Klaske van Sluis

36 2.4. Discussion 2.4 Discussion This systematic review underlines that the three main TL speech rehabilita- tion methods are acoustically and perceptually deviant from healthy speech. In PROs no comparison is made between the substitute speech rehabilitation groups and healthy speakers. Significantly better outcomes are reported for TES compared to ES for the acoustic parameters, fundamental frequency, max- imum phonation time and intensity. Perceptually, TES is rated with a signif- icantly better voice quality and intelligibility than ES and ELS. None of the speech rehabilitation groups reported evidently better outcomes in patient- reported outcomes. These outcomes need to be interpreted with caution. Only three of the 26 included studies are rated with a low risk of bias (level A). Most outcomes, thus, stem from level B rated studies. The included studies contain small numbers of patients, and inferential statistics is not always performed. In most studies the methodology of the acoustic measurements was not specified, leading to possi- bly incorrect outcomes. We found several extreme outliers in F0 and shimmer, that we had to exclude because of this [41, 44, 52]. Difficulties in reliable mea- suring intensity values are acknowledged, no absolute values are reported but only outcomes within studies. This systematic review shows once more that there is an urge for standard- ized measurement tools for evaluations of substitute voice speakers. Auditory- perceptual evaluations are often considered as being the gold standard for voice and speech evaluation. However, the great dispersion between raters has to be acknowledged. Researchers have proposed rating tools for standardized evalua- tions [7, 16, 18]. Nevertheless, these are not yet generally adopted. An interest- ing new approach is the development of automatic assessment tools, which are designed to provide objective outcomes, with some promising results recently being reported [56, 57]. Even though not all present automatic assessment tools seem suitable for analyzing substitute voices, in our opinion this is the most promising way to obtain objective voice outcomes. The number of PRO studies that could be included in this review is limited. The EORTC QLQ-H&N35 and EORTCQLQ-C30 were defined as relevant out- come measure but not reported in the included studies. We did not find studies which specifically report the outcomes on the speech domain of this question- naires for the different speaker groups. The VHI and V-RQOL are generally applied to evaluate vocal functioning after TL. The Communication and Par- ticipation Item Bank (CPIB) is a recently developed questionnaire [30], which is why it was not initially defined as an outcome of interest. However, the level A rated study of Eadie et al. [30] showed strong correlations between the VHI- 10 and CPIB short form scores. In this study, the speech rehabilitation groups were asked to judge their own voice quality and intelligibility. These outcomes also strongly correlated with the CPIB short form scores. Therefore, the CPIB short form can be seen as a useful tool to obtain patients’ opinion on vocal