Effectiveness of intensive, group therapy for teenagers who stutter

Authors


Abstract

Background

Treatment of adolescents who stutter is an under-researched area that would benefit from greater attention.

Aims

To investigate whether an intensive treatment programme for older teenagers who stutter, aged over 16 years of age, is effective in reducing overt and covert aspects of stuttering.

Methods & Procedures

A repeated-measures, single-subject experimental design was replicated across participants. The study consisted of a 5-week baseline phase, 2-week intensive treatment phase, 5-week consolidation phase and 10-month follow-up phase. Participants were asked to make ten video recordings at home during each phase, while completing a reading and a conversation task. Recordings were analysed in terms of the percentage of stuttered syllables using a simplified time-series analysis. Participants completed self-report questionnaires at predetermined times throughout the study. Data are presented for three males aged 17;7, 17;11 and 18;10.

Outcomes & Results

One participant completed all required recordings. Difficulties were encountered collecting follow-up data with the other two participants and data are available up to 5 months after the intensive therapy phase. A significant trend of reduced frequency of stuttering was found for all three participants during the intensive therapy phase. This trend continued throughout the consolidation phase and remained significant when available longer-term data were included in the analysis. Participants also reported increased self-efficacy about speaking and reduced overt and covert aspects of stammering.

Conclusions & Implications

Findings show that this therapy programme for teenagers had a significant treatment effect for the participants studied in the short- and medium-term, however longer-term data were not available for all participants. Issues in conducting research with this client group are discussed.

What this paper adds?

This paper outlines an integrated programme of therapy incorporating speech restructuring, cognitive behaviour therapy and communication skills training which has been implemented at the Michael Palin Centre. It adds to the current discussion about the use of CBT in the treatment of stuttering. It highlights some of the challenges involved in collecting long term data with adolescent research participants, however adds to the discussion of effectiveness of therapy with teenagers who stutter and underlines the importance of therapy resources being available for adolescents who stutter.

Introduction

The treatment of teenagers who stutter is an under-researched area (Bothe et al. 2006) but there are several reasons why therapy for this age group, and research into the effectiveness of therapy is important. Stuttering is increasingly viewed as a chronic disorder as adolescence progresses with the associated long-term risks of chronic stuttering including reduced psychological, social, vocational and economic opportunities (Blumgart et al. 2010, Yaruss 1998) and vulnerability to social anxiety (Kraaimaat et al. 2002, Mahr and Torosian 1999) which may persist across the lifespan (Bricker-Katz et al. 2009). There is evidence that adolescents who stutter are already vulnerable to these risks. They have been found to score more highly than adolescents who do not stutter on measures of communication apprehension (Blood et al. 2001) and social anxiety (Mulcahy et al. 2008), and to view themselves as being less able communicators than their fluent peers (Blood et al. 2001). Craig and Tran (2006) concluded that adolescents who stutter are more likely to become more shy, socially avoidant and fearful of communication than their fluent peers, and to develop pervasive negative attitudes and beliefs about themselves as communicators. The teenage years are therefore a potentially critical period for minimizing the severity and impact of chronic stuttering.

There are also developmental reasons why therapy may be particularly important at this time. Adolescence is a phase of rapid physical, social, cognitive and emotional change (Dahl and Gunnar 2009). Of particular interest from a clinical point of view are developments in executive function and construction of the sense of self. Executive function refers to the ability to direct thoughts and actions according to internal goals. Its development is particularly marked during adolescence and is demonstrated by increases in mental flexibility, ability to self-monitor accurately, appraise personal skills and performance, ability to plan tasks, work independently and problem-solve (Crone 2009). The way that the sense of self is constructed also alters during adolescence, with a shift towards this being based on how the individual believes that he or she is seen by others (Sebastian et al. 2008). The development of what is referred to as the ‘looking-glass self’ has a greater role during adolescence (Sebastian et al. 2008) and is associated with heightened self-consciousness, increased self-evaluation, and a greater awareness of, sensitivity to and concern about others’ opinions (Parker et al. 2006).

The normal maturation of adolescence may therefore result both in greater awareness of stuttering and increased self-consciousness about it, which can make it more complex to treat. However, adolescents also bring a set of cognitive skills that may help them to engage more successfully than when younger, in a process which requires self-monitoring, reflection and self-direction.

