Background:

The decision to admit a patient to the intensive care unit (ICU) is complex, reflecting both patient factors and available resources. Previous work has shown that ICU census and flow do not impact mortality of patients admitted to the ICU. However, the effect of ICU bed availability on patients outside the ICU is not known. We hypothesized that decreased ICU bed availability would be associated with an increased risk of cardiac arrest on the general wards.

Methods:

We conducted an observational study using prospectively collected data from a tertiary–care academic medical center with 68 adult ICU beds between January 2009 and June 2011. Ward occupancy and ICU bed availability data (which accounted for ICU census and nursing capacity) were obtained for the start of every 12–hour shift (7am to 7pm or 7pm to 7am). Cardiac arrests occurring on the ward during each shift were divided by the ward occupancy at the start of the shift to determine the cardiac arrest rate. These rates were analyzed by the total number of available ICU beds, as well as the subgroups of available medical ICU (MICU) and non–MICU beds. Poisson regression was used to adjust for potential confounders, including year, season, day or night and weekday or weekend.

Results:

Complete data were available from 1,716/1,820 (94%) shifts with a median ward occupancy of 217 [range 132–266]. The median number of total available ICU beds in a given shift was 5 [range 0–16]. The median number of available MICU and non–MICU beds was 1 [range 0–8] and 3 [range 0–14], respectively. During the study shifts, there were 96 ward cardiac arrests, resulting in a mean rate of 2.57 arrests/10,000 ward patients per shift. Of these arrests, 81 (84%) occurred in medical patients. There was no significant association between ward cardiac arrest rate and overall ICU bed availability (IRR = 1.06 [95% CI, 0.97 – 1.15], P = 0.19) or non–MICU bed availability (IRR = 1.00 [95% CI, 0.91 – 1.11], P = 0.94). However, there was a significant increase in the rate of cardiac arrest as the number of available MICU beds decreased (IRR = 1.27 [95% CI, 1.06 –1.54], P = 0.01). The figure shows associated unadjusted ward cardiac arrest rates and the frequency of MICU bed availability.

Conclusions:

Cardiac arrest on the general wards is associated with decreased availability of MICU beds, but not other ICU beds. This finding suggests that sicker ward patients may not be triaged appropriately when MICU beds are severely limited.

Figure 1Ward cardiac arrest rate by medical ICU (MICU) bed availability, n = number of 12–hour shifts.