Symptom variability during repeated measurement among hospice patients with advanced cancer

In this prospective study, we explored symptom variability in patients with cancer during repeated measurements.

Patients with cancer admitted to an inpatient hospice completed a daily questionnaire throughout their admission. The questionnaire consisted of 5 visual analogue scales (VAS) for anxiety, depression, nausea, pain, and sedation and 3 verbal rating scales (VRS) for depression, pain, and vomiting. Data from those who completed 5 consecutive days were used for the primary analysis. We used all available data points to compare VAS and VRS. An index was developed to assess for daily symptom variability.

A total of 125 hospice inpatients were enrolled; 46 (38%) completed 3 consecutive daily questionnaires and 30 (24%), 5 days. We found (1) a statistically significant decrease in severity of symptoms present on admission, (2) new symptoms developed, (3) consequently overall symptom prevalence on days 1 and 5 appeared unchanged, (4) high daily symptom variability as demonstrated by the variability index and also changing daily symptom interrelationships, (5) demographic characteristics influenced symptom patterns on admission and subsequently, (6) severe pain predicted more frequent and severe symptom burden only on admission, (7) severe depression predicted more frequent and severe symptom burden on admission and thereafter, (8) VAS scores for depression and pain did not correspond with discrete VRS categories (mild, moderate, severe).

(1) Symptom studies in advanced disease while difficult to conduct yield valuable information, (2) symptom relationships changed daily; strict timing of data collection is crucial for data analysis, (3) symptom monitoring following admission is an overlooked measure of risk assessment, (4) symptom prevalence studies alone for treatment follow-up may be misleading, (5) depression is an important predictor of symptoms and need to be more aggressively assessed and treated, (6) demographic characteristics may help identify symptom patterns and better direct treatment, (7) VRS rather than VAS was more reliable for assessing symptoms in hospice cancer patients.