Page 27 - Functional impairment and cues for rehabilitation of head and neck cancer patients -

Swallowing function after treatment for early-stage OPC
with swallowing liquids, pureed and/or solid food. Secondary functional outcomes were dry mouth, weight loss during the past week, use of tube feeding during the past week, pain (in mouth, jaw or throat), coughing, trouble opening mouth, bothersome appearance, and altered taste. These data were extracted from the European Organization for Research and Treatment of Cancer Quality of Life Head and Neck 35 questionnaire (EORTC-QLQ-H&N35), completed by participants at baseline and at 4- and 12-month follow-up (24). Questions with four possible answers were dichotomized into no (not at all and a little) or yes (quite a bit and very much) to increase interpretability of the risk ratio (RR). Analyses were repeated after dichotomizing the outcome into no (not at all) and yes (a little, quite a bit and very much). Also, research nurses asked participants whether they had a percutaneous endoscopic gastrostomy (PEG) or tracheostomy at 4- and 12-month follow-up. Research nurses assessed the presence of residual/ recurrent disease from the medical records at 4- and 12-month follow-up. All participants were flagged with the NHS Digital for 6-monthly updates on mortality and date of death.
Statistical analysis
Analyses were performed using IBM® SPSS® Statistics 24.0. First, intended and received treatment were described. Baseline characteristics and baseline functional outcomes were compared, grouped by treatment modality and whether data on (swallowing) function were available. The independent samples t-test was used to compare continuous variables of two groups, the one-way ANOVA was used to compare continuous variables of more than two groups, and the Chi-square test was used to compare categorical variables.
Differences in self-reported swallowing and secondary functional outcomes at both 4- and 12-months follow-up were compared using Poisson regression analysis with a robust error variance to estimate RRs and confidence intervals (CI). Poisson regression was used as odds ratios (obtained from logistic regression) are poor approximations of RRs if the outcome prevalence is high (25). First, the RR with 95% CI and p values, adjusted for age and gender only were calculated (minimally adjusted). Second, results after also adjusting for ACE-27, smoking status, oropharyngeal tumour site, TNM-stage, HPV-status, and pre-treatment swallowing problems were presented (adjusted). The minimally adjusted analyses were repeated on the participants included in the adjusted analyses to ensure that any changes in estimated RRs in the adjusted models were attributable to confounding rather than missing data.
Hazard ratios (HRs) were calculated using minimally adjusted Cox regression analyses (adjusted for age and gender only) as well as adjusted Cox regression analyses (adjusted for ACE-27, TNM stage, HPV-status, and smoking status also). Again, the minimally adjusted analyses were repeated on the participants included in the adjusted analyses. Survival was defined as time between date of consent and date of death or date of last mortality follow-up.
Patterns in baseline characteristics and functional outcomes of participants with T1N0 OPC only were compared to those of all early-stage OPC participants. Also, HPV-negative and – positive participants were compared.
2
25