DATA-LICENSE

Depending on the calculator, our dataset consists of notes that were either designed from templated-based functions implemented in Python, handwritten by clinicians, or taken from our dataset, Open-Patients.

Open-Patients is an aggregated dataset of 180k patient notes coming from three different sources. We have authorization to use the dataset from all three sources. The first source is the USMLE questions from MedQA which is released under the MIT License. The second source of our dataset are the Trec Clinical Decision Support and Trec Clinical Trial which are available for redistribution because they are both government-owned datasets released to the public. Lastly, PMC-Patients is released under the CC-BY-SA 4.0 license and so we have permission to incorporate PMC-Patients inside Open-Patients and MedCalc-Bench, but the dataset must be released under the same lisense. Hence, our source of notes, Open-Patients, and the dataset curated from it, MedCalc-Bench, are both released under the CC-BY-SA 4.0 license.

Based on the justification of license rules, both Open-Patients and MedCalc-Bench comply with the CC-BY-SA 4.0 license, but the authors of this paper will bear all responsibility in case of violation of rights.