Characteristics of potentially inappropriate, and inappropriately prolonged, ICU admissions in dying ICU patients: A retrospective cohort study

Little is known about the characteristics of potentially inappropriate or unnecessarily prolonged intensive care unit (ICU) admissions in Australia, nor the exposure rate of non-ICU clinicians to dying ICU patients. We conducted a single-centre retrospective cohort study at a university-affiliated hospital in Victoria, Australia, of patients admitted to the ICU between January 2022 and June 2023, who transitioned to end-of-life care during their ICU admission. Decisions regarding appropriateness were adjudicated during a bi-weekly morbidity and mortality meeting. Out of 287 patients 279 were included in the final analysis. One hundred and eight (39%) patients were deemed to have had a potentially inappropriate admission, and 37 (13%) were deemed to have had a potentially inappropriately prolonged admission. Significantly higher proportions of patients were admitted from either the ward (32.4% versus 22.4%, <i>P</i>=0.02) or another hospital (15.7% versus 6.4%, <i>P</i>=0.02) if they were deemed to have had a potentially inappropriate admission. Significantly higher proportions of patients deemed to have had a potentially inappropriately prolonged admission had treatment limitations (16.2% versus 40.5%, <i>P</i>=0.006), lower Australian and New Zealand Risk of Death scores (median score 27.2 versus 45.5, <i>P</i>=0.006) and a clinical frailty score of 5 or more (63.9% versus 45.1%, <i>P</i>=0.048). They also had a significantly longer median ICU length of stay (median length of stay 13.4 days versus 2.6 days, <i>P</i> <0.001) and received significantly higher rates of invasive supports such as tracheostomy (16.2% versus 1.2%, <i>P</i> <0.001). The four major themes linked to these admissions were 1) lack of planning/appropriate treatment limitations, 2) lack of recognition of dying, 3) issues with communication/consensus and 4) provision of highly invasive treatments. The median rate of exposure of individual ward-based clinicians was 1 dying ICU patient per 18 months. Early framing of goals of care, reassessment of treatment goals during an ICU admission, dedicated communication skills training, and embedded frailty assessments might reduce non-beneficial and prolonged ICU admissions.