Google co-founder Larry Page made a rare appearance at the TED conference and expounded on a few ideas he thinks will change the world. “Wouldn’t it be amazing to have anonymous medical records available to all research doctors?,” said Page, noting that it would save hundreds of thousands of lives.
There’s some great evidence to show that Page is right. One of the FDA’s researchers, Richard Pratt, surmised that if the public had had access to the millions of public health records, they could have spotted the deadly side-effects of Vioxx, preventing between 27K and 55K deaths.
Statisticians need a large sample to spot trends that are different from random chance (“statistical significance”). The more people in the dataset, the more patterns can be identified and the more lives can be saved.
Currently, the Health Insurance Portability and Accessibility Act places tight restrictions on the kinds of information that can be made public. “These are the consequences of HIPAA’s overcautious privacy rules,” wrote a team that manages the Harvard Info/Law blog, in the aptly title post “Death by HIPAA”.
In fact, opening up all health data could help spot all sorts of problems: we could conduct original research, look for hidden side-effects, and give users individual health recommendations based on their behavior. If Vioxx is any indication, Page is probably lowballing the magnitude of lives that could be saved or improved.
But, there’s a downside: it’d be near impossible to maintain our privacy. “We have been pretending that by removing enough information from databases that we can make people anonymous. We have been promising privacy, and this paper demonstrates that for a certain percent of a population, those promises are empty,” wrote John Wilbanks of Sage Bionetworks, a nonprofit that advocates for medical transparency.
In the aforementioned paper, researchers were able to identify some participants in an anonymous medical database when they matched it with DNA from a database of their relatives.
Statisticians are becoming increasingly sophisticated at identifying individuals with public and private data. And with the kind of data we’d need to make medical breakthroughs, it wouldn’t be that hard to unmask anyone.
Here are a few variables that would be helpful:
- Location: local toxins and pollution could contribute to health problems
- Age, Race, Height, Sex: certain demographics are prone to certain diseases
- My job: sedentary jobs are more prone to health problems
So let’s take myself as an example: There are only so many 5’4 31-year-old male Jews living in San Francisco’s Mission District working in the media. An amateur statistician could combine that data, scrape Twitter, and alert my friends to all my secret medical history in the time it’d take me to order a copy of 1984 on Amazon.
Page’s comments could be especially controversial in light of recent allegations that his company was aware of more NSA surveillance than they previously acknowledged. However innocuous, discussions about any privacy topic can get the company in hot water.
I support Page’s idea of open medical data. Then again, I’m the kind of guy who’s comfortable displaying data from my sex life in public. I suspect that in the future the benefits from transparent medical data will eventually override our sense of privacy. And Google could be the company to build the tool.