Sometimes an early diagnosis means the difference between life and death when it comes to treating serious illness.
Two Microsoft (MSFT +0.34%) researchers have written a paper showing the potential of using engagement with search engines to predict an eventual diagnosis. Written by Microsoft executives Eric Horvitz and Ryen White, working with Columbia University doctoral candidate John Paparrizos, the research showed how anonymized Bing search logs could be used to identify people whose search queries provided strong evidence that they had recently been diagnosed with pancreatic cancer, reported Microsoft's Next blog. The research trio then retroactively analyzed searches for symptoms of the fast-spreading disease, which is often diagnosed once it's too late for it to be cured.
They found that they could identify a pattern of queries likely to signal an eventual diagnosis. Results of the research were published in The Journal of Oncology Practice. The paper, Screening for Pancreatic Adenocarcinoma Using Signals From Web Search Logs: Feasibility Study and Results, shows that search engine data may be an effective tool for early diagnosis.
"We find that signals about patterns of queries in search logs can predict the future appearance of queries that are highly suggestive of a diagnosis of pancreatic adenocarcinoma," -- the medical term for pancreatic cancer, the authors wrote. "We show specifically that we can identify 5 to 15 percent of cases while preserving extremely low false positive rates" of as low as 1 in 100,000."
Just searching on Bing may help people learn they have a serious illness faster. Image source: Microsoft.
A new frontier in health tracking
Microsoft is not alone in trying to use data to keep people healthier. Apple (AAPL 0.39%) has built a health-tracking suite into its latest operating system that also aggregates data that can be used for research. Apple's Health app, however, is more of an active tool than what the Microsoft researchers are proposing. Essentially, the iOS product is an extension of fitness trackers that monitor basic data. Apple Health could be used to predict potential problems, but it's an opt-in solution -- users have to wear an Apple Watch and carry an iPhone, while in the case of the Microsoft research, potential illness can be diagnosed from web-search behavior when the consumer is not making an active effort to track anything.
While this study was solely linked to queries that help identify people with pancreatic cancer, in theory, Horvitz, managing director of Microsoft's Redmond, Washington, research lab, believes the research could apply to other illnesses.
"We are excited about applying this analytical pipeline to other devastating and hard-to-detect diseases," Horvitz said.
Both Horvitz and White told Next that this work was done as a proof of concept to show that a "different kind of sensor network or monitoring system" is possible. Microsoft does not plan to develop any products linked to this study, according to the researchers.
What does this mean?
Horvitz and White hope that their work inspires the medical community to discuss how a search-engine-query-based screening methodology might work.
"They suggest that it would likely involve analyzing anonymized data and having a method for people who opt in to receive some sort of notification about health risks, either directly or through their doctors, in the event algorithms detected a pattern of search queries that could signal a health concern," reported Next.
Like Apple's Health app, the research from the Microsoft team opens up a potential new way for people to learn about their health. The search engine query analysis is not a diagnosis, according to the researchers, instead it's a way to flag potential problems where medical professional can get involved and make the call.
"People are being diagnosed too late," Horvitz told Next. "We believe that these results frame a new approach to pre-screening or screening, but there's work to do to go from the feasibility study to real-world fielding."
