Building a Computer-Supported Quality Improvement System in One Year: The Experience of a Large State Psychiatric Hospital

Abstract

G. Pierce Wood Memorial Hospital, a large state psychiatric hospital, implemented a data-intensive quality improvement (QI) program as part of its efforts to attain Joint Commission accreditation. EARLY STEPS OF QI IMPLEMENTATION: With the December 1993 Joint Commission survey date in mind, the Executive QI Committee started weekly meetings in December 1992 to plan initiatives and establish a viable information management system for the QI process. A QI coach was hired to help plan all phases of the QI program, and a 40-hour training program on statistics and QI principles was provided to key middle managers. SOFTWARE CHOICE ISSUES: Software options can be examined in terms of data cleaning, data manipulation, statistical flexibility, graphics strength, word processor compatibility, and utility strength. SOFTWARE CONFIGURATION: The present configuration of our statistical software is relatively elaborate. Organizations new to using statistical software should start off modestly to avoid being overwhelmed by the infinite possibilities that large numbers of interlinked modules present. DATA NETWORK CONFIGURATION: The technical core of the computerized QI system and how it operates are discussed. The data interface between the local area network (LAN) and statistical system data is described in three stages: (1) the creation of the "system" file; (2) the generation of the raw data; and (3) the transference of the raw data into the system file. QI ACTIONS: A QI reporting form was developed on which each report user states his/her analysis of the monthly indicator data report and identifies root causes, describes the actions taken to resolve the problems, or evaluates the effectiveness of prior actions taken. In addition, each indicator manager presents a quarterly verbal report to the Executive QI Committee on QI occurrences in his/her area of responsibility. A nursing department indicator revealed an insufficient amount of supplies being placed in the code-blue ambulance. Determination of root causes led to the identification of a breakdown in the process whereby each shift adds and reviews supplies. As another example, the initiation of the stat IM (use of intramuscular injections or tranquilizing medications to restrain persons) indicator report led to a decrease in the number of IMs ordered. If we could have developed this statistical software-based QI program in only one year, then virtually any organization, large or small, can attain the technical capability to conduct "state of the art" QI.