Big Data in Health: Good for Anything Yet?

While the obstacles seem high right now the payoff is as well, and it's good to remember that never did technological disruption come without its challenges.
|

This article was originally published on Wamda, a platform that reports on and builds up, the startup and entrepreneurial ecosystem of the Middle East.

When it comes to using big data in healthcare, anywhere, it might be true to say no one seems to quite know what they're doing yet?

But Dr. Pushpavaly Ramasamy, medical informatics officer at the American Hospital Dubai (AHD), wants to know how, once data is collected, we then decide its value? And who is going to benefit from which part and how do we disseminate it?

Ramasamy, speaking about digital health and big data opportunities in the region at the Building Healthcare Middle East conference in Dubai, raised an issue that is felt globally: how best to leverage the data that is being generated.

The possibilities created by data are endless. As stated in a Wamda Research Lab and GE report, MENA'S Health Startups: Unlocking the path to scale and the future of healthcare, the impact of connected, data collecting devices on healthcare goes from monitoring and treating illnesses, to reducing direct cost of treatments, then to more industrial applications.

But only if the right approach is taken.

Collecting

From patient feedback on their hospital experience and blood test results on cancer patients, to data on a user's Fitbit, information is being scooped up by the healthcare industry constantly. It is estimated that as much as 15 billion terabytes of data have been created globally in healthcare alone.

One of the beauties of big data is that healthcare sector data is the same across the board and therefore can be shared and collated across organisations.

"When hospitals start combining multiple data sources like medicine data, treatment data, geographic patient data, social data of patients, data on doctors and in the future even genome data, the result is a lot better," says Mark van Rijmenam, founder of Datafloq, a Netherlands-based matchmaker of data scientists and industry stakeholders. "Treatments could result in lower readmission rates, lower costs and healthier patients."

Where to start

Dr. Aaron Han, AHD's acting chief medical officer, says the challenge is the amount of data. "You need to start somewhere."

Han says 'somewhere' is a data set which will have the most clinical impact.

This is going to differ whether you're a hospital or a provider. For example, in the UAE and wider MENA region Han says this could include data sets around diabetes and obesity.

"One area of interest is ongoing or continuous monitoring. Wearables is an area that is of interest to providers," he told Wamda. "It will be interesting to see how patient disease management and outcomes are impacted through these devices."

Garbage in, garbage out

Ver2 Digital Medicine CEO Brian de Francesca believes rubbish data is the biggest problem with big data in healthcare lies at the moment.

"If the data capture is not robust then the results are useless," he told Wamda.

Citing the IT adage 'garbage in, garbage out', de Francesca is adamant that all data collected across the globe is of no use to anyone if the processes of defining what needs to be collected and how it is collected are not aligned. Without the right collection process, the results will not be of use to anyone.

An example would be data on hospital infection rates. Each healthcare provider may have a different definition of this - an infection caught while being in the hospital versus an infection caught after being discharged. Indeed, it could render your IBM Watson healthcare cloud useless.

De Francesca is not sceptical though. "Before we leap into and onto big data, we need to get process literate," he said. "Establish that foundation first and build big data upon it. This region can lead the way - they have a blank canvas to work upon."

Leveraging

But even if you do get a clean data set "the collecting is one thing, it's the processing of it that is something else", says Ramasamy.

In order for big data analytics in healthcare to succeed it had to be packaged to be "menu-driven, user-friendly and transparent", said a report from the Health Information Science and Systems journal (HISS) in 2014.

Or, about "policy, people, process, platform", and not necessarily in that order, according to Han.

Ideally, he says, when you're looking for healthcare information it would be like looking online for a hotel - transparency helps you get the best bang for your buck - and we're not there yet with healthcare.

Success?

So, what is working right?

In the US the CancerLinQ initiative is an example of big data being used in the fight against cancer - the idea being to collect data on every cancer patient in the US in order to analyse and spot trends or patterns. Equally not-for-profit US-based Aurora Health Care has found success with collating data from its 15 hospitals, 185 clinics, and more than 80 community pharmacies. They have 18 different primary streams of data that are processed in near real-time. It is financial data, lab data, pharmacy data and procedure data.

And AHD is using big data to improve operations in relation to patient experience. Han says emerging data shows patient experience impacts on outcomes, so while they've always collected patient feedback they're now using it as a metric to measure health outcomes as well.

The benefits from big data will come from focusing as much on patient outcomes as on spending, and, HISS says, on minimising delays between capturing data and analysing it - a problem when healthcare providers don't even know which data to start collecting and sorting.

Where is the region at?

MENA isn't too far behind the rest.

Back in 2000 the US wasn't even in the top 14 of the world's best healthcare systems. In 2012 another McKinsey report said up to 70 percent of IT healthcare projects failed. And in 2014 the National Health Service (NHS) launched a programme to collate GPs' patient data into a central database, yet failed to convince the public it could be trusted with that data.

"We lost the trust of around two to three percent of the population, it ran into the sand," said Dr. Joe McDonald, chairman at NHS CCIO Network, in Dubai. The programme has since been suspended.

While de Francesca is adamant about the work still needing to be done big data's future success in healthcare in the region, Han believes, after being in the UAE for 11 years, that the Emirates is on the right track.

While the obstacles seem high right now the payoff is as well, and it's good to remember that never did technological disruption come without its challenges.