The Surprise Question and Identification of Palliative Care Needs among Hospitalized Patients with Advanced Hematologic or Solid Malignancies.

Abstract

Little is known about quality of life (QOL), depression, and end-of-life (EOL) outcomes among hospitalized patients with advanced cancer. To assess whether the surprise question identifies inpatients with advanced cancer likely to have unmet palliative care needs. Prospective cohort study and long-term follow-up. From 2008 to 2010, we enrolled 150 inpatients at Duke University with stage III/IV solid tumors or lymphoma/acute leukemia and whose physician would not be surprised if they died in less than one year. We assessed QOL (FACT-G), mood (brief CES-D), and EOL outcomes. Mean FACT-G score was quite low (66.9; SD 11). Forty-five patients (30%) had a brief CES-D score of ≥4 indicating a high likelihood of depression. In multivariate analyses, better QOL was associated with less depression (OR 0.91, p < 0.0001), controlling for tumor type, education, and spiritual well-being. Physicians correctly estimated death within one year in 101 (69%) cases, yet only 37 patients (25%) used hospice, and 4 (2.7%) received a palliative care consult; 89 (60.5%) had a do-not-resuscitate order, and 63 (43%) died in the hospital. The surprise question identifies inpatients with advanced solid or hematologic cancers having poor QOL and frequent depressive symptoms. Although physicians expected death within a year, EOL quality outcomes were poor. Hospitalized patients with advanced cancer may benefit from palliative care interventions to improve mood, QOL, and EOL care, and the surprise question is a practical method to identify those with unmet needs.