Health boost from big data

OPINION: The most negative aspect of the contested Medicare co-payment is not the unhelpful financial burden it seeks to impose on patients but the way it has so narrowed such a vital national debate. By characterising healthcare reform as a simplistic fiscal balancing act, we are overlooking the many emerging opportunities to find significant budgetary savings while simultaneously improving the quality of care. That is, opportunities for real, evidence-based reforms that can protect the health of every Australian into the future.  

Last year a major US report conservatively estimated that big health data – that is the detailed information held in millions upon millions of personal health records – could save Americans between US$300 and $450 billion in healthcare costs. For Australia, the potential savings are just as profound.   

In the history of global public health one of the biggest breakthroughs was achieved in the 1800s when the pioneering British epidemiologist, William Farr, thought to set up a system that routinely recorded causes of death, noting that as “diseases are more easily prevented than cured” the first step to prevention is discovering their causes. It was epidemiological data that famously finally proved that smoking causes cancer.

Today’s big data sets now give us extraordinary new opportunities to scrutinise virtually everything we are doing in healthcare, from determining the effectiveness of expensive 24/7 cardiac catheter labs for heart attack patients to working backwards to uncover the multitude of reasons for preventable, and again expensive, hospitalisations. In the UK, a new £34 million research network was recently established to promote exactly this sort of linked data research. In this broader context, both the Federal Government and many of its detractors are simply missing the point.

 In Australia, big data research is already producing useful insights. For example, the Sax Institute’s 45 and Up Study, is tracking the health of over 250,000 participants in NSW, by linking information about personal health and behaviours to electronic records relating to GP visits, medication use and hospital admissions. State of the art anonymisation techniques mean participants’ privacy is protected.

While we have long known, anecdotally, that many patients fail to stay on their prescribed blood pressure medication, we were able to identify the characteristics of those patients at highest risk of lapsing by cross referencing the various demographic measures and lifestyle indicators within the 45 and Up Study with medication records. This gives treating doctors and community health services the information they need to identify and target the right patients to follow up.

We are also currently cross referencing birth records with the medications mothers used during pregnancy to identify links between particular drugs and any adverse outcomes at birth and throughout a child’s development. Investigating pregnancy risks is traditionally very difficult – for example, few randomised drug trials include pregnant women – so by linking and mining data we can pinpoint risks we may not even be aware of to protect children during gestation.

The 2013 US report, The big data revolution in healthcare, by McKinsey’s Center for US Health System Reform, calculates that by using large scale health data to target high risk groups with simple interventions, like aspirin for coronary heart disease or early screening for those with a family history of high cholesterol, $38 billion could be saved in the downstream medical costs associated with heart disease in the US alone.

While well being and good health cannot always be reduced to dollars and cents, health systems across the globe are under unprecedented cost pressures. Healthcare spending in Australia has risen from about 8.4 per cent of GDP just over a decade ago to 9.5 per cent in 2001-12 and is projected to reach 12.4 of GDP by 2032-33, or $246 billion a year, mainly due to increasing demand driven by an ageing and growing population. Australia’s high quality healthcare system has many world class attributes, but anyone using it will have encountered the very human consequences of budgetary pressures and everyone working within it knows that they will be asked to achieve more with less into the future. Ironically, big data also tells us that the proposed Medicare co-payment could, in fact, prove counter productive by discouraging GP visits. Last year the Productivity Commission reported that 600,000 to 750,000 expensive hospital admissions a year could be avoided through effective community care.

Fortunately, Australia is well placed to take advantage of the opportunity big data research in health presents. This is partly a function of our large public health systems, and the vast array of data they collect as part of doing business, but also because of early interest and experience in big data research within our universities, often in collaboration with our health departments.

Inevitably, privacy will emerge as a valid concern. Australia has, however, already developed and built innovative secure systems for linking data that ensure that when our health records are used for research they include only unique ID numbers, not personal details, and are held only in highly secure computing environments. Recognition of this protection has led to calls by key Australian health consumer groups for more data linking and more use of big data to improve the efficiency and accountability of our health system.

All this makes imposing additional costs on patients look ill-informed and unimaginative when we have the capacity, skills and infrastructure for world-leading big data research that would more effectively direct our precious healthcare dollars.

Professor Louisa Jorm is Director of UNSW’s new Centre for Big Data Research in Health, an Australian first. Professor Peter Smith is Dean of UNSW Medicine.

An edited version of the op-ed was first published in The Australian.