Quality of life in adults who stutter

Quality of life in adults who stutter

A B S T R A C T

Although persistent developmental stuttering is known to affect daily living, just how great the impact is remains unclear. Furthermore, little is known about the underlying mechanisms which lead to a diminished quality of life (QoL). The primary objective of this study is to explore to what extent QoL is impaired in adults who stutter (AWS). In addition, this study aims to identify determinants of QoL in AWS by testing relationships between stuttering severity, coping, functioning and QoL and by testing for differences in variable scores between two AWS subgroups: receiving therapy versus not receiving therapy. A total of 91 AWS filled in several questionnaires to assess their stuttering severity, daily functioning, coping style and QoL. The QoL instruments used were the Health Utility Index 3 (HUI3) and the EuroQoL EQ-5D and EQ-VAS. The results indicated that moderate to severe stuttering has a negative impact on overall quality of life; HUI3 derived QoL values varied from .91 (for mild stuttering) to .73 (for severe stuttering). The domains of functioning that were predominantly affected were the individual’s speech, emotion, cognition and pain as measured by the HUI3 and daily activities and anxiety/depression as measured by the EQ-5D. AWS in the therapy group rated their stuttering as more severe and recorded more problems on the HUI3 speech domain than AWS in the non-therapy group. The EQ-VAS was the only instrument that showed a significant difference in overall QoL between groups. Finally, it was found that the relationship between stuttering severity and QoL was influenced by the individual’s coping style (emotion-oriented and task-oriented). These findings highlight the need for further research into stuttering in relation to QoL, and for a broader perspective on the diagnosis and treatment of stuttering, which would take into consideration quality of life and its determinants.

Learning outcomes: Readers will be able to:

(1) Understand how the Wilson and Cleary (1995) model of quality of life could be applied to comprehensively assess the quality of life in adults who stutter,

(2) describe how health related quality of life is impaired in adults who stutter,

(3) mention affected domains of functioning that are related to health related quality of life impairment in adults who stutter,

(4) describe the relationship between stuttering severity, functioning, coping and health related quality of life in adults who stutter,

(5) describe differences in stuttering severity, coping style, functioning and health related quality of life between adults who stutter who have registered for therapy and adults who stutter who have not.

Discussion

The objectives of the present study were (1) to investigate to what extent QoL is impaired in AWS and (2) to identify determinants of QoL in AWS. The latter was pursued by exploring relationships between stuttering severity, coping, functioning and QoL and by testing for differences in variable scores in two subgroups: the NT group and the T group. The results of this study show that stuttering severity affects overall QoL considerably. HUI3 derived QoL values were .91 for mild stuttering and .73 for severe stuttering. AWS who had just begun or were about to begin therapy rated their stuttering as more severe and recorded more problems on the HUI3 speech domain than AWS who were not in therapy. However, the results with respect to the differences in overall QoL between the T and NT group varied. While differences in overall QoL were not significant according to the HUI3 and the EQ-5D, according to the EQ-VAS they were. The effect size was .81, which can be considered as large (Cohen, 1988). The correlation analysis between stuttering severity and domains of functioning in the total group showed that a higher stuttering severity was mainly associated with limitations in the domains of speech and emotion. Lastly, regression analysis showed that the relationship between stuttering severity and overall QoL was influenced by task-oriented and emotion-oriented coping style.

With regard to the extent to which QoL in AWS is affected, our study could not confirm that the impact of severe stuttering on overall QoL was as great as suggested by Bramlett et al. (2006). QoL values for severe stuttering in our study ranged from .73 to .88, while Bramlett et al. (2006) found QoL values between .44 and .81 for severe stuttering. There could be two reasons for this difference. Firstly, this might be related to the somewhat wider range of stuttering severity in the Bramlett et al. (2006) study. Although a substantial number of participants in the current study had low scores on the SA-scale, which represents severe stuttering, none of the participants was classified as severe stuttering by the OASES-A. Secondly, the difference in QoL values may be due to differences in the way QoL values were obtained. Bramlett et al. (2006) derived their QoL values by direct valuation of vignettes describing stuttering: AWNS rated hypothetical

states of stuttering and their own health state. In the current study, QoL was indirectly assessed by using the HUI3 and EQ-5D. In this way QoL values (from the general public) were derived by applying a mathematical algorithm to the health states that were described by the AWS. These health states were generic, that is, they had no specific reference to stuttering. Therefore, the indirect instruments applied in the current study might not have been responsive enough to stuttering, resulting in an upward bias. In other words, the impact of stuttering on QoL might actually be greater than found in our study. Alternatively, it could be hypothesized that the absence of anchor points referring to other conditions worse than severe stuttering led to a downward bias in the direct assessment approach by Bramlett et al. (2006). This is known as contextual bias (Doctor, Bleichrodt, & Lin, 2008). The two studies have no singlemeasure in common to explore whether the negative impact on QoL has been underestimated in our study or overestimated in the Bramlett et al. (2006) study, or both.

