3071 - Identifying Socioeconomic Predictors of Voice Impairment in Long Term Survivors of Locally Advanced Head and Neck Cancer Treated with Radiation Therapy

Purpose/Objective(s):

Voice impairment is a known late effect in those receiving radiotherapy (RT) to the upper aerodigestive tract. Prior work has identified patient and treatment factors associated with voice impairment, but there is a lack of data investigating socioeconomic factors predictive of impairment. The purpose of this study is to characterize socioeconomic predictors of voice impairment using a validated patient-reported inventory.

Materials/Methods:

Long-term (>2 years) survivors of locally advanced (T3+ or N+) head and neck cancer (LAHNC) treated with =60 Gy at a single academic institution were enrolled. Patient clinical and demographic information was obtained through a comprehensive long-term follow-up questionnaire. Voice impairment was defined using the validated 10-question voice handicap index (VHI-10) with impairment defined by scores >11. Descriptive statistics were performed, and differences in impairment across multiple socioeconomic factors were tested using a Chi-squared test. Significant predictors for voice impairment were assessed with univariable logistic regression.

Results:

Of the 252 patients enrolled, 225 had complete voice and demographic data available for analysis. Median age was 65 years (IQR 59-73) with 71% male, 86% White, 85% living in urban counties, 43% T3/T4, and 71% N+. Mean time from RT completion was 7 years (IQR 3.9-8.5). Three most common disease subsites were oropharynx (50%), oral cavity (15%), and larynx (10%). At least one surgery or chemotherapy course was part of 62% and 60% of patients’ treatment, respectively. Voice impairment was noted in 35% of the cohort. Significant differences in voice impairment frequency were identified in patients of Black race (RR 1.88, p=0.006) insured through Medicaid (RR 2.06, p=0.02) who were unemployed (RR 2.20, p<0.001) with less than a high school education (RR 1.74, p=0.005) in the lowest tertile of household income (RR 1.96, p<0.001). Univariable logistic regression identified Black race (OR 3.09, p=0.004), Medicaid (OR 4.18, p=0.01), having less than a high school education (OR 2.52, p=0.003), unemployed status (OR 4.31, p<0.001), and being in the lowest income tertile (OR 3.00, p<0.001) as significant variables predictive of voice impairment.

Conclusion:

Voice impairment is a common late sequela of RT in long term survivors of LAHNC that disparately affects patients. Not only do clinical and treatment factors impact outcomes, but certain patient-specific socioeconomic variables can influence vocal toxicity. This study supports using these factors as tools to identify patients at risk for voice impairment to better direct resources toward those at greatest risk. This study also underscores the importance of enrolling socioeconomically diverse populations in trials to best understand outcomes in real-world patient populations.