The following HTML text is provided to enhance online
readability. Many aspects of typography translate only awkwardly to HTML.
Please use the page image
as the authoritative form to ensure accuracy.
Computational Technology for Effective Health Care: Immediate Steps and Strategic Directions
Four Domains of InformationTechnology in Health Care
Motivated by a presentation from Intermountain Healthcare’s Marc Probst, the committee found it useful to categorize health care information technology (IT) into four domains:
Automation. Automation is the use of IT to perform tasks that can be repeated with little modification—examples include bar code medication administration, generation of laboratory results, and issuing invoices for payment.
Connectivity. Connectivity begins with physical infrastructure—ensuring base-level electronic connections between various physical facilities so that data can be transmitted electronically. Examples might include high-speed fiber lines and routing capabilities throughout a physical plant, wide-area networks, and the deployment of wireless infrastructure. Connectivity includes interfaces that map data from one system into another. At the highest level, connectivity involves connecting people to systems and to each other.
Decision support. Decision support (DS) involves the use of IT-based applications to provide information at a high conceptual level to clinicians to facilitate or improve decisions made about care. For example, DS can include simple rule-based alerts such as reminders to physicians about possible drug interactions when medication orders are entered. DS can also involve the presentation of information to care providers in ways that make it easier for them to know how to direct their attention—a “dashboard” indicating patient status across an entire ward or for a physician’s 50 sickest patients would be an example of DS for presentation. Finally, DS can also refer to statistical and heuristic decision support reflecting an intelligent synthesis of information about the patient, information from the care setting, and biomedical knowledge—for example, a DS system might recommend a particular antibiotic based on the patient’s condition and a database of the recent sensitivity of microorganisms to different antibiotics in their hospital.
Data-mining capabilities. Data-mining capabilities use knowledge discovery techniques to analyze various similar or dissimilar datasets to recognize known or unknown relationships. Data mining converts raw data signals into clinical variables and models to provide a rich source for new approaches to evidence-based medicine and personalized care. Examples range from identification of a marker for breast cancer therapeutic response from microarray data, through mining the text literature for little-known drug-drug interactions, to mining multimedia electronic health records to identify a patient’s condition from a text note or a change in heart size from a sequence of images, and extracting ideas or relationships from a recent publication in a leading journal and pushing the information to the physicians who are treating patients who may benefit from those findings. Data mining provides many of the inputs needed for decision support.