Patients with head and neck cancer who resided in the most socioeconomically disadvantaged neighborhoods had lower pretreatment head and neck cancer–specific quality-of-life scores compared with those living in the least disadvantaged neighborhoods, according to findings published in JAMA Otolaryngology–Head & Neck Surgery.
Researchers evaluated 600 adult patients diagnosed with head and neck cancer from November 2015 to September 2022 who received care at a single urban tertiary care center. Prior to treatment initiation, patients completed the Functional Assessment of Cancer Therapy–Head and Neck (FACT-HN) questionnaire, a validated patient-reported measure of physical, social, emotional, functional, disease-specific, and overall quality of life (QOL). Higher scores indicate better QOL.
Neighborhood socioeconomic disadvantage was assessed using the state-level Area Deprivation Index (ADI), a composite measure incorporating area-level education, occupation, household income, housing, poverty, and access to utilities. Patients were categorized into 5 ADI quintiles, ranging from least disadvantaged to most disadvantaged.
The cohort had a mean age of 62.5 years, and 72% of patients were male. Overall, 52% of patients had stage III to IV disease at presentation, and the most common primary tumor site was the oropharynx.
In adjusted analyses, patients in the most disadvantaged ADI quintile had lower head and neck cancer–specific well-being scores than patients in the least disadvantaged quintile. The adjusted difference was −3.62 points, with a 95% CI ranging from −6.23 to −1.01. The head and neck cancer–specific subscale ranges from 0 to 40, and the cohort mean score was 27.7.
In unadjusted item-level analyses of the head and neck cancer–specific domain, greater neighborhood disadvantage was associated with lower reported ability to eat preferred foods, eat as much food as desired, swallow naturally and easily, and eat solid foods. Patients in more disadvantaged neighborhoods also reported greater tobacco product use. These item-level findings had small effect sizes and may be best interpreted as exploratory.
In the adjusted analysis, the models accounted for age, sex, race, relationship status, education level, smoking history, alcohol use, Charlson Comorbidity Index score, TNM stage, primary tumor site, and recommended surgery, radiation therapy, and chemotherapy.
The significant association was limited to the comparison between the fifth and first ADI quintiles. In the adjusted models reported in Table 3, the second, third, and fourth quintiles each had confidence intervals that crossed zero relative to the least disadvantaged quintile.
Researchers did not observe clinically meaningful adjusted associations between ADI quintile and physical, social, emotional, functional, or overall well-being scores. In Table 3, the adjusted overall well-being estimate for the most vs least disadvantaged quintile also crossed the null (β = −7.40; 95% CI, −14.81 to 0.01). A sensitivity analysis excluding recommended treatment variables showed similar results.
Interpreting the clinical magnitude of the main finding is complicated by the measure used. The minimal important difference the researchers cited (from Ringash et al.) was defined for the overall FACT-HN score—a decrease of at least 12 points—rather than for the head and neck cancer–specific subscale on which the primary finding rests. The authors reported that the unadjusted overall FACT-HN gap between the least and most disadvantaged quintiles was 12.3 points (mean) and 15.1 points (median). They cautioned that the significance of this gap was difficult to interpret because it compared scores between different patients rather than changes within individual patients over time.
The findings may help explain how quality-of-life differences observed among head and neck cancer survivors could begin prior to treatment. However, the cross-sectional design means the study cannot determine whether neighborhood deprivation caused lower head and neck cancer–specific well-being.
Several factors may affect interpretation. The adjusted models included approximately 494 to 497 patients rather than the full 600-patient cohort because of listwise deletion. Chemotherapy recommendation status was missing for 38% of patients and was handled as a missing category, and education level was missing for 10% of patients.
The cohort was also 78% White, and racial-minority groups were combined in the models because of small cell sizes. The researchers noted that, as a single-institution study, the results may not be generalizable to other patients with head and neck cancer nationally or globally.
Other limitations included reliance on neighborhood-level rather than patient-level socioeconomic data, potential mismatch between electronic medical record addresses and patients' true residences, possible changes in neighborhood deprivation over the study period, survey completion and health literacy bias, and the possibility that symptom burden could be related to delayed diagnosis or more advanced disease at presentation.
The authors wrote that these patients "may be identified in the future based on their address of residence, and their treatment could be augmented with screening for patient-level SDOHs."
Disclosures: Dr. Tam reported receiving grants from Pfizer and Genentech outside the submitted work. Dr. Ghanem reported receiving personal fees from Axogen for service on a medical advisory board outside the submitted work. Dr. Siddiqui reported receiving grants from Varian Medical Systems and personal fees from the Varian Noona Medical Advisory Board outside the submitted work. No other disclosures were reported.
