The term “Big Data” is one that was introduced only within the last few decades, and is a result of the incredible amount of information flowing around in our current digital age.
It is usually defined as extremely large sets of data that can only be analysed by computers (due its sheer volume) to reveal patterns, trends and associations, especially with regards to human behaviour.
University of Edinburgh’s Usher Institute of Population Health Sciences and Informatics director Prof Dr Andrew Morris, however, prefers the term “data science”, and with regards to healthcare, sees it as the intersection of mathematics, statistics, computer science and healthcare.
“It’s how we bring new technologies and new knowledge into healthcare,” he said, adding that healthcare is still very much paper-based.
Originally associated with information technology, economics and statistics, big data, or data science, has since expanded its influence into many other areas; however, healthcare seems to be one of the final holdouts.
Said Prof Morris: “I think it was Bill Gates who suggested that healthcare is the last major industry not to be transformed by the information age. And I think that it is time we address that collectively.”
Speaking at the Big Data in Healthcare Forum held recently at Taylor’s University’s Lakeside Campus in Subang Jaya, Selangor, the professor of medicine opined that the adoption of such technology is not only about delivering better quality care for patients, but also, the sustainability of healthcare systems on the whole.
“Healthcare systems are characterised by waste, variation and harm, and only with the application of information and data science in real time, will we drive up efficiency and effectiveness,” he said.
This is especially so with many people suffering from more than one disease as they age.
Quoting a population-wide study conducted in his home country of Scotland and published in The Lancet in May 2012, Prof Morris noted that at the age of 70, over 60% of Scots had two or more medical conditions, with some having up to 15 conditions.
“Our health systems are not fine-tuned to deal with multiple morbidities. Why? Because we don’t share information across the system,” he said.
Unifying patient data
Prof Morris observed: “Some countries are very fortunate because they have reams and reams and reams of public sector information, but it’s siloed and it’s not used.
“So, the question is, can we – in a trustworthy way – unleash the power of this big data?”
As Chief Scientist for the Scottish Government Health Directorates, he has seen firsthand the advantages of linking data across the many – usually fragmented – areas of healthcare.
“Our poor patients often have to go from general practice (GP) to hospitals to screening, from laboratory investigations to pharmacist, and the data does not follow.
“And he or she has to repeat themselves again and again, and it’s these gaps in care which is where the harm lies,” he said.
One system in Scotland designed to overcome this is the Emergency Care Summary.
All GPs in the country have electronic health records, which can be accessed by other doctors – with consent from the patient – in case of an admission to any Scottish hospital.
“We do it (accessing patient records) 3.5 million times per annum, and it changes management by 20%, purely and simply by having information at the point of care,” shared Prof Morris.
These records are updated twice daily from the GP systems, and include the patient’s medications and any drug allergies.
Another system is the nationwide PACS (picture archiving and communication system).
“Imaging is becoming a really important diagnostic tool of medicine, and the cost of imaging is going up and up and up,” said Prof Morris.
“So, again, we have a nationwide approach to all MRIs, CTs, ultrasounds, mammography, endoscopy, video imaging; and actually, by pooling together a nationwide service, it reduces reexamination by 10% – and this is in imaging duplication.”
In addition, there was a 100% reduction in film and chemical cost through the utilisation of this system.
The final example Prof Morris gave during his presentation was the Scottish database for diabetic patients.
“For the last 15 years, we have had a single clinical information system for every person living with diabetes,” he said, adding that this includes all their lab investigations, medications – past and present, assessments for potential diabetic complications, and imaging.
“What we’ve shown – and we’ve published on this – is that things get better.
“Amputation rates have gone down 40%, blindness has gone down by about 40%.
“If you have systematic approach to that journey of care, it makes a difference,” he said.
Affirming public policy
Big data can also be utilised in other areas, such as assessing the effectiveness of public health policy, rationalising screening intervals, and researching the effectiveness of drugs.
Scotland banned smoking in public places in 2006. “The question is, did the smoking ban work?” asked Prof Morris.
He explained that they used their already-available data on hospital admissions to check on the incidence of acute coronary syndrome and childhood asthma after the ban, with a comparison to England, which issued the ban later.
“So, the bottomline is, for acute coronary syndrome, admissions fell by 17%.
“Most interestingly, 67% of the reduction was in non-smokers, suggesting that the dangers of passive smoking were probably very much underestimated,” he said.
As for childhood asthma, the number of cases admitted to the hospital had been increasing by about 5.2% per year, while after the ban, it decreased by about 18.2% per year.
“So, it’s nice data to able to show the effectiveness of policy, to use data to inform public policy,” he said.
Another case study was related to the screening for diabetic retinopathy, which is supposed to be done once a year for all diabetic patients.
“The question is, why do we say it’s a year?” he asked, adding that it is possibly because a year is a convenient interval to remember.
“Using data we already had, we said, can we tailor pathways for individual patients?
“So what we showed was that, if you have type 2 diabetes and two examinations showing no eye disease, you can quite safely be offered two-yearly screenings.”
This resulted in 44% less people being screened, saving on time, resources and personnel, without any decrease in the quality of treatment.
Prof Morris also gave an example of precision medicine.
In the first population-wide study of pharmacogenetics, information from the electronic health records was used to look at therapeutic response to metformin.
“What we showed, and published in Nature Genetics, is that those who respond to metformin have a variance in an area called ATM, which is a gene.
“But it was terribly unknown to the diabetes community, but very well-known to the cancer community because it is known that ATM is cancer-protective.
“And this work actually led to about 200 clinical trials internationally on the effect of metformin – not on diabetes, but its medical effect on cancer,” he said, adding that it shows how genetic information is being used to unravel new insights into disease.
Driving the economy
Another area that can be improved with the use of data science, he suggested, is medical research.
“Classically, it is suggested that it takes at least 15 years for a basic science discovery to make it into the clinic. So, how can we use big data to drive and accelerate this pathway?
“And this is a governmental issue, it is a health system issue; we need to support systems that deliver better care and reduce cost,” he said.
As such, Prof Morris posed the question: “What’s an ecosystem that drives research, development and innovation within healthcare? Because this should be the core business for healthcare.
“So, every person or patient I see in my clinic, I should be thinking, how might they contribute to research and innovation so that it enhances the treatments for the folk of tomorrow?”
As the Scottish chief scientist for health, his office has come up with a policy to help answer this question, centred around six guiding principles.
The first is to capitalise on Scotland’s strong academic base. “I think Malaysia has two or three universities in the world’s top 200? So, you have a good academic base,” he said.
The second is to create collaborative partnerships with the missions of research, education and delivery of healthcare. “Research shouldn’t be on one side as sort of an optional extra, alongside quality care. It’s about bringing those worlds together.”
Thirdly is exploiting their ability to collect information through linking up data from health, non-health and social care databases to support better treatment, safety and research. “Data science should be at the heart of everything we do.”
Next is bureaucratic red tape. Prof Morris noted that regulations and governance of healthcare needed to be proportionate – neither too much nor too little.
“Fifthly, it’s about working with industry in a transparent way.
“And lastly, we’re trying to present Scotland as a single research site, which I think you are trying to do here in Malaysia,” he said.
Prof Morris added: “The underlying principle is to see the healthcare system as an asset, rather than as a liability.
“And if we get the alignment right, there is no reason why we can’t drive a strong, stable economy through the healthcare system,”
With the global healthcare market valued at USD1.8tril (RM7tril) and estimated to grow annually by 5% for the next couple of years, he noted: “If you have good products within that market, you can drive economic growth.”