Does health data aggregation lead to bias?

Health information is especially vulnerable to the kinds of subtle discrimination that can result from big data analysis, says a researcher advocating a framework to help people understand their legal rights.

Kate Crawford, a visiting professor at the MIT Center for Civic Media and a principal researcher at Microsoft Research, and Jason Shultz, of the New York University School of Law, published a proposed framework for "due process" for those whose privacy might be at risk from new methods of data aggregation.

Personal data itself doesn't have to be exposed to be at risk, they say. Search terms for disease symptoms, online purchases of medical supplies and even the RFID tags on drug packages can tell a lot about a person's health.

"When these data sets are cross-referenced with traditional health information, as big data is designed to do, it is possible to generate a detailed picture about a person's health, including information a person may never have disclosed to a health provider," they wrote.

Even tracking Facebook "likes" can reveal sensitive information, including sexual orientation, use of addictive substances, and parental separation, they said. In fact, a study from Children's Hospital Boston found that analyzing "likes" can help public health researchers predict, track and map obesity rates down to the neighborhood level.

The problem is that much information is collected about people without their consent--and it might be used against them, such as being denied health insurance or not being offered a job, without their knowledge, reported Technology Review.

In Crawford and Shultz's framework, a person subject to bias would have the right to learn how big data analytics were used. Social media, in particular, can get things really wrong, Crawford told a conference recently.

A recent article in JAMA Internal Medicine warned about tracking when people research health-related information. It found in an analysis of 20 health-related websites, that even using tools such as DoNotTrackMe and Ghostery failed to prevent tracking through some third-party element. Ditto for social media plug-in buttons, even if the person was not logged or didn't "like" anything.

Meanwhile, the Citizens' Council for Health Freedom has warned about information being collected by government entities as part of their pubic health initiatives--often without patients giving their consent.

To learn more:
- find the paper
- read the Technology Review article