Ask physicians about decision support systems, and you’ll probably get some groans and rolling eyes. Traditionally, decision support systems have amounted to little more than a set of annoying alerts, flagging issues that clinicians are already aware of and wasting precious time.

Early efforts to provide decision support have rarely surfaced the real insights that are needed for improving patient care, whether for an individual patient, or for a population of patients served by a practice or hospital. In fact, the National Research Council has concluded that the institution of healthcare is not adequately structured to help clinicians systematically improve their decision-making and practice, due in part to difficulties with data integration and sharing. Physicians develop a high comfort level with their decision-making processes, but they simply do not know what they do not know.

This no longer has to be the case because modern information technology has the potential to not just digitize record-keeping and paper-based processes, but to enable medical professionals and patients to leverage the wealth of available electronic medical information in a way that supports a new standard of analysis and decision-making.

The era of big data is no longer approaching; it is here now. An EMC-sponsored study found the world’s information is doubling every two years, with a colossal 1.8 zettabytes to be created and replicated in 2011. This is the equivalent of every person in the world having over 215 million high-resolution MRI scans per day. In medicine, the PubMed database sponsored by the National Institutes of Health now covers 4,600 journals that have pertinence for medical research publishing, and includes more than 22 million citations.

The volume of available medical data clearly exceeds the ability of individual humans to effectively process the information and consistently make superior decisions.

To be worth the time and effort of implementation, however, intelligent decision support technology must be able to use the analytical and predictive capabilities of high-power computing to help medical decision-makers go beyond record management to real insight and discovery. Enabling intelligent insight and discovery requires a robust integrated system capable of overcoming the challenges that exist in health IT today. We can understand the key components of such a system by outlining its five pillars:

• Using real-time data feeds and predictive analytics to facilitate advanced decision support across the healthcare spectrum, including both clinical and non-clinical settings.
• Facilitating the transition from legacy systems to modern integrated health data systems by creating a framework that provides a sensible, cost-effective, surround-and-supplement migration strategy and enhancing rather than interfering with workflows.
• Using metadata-based approaches to manage the massive data volumes available and to enable flexible data-management strategies that easily adapt to change.
• Encouraging innovation by leveraging a modern services-oriented architecture and through the use of a robust platform supporting multiple modular and portable applications.
• Ensuring high-level performance and results through comprehensive evaluation and control mechanisms.

In my next posts here I will explore these five pillars in turn, starting with the first pillar, real-time data and predictive analytics.

I’d also like to hear from readers who have been working with decision support systems, to know what has been successful for you and what has simply increased your workload without extending clinical insights. What has your experience been?