Mary Joanne Verhoef

Chapter 5 96 (from ED-visit) and the following predefined candidate predictors: SQ, poor performance status (ECOG 3-4), acute symptoms, palliative disease phase, symptom-management treatment, home care, ED-visits in the previous six months, limitations on life-sustaining treatments. Next, the SQ and predictors with a p-value of <0.10 in univariable analyses were entered in a multivariable analysis. Discriminative ability of the final multivariable model (including the SQ and any independent predictors) was estimated using Harrell’s c-statistic. Akaike’s information criterion was calculated to compare the ability of the predictive models with and without SQ to explain the variance in our dataset.42 Ethical approval and informed consent The Committee of Medical Ethics of the LUMC approved this study on May 27, 2013. According to Dutch Law (WGBO, article 458) and European Law (General Data Protection Regulation), written consent from included patients was not required. Data from this study were analysed after de-identification. RESULTS Patient and ED-visit characteristics ED-visits of 245 patients with advanced cancer were included. The attending physician at the ED answered the SQ with ‘No’ (NS) in 203 patients (83%), and ’Yes‘ (S) in 42 patients (17%) (Table 1). Overall median age was 62 years (interquartile (IQ-)range: 45-79 years) and 48.2% was male. Compared to S-patients, NS-patients more often received symptommanagement therapy (29.6% versus 16.7%, p=0.049), reported more main symptoms at the ED and had a worse clinical condition (ECOG 3-4 in 40% vs. 21%, p=0.034). Most frequently reported ESAS-symptoms were pain (overall 60.8%; NS 64.0% versus S 45.2%, p=0.023), nausea (29.8%; 31.0% versus 23.8%, p=0.46) and shortness of breath (29.8%; 31.0% versus 23.8%, p=0.46). During or after the ED-visit, changes emerged in limitations on LSTs: before the ED-visit, 64 patients (26.1%) had documented limitations on LSTs; after the ED-visit, this had risen to 104 patients (42.4%) (Table 2). Within 1 year, 193 patients (78.8 %) had died: 172 NS-patients (85%), and 21 S-patients (50%). At the end of followup, 233 patients had died (95%). Overall median survival was 3.0 months (IQ-range 1-11); 3.0 months for NS-patients (IQ-range 1-8) and 9 months for S-patients (IQ-range 3-28; log-rank p<0.0001, Figure 1).

RkJQdWJsaXNoZXIy MTk4NDMw