Analysis of Google searches and vast amounts of statistics are helping to uncover new medical treatments and ways to improve health.
Big data. It’s a term you may have heard thrown around, but what exactly is it? And how could it be making a difference to our health now and a few years into the future?
Big data is the broad name given to massive and complex sets of detailed statistics and information. From a health perspective, that data might be hospital admission numbers, the number of people diagnosed with a certain disease, or the number of times people do a Google search for a specific symptom they are worried about.
However big data is collected and whatever shape it takes, experts believe it promises to be a game changer in healthcare. It has the potential to improve and streamline healthcare systems and to help find ways to better prevent, diagnose, treat and cure disease.
“Big data is prevalent in our daily lives. When we consume services that are delivered digitally, we leave behind ‘breadcrumbs’ that can be analysed to predict preferences and provide more customised experiences,” says JobTech co-founder and Great Women Of Our Time 2018 Public Service & Education nominee, Charlotte Lim.
“You experience that with online retailers who promote specific products based on purchase history, and with streaming sites that provide movie and music playlist recommendations. To help people reskill and upskill to keep up with changes in the labour market, JobTech uses big data to provide transparency in terms of which jobs and skills are in high or low demand, and their projected growth rates.”
PATIENTS LIKE YOU
In the US, entrepreneurs Ben and Jamie Heywood are already harnessing the promise of big data. They started PatientsLikeMe – the “world’s largest personalised health network”. It has 60,000 members who use the network to report and share their experiences of some 2,800 different medical conditions.
Since it began, PatientsLikeMe has generated more than 43 million pieces of data about different diseases. It uses advanced mathematical techniques to analyse and finetune this data to help expand experts’ knowledge of diseases and to explore current and new treatments. Ultimately, PatientsLikeMe aims to improve healthcare and so lead to better outcomes for patients.
People can use the network and the information provided by other members to find answers to their own health questions and to track their progress, symptoms and treatments.
Last year, PatientsLikeMe used data from members to find out more about their healthcare experiences and what they regard as “good” care. The research found patients living with fibromyalgia, post-traumatic stress disorder and major depressive disorder are least satisfied with their healthcare, while people with amyotrophic lateral sclerosis (ALS), multiple sclerosis and Parkinson’s disease are most satisfied. The results flag that people with certain health issues may need greater support.
The big data also gives insights into the factors that patients believe are important in ‘good’ care. These include having a healthcare provider who fully explains treatment options and believing they are receiving the best possible healthcare for their condition.
YOUR HEALTH DOPPELGANGER
Isaac Kohane, a computer scientist and medical researcher at Harvard, is also pushing the boundaries of big data. He wants to organise and collect the health information of as many people as possible, so doctors can then search that data to find the “doppelganger” of a patient sitting before them.
Once that doppelganger is located, doctors can look at what successful treatments they may have been given, the outcome of those treatments, what didn’t work as well, and any symptoms or reactions they experienced. That information could then be used to develop more personalised and focused diagnoses and treatments for breast cancer patients, believes Isaac.
He argues that the datasets that doctors currently use to diagnose an illness are small – often based on the people they have treated themselves or what they have read in academic papers and medical journals. Much bigger datasets of patients’ overall health information would offer greater insights and, potentially, better treatments.
Seth Stephens-Davidowitz, a former Google data analyst who has spent years studying big data, adds,“If someone suffers depression or ADHD, there are great variations within those illnesses.” Seth, who is also the author of Everybody Lies: What the Internet Can Tell Us About Who We Really Are, says,“Some people with depression sleep all the time, some people can’t sleep, some people eat a lot, some people don’t eat much at all.
“There are different types of depression and the treatment that works for you may depend on the variant you have. If you can find your doppelganger – the patients who are most similar to you on many levels and who have the same health history and similar symptoms, you can see what has worked for them.”
REAL WORLD RESULTS
Big data research does not limit itself to use in the healthcare sector alone. Fashion, too, is making use of digital innovations and that includes apparel designers, manufacturers, and retailers tracking what customers want so they can improve their services, predict market trends and stay ahead of their competitors.
Fashion brands can measure customer reaction to samples and ideas, then adjust their products accordingly. In this sense, consumers are literally taking part in the design process. Such a high level of consumer buy-in all but guarantees a product’s success.
For the latest collection of her label ANNAKIKI, Chinese designer Anna Yang had a special collaborator: Huawei’s P30 Pro Smartphone. Pitched as the first fashion collection to be co-designed with Artificial Intelligence, it was unveiled at Milan Fashion Week in May to critical acclaim.
To create the unique 20-piece collection, Huawei specially developed an original Fashion Flair app with which Anna could interact, take inspiration from and design with. But this step towards a new smartphone-based approach to fashion design came with its own set of limitations. “The app in the end is still a machine, so despite the very good input it gave, it was still up to me to give the collection a human touch and add my own element of human creativity to it,” she shares of the design process.
- Charlotte Lim, co-founder, JobTech
Internet searches can also show how big data plays a role in improving health, says former Google data analyst Seth Stephens-Davidowitz. “Researchers at Columbia University and at Microsoft as well have found that you can frequently diagnose someone with pancreatic cancer based on the symptoms they search online,” he says.
Microsoft tracked data from tens of thousands of anonymous users of Bing, Microsoft’s search engine, and coded those given a diagnosis of pancreatic cancer. They then looked at the group’s searches for various health symptoms and found subtle patterns.
If someone searched for “indigestion” followed by “abdominal pain”, that was a risk factor for getting a diagnosis of pancreatic cancer within a short time. But if someone searched “indigestion” or “abdominal pain” alone, that wasn’t a risk factor for the cancer. Searching for “back pain” then “yellowing skin” also turned out to be a sign of pancreatic cancer, searching for back pain alone wasn’t.
"Artificial intelligence can be very useful but it can’t completely replace human creativity"
The Impact On Patients In Singapore
In Singapore, a recent example of big data at work is Changi General Hospital’s (CGH) analysis into patient waiting times at the emergency department. Dr Chow Wai Ling and her team combed through three years’ worth of data in a bid to improve the hospital’s operations.
Their research approach looked to answer a number of questions: Who were the patients coming to the emergency department and what were their numbers? On which days of the week and at what times of the day were they coming in? And, were they acute cases?
After analysing usage patterns, the team proposed a redistribution of manpower to better match the arrival patterns of patients. The emergency department then rejigged its roster for doctors, and saw astonishing results. Average median times for patients with more serious conditions shortened from 33 to 25 minutes, a 24 per cent improvement.
Doctors have also reported a qualitative improvement in their workload, says Dr Chow, Deputy Director of Health Services Research at CGH.
The hospital also dived into the data of thousands of diabetic patients under its Health Management Unit tele-health programme and found that it was patients with poorly controlled diabetes who benefitted most from the periodic phone calls from nurses. This enabled the hospital to revise the enrolment criteria and deploy resources more cost-effectively.