Finally, there is evidence that therapy with adolescents can be beneficial. Several studies have reported improvements after therapy based on speech restructuring using either smooth speech (Craig et al. 1996), prolonged speech (Hearne et al. 2008a) or gradual increase in length and complexity of utterance (GILCU) (Ryan and van Kirk Ryan 1995). Speech restructuring has also been shown to be effective when integrated with social skills training (Rustin and Purser 1983), avoidance reduction therapy (Boberg and Kully 1994), electromyography treatment (Craig et al. 1996), cognitive–behaviour therapy (CBT) (Blood 1995), and CBT plus communication skills training (Fry et al. 2009).

However difficulties engaging and retaining teenage clients in therapy, and in research projects, are often encountered and flexibility may be required in terms of the timing, context and scope of therapy. In their qualitative study of teenagers’ perspectives on stuttering and therapy, Hearne et al. (2008b) found that adolescents reported becoming more concerned about their stuttering at different ages, although there was a trend for this to occur in later adolescence as they approached a transition to the workforce or further education. They found that teenagers valued being able to engage in therapy at a time of their own choosing and the camaraderie of group therapy. Considerations such as age, maturity and readiness for therapy may therefore influence outcomes and, while it may not be the preferred option for all teenagers, group work has particular developmental relevance and may add to the appeal of therapy for this client group.

Aim

Fry et al. (2009) investigated the efficacy of a group therapy programme, based on the work of Rustin et al. (1995) and also described by Fry and Cook (2004), which consists of a 2-week, weekday only, intensive course for teenagers aged 16–19 years, integrating speech management skills, CBT and communication skills training. These authors used a repeated-measures, single-subject design, which is a methodology that has been used to investigate the efficacy of therapy for children who stutter (Matthews et al. 1997, Millard et al. 2008) and has advantages when research is conducted in real-world clinical settings. It allows for variations in therapy, which can be expected in response to the individual needs and preferences of clients, and the extensive baseline data means that the subject acts as their own control and a control condition is not required (Kazdin 1982, Pring 2005). Fry et al. (2009) found that the severity of overt and covert features of stuttering decreased for the subject studied, and reported self-efficacy related to speaking increased in response to therapy. Changes were shown to be durable, being maintained up to 10 months after therapy. However, while single-subject studies have high internal validity, results cannot be generalized, although this limitation can be addressed, and validity strengthened, by replication (Kully and Langevin 2005, Pring 2005).

The aim of this study was to replicate the work of Fry et al. (2009) and further investigate the efficacy of this therapy programme. Specific hypotheses tested were that teenagers attending this intensive therapy programme would demonstrate reduced severity of stammering and report reduced features of stammering as experienced in day-to-day life, reduced social anxiety and increased confidence about speaking.

Method

The setting

The study was conducted at a specialist, tertiary centre for stuttering that receives referrals for children and young people from speech and language therapists in the UK. Ethical approval for the study was granted by the Camden and Islington Community Research Ethics Committee.

Assessment

All adolescents aged 16 years and older complete a standard initial assessment consisting of a structured interview and a within-clinic analysis of stuttering severity using the SSI-3 (Riley 1994). Parents may request a parent session, but it is not required, and they may alternatively supply background information through a parent questionnaire. The aim of the assessment is to assess overt and covert aspects of stuttering, to understand the client's goals and preferences in terms of therapy, and to agree a course of action with the client.

Participants

All clients attending either one of two groups, each of ten young people, which took place within the 2-year time-frame of the study, were given information about the study and invited to take part. All those electing to take part were included.

The criteria for selection were that participants were aged between 16;0 and 19;0 years as this correlates with the age range of clients attending the course. The three participants reported on in this study could all be described as being in ‘late adolescence’, which is typically considered as being up to 19 years of age. All three were living at home, in full-time education and economically dependent, and from a lifespan perspective still preparing for emergent adult roles as is typical of adolescence. While falling in the older range of the clinical group as a whole, the convention of participants electing to take part in research, or not, may inevitably limit the degree to which specific ages can be targeted.

Selection criteria also included participants having been diagnosed as stuttering by a specialist speech and language therapist, having obtained a severity of stuttering rating on the Stuttering Severity Instrument—3 (SSI-3) (Riley 1994) of ‘mild stutter’ or higher and having been recommended for the intensive group based on clinical recommendations made independently of the study.

