How Big Data is Transforming Health Care
It used to be said that the pace of change in information technology was the fastest of any industry. Today there are competitors for the top spot, and health care is one of them. What follows is a short tour of some of the ways that Big Data is influencing how health care is delivered and managed.
Big Data: Coming Soon to Operating Rooms
Clinical Decision and Quality of Care Analytics
Clinical decision support applications cover a broad spectrum of potential uses. Most think of physicians as the primary consumers, and there is little doubt that a major impact can be made at that level. But the entire health care food chain — from patients through home health care aids to nurses and physicians — could be influenced. For instance, IBM’s Jeopardy champion Watson has been reworked to operate in hospital settings. McKesson’s InterQual Criteria application is used by the US Centers for Medicare and Medicaid to assess quality using condition-specific models.
Medical Data Streams
Think of this trend as Netflix for medical instrumentation. What was once batched or store-and-forward data will be pushed toward real time. Top of mind for many people is “biotelemetry,” or human sensor data streams. For instance, CardioNet offers what they call “next-generation cardiac monitoring service with beat-to-beat. Real time analysis, automatic arrhythmia detection and wireless ECG transmission.” An innovative student project was designed to capture fetal heart and uterine contractions. Some sensor systems may leverage wireless ad hoc network standards like ZigBee even while complaining that sensor speed limitations are impairing data quality objectives.
Whether deployed in rural areas due to a shortage of specialists, or in urban areas to provide certain types of care where patient transportation is either a risk or an additional expense, telemedicine is closely related to the emergence of medical data streams. Telemedicine represents a composite of data requirements. Prior to the start of a visit, information must be marshalled from multiple sources to assist remote providers. In addition, data must be relayed from patient monitoring devices back to the providers. Federated systems, not yet in wide use, should be anticipated.
Reporting to public tumor and disease registries are already mandated for acute and ambulatory care facilities, but soon that data will move more quickly to registries like the U.S. Surveillance, Epidemiology and End Results Program (SEER). Kentucky, which has the dubious distinction of having the highest incidence of cancer, is working to streamline the workflow of clinical data to epidemiologists. Institutions like the Mayo Clinic, Kaiser Permanente, the Veterans Administration and United Health Care have records for as many as 100 million individuals, including longitudinal health history on individuals. This de-identified data at registries and held in private repositories is rich for analysis.
A glance at the GE Health Care product suite offers a quick insight into the extent of image-enhanced applications for health care. Products groups include “interventional” suites to supplement human capabilities in radiology, cardiology, oncology, neuroradiology and electrophysiology. A Siemens software offering allows 3D visualization of radiological CT databases. It also facilitates display on mobile devices, and perhaps more significantly, such tools enhance collaboration by making it easier to share imagery with other caregivers and specialists.
Searchable Medical Imagery Seen as Boon to Radiologists
Others are working to make the already vast distributed database of radiological images fully searchable. This could improve outcomes for cancer patients by facilitating access to relevant research — based on image properties, not only image metadata.
Other Big Data Healthcare Initiatives
This short list omits significant Big Data projects. Consider the tracking of molecular diagnostic tests. One forecaster predicted that this is the fastest growing sector of clinical pathology laboratory testing. It is claimed that there are more than 3,000 molecular and genetic diagnostic tools currently marketed. There is the Big Data being accumulated as part of healthcare.gov. And new frontiers of drug administration, in which dosage is controlled based upon real time feedback from embedded patient monitors. Return on Investment for these innovations is difficult to assess.
Electronic Health Records were first proposed in the 60s as a timesaver as well as a source for quality improvement. Resistance to EHR implementation has been considerable, with medical and nursing schools lagging the technology. Now that EHR has been mandated for many institutions, the payoff is not always clear. Much of the valuable medical records data is unstructured (which makes it suitable for Syncsort Hadoop connectivity and data transformation for analysis. Unfortunately, most of that is still offline and unavailable for analytics or clinical decision-making.
Biggest Data Yet to Come
Big Data innovations are poised to transform many aspects of how health care is delivered. Which Big Data initiatives are worth the cost and inevitable privacy tradeoffs? This too is a Big Data question, which can only be addressed by gathering cost and effectiveness information on a massive scale. Beyond data availability, these changes may be so transformational that it could require a new generation of physicians and specialists whose training more fully integrates the use of Big Data.
Mark Underwood writes about knowledge engineering, Big Data security and privacy.
Meta: Despite a slow start and growing pains in many sectors, health care is one of the fastest growing sources of Big Data.