Through my recent book and research, I have had the opportunity to speak at a number of conferences and talk to dozens of experts with interests in tapping the power of data. Six months ago, McGraw-Hill published my book Open Data Now – an overview of data-driven innovation that’s been previously featured in YourStory.
Since then, I’ve been directing the Open Data 500 study at the GovLab at NYU, a project to learn how hundreds of companies are using government open data as a key business resource. We often talk about where the data revolution is likely to have the most disruptive impact. And we often come up with the same answer: healthcare.
Our growing ability to analyze huge amounts of healthcare data is likely to revolutionize the practice of medicine. We can expect much more efficient systems for tracking patients and their care, leading to lower costs and fewer medical errors. We can look forward to more data-driven diagnostics, treatment plans, and predictive analytics to determine the best treatments more scientifically. And we will see a new era of personalized medicine, where data about an individual – ranging from genetic makeup to exercise habits – becomes data to help algorithmically determine a strategy for care.
Early in June, all of these possibilities and more were discussed and demonstrated at the Health Datapalooza in Washington, DC. ‘Datapalooza’ is an Americanism – it literally means ‘an all-out crazy party of data.’ The Health Datapalooza has grown from a small group meeting convened in 2010 by the US Department of Health and Human Services, to an annual event that gathers about 2000 people. It brings together dozens of companies, government agencies, and non-profit organizations using data in innovative ways to improve health. Their work shows how four different kinds of data – Big Data, Open Data, Personal Data, and Scientific Data – can be used together to great effect.
I. Big Data
By analyzing Big Data – voluminous, complex collections of data of all kinds – healthcare analysts are now able to find patterns in public health, healthcare costs, regional differences in care, and more. Much of this analysis involves data that is privately held by insurance companies, hospitals, or other healthcare organizations, or includes confidential patient records. It’s not available for everyone to analyze and examine.
For those who have access to it, however, this kind of Big Data can lead to new insights in patient care, and new strategies for providing better, more efficient healthcare.
II. Open Data
Open Data on healthcare can be used much more widely. While Big Data and Open Data overlap, they are not the same. Open Data, as I’ve described in my book, is accessible public data that anyone can use to start new ventures, analyze patterns and trends, make data-driven decisions, and solve complex problems. Not all data can be ‘opened up’ easily or even legally – patient records, for example, should not be made public – but the release of more Open Data is creating new possibilities in healthcare. And when Big Data is also Open Data, the opportunities are especially great. Both the US and the UK governments, which are leaders in releasing Open Data, now provide big, open datasets that are starting to transform healthcare in both countries.
In the US, with its unusual and sometimes dysfunctional mix of public and private healthcare coverage, the Centers for Medicare and Medicaid Services (CMS) have become a major source of this kind of data. CMS, which collects lots of big data about publicly financed healthcare, made headlines recently by releasing records of Medicare billing that named the doctors involved, and showed that some were making millions or even tens of millions of dollars a year by manipulating the system. Those revelations were only part of a larger program to release data that is shedding light on hospital quality and safety, healthcare costs, and other issues. As one speaker noted at the Datapalooza, “Medicare has released more data in the last five years than they did in the first 50.” Other government agencies are following suit: The first day of the Health Datapalooza, the US Food and Drug Administration launched openFDA, a new website that makes data on adverse drug reactions readily available for the first time.
In the UK, where the National Health Service has centralized healthcare, there is even more potential to use big, open data to improve health. A recent analysis of drug prescribing patterns showed widespread variation across the country, with doctors in some areas prescribing more brand-name than generic drugs at a higher price with no clinical benefit. The researchers estimate that changing doctors’ prescribing habits, based on this information, could save hundreds of millions of pounds per year.
The GovLab at NYU, where I serve as senior advisor, has now released a new study on the potential power of Open Data from the National Health Service.
III. Personal Data
Personal Data is the third piece of the puzzle. People are getting more control of their own health data to improve their own healthcare. In the US, the Blue Button program, which started in the Veterans Administration, has made it easy for people to download their own medical records. That program has now been adopted by private healthcare providers to make medical records available to 150 million Americans.
Add to this the new quantified-self applications, which enable people to track their blood pressure and other vital signs continuously, and you have the prospect of more personalized, and effective medical care. Adriana Lukas, head of the London Quantified Self Group, made a prediction at the Datapalooza: “In the future, my doctor will prescribe apps and then see the data on a dashboard.”
IV. Scientific Data
Scientific Data is the last piece of the data revolution in healthcare. In addition to new data about doctors, treatments, and patients themselves, we now have more open scientific data, particularly data about the human genome that can be used to improve health. The Human Genome Project, which ran from 1990 to 2003, pioneered a new way of doing science: The scientists involved made a commitment to share their data openly and early in the interest of rapid progress.
In addition to collecting massive amounts of data on the genome, scientists have now developed much cheaper ways to analyze an individual’s own genetic makeup: the price of sequencing a single person’s genes has plummeted from $100 million in 2001 to about $4000 today, and may drop to $1000 by the end of the year. We’re heading into the era of pharmaco-genomics — using all this genomic data to tailor treatments to the individual more precisely and more effectively than ever before.
Where does this all lead? In a keynote at the Datapalooza, Vinod Khosla, a leading tech venture capitalist and the former CEO of Sun Microsystems, laid out his vision. Over the next decade or two, he said, “Data science will do more for medicine than all the biological sciences combined.” Khosla argues that we’ll have to adopt more data-driven medicine to make healthcare more accurate: In the US, 210,000 people die each year from preventable medical errors. “The error rate in medicine,” said Khosla, “is the same as if Google was allowed to have a driverless car that had one accident a week.”
Human doctors are not computers, and Khosla argues that computers are better at avoiding not just obvious errors but more subtle ones: They excel at analyzing data to predict the likely benefit that a procedure or treatment will have for a specific patient with a specific medical history. After 10 or 20 years of training by expert doctors, Khosla believes that computers will help us move from “the practice of medicine” to “the science of medicine.” The doctor’s role will shift from an analytic one to a more human, and humane, role of helping patients make choices and use the medical care that’s available to them.
The vision of a new data-driven medicine is still evolving, and the results remain to be seen. But the vast amounts of newly available health data are already driving a lot of investment in new companies that are putting it to use.
Stay tuned for the coverage of 20 of the most interesting companies in my next post for Your Story!