Abstract
Objectives/Hypothesis:
To identify factors associated with nonparticipation in long‐term quality‐of‐life (QOL) analysis in head and neck squamous cell cancer (HNSCC) patients.
Study Design:
Restrospective analysis of prospectively collected data.
Methods:
Two hundred sixty‐one previously untreated HNSCC patients were prospectively evaluated using the University of Washington QOL questionnaire. Questionnaires were administered before treatment and at 1 year following treatment. Patients without data at 1 year were classified as nonparticipants.
Results:
At 1 year, 152 patients (58%) had complete QOL data. Nonparticipants differed from participants by stage, comorbidity, and disease status, with a greater proportion of patients with advanced stage disease, advanced comorbidity, and recurrent disease participating in QOL analysis at 1 year (P < .05). No significant differences were found between nonparticipants and participants with respect to age, race, primary site, treatment, travel distance, or socioeconomic variables. Multiple linear regression analysis showed differences in initial QOL scores between participants and nonparticipants only for activity, with lower mean initial scores for nonparticipants (β = −9.9, 95% confidence interval, −16.8 to −3.0; P = .005). Multiple logistic regression demonstrated significant differences in the odds of nonparticipation at 1 year for males (odds ratio [OR] 0.4, P = .01), T4 disease (OR 0.3, P = .001), uninsured patients (OR 2.8, P = .007), recurrent disease (OR 0.2, P < .0001), and comorbidity (OR 0.5, P = .025), after controlling for all other variables.
Conclusions:
Long‐term QOL analysis may be biased by greater participation from patients who require greater medical attention because of advanced tumor stage, recurrent disease, or comorbidity, whereas patients without insurance are under‐represented. These findings suggest limitations to the interpretation and application of long‐term QOL data. Laryngoscope, 2010