JCPSLP Vol 18 no 2 July 2016

Age at onset Yairi et al. (1996), in comparing groups of children who recovered from stuttering with those whose stuttering persisted, reported that children who continued to stutter started stuttering 5 to 8 months later (onset at 39 months of age) than those who recovered (onset from 30 to 33 months of age). Time since onset The time since onset of stuttering has been found to be a prognostic factor for persistent stuttering, with the risk of persistent stuttering being greater in individuals who have been stuttering for more than one year than in those who have not (Ambrose & Yairi, 1999). Neurology Chang and Ludlow (2010) reported further neurological data to identify factors of persistence. Specifically, children who had a persistent stutter had reduced white matter integrity in the left hemisphere areas of the brain involved in speech compared to those who recovered. Language ability There are inconsistent findings for status of speech and language skills in predicting persistence of stuttering. For example, Yairi et al. (1996) reported that better speech and language skills may be related to recovery from stuttering, whereas Watkins and Yairi (1997) found that children whose stuttering persisted had typical to advance language skills. Additional factors Other behavioural factors that have been found to be different between children with persistent stuttering and children who recovered include differences in second formant transitions (Subramanian, Yairi, & Amir, 2003), higher variability of articulation rate for persistent stuttering (Kloth, Kraaimaat, Janssen, & Brutten, 1999), and poorer phonological and speech production abilities for persistent stuttering (Spencer & Weber-Fox, 2014). Howell and Davis (2011) found that symptom severity, as measured by the Stuttering Severity Index, of children who stutter at 8 years old was the only significant factor that was able to predict persistence of stuttering in the teenage years using logistic regression analysis. However, Yairi and colleagues (1996) did not find that initial stuttering severity predicted persistence in their study of younger children (under 6 years). Interpreting risk and prognostic factors In summary, common prognostic factors of stuttering onset and of persistent stuttering include positive family history of stuttering and age. An older child has less risk of stuttering onset, though if a child is older when onset does occur, the risk for the stutter to be persistent is higher. Additional factors associated with persistent stuttering (i.e., behavioural factors) have been identified; however, caution is required when interpreting such findings. For example, replication of findings is needed to establish reliable indicators of persistence, since current results are based predominantly on single or few studies, with many having small sample sizes (e.g., Chang et al., 2005, n = 14). A number of studies are retrospective in design and it is recommended that prospective studies are more appropriate to answer questions of prognosis (Moons et al., 2009). While there have been comprehensive reviews of prognostic factors of stuttering onset and persistence, none were systematic reviews. A systematic review would synthesis and evaluate available information on this topic

allowing for the interpretation of data from large bodies of information (Petticrew & Roberts, 2006). Predictive factors for stuttering treatment Findings from systematic reviews There is much to learn from investigating the predictive factors of stuttering treatment outcomes. With improved measures of treatment outcomes encompassing the impact of stuttering on an individual’s quality of life, as well as the more reliable identification and measurement of relevant client and clinician factors, it may be possible one day to develop predictive models of treatment success. Previous systematic reviews of stuttering treatment include one by Bothe, Davidow, and Bramlett (2006). They conducted a qualitative systematic review of behavioural, cognitive, and related approaches to stuttering therapy across a range of age groups. The prolonged speech approach, treatment targeting self-management, and treatment using response contingencies were found to have the strongest evidence for adults who stutter (AWS). For early stuttering, the authors concluded response- contingent-based therapies had the strongest evidence. Herder, Howard, Nye, and Vanryckeghem (2006) reported results from a systematic review and meta- analysis of behavioural stuttering treatments and concluded that treatments resulted in positive therapeutic gains, but there was no one treatment approach that was significantly better than the others. Subsequently, a systematic review by Nye and colleagues (2013) reported that the current best evidence for children who stutter up to 6 years of age was the Lidcombe Program (LP) developed by Onslow and colleagues. Of these systematic reviews of treatment in stuttering, none specifically addressed predictive factors of treatment outcomes. Nevertheless, the importance of finding out more about predictive factors was recognised (Herder et al., 2006; Nye et al., 2013). Nye et al. (2013) further acknowledged that there were limited data available to make any useful conclusions. Herder et al. (2006) noted a lack of understanding of how each factor or sets of factors may impact on treatment outcome. Interestingly, they concluded that differences in the effectiveness of intervention approaches were unlikely due to the nature of the intervention strategies themselves or to participant characteristics. Instead they hypothesised that clinician impact could have played a role, stating that “it might well be that the clinician represents a helping individual who is perceived to have the knowledge and skills to bring about a change in the speech behavior of a person who stutters” (Herder et al., 2006, p. 70). Adults who stutter Results from the systematic reviews demonstrate that treatment for stuttering is generally beneficial. However, predicting which individuals will benefit most, as well as knowing how to support those who may have higher risk of regression, is a pivotal for clinicians prior to commencing intervention. The potential to relapse is of interest and concern in the provision of treatment to AWS, with up to 72% of adults relapsing post-treatment (Craig, 1998). Relapse is defined as “stuttering to a degree which was not acceptable to yourself for at least a period of one week” (Craig, 1998, p. 3). Stuttering severity The definition of a predictive factor provided at the beginning of this review stated that it relates to client

96

JCPSLP Volume 18, Number 2 2016

Journal of Clinical Practice in Speech-Language Pathology