- Geisinger Health System has launched an initiative integrating its existing analytics and management systems with point-of-care data in effort to use big data, system officials wrote recently in the Harvard Business Review.
- The system, called a Unified Data Architecture (UDA), not only uses the simple numerical data available in standard databases, it can process more nuanced information like the free-text language in physician reports.
- Processing of this data can yield insights about individual patients as well as patient and provider trends.
Although a growing number of medical researchers have begun using big data, those directly involved in patient care have found it challenging to go much beyond integrating EHRs into their routines. Add EHR challenges to the siloed legacy data systems within many hospitals and a lack of computing power and you find an environment that is hardly conducive to processing huge amounts of integrated data.
In spite of these obstacles, Geisinger Health System has implemented a computing solution that is allowing them to leverage their existing patient and provider data. Using this UDA, Geisinger staff can track and analyze patient outcomes, compare genomic sequences to clinical care and examine data across patient and provider groups. Ultimately (and with patients' permission), the system will also be able to access and integrate data from third-party systems, such as mobile applications and grocery reward cards.
Although they are still in the process of integrating data into the UDA, Geisinger staff have already experienced an early success: Identification of patients with life-threatening abdominal aortic aneurysms, found via imaging tests that were intended to evaluate traumatic injuries. Such findings in the trauma setting tend to drop in priority due to an intense focus on severe, acute injuries. However, analytics performed on free-text imaging reports pick up on the signs, and staff can then offer surgery to the highest-risk patients. Geisinger reports that this program, which they call “Close the Loop,” has already saved lives.
So far, other uses of the data include tracking waste in operating rooms, offering comparative surgical cost and outcome data and identifying patterns in the use of supplies.
The analysis of free-text is seen as a means to help create the whole picture of a patient as important information, such as social behavior determinants, often fall through the cracks of EHR systems in unstructured text. Being able to analysis and catch such free text can help care delivery members provide the best care for a patient.