Institution-free days after critical illness: A multicenter retrospective study

Background: Patient-centered outcomes beyond mortality such as institution-free days (IFD) are becoming increasingly relevant in critical care trials.Methods: We calculated IFD using three definitions which differed in the way death and censoring of after -hospital deaths were handled analysing data from adult patient databases admitted to four ICUs of North Bris-bane, Australia. Differences in distribution of IFD using different definitions were explored with descriptive statistics and histograms. Six pre-specified variables (age, illness severity, comorbidities index, elective status, surgical/medical admission and treatment limitations) were assessed and reported as determinants of IFDs for the proposed definitions.Results: Data from 25,371 ICU admissions was analysed. The density distribution of IFD was bimodal with a peak at 0 days and a variable right-sided peak depending on the definition used. The mean IFD varied from 253 (standard deviation(SD) 151.3) to 295 (SD 116.2) depending on definition used. Multivariable zero-inflated negative binomial regression modelling showed that the six pre-specified variables had significant associations with IFD and their magnitude of effect varied with the definition used.Conclusions: IFD is a simple, easily measurable patient-centered outcome that varies depending on the definition used. Patient input should be sought to define the optimum definition and clinical use of IFD.

Automated summary

Patient-centered outcomes such as institution-free days (IFD) are important in critical care trials. The study found that different definitions of IFD lead to different results, depending on how death and post-hospital deaths are handled. The study also identified significant associations between IFD and factors such as age, illness severity, comorbidities index, elective status, surgical/medical admission, and treatment limitations.