A Large Scale Oracle-Based Severity and Outcomes Database for Intensive Care

M. Michael Shabot
Cedars-Sinai Medical Center
Surgery & Info Services

The clinical data within CareVue, when combined with hospital outcome data, forms the basis of an extremely valuable long term CQI database. Over the course of the last 14 years we have created such a database, which currently resides in Oracle. This database contains CareVue, severity, procedure, diagnosis and outcome data for nearly 30,000 consecutive Surgical ICU patients. Our ICU data collection started with information exported from HP's PDMS system and continued without interruption when the SICUs converted to CareVue in 1991. CareVue's dbExport facility is used to extract data from CareVue, although conversion to HP's Database Archive (DBA) system is expected in the near future. Complete demographics, flowsheet, laboratory, history and physical and other data are exported from CareVue daily, from which three severity of illness scores are computed: the Simplified Acute Physiology Score (SAPS), APACHE II and a quantitative TISS score (QTISS). The daily data are reorganized into longitudinal patient files with C and C++ software written at Cedars-Sinai. Procedure, diagnosis, financial and outcome data is held in a separate database maintained by the Medical Center's Information Services department. Outcome data sets for SICU patients are extracted from the hospital system on a periodic basis and merged with the data exported from CareVue. In this way the outcomes of large sets of SICU patients have been analyzed from the merged database and used for internally and externally published CQI and outcome studies (Refs 1-15). These studies have provided the "evidence" needed to develop the pathways and guidelines used to triage patients into and out of the SICU, and for many other quality management purposes. Ongoing CQI studies of SICU severity, mortality, readmissions and other parameters are routinely extracted from the database and evaluated at monthly SICU Committee meetings. Recently the Medical Center's outcome database was migrated to Oracle, for which modern data viewing and analysis tools were also implemented. Our work at the present time is to automate the transfer and merge of this outcome data into the data extracted from CareVue, which will soon come from the Oracle-based Database Archive. This will result in a severity-outcomes database which will be updated daily, such that it will always be up to date. Our goal is to provide ongoing real-time outcomes analysis in order to continuously improve the process of surgical critical care.


Back to Expression Arrays Technologies and Methods of Analysis