Use cases and technologies supporting medical data, privacy, and anonymization

Health data accounts for more than 30% of global data (Intel, 2022) and is growing every year. These data are not just tangible, direct information, such as vaccination dates and test results. It also includes contextual data such as community quality, diet trackers, exercise trackers, sleep monitors, and more. Abhishek Khowala, chief health engineer at Intel, and Agata Chudzinska, head of artificial intelligence at TheBlue.AI, discuss the need for anonymously sharing health data.

With global provider shortages, an aging population, and an increase in multiple chronic diseases, it is more important than ever to share, understand, and learn from health data. Shared data can help speed drug development, transfer care more easily, and strengthen collaboration among research institutions.

“Structured data can be presented in a relational database. We almost know which fields are relevant for privacy, and it’s a little easier to edit those fields or remove them from the data. However, the vast majority — over 80 percent of data — is Structured; it’s more difficult to analyze this unstructured data and find out exactly where any type of identifiable information is located,” Khowala said.

One solution to this problem is natural language processing, an artificial intelligence function applied to clinical records to ensure anonymity. The technology can understand text and remove identifiable information.

Intel and The Blue.AI implemented a mixed reality solution at a Florida hospital. The solution includes multiple sensors and cameras that collect data about patients and staff, who then use computer vision technology to remove any personally identifiable information.

“In all these cases, we always need to make it possible for humans to examine it,” explains Chudzinska. An example is a photo of a clinician consenting to surgery after surgery. However, a screen can be seen in the background showing the name and patient information of the patient who has undergone surgery. Chudzinska continued, “It’s not always obvious that the information that led to a particular individual is visible.”

Protecting personal identifiers when sharing health data is as necessary as sharing the data itself. “What’s the use of not sharing medical data?” Howara asked. Ensuring protection is a barrier to data sharing. “More than 70 percent of the world’s countries — about 140 countries — have some sort of regulation in place to protect an individual’s data and privacy,” Chuzinska said. AI simplifies the process of securing facilities and can delete personal information in minutes and hours, rather than days or months through human analysis.

Follow AI technology leaders Agata Chudzinska and Abhishek Khowala on LinkedIn. Learn more about the Intel Applied AI Health Data Case Study on the website.

Subscribe to this channel on Apple Podcasts, Spotify and Google Podcasts to learn more from Intel Health and Life Sciences at the Edge.

Source link