Can’t see the data forest for the MyHealthRecord trees

Think small to get big data going in health say the experts and bypass the MyHealthRecord!

---

 

You’ve heard the quote: “Big data is like teenage sex: everyone talks about it, nobody really knows how to do it, everyone thinks everyone is doing it, so everyone claims they are doing it.”

In healthcare the same quote might go: “Big data is like Bigfoot: some people swear they have seen it and it’s spectacular; most people don’t believe those people; those that do are scared of it; it does look weird and scary; but, it’s generally seen as a figment of crazy people’s imagination.”

Dr Thomas Handler, vice president and head of healthcare research with global tech research group Gartner, is worried that in Australia we may be putting the cart before the horse in terms of healthcare data. He thinks we need to move away from our obsession with collecting the data and focus first on what insights such data might actually provide us. That, he suggests, might lead us to more focus on how we collect it and avoid the age-old problem of collecting a tonne of data which, because of a mix of outdated, propriety and eclectic collection systems, can’t be shared or analysed, and therefore is pretty useless.

The potential of “big data” in healthcare is, of course, massive. If we could share and join data even mildly well we might make huge inroads into many health issues, from infant mortality, to dementia, depression, diabetes, asthma and cancer.

Connected epidemiological data might give prevention the place it deserves in managing our national health budget.’

But is it all just a dream in the heads of visionaries that we will never realise?

Think small to realise big (data)

There are some signs, albeit slightly left-field ones, that persisting with the “big data” in health agenda, could pay off. And it looks like we may be thinking just a little too big.  Some smart small ideas which use cheap new technology are having some big wins.

Amazon Web Services, Vodafone Australia and Sydney’s Garvan Institute last year launched a mobile app based in the cloud, or shared computing model, whereby mobile users consent to cancer researchers accessing unused data resources while their phones are off or otherwise not being used. The objective is to vastly boost the computing resources available for complex genome sequencing, one of the most exciting goals of big data.

Some of the world’s largest companies have skin in the health and cloud computing game. When you have the likes of AWS, IBM, Microsoft, Google, SAP, SAS and others all eager to help crunch the medical data, you are playing in the right space.

The Garvan Institute claims to have the largest database of genome data in Australia and is a world-leader in the application of bioinformatics, or technology in biology.

When the human genome was mapped in 2003, many were moved to declare the end of all disease and human suffering was nigh. Things didn’t quite work out that way. But one of the truly compelling and real possibilities that will likely emerge from that endeavour is that of personalised medicine.

One real-life example of this includes gene assays that can identify which women with early breast cancer may safely avoid chemotherapy.

And while a recent stoush has erupted over the usefulness of DNA tests being spruiked by a pharmacy chain to guide prescribing, this is a field that holds great promise.

Think and spend small to go big would be an apt war-cry of a locally run data analytics competition held out of Canberra called GovHack. The competition, held annually, attempts to provide high level lateral solutions to government problems, in a very short timeframe, using top teams and no money. One thing they came up with last year was the slightly macabre “Death and Taxes” project.

The objective of  “Death and Taxes” was to identify the main demographics of people who contribute most to the ATO’s coffers over their lifetime, in order to target them with more effective healthcare to keep them alive, and paying taxes, longer.

Researchers set out to correlate data from the ATO with deaths recorded by the Australian Institute of Health and Welfare’s  (AIHW) mortality figures for 2012. They also looked at the AIHW reports for cancer mortality and health expenditure.

In the same year, another GovHack team worked on a solution to address costly and unnecessary hospital emergency department visits. Developed by South Australian tech firm Chamonix, HealthBuddy is a prototype mobile app drawing together information such as government checklists advising what conditions should be treated at which facilities, travel and wait times, capacity, as well as travel options and directions.

In charge of the project was Chamonix’s lead for platforms and solutions, Ashleigh Green. “Say you had five ambulances travelling to Flinders Hospital (in Adelaide) where there are eight beds available.  The app would allow health coordinators to make accurate and quick decisions as to whether ambulances should be diverted to other hospitals or other types of facilities,” he explains. The list of facilities includes general practices, he adds, with scope to have specific areas of expertise recorded and retrievable from the system, as well as the location of GP practices and how to get to them.

A purpose-built version of the app, called HIPSMobile (Health Identifier PCHR System), is currently being trialled by the Northern Territory Department of Health, with live deployment expected around March or April this year.

