Konplik Health’s anonymization system for EHRs
Removing personal information from electronic health records is the first step in any health data analytics process.
Konplik’s de-identification algorithm can accurately identify and hide any patient name, doctor name, telephone number, email address, medical address, healthcare provider name, and any other type of personal information appearing in health records.
HIPAA rules require all holders of healthcare data to de-identify any data shared with third parties to assure patients’ privacy.
Electronic Health Records contain non-structured text in which personal information can be found. If those records are going to be shared with third parties, reviewing them by hand is hard work. Natural Language Processing technology can help.
Konplik has worked with MeaningCloud, our NLP sister company, to develop a deep learning-based model.
This model has been built using the BioClinicalBERT pre-trained model, which has been fine-tuned for a sequence labeling task. The training carried out was based on collections of medical records. This model is combined with MeaningCloud’s knowledge-based Insights Engine API to increase recognition coverage.
Every so often, in a production environment, the precision of the model must be validated and the model re-trained if needed. At Konplik, we have a methodology to ensure precision is maintained at its best at all times.
Konplik has deployed a demonstrator you can use to test our anonymization algorithm firsthand.
Konplik provides this de-identification functionality through an API, making it really easy to integrate with any Electronic Health Record management system.
So, what are you waiting for?
Check out our demo !! And get in touch with us if you want to know more.
Konplik Health™ makes the most of cutting edge technologies including proprietary software and innovative self-developed data processing tools and algorithms to extract value from unstructured data in healthcare. Konplik Health™ will help you discover new purposeful and actionable sources of revenue
Wilmington, DE 19801
We just released a new microsite for the voice of the patient at https://konplik.health/voice-of-the-patient/ It includes a video showcasing the experience of a European pharmaceutical company that created a project aimed at collecting and analyzing the voice of women...