One of the striking examples of the social value of the new techniques of data aggregation and analysis is described in Big Data: A Revolution That Will Transform How We Live, Work, and Think by Viktor Mayer-Schonberger and Kenneth Cukier. Big data is a business buzzword, but the underlying reality it describes is real. The three Vs of big data – variety, velocity and volume – represent the new world in which data in a variety of formats, including unstructured data like video or text, come at a researcher in enormous quantities and in a constantly changing stream. Add to this new world of data a range of advanced analytical techniques that can detect novel correlations in data without the need for a prior causal hypothesis, and the result is truly something new under the sun – a way to discover unsuspected and unanticipated insights into the world that simply could not have been uncovered with unaided empirical observation.
The utility of these techniques is often not appreciated in policymaking circles. But Mayer-Schonberger and Cukier provide an example of how big data analytics literally saves lives.
Dr. Carolyn McGregor and a team of researchers at the University of Ontario Institute of Technology and IBM are working with a number of hospitals on software to analyze data generated from premature babies. Monitors track the babies’ vital signs such as heart rate, respiration rate, temperature, blood pressure, and blood oxygen level. Nurses pay close attention to the information generated by these monitors – and when they see signs that the baby is suffering from an infection or fever they rush in to begin treatment. But the data itself is often just thrown away.
The medical researchers working with IBM retained the data and subjected it to rigorous analysis and the results were startling. Patterns were detected that could predict the onset of a fever as much as 24 hours in advance, thereby allowing medical intervention well before a crisis had developed.
Moreover, some of these advance indicators of a problem were counterintuitive. Researchers found that the simultaneous stabilization of vital signs was an advance warning of an infection to come, a sort of calm before the storm. No one would have predicted this in advance. Even now that the correlation has been established, the causal mechanism is a matter of speculation. Perhaps the baby’s body senses the infection coming and tries to batten down the hatches in preparation for the rough ride ahead.
The utility of the correlation does not depend on knowing the causal mechanism. Doctors and nurses can begin to intervene on the basis of a reliable indicator that trouble is coming.
Further research to uncover the unknown causal mechanism is obviously needed to have reliable way to know when to act on the predictive analytics. But big data has uncovered a useful fact about the onset of fevers and infections in premature babies that can become standard practice for early intervention.
The big picture here is this. Retaining data and using it for a purpose for which it was never intended can sometimes create enormous social benefits. In this case, these techniques of data retention and analytical reuse save lives.