Abstract
I read the article by Dr. Coulon and colleagues with interest (Palliative Medicine Reports Volume 5.1.2024). They explored the use of the Surprise Question in geriatric patients admitted through the emergency department and compared surprise question assessments and mortality outcomes between nurses and physicians. The findings are very much in line with a review we previously published.
I am particularly intrigued by these findings and am eager to discuss them further. 1 It is also consistent with the data from our medical center. Their positive likelihood ratio was approximately 2.4 to 2.3. This translates to an additional 15% mortality prediction. A positive likelihood ratio of 2–5 has weak evidence to rule in mortality. A negative likelihood ratio of 0.1–0.2 has moderate evidence against mortality, and a ratio of 0.2–0.5 is weak evidence. 2 The Surprise Question, when used alone, is a weak prognostic tool. Two recent systematic reviews used specificity, sensitivity, predictive values and not likelihood ratios.3,4 Predictive values are highly dependent on the prevalence of mortality in the population and, hence, are not an accurate means of determining the diagnostic capability of a tool. The clinical relevance to the C-statistic is abstract, unlike the likelihood ratios. The original intent of the surprise question was a referral to palliative care if the question was answered “positive,” and yet most studies have used it as a prognostic tool. There is no information on the palliative and economic outcomes of using the Surprise Question as a referral. The Surprise Question may add to other prognostic factors. 5 A retrospective study of 535 patients with heart failure found that combining the Surprise Question and the number of heart failure admissions improved the specificity compared with hospitalizations and the Surprise Question alone. 6 Future studies should explore the use of the Surprise Question with other prognostic factors. Clinical care should not be based on the Surprise Question until patient, caregiver, and health care outcome benefits are realized when using the Surprise Question.