Five participants were ultimately recruited, three from the first year's intensive course and two from the second. Two withdrew during the baseline phase while continuing with the therapy programme, leaving two participants from the first year and one participant from the second. These participants were males aged 17;7, 17;11 and 18;10, monolingual, English speakers, in full-time education and living at home. All gave informed consent to take part.

  • Participant 1 (P1) was aged 17;11 at the start of the study. His parents reported that he started to stutter at 10 years of age and that there was no family history of stuttering. He had no additional, diagnosed, speech and language difficulties, although was described, informally, as having ‘some features’ that were consistent with Asperger's syndrome. These included a tendency to interpret information literally, a precise style of communication and a strong interest in a particular subject. Prior to the course he had received ‘intermittent’ speech and language therapy locally with no therapy in the 12 months before the course. On initial assessment P1 obtained an SSI-3 score of 27 and his stammer was described as moderate using the criteria of the SSI-3.
  • Participant 2 (P2) was aged 17;7 at the start of the study. His father reported that he started to stutter at approximately 2 years of age. There was a family history of both recovered and chronic stuttering paternally. P2 had no other speech and language diagnoses. Prior to the course he had received ‘intermittent’ individual therapy locally between the ages of 8 and 10 years and had also attended three intensive, group therapy courses since the age of 10 years. He had received no substantial or regular therapy in the 10 months before this study. On initial assessment he obtained an SSI-3 score of 46 and his stammer was described as very severe using the criteria of the SSI-3.
  • Participant 3 (P3) was aged 18;10 at the start of the study. He started to stutter at approximately 5 years of age and reported a family history of stuttering. P3 had no other speech and language difficulties and had received no speech and language therapy, although he had tried hypnotherapy for stuttering several years before attending the centre. On initial assessment he obtained an SSI-3 score of 9 and his stammer was described as very mild using the criteria of the SSI-3.

Description of the programme

The programme has three components: speech management skills, CBT and communication skills training. The speech management component integrates speech modification and fluency shaping skills. Speech modification focuses on identification, desensitization and learning how to modify the moment of stuttering. Therapy tasks include group discussions and reflection on personal stuttering behaviours, practice in identifying and describing moments of stuttering using visual feedback, practice in modifying stuttered moments through controlled reduction of tension, exploring the use of voluntary stuttering, and increasing openness about stuttering. Fluency shaping skills include rate reduction, use of light articulatory contacts and flowing words together. A structured programme is used to help clients gradually increase their confidence in using the skills that they personally find most effective and appealing.

The CBT component of the programme is based on the work of Beck (1976) and contemporary theory and treatment of social anxiety (Clark and Wells 1995). CBT is increasingly referred to in the treatment of stuttering and has been found to reduce social anxiety, although not overt stuttering, with adults who stutter (Menzies et al. 2008). Participants are introduced to a cognitive model that explains the links between thoughts, feelings, physiological responses and behaviours. They are helped to personalize the model by identifying key mechanisms that explain their own experience, and in particular to identify patterns of unhelpful pre- and post-event thinking and use of safety-seeking behaviours (Clark and Wells 1995). They are helped to develop cognitive restructuring skills, namely identifying and challenging negative automatic thoughts, to explore dropping their usual safety behaviours and increase their problem-solving skills in relation to real-life speaking situations.

The CBT component is integrated with all aspects of therapy. For example, participants are invited to check for negative automatic thoughts before any tasks that might trigger heightened anxiety, such as observing themselves on video-recorded tasks, giving a presentation to the group or completing fluency practice. The emphasis is on real-world application of the principles of CBT, using the group perspective whenever relevant to provide balanced feedback or alternative perspectives.

Rustin and Purser (1983) found that communication skills training augmented treatment gains and changes to communication style can be understood in terms of CBT theory. When stuttering is anticipated, people who stutter tend to use coping strategies, such as avoiding situations or speaking less, which have an unintended negative effect on their overall communication (Bloodstein 1995). For example, reducing eye contact, saying less, taking longer turns when ‘on a roll’ with fluency or scanning utterances for possible stuttering may result in the individual being less observant of his or her conversational partner, reading turn-taking cues less well, listening less effectively or appearing disengaged, which may in turn result in undesired listener responses. In addition, a focus on fluency as the communication ‘be all and end all’ may result in other aspects of communication being under-valued. A holistic awareness of communication is encouraged through discussion of specific skills, including observation and eye contact, listening, turn-taking, praise and self-reinforcement, problem-solving and negotiation. Video feedback is used to help clients identify their communication strengths as well as potential changes that they view as desirable. These are then reinforced through individual and group activities. The work is integrated with CBT, enabling participants to understand the links between anticipated stuttering and any associated fears, their behavioural reactions or coping strategies, and the often unintended impact these can have on interpersonal effectiveness.

