The DYMUS questionnaire can provide clinicians with information about the risk of dysphagia in patients with MS. Our study brings novel data about dysphagia in a sample of Czech patients with MS and provides a validated Czech version of the DYMUS questionnaire, supplemented by basic normative data.

Our study provides a basis for the evaluation of dysphagia by an electronic version of the DYMUS questionnaire, with a possibility of a paper-based backup, if necessary. In our study, the paper-based backup was requested especially by older participants who are not digital natives. Although we did not directly compare the electronic and paper-based DYMUS, there were no significant differences between the versions. In our sample, the DYMUS questionnaire results were not related to the type of version administered. Furthermore, both versions achieved similar item and scale statistics.

Based on the PCA, we saw a similar two-component questionnaire structure (dysphagia for solids and liquids) as in the original English version of the questionnaire [13]. However, a few items did not completely correspond to the original factor loadings as proposed in the English version. This corresponds to other studies using PCA [12, 17] and suggests that the discriminative validity of the DYMUS questionnaire for dysphagia to liquids or solids needs further research attention. Moreover, our results support previous findings [12, 16,17,18,19, 21 showing poor functioning of item Q10 (loss of weight); therefore, we have proposed a shorter 9-item version of the questionnaire.

Concerning item Q10, some authors have argued that the item should be preserved because a loss of weight is one of the most endangering symptoms of dysphagia [28]. However, others have pointed out the item’s low specificity because a loss of weight can be caused by many other medical conditions, including malnutrition due to MS [12]. After pilot-testing, we tried to reword the item to exclude a voluntary loss of weight from the question (e.g., weight-reduction diet). As our results showed, this attempt did not improve the item’s functioning. Considering the original purpose of the DYMUS questionnaire as a screening tool before performing detailed dysphagia instrumental assessment [1, 29,30,31], we argue that item Q10 with its low specificity is rather problematic and should be dropped. It would be an essential question only in a population with already instrumentally confirmed dysphagia.

Similarly to other study [18], our normative data suggested the cut-off value for symptoms of dysphagia to lie between total scores of 1 and 2 for both the original 10-item and the shorter 9-item versions of the questionnaire. Using the newly proposed criteria, the prevalence of dysphagia in our sample of patients with MS was around 11%. Even with the originally recommended one positive answer criterion [13], only 23% of people with MS in our Czech validation study showed a risk of experiencing dysphagia. This contrasts with other validation studies which have described a prevalence of dysphagia between 31 and 58% [16, 18, 19, 32]. This difference might be explained by a lower EDSS (the median EDSS score was 2.0) in our sample, which is further supported by previous studies in which the incidence of dysphagia in people with mild disability was similar to our current findings – that is, 17%–21% [2, 5].

Although we did not assess test–retest reliability, we tested long-term stability of the results in patients with MS after 1 year. Kendall’s tau-b of 0.788 is very satisfactory, especially considering the longer 1-year retest period, possible fluctuations in the results due to MS and the rather conservative correlational measures applied. Our data on long-term stability are comparable to classical test–retest measures presented in other validation studies [12, 14, 17,18,19].

We also proved the criterion validity by comparing the results of patients with MS and HC, and by a small positive correlation between the DYMUS score and disease severity measures such as the EDSS, and the level of anxiety, depression and subjective perceived fatigue. Similarly to dysphagia, these measures worsen in people with greater disease burden [33]. The EDSS has proved to be correlated positively with DYMUS scores across multiple studies [14, 16,17,18,19, 28]. However, in contrast to previous findings [2, 14, 28], we found no correlation between the DYMUS score and disease duration. This is not that surprising given the fact that disease duration does not necessarily provide information about disease severity, and it is in line with other authors with similar results [17, 18, 32, 34]. A larger sample that is more representative of people with more severe MS could generate stronger correlations with other disease severity measures, also including disease duration.

While preparing this manuscript, validation of a second Czech version of the DYMUS questionnaire was published [28]. This is a rather unique situation given the fact that both teams received the DYMUS author’s permission to translate the questionnaire. Both studies showed strengths of the Czech DYMUS translation. There was satisfactory internal consistency comparable with other language versions [11, 12, 14,15,16,17,18,19, 21] for both Czech versions. The version of Kolčava et al. [28] was validated on a smaller sample of patients with MS and HC; however, the authors provided data on excellent test–retest reliability (assessed telephonically after 5 days; Spearman’s r = 0.94, p < 0.001). They also confirmed the translation’s criterion validity through correlations with the EDSS, disease duration and the 100 ml water swallowing test. On the other hand, Kolčava et al. [28] did not assess dimensionality and did not provide detailed normative data.

There are three minor content discrepancies between the Czech translations worthy of discussion. Specifically, they are in items Q4 (Kolčava et al. [28] does not follow the original English version so precisely and asks about the ‘feeling of’ food stuck in your throat), Q7 (our version does not mention literally ‘solid food’ as the English original does) and Q10 (based on pilot-testing, we modified Q10 by asking for ‘unintentional’ loss of weight, to eliminate an affirmative answer by people on voluntary weight-reduction diets). The discrepancies between the versions are rather small, yet it would be beneficial to propose a merged consensual Czech version in the future.

There are some limitations to this study. First, we did not validate the DYMUS questionnaire with objective instrumental tests such as a videofluoroscopic swallowing study (VFSS), fibreoptic endoscopic evaluation of swallowing (FEES) or EMG activity. Second, the sample in our study was not random. We distributed the DYMUS questionnaire as a part of the neuropsychological assessment in larger observational prospective studies. Even though our MS sample chosen by judgemental sampling represents all major MS disease subtypes, patients with a low burden and shorter disease duration are overrepresented in our sample. Third, because of COVID-19 preventive measures, we decided to discontinue questionnaire distribution early. We believe it did not cause any bias because patients were assigned over time independently of their demographic or disease characteristics. Fourth, this validation study was performed in a single centre; however, the main results can be applied to other Czech-speaking MS centres. Fifth, in patients with MS the administration was part of their routine visit to the MS Center, a factor that could negatively affect their motivation, concentration or time possibilities. This could influence the different response rate in patients with MS and HC. If there was a lack of time, patients were asked to complete the questionnaire remotely. However, the size of this sample did not allow us to control for differences between on-site and remote distribution.