Screening tools to identify patients with unmet palliative care needs in the emergency department: A systematic review.

Objectives: This systematic review identified and assessed psychometric properties of the available screening tools to identify patients with unmet palliative care (PC) needs in the emergency department (ED).

Methods: A comprehensive search of electronic databases and the gray literature was conducted. Two independent reviewers completed study screening and inclusion, data extraction, and quality assessment. A descriptive summary of the results was reported using median of medians and interquartile ranges (IQRs).

Results: A total of 35 studies were included, involving the assessment of 14 unique screening tools. The most commonly used screening tool was the surprise question (SQ; n = 12 studies), followed by the Palliative Care and Rapid Emergency Screening (P-CaRES) tool (n = 8), and the screening for palliative and end-of-life care needs in the emergency department (SPEED) instrument (n = 4). Twelve of the included studies reported on the psychometric properties of the screening tools, of which eight of these studies assessed the performance of the SQ to predict patient mortality. Overall, the median sensitivity (63%, IQR 38%-78%) and specificity (75%, IQR 57%-84%) of the SQ to predict mortality at 1 or 12 months was moderate. While the median positive predictive value of the SQ was low (32%, IQR 16%-40%), the median negative predictive value was high (91%, IQR 88%-95%). Across the studies, the proportion of patients identified as having unmet PC based on the criteria of the screening tools ranged from 5% to 83%.

Conclusions: This review identified 14 unique screening tools used to identify adult patients with unmet PC needs in the ED. One screening tool, the SQ, was found to have moderate sensitivity and specificity to accurately predict future patient mortality. Additional research is needed to better understand the clinical value of this and the other available tools prior to their widespread implementation.