• © 2011 by American Society of Clinical Oncology

Reply to J.-J. de Arriba, et al

In the letter by de Arriba et al,¹ the authors raise questions about how delirium affects the results of our study² that examined the trajectory of cancer symptom severity and performance status, measured using the Edmonton Symptom Assessment System (ESAS) and Palliative Performance Scale, respectively, in the last 6 months of life. Particularly, the authors inquire if delirium was screened for and how, and how many patients with delirium were excluded from the study or had assessments completed by a proxy.

ESAS does not measure delirium and no screening for delirium was performed. We were unable to assess definitively how many patients failed to complete the ESAS because of delirium. However, because the large majority of our patients completed assessments in a cancer center (82%) versus the home (18%), our cohort consisted mainly of ambulatory patients with cancer who were healthy enough to attend the clinic and were thus less likely to suffer from delirium. We suspect that patients with delirium are more likely to receive treatment at home or in an emergency department rather than in an outpatient clinic. We could not find any studies that reported the prevalence of delirium in an ambulatory population with cancer, although we surmise that the rate is much lower than that found in a palliative care unit.

Our data sources were not able to determine who completed the assessment. We believe that patients themselves completed the majority of assessments, and this is corroborated by a review of our data: of the nearly 57,000 ESAS assessments, 91% were fully complete, and less than 2% had ≥ 4 symptom scores missing (the missing data may be a consequence when a proxy is asked to complete the subjective symptom scales). We agree with de Arriba et al¹ that those visiting the cancer center with delirium likely would not complete the voluntary assessment. Moreover, the large decline in the number of assessments completed in the final 2 weeks of life is largely a result of the challenges of collecting data from an especially sick population; delirium is likely to be a contributing factor.

Despite the limitations in delirium assessment noted above, a major contribution of our results is the fact that our cohort was not clearly identified as a palliative cancer population, but rather an outpatient population. Our study showed that there were significant unmet symptom needs among patients with a performance status score in the transitional (score, 70 to 40) and end-of-life (score, 30 to 10) ranges. Thus, in an ambulatory cancer population, it would be reasonable to consider referral to palliative care when the Palliative Performance Scale score reaches a transitional range so that palliative care services can be integrated earlier in the care continuum. These palliative care services should include more comprehensive assessments of other symptoms that are prevalent at the end of life, such as delirium.