Caregivers' accuracy in reporting patients' symptom: A preliminary report.

64 Background: Improving symptom control during end of life (EOL) is a core goal of palliative care. When patients are not able to report their symptoms, caregivers (CG) report symptom intensity as surrogates. Data show that there is good agreement between patients and CG in reporting symptom intensity. However, little is known about factores that influence CGs' accuracy. The aim of the study was to determine CG accuracy of advanced cancer patients' symptoms and to identify CG factors that could modify it.

METHODS: In this prospective study, patients with advanced cancer enrolled in the National Program of Palliative Care at a public Hospital in Santiago, Chile and their CGs independently scored ten patients' symptoms using the Edmonton Symptom Assessment Scale (ESAS). Correlation between patient and CG scores were estimated for each symptom. Differences between patient and CG scores were calculated for each symptom and were transformed into positive values. A continuous variable was created with the sum of all the differences as an indicator of caregiver overall accuracy, with smaller scores meaning better accuracy. CG depression, anxiety (HADS) and burden (ZARIT) were also assessed.

RESULTS: 36 patients and their CG were included in this preliminary analysis. Mean patient age was 64, 20 (56%) females, 13 (36%) had GI cancer, 7 (19%) lung cancer and 16 (45%) other. Mean caregiver age was 53, 25 (69%) female. We found positive correlations between patients' and caregivers' assessment of pain, fatigue, nausea, anorexia, dyspnea, depression and insomnia (r > 0.3 and p < .05 for each symptom). CG accuracy ranged between 10 and 44 points (mean 25, standard deviation 9) and was not associated with CG age, gender, depression, anxiety or burden. CG accuracy was negatively associated with CG worrying thoughts as assessed by one of the HADS questions (Coef -3.99, p = .015), meaning that CG were more accurate when their worrying thoughts were higher. This association remained significant when adjusted by CG depression, anxiety and burden.

CONCLUSIONS: CG are more accurate in reporting patient symptoms when their levels of worrying thoughts are higher. This information may have implications in interpreting CG report during EOL.