Objectives  

Intensive care is an expensive and limited resource, and when a demand supply mismatch between available beds and influx of patients occurs, one temporary measure is to discharge a patient to make room for the new admission. Sometimes the patient is discharged sooner from its original ICU than ideal; i.e., a so-called ‘premature discharge’. This could be either to a different ward within the same hospital if the patient is deemed well enough to cope with a lower level of care, or to another intensive care unit if critical care is still to be provided. Data from the Swedish intensive care register (SIR) showed that there was a high incidence and increased mortality of patients discharged at night. There were also differences in mortalities between patients that were transferred from one ICU to another. I have analysed the mortality associated with different types of ICU-to-ICU transfers and control groups and examined a national quality improvement project regarding discharges at night to see if mortality, incidence, or discharge culture could change.  

Methods  

All three studies are conducted with data from the Swedish intensive care register and vital status was ascertained by linking SIR to the Swedish population register. Study I consisted of two parts: mortality, and incidence of night-time discharge. The quality improvement project in Study I was analysed in a before and after approach with local improvement projects at different ICUs. In Studies II and III, transfers were grouped by the attending intensivist according to SIR guidelines into one of three defined categories: capacity transfer, clinical transfer, or repatriation. The groups were compared to each other in Study II, and capacity transfers were matched to a control group that remained in the ICU in Study III. Multilevel logistic regression was used, and all studies contained some statistics using individual ICUs as a random factor. Life sustaining treatment limitations were included in Studies II and III. 

Results  

In Study I, there was a decrease in night-time discharges during the study period. The incidence fell from 7.0% in 2006 to 4.9% in 2015. Alongside this, the mortality associated with night-time discharge was reduced, the odds ratio fell from 1.20 to 1.06 with a loss of significance. All this coincided in time with the national improvement project. Study II showed that 14.8% of all discharges from a Swedish ICU ended with transfer to another ICU, and that an increased mortality rate was associated with ICU-to-ICU transfers during periods of demand–supply mismatch. Capacity transfers were 15.8% of all transfers accounting for roughly 2.0% of ICU survivors. One in four capacity transferred patient died within 30 days of discharge, compared to one in seven for transfers due to clinical reasons. The third study showed that capacity transfer was associated with an average risk increase in 30-day mortality of 4.7%, and a 180-day mortality of 4.9% compared to non-transferred patients when analysed using a potential outcomes framework.   

Conclusion  

The studies concludes that patients experiencing a capacity transfer are exposed to increased mortality risk, both when compared to other types of inter hospital ICU-to-ICU transfers as well as when compared to patients that were not transferred. The increased risk appeared to be unrelated to patient characteristics and illness severity as well as many additional factors measured in the referring ICU. The studies also suggest that a suboptimal outcome after premature discharge at night can be changed and that a national project to adjust outcome and incidence can be undertaken with positive results.