Anatomical Therapeutic Chemical Code.