JCPSLP Vol 18 no 2 July 2016

Aims of this review A comprehensive systematic review of the prognostic and predictive factors of stuttering is beyond the scope of this article. Instead, the aim is to provide a synopsis of prognostic and predictive factors, and to present an argument for why there is a need for a comprehensive systematic review of the topic to be conducted. There are no previously published reviews of predictive factors of treatment outcomes of stuttering. In contrast, there are some previous published reviews of stuttering prognostic factors, for stuttering onset and for persistence of stuttering without treatment. For example, Yairi and Ambrose (2013) discuss the incidence, prevalence, natural recovery and persistency, and subtypes of stuttering in light of recent research advances. While the literature in this area has not changed dramatically in the past 25 years, more recent studies, including a prospective, community cohort study by Reilly and colleagues (Reilly et al., 2013), are contributing new insights into prognostic factors, outlined below. Prognostic factors associated with stuttering onset and persistent stuttering Stuttering onset The onset of stuttering usually occurs between 2 and 5 years of age (Bloodstein & Bernstein Ratner, 2008). The prevalence of stuttering, or percentage in a particular population at a given time, is just below at 0.72% (Craig, Hancock, Tran, Craig, & Peters, 2002). Recent incidence data of the lifetime risk for stuttering indicate a rate of between 8% (Dworzynski, Remington, Rijsdijk, Howell, & Plomin, 2007) and 11% (Reilly et al., 2013). Accurately predicting who will stutter is challenging given there is no single known cause. However, in recent years, there has been converging evidence to indicate that stuttering is a complex neurological disorder of speech motor control with genetic influences (Dworzynski et al., 2007). Genetics Regarding prognostic factors of stuttering onset, there is a tendency for stuttering to run in families, with approximately 30%–50% of people who stutter reporting a positive family history (Bloodstein & Bernstein-Ratner, 2008). Stuttering is more common in monozygotic twins (52%) than in dizygotic twins (12%) (Dworzynski et al., 2007), but the specific role of genetics in stuttering onset is still somewhat unclear. Linkage studies have analysed the genetic marker(s) for stuttering in affected and unaffected members in families (Yairi & Ambrose, 2013). Such studies have reported multiple genes that could be related to stuttering. However, the findings are inconsistent and need replication. In Reilly et al.’s (2013) study, logistic regression analysis found that being a twin was a significant prognostic variable and family history of stuttering was close to significant ( p = .07) for predicting stuttering onset by 4 years of age. Age Age is a relevant prognostic factor for onset of stuttering. The older a child is, the less risk they have of beginning to stutter. Reilly et al. (2013) reported that the incidence for stuttering onset slowed markedly after 3.6 years of age. Similarly, Craig et al. (2002) found that a child of 6 years or older is 75% less likely to start stuttering when compared with younger children.

Gender There are conflicting findings regarding gender as a prognostic factor. Yairi and Ambrose (2013) concluded that only small differences have been found between males and females for stuttering onset, whereas Reilly and colleagues (2013) reported that being male was a significant risk factor for stuttering onset by age 4. Neurology Neurological studies of factors associated with stuttering onset are emerging. To illustrate, brain scans of preschool aged children, taken soon after the onset of stuttering, have revealed deficiencies in left grey matter volume when compared to fluent controls (Chang, Erickson, Ambrose, Hasegawa-Johnson, & Ludlow, 2005). Maternal education Reilly et al. (2013) reported that by age 4 ( n = 1619), higher maternal education was an additional significant prognostic variable for stuttering onset using logistic regression analysis. The authors speculated that this may be due to more reports of stuttering in children from mothers who are more highly educated and aware of stuttering. Persistent stuttering The difference between the incidence and prevalence rates has been attributed to the rate of natural recovery in a large proportion of people who stutter. Some disagreement as to the definition of natural recovery exists (Bloodstein & Bernstein Ratner, 2008), but it is generally considered to be recovery without clinical intervention (Yairi & Ambrose, 2013). Alternatively, treatment assisted recovery is sometimes referred to, and the distinction between this and natural recovery should be made clear in studies. At times it is not, and this can be problematic in defining relevant prognostic factors to predict those who recover or persist with stuttering without clinical treatment. It has been suggested that up to 74% to 83% of children who start to stutter recover naturally (Ambrose & Yairi, 1999; Dworzynski et al., 2007), but this also means that approximately 1 in 5 of these children will develop a persistent stutter. It is of interest to be able to predict if a child who has started to stutter will naturally recover in order to accurately prioritise therapy services to those at Persistent stuttering has been found to relate to positive familial history of stuttering (Yairi & Ambrose, 2013) with individuals who stutter and have a family member who stuttered but recovered being more likely to recover themselves (Dworzynski et al., 2007; Yairi, Ambrose, Paden, & Throneburg, 1996). Gender The sex ratio of stuttering indicates that gender in itself is a risk factor for persistent stuttering. Ambrose and Yairi (1999) conducted a longitudinal study of 147 children, collecting data from when the children were close to stuttering onset. They found that 84% females recovered versus 77% males, and that females who recovered did so at a younger age than males who recovered. Contrary to findings from Yairi and colleagues, Reilly et al. (2013) reported a higher rate of recovery within 12 months for males compared to females. However, as the number of recovered children was low (n = 9), the authors stated that they could not examine predictors of recovery appropriately. greater risk. Genetics

95

JCPSLP Volume 18, Number 2 2016

www.speechpathologyaustralia.org.au