Abstract
CONTEXT: Parkinson's disease and related disorders (PDRD) are fatal neurodegenerative disorders characterized by a fluctuating course that can complicate prognostication. The "surprise question" (SQ: "Would you be surprised if your patient died in the next year?") has been used to identify patients with limited prognosis but has not been assessed in PDRD. OBJECTIVES: To determine the validity of the SQ in predicting 12-month mortality in PDRD. METHODS: Data was analyzed from 301 patients and 34 community-based neurologists who were participating in a clinical trial of outpatient palliative care for patients with PDRD. Clinicians answered the SQ for each patient at baseline. Descriptive statistics at baseline, chi-square tests of independence, 2 × 2 and 2 × 3 cross tables were used. Survival analysis compared SQ responses using Kaplan-Meier curves. Risk estimate analyses identified patient characteristics associated with clinicians' responses. RESULTS: Mortality was 10.3% (N = 31) at 1 year. The sensitivity and specificity of the SQ was 80.7% and 58.9%, respectively with AUC = 0.70, positive predictive value of 18.4% and negative predictive value of 96.4%. Older age, atypical parkinsonism, and dementia were associated with responding "no" to the SQ. CONCLUSION: The SQ is sensitive to 12-month mortality in PDRD, with a high negative predictive value. The SQ may be useful for identifying patients less likely to die within a year and may be useful for identifying patients with palliative care needs outside of end-of-life care. This latter use may assist in mobilizing early and timely referral to specialist palliative care.