Identification of Baseline Characteristics Associated With Development of Depression Among Patients With Head and Neck Cancer

Abstract

Patients with head and neck cancer (HNC) experience increased risk of depression and compromised quality of life. Identifying patients with HNC at risk of depression can help establish targeted interventions. To identify factors that may be associated with the development of moderate or severe depression during treatment of HNC. This is a retrospective, ad hoc, secondary analysis of prospectively collected data from a randomized, double-blind, placebo-controlled clinical trial. Patients were screened at academic- and community-based tertiary care HNC centers from January 2008 to December 2011. Of the 125 evaluable patients with stages II through IV HNC but without baseline depression, 60 were randomized to prophylactic antidepressant escitalopram oxalate and 65 to placebo at the time of the initial diagnosis. Data analyses were conducted from May 2016 to April 2017. Depression outcomes were measured using Quick Inventory of Depressive Symptomatology-Self Report (QIDS-SR) scores (range, 0-27 with a score of 11 or higher indicative of moderate or greater depression). Factors that may be associated with development of moderate or severe depression were assessed, including patient demographics; cancer site and stage; primary treatment modality (surgery or radiotherapy); history of depression or other psychiatric diagnosis; previous treatment of depression or suicide attempt, family history of depression, suicide, or suicide attempt; and baseline score on the QIDS-SR and clinician-rated QIDS instruments. Participants were stratified by study site, sex, cancer stage (early [stage II] vs advanced [stage III or IV]), primary modality of treatment (radiotherapy with or without chemotherapy vs surgery with or without radiotherapy), and randomization to placebo or escitalopram and balanced within these strata. The mean (SD) age of the 148 patients in the study population was 63.0 (11.9) years; 118 (79.7%) were men, and 143 (96.6%) were white. In the evaluable population of 125 patients, receiver operating characteristic analyses assessing the area under the curve for baseline QIDS-SR score (0.816; 95% CI, 0.696-0.935) and for initial radiotherapy-based treatment (0.681, 95% CI, 0.552-0.811) suggested that these 2 variables were associated with the likelihood of developing moderate or greater depression during the study period among patients who did not receive prophylactic antidepressants. The diagnostic sensitivity for identifying patients at risk of depression using the baseline QIDS-SR score improved to 100% at a threshold of 2 from 94% at a threshold of 4. Baseline symptoms and initial radiotherapy-based treatment may be associated with development of moderate or greater depression in patients with HNC. Patients with QIDS-SR baseline scores of 2 or higher may benefit the most from pharmacologic prophylaxis of depression.