In the current study, as in the study of Bramlett et al. (2006), substantial differences in QoL values were established using different instruments. Comparing the three QoL measurements for the most severe stuttering state, the impairment on the HUI3 was greater than on the EQ-5D and EQ-VAS. This difference may be explained by inclusion of the speech domain in the HUI3, which improves its responsiveness to stuttering. This might also clarify why the HUI3 measurement showed QoL impairment for the mildest forms of stuttering, but EQ-5D measurement showed relatively little or none. Ceiling effects for the EQ-5D, as reported in other relatively healthy populations (Kopec & Willison, 2003; Lamers, Bouwmans, van Straten, Donker, & Hakkaart, 2006), may have contributed to a limited responsiveness of this instrument in AWS. Accordingly, the EQ-5D might have overestimated QoL, although the alternative hypothesis, that the HUI3 has underestimated QoL, cannot easily be abandoned. By inclusion of speech as a domain, the emphasis on the speech problems may be larger than their impact on QoL warrants.

In theory, EQ-VAS outcomes could help to identify whether QoL was underestimated by the HUI3 or overestimated by the EQ-5D, since the EQ-VAS measures QoL directly and not via its impact on basic domains of functioning. Therefore, the VAS scale is not prone to possible misrepresentation of QoL, which could occur if the HUI3 and the EQ-5D do not include all the relevant domains. In addition, the EQ-VAS values QoL from the perspective of the respondent himself instead of the general population. However, neither hypothesis could be supported, since the results indicate that the EQ-VAS was less responsive

than both the EQ-5D and HUI3 for changes at the symptom level. An ‘end of scale’ bias might have limited the responsiveness of the EQ-VAS. Subjects tend to avoid using scale ends (Drummond, Sculpher, Torrance, O’Brien, & Stoddart, 2005; McCabe et al., 2006), which implies that the QoL effect of mild health problems is difficult to measure on a VAS scale. Support for this hypothesis is found in the result that EQ-VAS scores were limited to a smaller range of the scale than HUI3 and EQ-5D scores.

Thus, unfortunately, the EQ-VAS does not provide the key to whether the EQ-5D overestimated QoL, or the HUI3 underestimated it.

Our findings that stuttering affects functioning in a negative way are in line with the results of other studies (e.g. Andrade et al., 2008; Craig et al., 2009; Klompas & Ross, 2004). The domains that significantly correlate to stuttering severity in our study correspond to a great extent with the domains affected in the Craig et al. (2009) study, that is mainly social and psychological dimensions. An interesting finding of the current study is the positive correlation between stuttering severity as measured by the OASES-A and the pain domain of the HUI3. This result may reflect the broad definition of the HUI3 pain domain, which covers pain and discomfort. Alternatively, AWS reporting physical pain, especially in the breast region, when asked what they feel in their body when they speak, stutter or try to avoid stuttering, is a quite common response in the clinical experience of the third author. Besides, it may be hypothesized that stuttering affects physical well-being because of higher stress levels associated with the experience of social anxiety (Menzies et al., 2009). There is evidence for a common neural basis for regulating social pain and physical pain (Macdonald & Leary, 2005). As a result, the physical pain threshold can be triggered by social pain.

The regression analyses into the relationships between stuttering severity, coping and overall QoL identified coping as a mediating factor in QoL in AWS, in addition to stuttering severity and demographic variables. The results of the HUI3 regression analysis suggested that both stuttering severity and coping style can be directly related to QoL in equal measure.

