We conducted a retrospective, cross-sectional analysis of 411 HNC survivors who completed PROM questionnaires on neck disability and insomnia during their visit to the UPMC Head and Neck Cancer Survivorship Clinic between August 2018-March 2021. All adults (≥ 18 years) who completed treatment for squamous cell carcinoma of the oral cavity, oropharynx, and larynx/hypopharynx at their initial visit were included in the analysis. Survivors with a history of recurrence, second primary carcinomas, or distant metastasis were excluded. The University of Pittsburgh’s Human Research Protection Office approved this study.
Demographics and clinical characteristics were abstracted from the medical record, including variables of age, sex, race (White, Other [e.g., African American/Asian]), marital status, tumor site (i.e., oropharynx, oral cavity, larynx/hypopharynx), American Joint Committee on Cancer (AJCC) staging (7th and 8th edition), treatment modality (i.e., surgery alone, non-surgical, and surgery plus adjuvant), and time since treatment completion. Analysis was completed for survivors both immediately and over six months since treatment completion.
Neck Disability Index
The Neck Disability Index is a 10-question measure of the impact of disability on functional activities, pain, personal care, sleep, and movement, with higher scores indicating more severe disability [45]. Each item is scored from 0–5, increasing with severity, with scale scores ranging from 0–50. The minimal clinically important difference (MCID) is reported in prior studies as approximately 10 points [46–49]. The NDI has been used to measure neck dysfunction reliably and consistently in HNC patients [22, 45, 50–53]. In our study, NDI was used as a continuous scale to evaluate the relationship with insomnia, with higher scores indicating increasingly severe neck impairment, and categorized into three groups: no disability (≤ 4), mild disability (5–14) or moderate-complete disability (> 15). Each NDI subsection was analyzed for its relationship with the insomnia category. For the univariable and multivariable analysis, the question related to sleep was removed to prevent confounding the relationship.
Insomnia Severity Index
Insomnia was measured using the insomnia severity index (ISI), a seven-item questionnaire with questions on sleep quality, maintenance, and interference with daily functioning [54]. It is a reliable and valid questionnaire in the HNC and cancer survivor population with questions referring to symptoms within the prior two weeks [55–57]. MCID for the ISI is reported in other studies as 6 points [58]. Total score ranges from 0–28, and measurements were used on a continuous and categorical scale, with higher scores indicating more clinically severe insomnia. For ordinal regression models, patients were stratified into three groups: no insomnia (score 0–7), subthreshold insomnia (score 8–14), and clinical insomnia (score > 14).
Statistical Analysis
All statistical analysis was performed using RStudio (1.1.456; RStudio, Inc, Boston, Massachusetts) and SAS v9.4 (SAS Institute, Cary, North Carolina). We calculated frequency (percentage) for categorical variables and mean ± standard deviation (SD) for continuous variables for the descriptive analysis. Univariable linear regression was performed to analyze the association between independent variables, including age, time since treatment, sex, race (White, Other [e.g., African American/Asian]), marital status, AJCC 7th edition staging [i.e., early (stage I/II) or advanced (stage III/IV)], tumor site (i.e., oral cavity, oropharynx and larynx/hypopharynx), HPV status (i.e., positive, negative, unknown, or not applicable), and NDI, and insomnia category. Considering both statistical and clinical significance, the multivariable linear regression models included age, race, time since treatment, cancer site, staging, and NDI category. Variables with a p-value less than 0.05 were considered significant.