Chapter 5 94 predictive value 37.1%, negative predictive value 93.1%).21 The SQ may be more accurate combined with other indicators of palliative care needs, such as symptoms prevalent in the palliative phase, performance status and indicators of increased utilization of formal and informal care.11,13,30-34 NECPAL and GSF-PIG combine the SQ with these indicators for screening of patients with palliative care needs; however, it is not yet published whether the combination of these indicators with the SQ improves the performance of these screening instruments.13,30 At the ED, the SQ has been tested in patients with heart failure and in elderly.35,36 In elderly, SQ’s c-statistic increased after adding the physician’s working experience and the PREDICT-criteria for identification of elderly with a limited life-expectancy.36 To our knowledge, the performance of the SQ with and without other indicators of palliative care needs has not been studied yet in adult patients with advanced cancer visiting the ED. Objectives of this study were to evaluate the prognostic value of the SQ in patients with advanced cancer visiting the ED; and to study the yield of adding other predictors for approaching death. PATIENTS AND METHODS Setting and patients This cohort study took place at the ED of a Dutch academic medical centre, where acute care is offered 24/7. On average, eighty patients are seen per day for both oncological and non-oncological problems. All ED-visits from May 2013 to July 2014 were prospectively screened. Inclusion criteria for patients were: diagnosis of cancer in the palliative phase2, i.e. metastasized or incurable cancer; who were at least 18 years of age. Patients with a haematological malignancy or who were not admitted to the ED for the Medical Oncology Department were excluded. Only the first ED-visits of patients were included; subsequent ED-visits by the same patients were excluded. Follow-up on overall survival was updated until February 2019. Survival data were obtained from the EPRs, which is monthly updated via the Dutch population register. Data collection After the ED-visit, a list of patients who visited the ED were screened if they met the inclusion criteria by two trained research assistants; if they did, attending physicians were sent an e-questionnaire with the SQ. E-questionnaires were sent within 1 working day to limit recall bias. Patients were divided into two groups according to the SQ: ‘Would I be surprised if this patient died within one year?’; ‘NS-patients’ of whom the physician would not be surprised, and ‘S-patients’ of whom the physician would be surprised. For NSpatients, the e-questionnaire proceeded with the Eastern Cooperative Oncology Group
RkJQdWJsaXNoZXIy MTk4NDMw