We hypothesize that when using a clinician order entry system (COES), nurses and physicians encounter specific information needs that can be predicted based on the information they are reviewing. By automatically providing links to on-line resources designed to resolve those needs, we can decrease the rate at which information seeking is deferred. This should, in turn, increase the rate at which information seeking is successful, leading to better-informed ordering. We will study this problem using programs, called """"""""infobuttons"""""""" that use context-specific information to anticipate information needs and automate retrieval from appropriate resources. For example, an infobutton placed next to patient data in a COES display can use the patient data itself to facilitate information retrieval. We believe it is time to study this approach for COES, at multiple institutions. Our approach will be to extend our current research to: . Study information needs that arise when clinicians use commercial or local COES's at three institutions . Build infobuttons that address the specific needs that are identified . Integrate links from the COES's to the infobuttons . Study the ability of the infobuttons to improve access to information Our study will involve the direct observation of clinicians, using a portable usability laboratory, as they use Web-based COES's at New York Presbyterian Hospital, LDS Hospital and Regenstrief Institute during routine patient care and """"""""think aloud"""""""" about their information needs. We will analyze these activities to determine, for a given context, the information needs most likely to arise. We will then construct institution-independent infobuttons to automate the retrieval of specific information relevant to specific needs. The infobuttons will be integrated into the COES and we will study their use by clinicians to determine if they are usable and useful in the process of actual patient care.
Showing the most recent 10 out of 40 publications
