Using computers for intensive care unit research

A computerized clinical information system (CIS) is potentially a very important information tool for research and improving health care processes as well as for optimizing data management and thereby minimizing health care costs. The newest CISs automatically collect patient data from various sources, including monitors, the laboratory, radiology, and patient notes, and make the data highly organized and readily accessible. In the future CISs may be able to conduct signal analysis, assist in care decisions, provide advanced graphical data presentation, and generate warnings to clinicians. Most CIS systems include large databases, and the advent of relational databases has improved data retrieval and manipulation and thus made the data a powerful tool in outcomes research. On the whole CISs collect more frequent and more accurate data than do clinicians using paper-based data collection systems, but research continues on how accurate CIS data is, how to improve that accuracy, and how much data checking and correction is needed. At my institution we have used CIS data to study changes in patients' code status and to evaluate a protocol for arterial blood gas (ABG) testing. The primary challenges to optimizing a CIS are ensuring accurate data entry, learning to query the data so as to avoid misleading conclusions, and to administer and maintain the hardware and software so as to minimize the chance of data loss and system down time. CISs are in a relatively early stage of their development, and engineering improvements will eventually make CIS data highly accurate and easily accessible and queryable so that CISs become even more valuable for research.