Chamonix director and serial Australian tech entrepreneur Geoff Rohrsheim believes HIPSMobile will play a major role in helping NT Health move to paperless systems and become a test case for what he terms mHealth, or mobile health.

“This going to help address one of the biggest problems in healthcare, which is scheduling,” he says. “You can’t schedule with paper.”

Somehow the discussion on big data in health always seems to get dragged back to what some commentators think is the lowest common denominator and potentially the problem, not the solution:  the MyHealthRecord.

Even the highly visionary Rohrsheim can’t get past this pesky government white elephant project that has so far cost Australian taxpayers more than $1 billion yet has barely crawled off the starting line. He says that the ability to quickly retrieve and share patient data will not only lead to more efficient patient treatment at hospitals, but for GPs it will mean that for the first time they will have immediate access to accurate hospital records for patients.

“GPs had been complaining that without hospitals connected [to MyHealth] ‘why should we bother?’,” says Rohrsheim.

Many hospitals today still rely heavily on fax machines to transmit patient information, which means GPs find it hard to know what actually happened in hospital. Systems such HIPSMobile could change this, Rohrsheim says, and, in turn, also make patient information such as imaging results, medications, and blood tests results more easily accessible for hospital staff.

“Eventually it will all come down to, ‘what are the actionable events you want to be able to do with your thumb on your phone’?”, says Rohrsheim.“Do doctors want to increase observation frequency, order another test or approve a discharge while they’re walking along with their coffee?”

Marcel Dinger, head of The Kinghorn Centre for Clinical Genomics housed within the Garvan Institute, believes the proliferation of genome data will likely be the saviour of My Health Record and equivalent digital initiatives around the world.

“The My Health system hasn’t worked because no one was incentivised to do it,” Dinger says. “But with the recognition that phenotypal information is so important we might see genomics act as a catalyst to draw together large amounts of detailed and properly curated health information.”

Sounds great. But with the vision comes the controversy.

Rohrsheim believes that medical entrepreneurs should be allowed to use the data in line with their own financial interests provided those interests also meet those of patients and the community.

He provides the example of a surgical practice wanting to increase the number of knee reconstructions it performs over a given period. “They could then ask for data on referrals for knee reconstructions and coordinate their practices accordingly.”

Following the money, not spending it 

Russian-born data scientist, Inna Kolyshkina, says that financial services is by far the most advanced industry sector when it comes to data analytics. And the reason should come as no surprise to anyone: money. In her view, this is how all the various stakeholders will likely achieve the much-needed incentivisation that Garvan’s Dinger and others see as missing.

Kolyshkina, a founding member of the Institute of Analytics Professionals Australia, advises government and private sector organisations on how to implement effective data analytics programs. The South Australian Department of Health is one of her clients.

“If we as a nation want to decrease our spending on health without compromising the quality of health we need to get rid of the situation where a lot of resources are wasted on finding out information we don’t need or conducting recommended procedures that are not necessary,” she says.

Kolyshkina proved the point in 2013 by putting together a team that within 46 hours hacked together several data sets on early childhood development, socioeconomic disadvantage and health, to produce data that provided the South Australian Government with three key areas to focus on to improve the life of kids in South Australia:

• Maternal smoking during pregnancy
• Lack of motor vehicle in a household
• Medication affordability

In a separate study conducted for the Actuaries Institute Australia, Kolyshkina and a group of researchers presented some illuminating data regarding workers taking time off for depression. Using employment data and patient records, the researchers found a wide disparity in the amount of time taken off by people who received treatment for depression from a GP and those who saw psychiatrists.

What was the finding??

She cites the sharp increases recently in money spent raising awareness about depression.

“How effective is this really? We don’t know,” she says. “Did anyone ask, has anyone asked ‘how do these programs and advertisements help a depressed person sitting there?’ It annoys me greatly. The beauty of data is that it’s completely objective,” Kolyshkina says.

“More data and better methods for analysing it will allow us to discover solutions instead of wasting money on unnecessary procedures.”

But, she says, you need to think about what data you need first, not rush madly into creating giant new data sets.

Like Kolyshkina, Gartner’s Handler, is wary of a mad rush to create data without much work on what data should we be trying to create, then connect, to create real value.

But he still supports the My Health Record opt-out decision, made late last year by the Federal Government. If you are going to try and collect all that data, you had better do it in a sensible way, he says.

“It was entirely insane to spend a fortune on these systems and make it opt in,” he told The Medical Republic.