A recent survey conducted by the eHealth Initiative and the College of Healthcare Information Management Executives revealed while 80 percent of CIOs and other healthcare executives believe data analytics are important to their organizations' strategic goals, 84 percent said using big data presents a challenge.
Some of this challenge is just how big the data has become and how rapidly it's growing, says Stephen Gold, vice president of the Watson Ecosystem at IBM's Watson Group. "You want to consume [all of the available data] to help you make an informed decision, but if you consumed everything tomorrow and then got to work you'd soon realize the volume of data is doubling every couple of years," he says.
As the amount of available data expands, so does the variety. For example, electronic medical records provide hospitals with large sets of structured data, but also include unstructured data such as transcribed patient histories. And, for a clinician looking to make an evidence-based medical decision, medical literature, clinical trial information and other such new and unstructured data should be considered as well. Of course, it would be impossible for a clinician to keep up with all these new studies and articles and incorporate the findings into care decisions, says Mr. Gold.
The same challenges are faced by hospitals looking to use big data to improve operations decisions, such as supply chain purchasing. For example, a hospital trying to decide what defibrillator to acquire would want to examine recent outcomes from different models, consult recent literature and solicit input from physicians. "Traditionally, research like that takes weeks, but the information moves in real time," says Mr. Gold. In order to make an efficient decision, hospitals need to use an analytics solution to harness the information and make it useful.
Hospitals have made "tremendous progress" in using structured data from within their organization in decision-making, says Mr. Gold. Where there is room for improvement is embracing the increasing variety of data, particularly data that comes from outside the hospital or health system's walls. "We're starting to see more collective organizations look into data-sharing and some really advanced analytics," he says.
One of these organizations is the American Joint Replacement Registry, a national database containing information on hip and knee replacement procedures from across the country. The database is currently collecting information from 297 hospitals from 48 states, with the ultimate goal of eventually housing data on 90 percent of the hip and knee replacement procedures performed in the country, says Caryn D. Etkin, PhD, the AJRR's director of analytics.
The registry currently collects Level I data, including basic information on the patient and implant and the diagnosis code used. The registry will soon collect Level II data, including complete information on complications and patients' comorbidities, though through the same electronic submission method the AJRR currently uses that is designed to minimize providers' reporting burden, says Randy Meinzer, the AJRR's director of IT.
The AJRR has offered this data to providers in the form of individual reports that show organization-specific results and metrics, and it has produced national reports with de-identified data to highlight trends across the country. The AJRR is in the process of launching electronic dashboards that will give participating providers access to this information more quickly in an easy-to-access format.
The AJRR hopes this information will help providers select implants and make other decisions that will improve care. "The AJRR is patterned after other registries that have been in operation for over 20 years, Swedish registries, and what they learned is that once they started collecting this information the revision rates decreased by as much as 50 percent," says Mr. Meinzer. "That's what we expect in the U.S. — maybe not as dramatic, but an improvement in the revision rate."
Dr. Etkin says she believes the AJRR dashboard will lead to lower costs as well. "Any decrease in the revision rate will mean substantial savings for the U.S. government and other payers, and that's a big point," she says. She also expects the addition of Level II data into the database will provide hospitals with more information to help hospitals track, and improve, other quality measures as well.
Another national database has the same goals, but a different data source. FORCE-TJR, funded by the Agency for Healthcare Research and Quality and run by the University of Massachusetts' department of orthopedics and physical rehabilitation, is collecting information directly from a representative cohort of 30,000 total joint replacement patients of more than 150 surgeons in 22 states.
FORCE-TJR researchers are checking in with each patient in the cohort annually, and will continue to do so indefinitely, to gain a complete picture of the patients' postoperative experiences. Because more than 25 percent of early post-TJR medical events are treated at a hospital that is different from where the surgery occured, by going directly to the patient and not using data from the hospital or surgeon, the researchers can keep tracking each patient even if they move or visit another provider and get a more complete picture of the patient, says Patricia Franklin, MD, MBA, MPH, a professor at the University of Massachusetts Medical School and one of the principle investigators on the project.
Providers that partner with FORCE-TJR are given access to the data to help them get a better picture of their patients' long-term outcomes. "They don't want to be surprised by CMS' reports" and subsequent reimbursement reductions, says Dr. Franklin, "so by working with us, they get the data in real time and can understand how they compare to the rest of the nation."
Having access to clinical, rather than just claims data, is empowering to providers, says Dr. Franklin. Claims data tells providers what is happening, clinical data like FORCE-TJR's can help providers understand why. "And by understanding why, they can start making changes to improve outcomes," she says.
Mr. Gold, Mr. Meinzer, Dr. Etkin and Dr. Franklin will join David Rodriguez, vice president of MD Buyline, to discuss how to use big data to optimize purchasing decisions in the orthopedic service line during a July 31 webinar. For more information or to register, please click here.