Australia spends about $2 billion (1.3%) of its national health budget of $170 billion on preventative health programs. While precise comparative data is difficult to find, public researchers have identified that Australia under invests in preventative health compared to its OECD peers.
Certainly previous research suggests strong cost and outcome benefits from taking a preventative approach to health and wellbeing.
As Australia’s population increasingly demands more sophisticated clinical responses to chronic disease, preventative strategies offer governments a potentially robust intervention to help control ballooning health care costs.
Dr Wouter Hoogland, a medical doctor who now leads McKinsey’s healthcare analytics teams in Asia, says preventative medicine is already an important part of the healthcare system (consider vaccines), but it won’t reach its paradigm-shifting potential until it becomes personalised.
Dr Zaana Howard, a Senior Expert and Leader in McKinsey’s Asia Pacific Design practice, agrees. “To really personalise healthcare, we need to combine powerful analytics with human centred design. Analytics provides pattern recognition of ‘what’ is occurring and insight into ‘how’ it is happening, since it builds off a rational logic base.
Human centred design augments this by focusing on people’s behaviours, emotions, and motivations, digging deeper into how and ‘why’ people make decisions. This is a critical to acknowledge, as our behaviours and beliefs aren’t always based on a rational or infallible premise,” she said.
Combining logic with behavioural science makes sense as a guiding principle for preventative healthcare, but why is personalisation so important? Hoogland explains. “Most people have accepted that general preventative measures such as a balanced diet and moderate exercise will decrease their risk of illness, but how many actually make these lifestyle changes?
“The research is convincing enough, but it still describes risk and correlation at an abstract population level, not at the ‘should I have a side of veggies or the chocolate lava cake?’ decision point. Would you make a different choice, if you knew that the combination of your genetics, trends in your blood pressure over time, and recent changes in your exercise habits gave a 70% probability that this particular dessert would clog a coronary? You might rather not know, but that’s where we’re headed. Without a way to estimate the consequences of our own decisions, characteristics and external risk factors, it’s no wonder we’d rather just roll the dice and let our excellent reactive healthcare industry pick up the pieces. Prevention might be better than cure, but cure seems a lot more convenient.”
Unfortunately, in healthcare the cure is usually far more expensive and painful than prevention, and with an ageing population and the costs of “gold standard” care rising inexorably, the only way to keep healthcare sustainable is to keep people healthy. In Hoogland’s view, prevention addresses both the cost and the quality of healthcare, but its effectiveness depends on providing people with a personally meaningful, data-driven choice between individually tailored interventions.
“The first part depends on our rapid adoption of evolving technology – one area where we need courageous policy-making, particularly around public-private data sharing and innovation,” he said. Howard adds “and for the second, human part, we need to continue the ongoing mindset shift in government, and double down on designing and implementing preventative strategies that are directly based on citizens’ needs and behaviours”.
“Think about preventative healthcare as a prediction problem,” says Hoogland. The better a government can predict the health risks of a person or population, the better it can target policy for a given health intervention – be that in-person dietary counselling or a prophylactic lipid-lowering therapy.
At the same time, once you have accurate predictions of the health benefits individuals would derive from an intervention, you can encourage only those likely to see positive outcomes to use it; avoiding significant cost and unpleasantness from treating those who won’t benefit at all.
“Historically we’ve improved our ability to predict by running massive, multi-year clinical and cost-benefit studies, such as those that inform the QALY calculations used by the UK’s National Institute of Clinical Excellence,” says Hoogland. “Today though, we can combine those with big data and advanced analytics to make our predictions much more detailed and accurate.”
There are many ways analytics could help, Hoogland continues: “geospatial analysis and clustering methods can reveal patterns of disease, and how these correlate with location, healthcare access and social determinants of health. Regressions, machine learning and neural network algorithms can predict the risk of a new disease, complication or behaviour in an individual from a given cluster.
By combining methods, it is possible to estimate not only who would benefit from a health intervention, but how much. Though analytics can be used to optimise for particular policy goals, it can’t solve the age-old political quandary of how to choose exactly who to give exactly what level of support – but it does allow for far better-informed decisions”.
Howard believes this is where human centred design can play a critical role. “When we have an idea of who would benefit and by how much, we can then engage people across the health system – citizens, health care professionals, health care clinics and government to co-design solutions together with much higher rates of positive behaviour change. It’s important to remember that supporting behaviour change doesn’t necessarily require high tech or a high price tag.