Two types of coping were associated with QoL. Higher scores on the CISS-E (emotion-oriented coping) were correlated with lower QoL. While it is known that dealing with emotions in a constructive way positively influences the adjustment to a chronic disease (De Ridder et al., 2008), higher CISS-E scores reflect a more negative way of dealing with emotions (e.g. denial, mental or behavioral distance, brooding), presumably resulting in a greater psychological impact and a lower QoL (De Ridder & van Heck, 2004). The regression analysis also revealed that higher task-oriented coping scores were associated with better QoL, reflecting that task-orientation is an active and adaptive way of coping which influences QoL in a positive way (Lazarus & Folkman, 1984). QoL might be maximized by individuals who apply the various strategies flexibly depending on the circumstances that they have to deal with (Lazarus, 1993).

The differences in the results between the therapy group and non-therapy group in this study provide further insight into the underlying mechanisms of QoL in AWS. The groups differed significantly in stuttering severity, in score on the speech domain of the HUI3 and in overall QoL as assessed by the EQ-VAS. There were no group differences in coping scores. The regression analysis with the EQ-VAS as dependent variable was the only analysis that revealed group ID as a significant predictor of overall QoL. These results suggest that AWS who seek treatment do this because they desire symptom relief, and not because they are poor at coping.

Elements in our study design that might evoke questions about the external validity are related to the choice of including a T and NT group of AWS and to the use of self-assessed measures to establish stuttering severity. The NT group was included because we wanted to cover the maximum range of QoL values in the group of AWS and hypothesized that QoL might be higher in AWS not seeking treatment and/or that relationships between stuttering, coping and QoL might differ between groups. The representativeness of the NT group cannot be established, due to the lack of detailed information about the Dutch AWS population not receiving treatment. Furthermore, the results show that there are between group differences, namely a lower stuttering severity and a better subjective QoL, as measured with the EQ-VAS, for the NT group. The difference in stuttering severitywas also reflected in a betterHUI3 speech QoL value for theNT group.

These results imply that outcomes obtained in clinical populations cannot simply be generalised to the population of AWS as a whole and vice versa. With regard to the applied speech measures, we are confident that self-identification of stuttering in the NT group and self-assessed stuttering severity has not negatively affected the external validity, because 81% of the AWS in the study reported having been previously diagnosed as stuttering by a professional. Furthermore,

Huinck and Rietveld (2007) showed that correlations between a self-assessment scale of speech satisfaction and measures which reflect overt stuttering behavior are relatively strong, indicating a high validity of a simple and cost-effective speech rating scale. This suggests that our study results would provide a valid estimation of QoL in all AWS.

Our study presents evidence that stuttering in adults is a serious problem affecting health. A broadly based outcome measure such as QoL could provide a means of evaluating the impact of stuttering on daily life. QoL measures could therefore be applied in therapy evaluation studies, or in evaluating the relationship between the cost and benefit of stuttering interventions. Furthermore, the relevance of coping for QoL in AWS, which was demonstrated in this study, shows that a good understanding of the determinants of QoL is essential to develop rational and cost-effective treatments: ‘‘The development of treatment strategies requires not only that we identify the key factors that combine to determine function and quality of life, but also that we understand their relative importance and the degree to which they can be altered or modified’’ (Wilson & Cleary, 1995, p. 63). Our study is a first step in exploring the determinants of QoL in relation to stuttering. The effect of coping on the relationship between stuttering severity and QoL which was established in this study suggests that addressing coping style could be a useful component in the process of diagnosing and selecting treatment approaches for AWS. Using a coping instrument during the assessment phase indicates how an individual copes with stressful situations in daily life. If an AWS is using an inadequate coping pattern, therapeutic goals could be identified which would enable the AWS to change his personal coping style to deal more effectively with stressors that provoke stuttering or the stuttering behavior itself, thereby reducing its negative impact on QoL. For instance, if a client displays relatively high scores on the emotionoriented coping scale and low task-oriented coping scores, treatment goals might be focused on learning task-oriented coping strategies and becoming less dependent on emotional ways of dealing with stress. This idea is supported by Hayhow et al. (2002) who showed that AWS have the desire to get help in managing their stuttering and in developing coping strategies. We would therefore recommend that more studies be done on coping in relation to stuttering, such as the ones recently reported by Plexico and colleagues (Plexico et al., 2009a, 2009b).

In conclusion, by using generic QoL measures, it was shown that the health condition of moderate to severe stuttering substantially reduces the QoL in AWS as compared to the perfect health state. This result, and the significant relationship between stuttering severity, coping style and QoL, highlights the need for further research in order to clarify the conceptualization of QoL in relation to stuttering, as a foundation for the further development of effective therapies for the

disorder of stuttering.

PDF>>