Big data and big science are buzz phrases in health research at the moment. It is not at all apparent what the exact definition of these are or should be and whether they will be short lived in our lexicon, but I think it reasonable to assume that where there is buzz there is honey.
I think of big data in health as information routinely collected by our interaction with health systems, both formal (eg GPs or hospitals) and informal (eg networked devices that continuously monitor our heart beat). Through ever improving connectivity such data may become available (anonymously) for the health researcher and policy maker. The statistical tools needed to analyse this volume of data without producing spurious correlations are still being developed and there are some genuine ethical concerns that must be addressed.
Within New Zealand we have a unique alpha-numeric identifier for anyone who has encountered our formal health system. This is very unusual internationally and puts us in a good position to pull data together from multiple sources and to monitor change over time. Recently I have used this system to assess the performance of new emergency department chest-pain pathways at multiple hospitals throughout the country. These pathways had been developed in research programs in Christchurch and Brisbane. Following a Ministry of Health initiative for each emergency department to adopt such a pathway, and with the financial support of a Health Research Council grant (and my personal sponsors), we were able to establish efficacy and safety parameters of the change in practice.
If we had used a traditional model of employing research staff at each hospital the costs would have run into many millions and would simply not have been possible given how health research is financed in this country. This model of monitoring changes made to how health care is delivered is both pragmatic and affordable. It is also necessary if we are to be reassured that change is really improving practice. We expect to see more big data used in this way.
Big science is often thought of in terms of hundreds or thousands of researchers in facilities like CERN costing hundreds of millions of dollars. I think big science need not be so large or expensive. Rather it is large international collaborations whereby sufficient good quality clinical research data is gathered to answer important clinical questions. The key is “sufficient”. Because of the prevalence of a disease or the size of a population base any one research group may not be able, in a reasonable time frame, to collect sufficient data to answer the important questions.
Over the past two years I have been involved in several international studies where we have pooled data, some of which our group has led, some of which are led by colleagues overseas. We are now formalising a “consortium” to further ensure data is well and appropriately used and collected. This move had been particularly important as even million dollar studies of a thousand patients do not have sufficient data to answer some of the key safety questions around the diagnosis of heart attacks (my current focus). A criticism of much academic clinical research is that it is just not useful1. This is in large part because the studies are too small to give results that would change practice. They are also often not pragmatic enough (eg by excluding significant portions of patients likely to be assessed or treated by the intervention under study).
Recognition that it is through large collaborative studies that useful practical change can occur will lead to more such collaborations. They require people to be involved with a slightly different skill set than those whose research is purely local – in particular the “people” skills required to form productive and lasting cross-cultural relationships. They also require flexibility in funding which may lead to how rules for some grants change (eg by allowing some portion of funding to be spent offshore).
The era of Big data and Big science for Big health is both daunting and exciting. While there will no doubt be blind alleys and false starts as with any research or new venture, there will also be practical and meaningful evidence based changes to health delivery. Something to look forward to.
- Ioannidis, J. P. A. (2016). Why Most Clinical Research Is Not Useful. Plos Medicine, 13(6), e1002049. http://doi.org/10.1371/journal.pmed.1002049.t003
The Sciblogs Horizon Scan
This post is part of the Sciblogs Horizon Scan summer series, featuring posts from New Zealand researchers exploring what the future holds across a range of fields.
Featured image: thainetizen.org CC BY SA 3.0