Data Storage

Created by Celine Lin, Modified on Mon, 11 Nov, 2024 at 12:50 PM by Celine Lin

How does Empathia AI ensure data privacy while improving transcription and note quality?


At Empathia, we do not use patient data for training purposes. Instead, we use clinician feedback, edits, and smart edit requests to train the AI on personal preferences and improve transcription accuracy.


To enhance our AI transcribing accuracy, we focus on two main steps:

  1. Voice-to-Text Transcription: We work closely with champion users, especially those who speak other languages, to validate the transcription and translation quality. Importantly, no patient information is involved in this process.

  2. Medical Reasoning

    1. To improve note quality, the AI learns from clinician feedback to understand where information should be placed in different sections of the note, such as the History of Present Illness (HPI), Review of Systems (ROS), and Assessment. This helps ensure that the AI aligns with clinical practices.

    2. Additionally, we utilize a side content database containing the latest medical textbooks, research papers, and clinical guidelines to continuously validate and train the AI and enhance note quality.


To ensure high accuracy and reliability, we might use de-identified patient encounter data for testing purposes.  This data is solely used for validating and evaluating the performance of our AI model, without any direct learning or model adjustments based on the test data. This means the AI does not alter its behavior or 'learn' from these encounters, but rather, they serve to verify that the model accurately handles real-world scenarios. By maintaining this separation between testing and training, we ensure the privacy of patient information while rigorously upholding quality standards. All the data is encrypted and securely stored in designated servers. 



We recommend keeping the encounter data (transcript and notes) for 2-4 weeks, in case you need to retrieve any encounter information. 


Was this article helpful?

That’s Great!

Thank you for your feedback

Sorry! We couldn't be helpful

Thank you for your feedback

Let us know how can we improve this article!

Select at least one of the reasons
CAPTCHA verification is required.

Feedback sent

We appreciate your effort and will try to fix the article