Usability of the surprise question by nurses to identify 12-month mortality in hospitalized older patients: A prospective cohort study

Background: Palliative needs in older patients are often not timely identified. The Surprise Question (SQ) ‘would I be surprised if this patient died in the next year?’ is a well-researched tool that could aid in this effort. Most studies thus far involved physicians or specialist nurses, however the predictive value of the SQ when used by general nurses caring for hospitalized older patients is unknown. Objectives: To assess the predictive value of the SQ when used by general nurses and student nurses, in determining one year mortality in acutely hospitalized older patients. Design: Observational cohort study with an one year follow-up. Setting: One academic and one regional hospital in the Netherlands. Participants: Patients ≥70 years acutely hospitalized for at least 48 hours. Methods: Registered nurses and student nurses answered the SQ with ‘No’ (a positive SQ), ‘Yes’ or ‘Don't know’. Data on student nurses was analysed separately. The sensitivity, specificity, negative- and positive predictive values were calculated. Furthermore, logistic regression was performed to determine the odds of death. Results: 66 registered nurses answered the SQ for 252 patients of whom 77 (30.6%) died in the year after inclusion. Respectively, 44%, 14% and 22% died within the ‘No’, ‘Yes’ and ‘Don't know’ group. 85% of patients who died during admission or in the first three months post-discharge were identified. The sensitivity and specificity were 76.7% and 56.6%. The positive and negative predictive values were 43.7% and 84.6 %. Compared to persons in whom the SQ was answered with yes, a no answer was associated with an 4.7 times increased odds of dying in the next 12 months (odds ratio 4.71, 95% CI 2.43-9.12, p<0.001). Additionally, 20 student nurses answered the SQ about 73 patients; sensitivity and specificity were 46.7% and 72.1%, with a positive and negative predictive value of 53.8% and 66.0% respectively. Conclusion: The usability of the Surprise Question in predicting 12-month mortality in older acutely admitted patients is limited, due to the high false positive rate. The SQ when used by non-specialized nurses identifies vulnerable patients with an increased mortality risk and can be used as a first step in assessing a patients’ palliative needs, but has limited use as a single criterion for referral to specialist palliative care.