It’s no surprise that keeping people healthy is costing more money. From the price of medications and the cost of hospital stays to doctors’ fees and medical tests, health-care costs around the world are skyrocketing. The World Health Organization attributes much of this to wasteful spending on such things as ineffective drugs and duplicate procedures and paperwork, as well as missed disease-prevention opportunities.
“All countries can do something—many of them a great deal—to improve the efficiency of their health systems,” reports the WHO, “thereby releasing resources that could be used to cover more people, more services, and more of the costs.”
It’s estimated that the health-care industry could save billions by using big-data health analytics to mine the treasure trove of information in electronic health records, insurance claims, prescription orders, clinical studies, government reports, and laboratory results.
Analytics could be used to systematically review clinical data so that treatment decisions could be based on the best available data instead of on physicians’ judgment alone. Long waits at hospitals for a room could be reduced once calculations can be made to predict when beds might become empty. Flu outbreaks could be contained if health authorities could track the numbers and locations of those who contract the illness.
And finally, ordinary people will gain more control. “We have to figure out how to use these data and technologies to help people make health-enhancing choices,” says IEEE Senior Member Michael S. Johnson. He is director of utility-care data analysis for the Kaiser Foundation Health Plan, in Oakland, Calif. Part of the Kaiser Permanente health system, it has more than 9 million members, 17,000 salaried physicians, 611 medical offices, and 37 hospitals. Johnson was also a speaker at the IEEE Computer Society’s 2013 Rock Stars of Big Data event, held last October in Mountain View, Calif.
“The data accumulating in health records are an unbelievably rich source for improving public health, for communities to understand their own needs, and for spotting inequities and disparities in care within society as a whole,” he says.
Several health-analysis initiatives are under way at Kaiser and other organizations.
Kaiser has the most electronic health records in the United States. Occupying about 30 petabytes of storage, these records double almost every two years, according to Johnson. Although Kaiser began creating electronic records about a decade ago, it just recently developed analytical tools to make sense of the information to improve care.
For example, the company searches through the electronic medical charts of the 1,500 or so individuals who visited or contacted a Kaiser facility the previous day. It then produces a daily report of patients who require follow-up care, such as blood tests or immunizations. The program also looks for gaps in care using evidence-based clinical rules that govern best-care practices, such as how often to monitor the blood sugar levels of diabetics. A record may be flagged and reviewed by the patient’s doctor, who can then schedule a procedure if necessary.
“Instead of seeing only 20 patients a day, doctors are able to see 75 to 100 people and get ahead of the wave,” Johnson says. “You can imagine how satisfying that is for a physician. We believe these types of physician-support tools, used side by side with electronic medical records, are the future of health care.”
Kaiser is also using predictive health analytics to improve procedures in its hospitals, he says, because care there is “measured in minutes, and it’s critical to do the right thing right now.” It is developing programs to prevent falls by patients in the hospital, predict the length of hospital stays, create early-warning systems to spot complications after a procedure, and reduce the number of people being readmitted for the same condition.
Several other health analysis projects were outlined in “A Look at Challenges and Opportunities of Big-Data Analytics in Health Care,” written by researchers from Cisco Systems, including IEEE Senior Member Raghunath Nambiar, the company’s chief architect for big-data solutions. The article is available in the IEEE Xplore Digital Library.
Several big-data projects aimed at battling the flu were outlined in the Cisco researchers’ article. Seasonal flu spreads easily and can quickly sweep through schools, nursing homes, and businesses, hospitalizing millions and killing upwards of 500,000 people worldwide. It’s important to contain the flu’s spread and lessen the chances of a pandemic like the one caused by the H1N1 virus in 2009.
Each week, the U.S. Centers for Disease Control makes available for analysis the more than 700,000 weekly flu reports it receives from health-care providers that contain details about the illness, treatments given, and whether they were successful. The CDC FluView application sifts through and organizes this data to provide a picture of how the disease is spreading as well as what vaccine strain is working best. The information is posted on the CDC website.
The WHO’s FluNet compiles data provided by the National Influenza Centres of the Global Influenza Surveillance and Response System and other national flu-tracking laboratories. That information is uploaded and used for tracking the movement of viruses globally and interpreting the epidemiological data. The real-time data are publicly available and presented in various formats, including tables, maps, and graphs.
Meanwhile, computer scientists at the University of Southern California, in Los Angeles, and medical experts have teamed up to use data to better treat patients with Parkinson’s disease, a progressive disorder of the nervous system.
The team created an algorithm that analyzes data from sensors that track a patient’s movements, including 3-D sensors, similar to those used in the Microsoft Kinect gaming system, a smartphone, and body sensors. The sensors monitor the disease’s progress and the treatment’s effectiveness in real time. If decreased range of motion or flexibility is spotted, caregivers are alerted so they might prescribe different medications or have the patient try other muscle-strengthening exercises.
According to Johnson and the Cisco researchers, with the help of big data, more personalized medicine that uses patient-specific data, including genomics, will be the future of patient care. Individualized treatment plans will control costs and improve quality of life in at least three ways—by reducing trial-and-error prescribing, avoiding adverse drug reactions, and preventing unnecessary hospitalizations.
“Today the health care industry is just beginning to understand all the innovative things that can be done with big data,” wrote the Cisco researchers. “Integrating data from various sources can build predictive models that can lower overall cost and improve quality of care significantly. New data sources and analytics technologies are expected to emerge in the near future that will change the way medicine is practiced.”