Assessment of Clinical Palliative Care Trigger Status vs Actual Needs among Critically Ill Patients and their Family Members

Background: There is currently no consensus on how to identify ICU patients most likely to benefit from palliative care.  Trigger-based consultation protocols have become common, where patients are identified for specialty palliative care consultation based on clinical characteristics associated with high mortality and resource utilization.  However, it is not known whether these models’ triggers are suitable proxy measures for palliative care needs. Further, these triggers will often identify many patients and exceed the availability of specialty palliative care services.

Design and Participants:  The hypothesis was that higher levels of family-reported palliative care needs would be seen in ICU patients who had a clinical palliative care trigger compared to patients without a trigger.  This was a prospective observational cohort study of 257 patient + family member dyads at 6 adult ICUs in 2 hospitals (one academic and one community) within the Duke University system.  Included patients were adults (18+ years) who received mechanical ventilation for at least 48 hours and who lacked decision-making capacity and who had an available/involved English-speaking family member.  Patients who were expected to die within 24 hours and those with imminent plan for comfort care were excluded.

The primary outcome was the Needs at the End-of-Life Screening Tool (NEST), a 13-item electronic assessment of needs in 8 domains of palliative care, which was administered electronically to family members.  The primary exposure was any of 9 clinical triggers for palliative consult present in the first 48 hours of ICU admission.  Additional patient demographic and clinical characteristics were also recorded.  NEST score threshold for “serious palliative care need” was defined a priori based on prior work.

Results: 1179 patients were screened, yielding 360 eligible patient-family dyads. 262 dyads enrolled, and 257 ultimately had complete data for inclusion. Median patient age was 58, with 51% men, 33.5% Black, 57.2% White; 57.6% were in MICUs and 42.4% in SICUs.  76.7% of the family members were female, and 44.4% were the spouse/partner of the patient. No dyad had received a palliative care consultation before survey completion (though notably this was not an exclusion criterion.)

44.7% of patients met at least one clinical palliative care trigger, with the most frequent one being worsening organ dysfunction (47.8% of patients meeting a trigger).  Patients who met a trigger were older than those who did not meet a trigger (median age 62 vs 54.5), but otherwise similar. There was no difference in median NEST score or in prevalence of NEST score above threshold indicating serious needs between dyads with triggers present or absent. Median score (scale 0-130, with 0 indicating no needs and 130 indicating highest needs) was 21 vs 22.5, P=0.52, and prevalence of serious needs (NEST score 30 or higher) was 33% vs 34.5%, P=0.81, in dyads with present and absent triggers, respectively. The presence of a trigger demonstrated low sensitivity (44.7%), specificity (55.2%), Positive Predictive Value (33%), Negative Predictive Value (66.9%), and accuracy (51.8%) for identifying serious palliative care need.  Positive and negative likelihood ratios were both 1.0, and C statistic was 0.5, all indicating that trigger status did not change likelihood of serious need. Various secondary outcomes were measured. There was no difference between groups in frequency of family-reported goal-concordant care, communication quality, or psychological distress, or in ICU or hospital length of stay.  In dyads with present triggers, family more frequently rated quality of relationship with the ICU physician favorably and less frequently believed patient survival was likely.

Commentary: This study demonstrated that family-reported palliative care needs did not correlate with presence of common triggers for palliative care consultation, and presence of a clinical trigger was no better than chance at identifying patients with the most serious needs. These results do suggest, however, that palliative care needs can be measured simply via electronic family surveys, and that incorporating a standardized assessment of needs could increase the likelihood that needs will be recognized and addressed in the ICU. Further, standardized needs assessment could better activate specific IDT members depending on the type of needs identified (e.g. identification of a spiritual need could trigger a chaplain, etc.)
One major limitation of this study is that family-reported needs may not be an accurate proxy for true patient needs.

Bottom Line:  In the ICU, presence of clinical palliative care consultation triggers is not associated with higher levels of unmet palliative care needs.

Source:  Cox CE, Ashana DC, Haines KL, et al. Assessment of clinical palliative care trigger status vs actual needs among critically ill patients and their family members. JAMA Netw Open. 2022;5(1):e2144093. doi:10.1001/jamanetworkopen.2021.44093