Interventions can be relatively simple. Think of the Australian government’s Health Star Rating System – that’s a human centred design effort to make you consider your veggies-or-cake decision just a little more carefully. It presents you with relevant data when and where you need it – the rest is up to you.”
Howard adds that “co-designing with and for people is an important consideration here: by continuously involving the people who would be most impacted by the solution in the design process – through ethnographic research, idea creation, prototyping and testing interventions with them – we can ensure the solution is designed to be simple, usable, and meet their needs. That results in real change”.
Both Howard and Hoogland acknowledge the Australian government’s efforts in their respective fields of interest, as well as the structural barriers inherent in our healthcare funding systems. Hoogland notes “the government is already using many of these techniques across its healthcare portfolio, and has shown a real appetite for change through its internal behavioural analytics capability-building initiatives, as well as its major investment in the Digital Health CRC.
It’s also worth noting that Australia’s highly federalised system of healthcare funding can muddy the incentives for investing in preventative care. With healthcare paid for by a complex mix of federal and state governments as well as private health insurers and patients themselves, the benefits of prevention might not accrue to those who invested in it – and that can be tricky to explain to taxpayers or shareholders”.
Of course there remains plenty of room for improvement. Human centred design can deliver significant impact anywhere there is access to consumers, whether at a small or large scale.
By contrast, analytics are only as useful as the data that power them. In the long run, technological developments in predictive algorithms and non-invasive diagnostic sensors (perhaps embedded in smartphones, contact lenses, lavatories, or blood-borne nanobots), will provide data that enable truly personalised preventative care. In the meantime, though, our most informative data are the medical records, healthcare claims and social data held by government, hospitals and insurers – and these tend to be very difficult for innovators to access.
As Hoogland points out, “Data is the real key here, and one area where forward-looking policy can make all the difference. There are four elements to consider: quantity, quality, variety and accessibility. With a single national public payer system covering over 20 million people, we already have quantity, and improving data quality is a major area of focus across healthcare. Variety comes from linking public sector records held in MBS, PBS and welfare datasets with data held by hospitals, private health insurers and other non-health partners, to form a detailed end-to-end view of a person’s lifestyle and healthcare needs. Accessibility? That’s the tough part”.
“Maintaining the privacy and security of healthcare data is absolutely critical to earning citizens’ trust and support in public health. At the same time, locking data away in government servers puts the entire burden of keeping pace with global innovation on the public sector”, Hoogland says. “We need a vehicle for private commercial entities to securely access deidentified government healthcare data. This could be through a commercial model, as with the Truven datasets in the US, through a carefully vetted deidentified data matching program similar to Data Republic in the Australian private sector, or even through a more innovative model, such as a government-owned, cloud-based, public-private data science collaboration platform”.
Howard points out that public opinion is also shifting in this regard, raising thorny questions of data ownership and accountability: “People are more willing to share their health data if they believe they will gain direct health benefits from doing so, and even more so if it will support the broader community at the same time. People are starting to realise that this is already happening commercially, through their usage of health tracking wearables such as FitBits and Apple Watches. Those devices provide an engaging design experience and transparent health or activity metrics to their users in exchange for their data.
Those data are then used to develop new commercial products, but also to further the public good: take for example Apple’s collaboration with Stanford University to detect and predict irregular heart rhythms using Apple Watch data. These sorts of research are driving a seminal question of our age, embodied in the ongoing controversy around My Health Record: as long as the benefits and costs to both parties are made clear, why not let people choose to share data about themselves – or indeed, actively opt-out in cases where they don’t want to share?”
“Data aside, Australia is already in a good position,” Hoogland advises. “The Medicare-funded primary healthcare system offers citizens an invaluable portal to policy-driven preventative care. The Department of Health’s recent Medicare Benefits Schedule review has further entrenched a culture of awareness around medical inefficiency and evidence-based practice.
The private health insurance industry can’t refuse coverage or charge higher premiums even if it knows an enrolee has a pre-existing illness. The My Health Record system has the potential to collect the kind of detailed medical data needed to inform truly individualised healthcare.”
Howard adds “Government is also increasingly focused on using customer and citizen experience to deliver better care to Australians. Now all that’s needed is to put the pieces together: power better analytics through secure public-private data sharing and collaboration, and apply privacy-conscious human centred design to the output, to produce a new generation of highly effective preventative health policies. That’s the path to innovating our national health through preventative healthcare – without adding to the tax bill”.