Google has been training artificial intelligence machines to quickly analyze a patient’s health record to help doctors make a more accurate diagnosis. In some cases, this can mean that the algorithm’s capabilities range from calculating a patient’s chance of survival to their odds of readmission. While this technology is impressive, Google has been proceeding with caution to ensure it does not repeat previous errors regarding patient privacy laws.
What Can The Technology Do?
Bloomberg reports the case ofa woman with late-stage breast cancer who came to a hospital with fluid in her lungs. When doctors used hospital computers to read her vitals, they estimated she had a 9.3 percent chance of dying during her stay. However, when Google’s algorithm ran through 175,639 data points, her death risk shot up to 19.9 percent. A few days later, the woman passed.
Google’s health research unit, Medical Brain, has been developing this algorithm. It uses a “neural network” form of artificial intelligence software to advance predictive modeling with electronic health records. In a paper published in Nature Research Journal, a team of Google researchers reported that their “models outperformed traditional, clinically-used predictive models in all cases. We believe that this approach can be used to create accurate and scalable predictions for a variety of clinical scenarios.” In their case study at two hospitals, they were able to have a 95 percent accuracy for predicting inpatient mortality at Hospital A and a 93 percent accuracy for Hospital B.
The Bloomberg article reports how Google’s algorithm was able to “sift through data previously out of reach: notes buried in PDFs or scribbled on old charts.” According to Nigam Shah, an associate professor at Stanford University and co-author of the Nature Research Journal paper, “As much as 80 percent of the time spent on today’s predictive models goes to the ‘scut work’ of making the data presentable.”
Google’s History with Data Collection
While Google’s artificial intelligence could help revolutionize the health care industry, there are ethical concerns surrounding data and privacy. Andrew Burt, the chief privacy officer of data company Immuta, says “Companies like Google and other tech giants are going to have a unique, almost monopolistic ability to capitalize on all the data we generate.”
When dealing with consumer data, Google does not have a perfect track record. In 2017, Google got in hot water when its UK-based artificial intelligence company, DeepMind, accessed 1.6 million patient records without the knowledge or consent of patients. According to BBC News, the Royal Free NHS Foundation Trust, which runs three hospitals in London, handed over the information to DeepMind “to develop and refine an alert, diagnosis and detection system that can spot when patients are at risk of developing acute kidney injury (AKI).”
However, according to New Scientist, the data shared between DeepMind and Royal Free included patient data from the last five years but also “logs of day-to-day hospital activity, such as records of the location and status of patients – as well as who visits them and when.”
DeepMind was eventually charged for using the data for purposes that regulators had not approved. The transfer of data was not illegal within itself, but according to BBC News, the UK’s Information Commission ruled that the “hospital did not tell patients enough about the way their data was used.”
Given this history, Google has been careful to make sure they won’t be overstepping privacy laws. According to Bloomberg, “Google and its hospital partners insist their data is anonymous, secure and used with patient permission.” For Google’s recent research, it gained access to 46 billion pieces of anonymous patient data from the University of California, San Francisco, and the University of Chicago.