Long-Term Objective: The duration of hospitalization, commonly known as length of stay (LOS), is a major determinant of hospital costs. The marked variation in LOS between hospitals for groups of patients with similar illnesses suggests there is room for improved efficiency. Our long-term objective is to create a software application linked to the hospital information system to enable hospitals to enhance their efficiency. The major innovation in this proposal is the use of severity of illness, type of surgical procedure, and perioperative complications to predict LOS, which can then be used for shorter-term bed management, longer-term projection of bed capacity and personnel, and compare LOS across providers.
Specific Aims :
The specific aims of this proposal are to construct, test, and evaluate a software module, the Surgical Bed Management System (SBMS), which will: 1) Use predicted LOS to manage admissions and discharges within a hospital;2) Use predicted and actual LOS of a historical cohort of admissions (e.g., previous fiscal year) in a simulation model for strategic planning of bed capacity needs (e.g., the next fiscal year);and 3) Adjust LOS for severity of illness, comorbidity, and surgical procedure so that this measure of efficiency may be credibly compared across physicians and hospitals. Research Design and Methods: The SBMS will be a web-based software module linked to the hospital information system that facilitates gathering the necessary patient- and hospital-level data from the hospital's health information system and by direct data entry. The SBMS will estimate each patient's predicted LOS on admission and again shortly following surgery, and will aggregate and present this information as an electronic Predicted Bed-Occupancy Board for use in planning near-term admissions. Additionally, the SBMS will use hospital-specific- historical data in a simulation exercise to forecast long-term bed needs, and will compute risk- adjusted LOS for use in comparing efficiency with other hospitals. This is a two and one-half-year proposal, in which we propose to design and program prototypes, and secure user feedback in first year. We will implement SBMS and train users at the University of Michigan Hospital and conduct alpha testing during months 7 - 15, followed by use for all inpatient general and vascular surgery during months 16 - 30. We will then follow the same steps in a second hospital, Newton-Wellesley. During months 25 - 30, we will collect and analyze data from both hospitals to assess our outcomes. The primary outcome will be the change in risk-adjusted LOS during the last six months of implementation in the two test hospitals compared with the six months immediately preceding implementation. Secondary outcomes will be user satisfaction assessed with a survey instrument and an assessment of marketability based on structured interviews with key hospital leadership personnel.
