Speech Evaluation in Patients Treated for Oral Cancer: a 5-year Prospective Study

ABSTRACT Background: Speech difficulties are found in 34% to 65% of patients with oral cancer and perceptual assessment of these difficulties is common practice among speech and language therapists (SLTs). A longitudinal study using a validated assessment tool regarding speech quality in patients with oral cancer is missing. Aim: Firstly to obtain insight in the development over time regarding speech quality up to five years after oral oncological treatment. Secondly to identify the clinical factors during a 5-year period that are of influence on speech quality in patients treated for oral cancer. Finally to determine the reliability of the parameter ‘grade’ of the London Speech Evaluation (LSE). Methods: Perceptual assessment of speech using the parameter ‘grade’ of the LSE was carried out 4-6 weeks before treatment, 4-6 weeks after surgery and/or 4-6 weeks after radiotherapy, and 6, 12, and 60 months after treatment. A generalized linear mixed backward stepwise model for ‘grade’ was constructed. To determine intra- and interrater level of agreement weighted Kappa’s were calculated. Results: Speech quality decreased significantly after intervention and increased significantly between the 6 and 12 months assessments. Assessment period, age shortly before oncological intervention, tumour location, tumour size, reconstruction method and dental status had a significant effect on the parameter ‘grade’. Interrater level of agreement ranged from moderate to substantial, intrarater agreement ranged from moderate to almost perfect. Conclusion and implications: The development over time in patients with oral cancer has been obtained. Quality of speech is influenced by the effect of assessment period, age shortly before the oncological intervention, location of the tumour, tumour size, reconstruction method, and dental status. The parameter ‘grade’ of the LSE has proven to be a valid manner to assess speech quality. This parameter can be used by SLTs to monitor effects of their intervention