Tag Archives: small data

Healthcare in Australia and NZ only has ‘Small Data’ – Really?


‘Big Data’ is not about BIG nor is it about DATA… but one thing I’m certain of is that these technologies and methodologies will accelerate discoveries, improve patient outcomes and dramatically reduce healthcare costs.  (The one problem is that IT vendors chose the wrong name!)

Before you stop reading, let me convince you of the merits and applicability to healthcare. Consider hip replacements, if a way was found to replace a hip so that it would never have to be redone, both patient outcome and cost would be dramatically improved. (Example from this blog post where the TED talk cites a group of doctors who collaborated, gathered data and found a pattern, which resulted in these outcomes.) Now if you could watch many hip replacements and follow the patients, given a large number of procedures, you would start to detect which ‘techniques’ resulted in the best outcome.  This is the idea of “big data”, to find these patterns automatically using computers and the available data.

What’s changed? Over the last decade technologies that can economically store and reason over disparate data types have been developed. (By different data types think about structured data, the data in a spreadsheet/database that were invented for computers, and natural data called un-structured, such as pictures, X-Rays and ECG waveforms which humans quickly make sense.) The power of these new technologies is that they bring all this information together and provide the analytic tools to find these patterns and correlations and/or create predictive models.

Sounds complicated but if there was a way to capture the data about procedures and the patient outcome overtime, ‘big data’ could find these patterns which result in the best overall outcome. Immediately the cry goes out that the healthcare professionals cannot spend time inputting more data! And they are absolutely right, these systems should aid and assist the practitioner, but let me suggest that a great deal of the data currently exists in the disparate computer systems, within monitors and the various imaging and measuring modalities, as well as on paper. While its up to the IT industry to provide the ways to extract all this information in a secure and controlled way, there are emerging technologies which will take this idea further.

One interesting and perhaps confrontational technology is video analysis. Today video analysis is used to detect ‘suspicious’ behaviour in public places, (Boston example here), helps major stores detect potential shoppers needing assistance and improve workplace practices to reduce accidents. So it is conceivable that a video of a surgical procedure could be analysed and compared with others, to provide input into the improvement cycle! Or similarly a radiologist with a tricky image could be presented with similar x-rays and the diagnoses he peers made.

In summary ‘Big Data’ is about using available data to improve processes, understand trends, find correlations and develop predictive models, while you don’t need huge amounts of data, you do need the vision to make it happen! While Australia and New Zealand lag behind in this area, I wonder if we can learn from what has been done in the rest of the world and leapfrog them?