De-identification Service in Azure Health Data Services

The authors of this blog are: Kimia Mayborn.

Machine learning and analytics are increasingly being used to improve health outcomes, enhance patient and clinician experiences, and optimize organizational performance within healthcare systems. The foundation for these solutions is data, which continues to grow at an unprecedented rate, especially in unstructured documents. Healthcare organizations that wish to utilize this data for machine learning, analytics, or other purposes other than clinical care may need to de-identify their patients’ health information.. However, manually de-identifying unstructured patient health records is time-consuming and expensive. Moreover, many automated methods do not meet the stringent requirements for medical data privacy, making them unsuitable to support medical advancements.

Today, Microsoft is excited to offer a new anonymization service in Azure Health Data Services. This allows organizations to securely anonymize clinical data while maintaining its clinical relevance and adhering to the rigorous standards of the HIPAA Privacy Rule.

The de-identification service consists of three operations: ‘TAG’, ‘REDACT’, and ‘SURROGATE’. Proxy functions replace PHI elements with realistic surrogates, maximizing the balance between privacy and utility. This process creates de-identified synthetic data that closely resembles the original data and allows analytics and machine learning models to interact with de-identified, realistic data found in production environments or in inference.

De-identification services allow healthcare organizations to leverage data in de-identified format to:

• Train private machine learning models, including generative models, using de-identified data.
• Develop analytics dashboards to drive data-driven decisions.
• Generate synthetic test data to solve problems that are difficult to reproduce in a test environment.
• Facilitating data sharing between collaborating institutions facilitates the creation of broader data sets and opens opportunities for clinical research and discovery.
• To conduct longitudinal studies to assess the predictive value of risk factors for disease without disclosing patient data.

Organizations in the healthcare sector can benefit from de-identification services, and early adopters are already planning to leverage them to advance their most prominent use cases.

The joint research groups of Professor David Eyre (Professor of Infectious Diseases), Big Data Institute, and Dominic Furniss (Professor of Plastic and Reconstructive Surgery), Nuffield Department of Orthopedics, Rheumatology and Musculoskeletal Sciences, University of Oxford have been carrying out the following research: – It is an identification service to support clinical research in the UK National Health Service (NHS) and the AHDS de-identification service has been recommended for its performance in protecting NHS patient data.

Dr. Rachel Kuo and Dr. Andrew Soltan are developing multimodal-based models aimed at advancing medical diagnosis and treatment in plastic surgery and oncology. Dr. Kuo and Dr. Soltan work closely with their patient partners and call for strong de-identification for both patients and researchers. Large amounts of clinical data are required to train models, and automated and efficient de-identification is essential to expand data availability. By first de-identifying the vast amounts of clinical data needed to train these models, Dr. Kuo and Dr. Soltan obfuscated the training data so that the models could not reveal patient identifiers, protecting patient privacy and protecting the models from memorization attacks. protects.

At Microsoft, we work to empower healthcare providers, payers, scientists, and life sciences companies. Accelerate your data and AI journey while We maintain a strong commitment to protecting patient privacy.

Microsoft Cloud for Healthcare helps organizations create a healthier future through data and AI.

We are excited to strengthen our investments in data and AI. Microsoft Cloud for Healthcare. Our healthcare solutions are built on trust and Microsoft’s Responsible AI principles. Through these innovations, we’re making it easier for our partners and customers to create connected experiences at every point of care, empower the healthcare workforce, and unlock value from their data using data standards that matter to the healthcare industry.

Learn more: