Klaske van Sluis

58 3.4. Discussion the combination of instruments which are used, effectiveness and responsiveness of the instruments for changes over time are evaluated. By conducting this study in five hospitals in two countries, a variety of patients, languages, and treatment strategies are involved. We evaluated effects of medical detail on voice outcome, and although the number of participants of our study led to no significant results in medical history factors, this framework is useful for ongoing work. This study has some limitations. Due to the small sample size, multiple assessments, and the variety of outcome measures we were forced to perform the LME modelling on summarized dichotomized variables. With restructuring variables into dichotomous variables information about details in the surgery are lost, e.g. Major Reconstruction is used as a summarized variable which orig- inally included details on type and extent of (flap) reconstruction. Although all evaluation tools are widely used, they are not validated for use after TL. By conducting this study as a prospective cohort study, we aimed to overcome selection bias; nevertheless, a number of participants were not included, as- sessments were missed due to logistic reasons and medical complications, and participant mortality were excluded from the study. Therefore, outcomes are collected from patients who are alive and willing to fulfil study related proce- dures, which may lead to overestimation of the outcomes. We anticipated on evaluating different voice methods, e.g. esophageal speech and electrolarynx speech. In this cohort, however, no esophageal speakers were present and only two participants used electrolarynx speech. Therefore, no sub-group analysis between voice methods could be performed. 3.4.2 Recommendations for clinical practice and future research Thirty per cent (n=13) of participants did not complete the study due to mortality. Sadly, nine participants did not reach acceptable (TE-)speech and had to depend on augmentative alternative communication methods such as typing, writing and mouthing in the palliative phase of their life. For clinical practice, it is recommended to inform patients about the possibility to end-up without sufficient tracheoesophageal speech, especially when prognosis is poor. The instruments in this study have shown to be useful to detect a differ- ence over time from pre- to one year post-surgery. Former studies that have evaluated voice outcomes after total laryngectomy utilise a wide variety of mea- surement tools and time points after surgery [2, 4]. AVQI, VHI-10, EQ-5D-5L, and VAS scales for perceptual ratings, used in our study, proved to be sensitive to detect differences over time from pre- to post-surgery. Sensitivity is lacking when differences between treatment groups and over time post-surgery have to be detected. Continued efforts are needed to establish the optimal tools, and validate these instruments for research and clinical practice in this population. Improvement for patient-reported voice functioning and QoL at twelve months post-surgery was found, whilst AVQI score remains altered (Figure 3). This