Big Data: ONC Bullish, Providers Cautious
The Office of the National Coordinator for Health IT (ONC) is big on big data. In fact, it recently proposed four pieces of legislation aimed at boosting “the interoperable sharing of electronic health information to improve the health of our nation.”
The first would prohibit information blocking, both by providers and health IT developers, which ONC says poses a threat to healthcare delivery system reform. The second would require transparency in health IT products and services (ONC says more detailed and meaningful information about certified products is needed). The third would address the business policies, practices, and behavior of health IT entities to ensure seamless, secure information exchange. The fourth would establish a public-private partnership to encourage better reporting and analysis of health IT-related safety events.
“As with all technology, health IT will continue evolving rapidly, and we must be able to keep pace with the needs of the ever-changing landscape,” write ONC’s Karen DeSalvo and Lisa Lewis. “These legislative proposals are a critical component of our efforts to do so, and to enable truly interoperable health information exchange that benefits all individuals and the health system at large.”
A hybrid approach
While there’s little question of the benefits of information exchange, some are questioning over-reliance on big data when it comes to care management. Despite the wealth of clinical and financial data available, provider insights and patient-reported data must not be overlooked, say Dr. James Colbert, senior medical director for population health at Verisk Health, and Dr. Ishani Ganguli of Mass General Hospital Ambulatory Practice of the Future.
Doctors Colbert and Ganguli wrote a blog post for Health Affairs pointing out that while big data analytics tools are excellent at stratifying a large patient population to identify at-risk patients, they tend to over-represent patients with transiently high costs due to a catastrophic event. They may also miss social factors that play a role in health outcomes.
For example, an elderly man with diabetes and congestive heart failure seems the perfect candidate for care management, but the care manager may find non-health issues such as not having transportation to get to his cardiology and primary-care appointments are the main factors in his frequent hospital admissions.
Colbert and Ganguli advocate a hybrid approach that starts with a population review using claims and EHR data, then moves to a physician review to make sure details such as living situation, social support networks, and ability to follow through with personal health goals are considered. The authors also stress the importance of patient insights, pointing to a study of Medicare beneficiaries that uncovered three patient-reported questions that accurately predicted hospital admission and identified the costliest patients.
Help for data challenges
Despite the government’s bullish attitude toward big data and technology’s potential to identify at-risk patients, many hospitals continue to struggle with data analysis. “The myriad data sources and formats necessary to fully use Big Data challenge healthcare organizations,” writes research analyst James Haight of Blue Hill Research in “Bridging the Data Preparation Gap: Healthcare.”
Haight says a lack of experience handling the unstructured data types associated with critical healthcare data is the biggest issue. He points to software aimed at bridging the gap, a new breed called data preparation solutions that allow data analysts to build and manage data products faster and more effectively.
The bottom line for hospitals seeking improvement is an ability to integrate data, efficiently use existing data resources, and give business teams user-friendly tools that let them explore hypotheses without relying on IT.
Definitive Healthcare has the most up-to-date, comprehensive and integrated data on hospitals, physicians and other healthcare providers. We track dozens of big data vendors and applications, including EHRs, business intelligence, data sharing, patient management, and more.