Big Data, Cognitive Computing, & Healthcare

Everybody seems to agree on two things when it comes to healthcare — it should be affordable and effective. Professionals hoping to achieve this “doing more for less” objective are hoping that big data analytics can help. The most obvious applications of big data analytics are helping medical researchers unlock the secrets of troubling diseases and then helping develop treatments for them. Last year, Jennifer Bresnick (@JenRoseBresnick) reported, “The National Institutes of Health (NIH) isn’t just ready to mine the big data mountain: they’re already swinging the pickaxe. The Big Data to Knowledge (BD2K) initiative is ready to maximize the usefulness of biomedical data for research and analytics, … by enlisting the cooperation of stakeholders across the industry to access, harness, and extract value from the copious amounts of data being produced by healthcare organizations and researchers each and every day.” [“NIH develops big data plan to encourage research, analytics,” HealthITAnalytics, 14 July 2014] NIH researchers note:

healthcare and cognitive computing 01“While research has always involved the collection and organization of data, the volume, variety, and velocity of current ‘big data’ production presents new opportunities and challenges in both scale and complexity. At the same time, there is a broader cultural shift underway from approaches that kept data mostly private with sharing of resultant knowledge in the form of publications to an information-based culture that dynamically engages the scientific community through the active sharing of both data and publications. Big data are not only a new reality for the biomedical scientist, but an imperative that must be understood and used effectively in the quest for new knowledge. Needed are new approaches for data management and analysis that allow scientists to better access and extract value from data so as to advance research and discovery.”

One of the “new approaches for data management and analysis” that has emerged is cognitive computing. Cognitive computer systems are able to ingest vast amounts of data and, using machine learning and natural language processing, discover insights and relationships that would have otherwise remained hidden. One of hurdles to medical research has been and remains patient privacy. Nothing is more personal or intimate than one’s health — and there are concerns that medical information made available for research could be used in other ways that could negatively impact a patient’s life (e.g., in areas such as employment and insurance). Nevertheless, researchers are hoping that patients see the bigger picture and voluntarily make their data available for analysis.

Data analysis is not just going to help researchers in large studies it will help medical personnel improve personal care. Duke Medicine researchers assert, “The changing dynamic of health studies driven by ‘big data’ research projects will empower patients to become active participants who provide real-time information such as symptoms, side effects and clinical outcomes.” [“New technologies fuel patient participation, data collection in research,” R&D Magazine, 10 July 2014]

“Fueled by new technologies — including electronic health records and monitoring devices that people can wear as clothing or accessories — health studies are now poised to integrate data from a much larger pool of information. The new data is immediate and actionable, providing not only research material, but also clinical information that can improve the patient’s care in the short term.”

One man who believes this kind of data sharing and analysis represents the future of healthcare is Steven Keating, a graduate student at the MIT Media Lab and a brain cancer survivor. New York Times‘ columnist Steve Lohr (@SteveLohr) writes, “The young scientist’s collection and analysis of his own data makes him an extraordinary exception today, but physicians and health care experts say he is a sprinter along a path others are walking — toward consumers taking a more active interest in gathering, studying and sharing their medical data.” [“Using Patient Data to Democratize Medical Discovery,” The New York Times, 2 April 2015] Lohr continues:

“Better-informed patients, they say, are more likely to take better care of themselves, comply with prescription drug regimens and even detect early-warning signals of illness, as Mr. Keating did. But there is another big potential benefit of the wider use and sharing of patient information — medical research. Mr. Keating himself calls it a ‘huge crowdsourcing opportunity for research.’ Not surprisingly, he has offered up his own data for several research projects. Research organizations are beginning to look at the opportunity in new ways as well. Traditionally, patient information has been tapped for research in large pooled databases. Individual contributions only made it in as part of these centralized efforts. That is still the norm, but change is coming. Institutional practice, individual attitudes and new technology are the reasons.”

In another article, Bresnick highlights how big data analytics has been used to help doctors and patients in the area of cardiology. [“Case study: Big data improves cardiology diagnoses by 17%,” HealthITAnalytics, 7 July 2014] She reports:

“In [a] case study examining the impact of big data analytics on clinical decision making, Dr. Partho Sengupta, Director of Cardiac Ultrasound Research and Associate Professor of Medicine in Cardiology at the Mount Sinai Hospital, has used an associative memory engine from Saffron Technology to crunch enormous datasets for more accurate diagnoses. Using 10,000 attributes collected from 90 metrics in six different locations of the heart, all produced by a single, one-second heartbeat, the analytics technology has been able to find patterns and pinpoint disease states more quickly and accurately than even the most highly-trained physicians.”

Dr. Sengupta explained to Bresnick that one reason “learning intelligence platforms” (i.e., cognitive computers) are making a significant impact in medical research is because they can integrate data from numerous sources. “The problem is that the data is scattered everywhere,” Sengupta stated. “It’s in the EMR, but everything is still in siloes.” He went on to note that in his study, the associative memory engine was able to diagnose different heart diseases with greater accuracy than humans.

“In the initial pilot phase, when I did my own statistical algorithms, we had about 73% ability to differentiate the two diseases. But when the initial pilot run happened, we were very pleased to see that there was a discrimination of 90% between the two datasets and without any human intervention. What that means is that the highly complex analyses that were done produced a discrimination which exceeded human ability to diagnose the two conditions. Having said that, you have to be extremely cautious, but it’s very exciting that machine learning and learning intelligence platforms can reach the ability to do this differentiation, if not exceed it.”

In another unique big data approach, the Huntsman Cancer Institute at the University of Utah is using genealogical records to complement health records in its attempt to find cures for cancer. Not only can big data analytics save lives, but, according to 2011 report by the McKinsey Global Institute, big data analytics could save the American health care system $300 billion per year. The upside to big data analytics far outweighs the potential downside. Lohr reports that reluctance to share personal health information is declining and that could be good news for all of us.

Follow me on Twitter