Hospital occupancy and patient mortality

Research Objective:

The Institute of Medicine's report, Hospital-Based Emergency Care: At the Breaking Point, highlighted the impact of hospital overcrowding facing patients in the U.S. As patients move through the hospital, their care is influenced by the availability of resources, staff, space, and organization of other units of the hospital. If there are inadequate beds, staff, or inefficient uses of them, particularly during periods of high volume, bottlenecks can cause crowding, reduced quality of care, and reduced community access. This study examined the effect of high inpatient occupancy rate on the day of admission on risk-adjusted all-cause in-hospital mortality.

Study Design:

A retrospective analysis of cross-sectional data was carried out to determine the effect of hospital occupancy rate on the day of admission on risk-adjusted all-cause in-hospital mortality among adults in community hospitals. The sample (n = 950,085) of adults age 18-85 was derived from discharge abstracts of all adult, non-federal, acute care hospitals obtained from the Pennsylvania Healthcare Cost Containment Council (PHC4) in 1999 that were matched with the American Hospital Association (AHA) Annual Survey file (n = 172). The entire population (N = 1,411,000) of discharge abstracts from the sample hospitals was used to derive the occupancy variables based on patient admission and discharge data. Additional hospital characteristics such as size, teaching status, technological status, and rural/non-rural location were included in the analysis. Logistic regression models using robust procedures were estimated.

Principal Findings:

Analysis of variance with adjustments for multiple group comparisons demonstrated that the mean occupancy rate for hospitals differed significantly by hospital size, teaching status, technological status, and non-rural location. At the patient level, the risk-adjusted likelihood of all-cause in-hospital death for adults increased 4% (OR 1.04, SE.01, 95% CI 1.02-1.05) for each 10% increase in hospital occupancy on the day of admission controlling for hospital characteristics.

Conclusions:

Hospitals differ in their mean occupancy and occupancy rate by various hospital characteristics and there is cyclic variation across time. Patients admitted to the hospital on days when capacity was strained had an increased likelihood of risk-adjusted all-cause death while hospitalized. Policy mechanisms and financial incentives that balance the tension between the need for adequate capacity for increases in workload and the need to maintain occupancy should be considered. Factors mediating the relationship between high occupancy and poor outcomes such as nurse staffing and practice environment characteristics should also be explored.