The therapeutic style throughout the programme is collaborative, with the aim of encouraging clients to define their own goals and develop a repertoire of self-help skills that best fit with their personal goals, decisions and values. The group setting allows clients to share experience and support each other, as well as benefit from each others’ perspectives and observations. Participants are helped to reflect on the course and construct a personal action plan that summarizes key speech, communication and CBT skills, the rationale for their use, a schedule of practice activities and reflection, and guidelines for managing setbacks. This helps clients to continue generalizing their skills during the consolidation period that follows the intensive component of the programme. There are 4 follow-up days throughout the ensuing year. These are client-led, however they typically involve a refresher of all components of the course with an emphasis on practical work and problem-solving. Additional individual therapy sessions are arranged during the year if these are requested.

This programme is recommended for adolescents who have sufficient language and cognitive skills to enable them to access the course content, who are able to make the necessary commitment in terms of attendance, completing homework assignments and attending follow-up meetings, and who find the prospect of working in a group appealing.

Design

This was a single-subject experimental study that was replicated across participants. It had four phases:

  • Phase A consisted of a 5-week baseline during which participants did not receive therapy.
  • Phase B was a 2-week, intensive treatment phase with therapy taking place daily excluding weekends.
  • Phase C was a 5-week consolidation phase during which treatment was self-managed.
  • Phase D was a 10-month maintenance phase. Four 1-day, group follow-up sessions were held at intervals during phase D.

Measures

Percentage syllables stuttered

Participants video-recorded themselves at home speaking in conversation with another person for 5 min and reading aloud for 2–3 min. Recordings were made twice a week in phase A, each day in phase B, twice a week in phase C and once a month in phase D, yielding a possible 40 recordings. The frequency of stuttering was measured by calculating the percentage of stuttered syllables (%SS) for each speaking task. Recordings were coded and randomized so that first and second raters were blind to the time of recording. Recordings for P1 were analysed by a postgraduate, student speech and language therapist who had been trained in identifying and counting stuttered syllables. Recordings for P2 and P3 were analysed by a specialist therapist at the centre who is experienced in assessing stuttering, and who was independent of the treatment programme. Stuttering was defined as repetition of part-word or single-syllable whole words, prolongations of sounds, and speech blocks. Phrase repetitions and tension-free pauses were not counted as stuttered events.

Statistical analysis of %SS was carried out using a simplified time-series analysis (Blumberg 1984, Tryon 1982), in which Young's C-statistic is used to determine whether or not a time-series contains any trends, or systematic departures from random variation. It allows comparison of the baseline with subsequent phases, and enables comments to be made about the impact of treatment. The first step is to evaluate baseline data and determine whether or not the baseline contains a statistically significant trend. If this is ruled out, the C-statistic is then used to determine whether or not a trend exists across subsequent phases when these are aggregated. A significant result at this point is interpreted as evidence that the treatment series departs from the baseline series or, in other words, that a treatment effect is present. In this study, the baseline phase was analysed and then phases A–C were aggregated. Phase D was subsequently included in the analysis to determine the durability of any identified trends.

A total of 30% of each participant's recordings was randomly selected for blind analysis by a second rater who was a specialist speech and language therapist.

Intra-judge reliability was also examined on 25% of P1's recordings, on 33% of P2's recordings and on 30% of P3's recordings. Intra-judge reliability was calculated using the Smaller/Larger index or marginal agreement index (Frick and Semmel 1978), which can be used when there are two observers of a phenomena. The index is calculated by dividing the smaller of the two values of the occurrence of the target behaviour by the larger value, resulting in an index ranging from 0.00 to 1.00.

Self-report questionnaires

Participants completed four self-report measures. These were completed at the beginning of the baseline phase, on the first and last day of the intensive phase, at the end of the consolidation phase, and then at 3, 6 and 10 months post-treatment:

  • The Self-Efficacy Efficacy Scale for Adolescents (SEA-Scale) (Manning 1994) measures self-efficacy for verbal fluency, or confidence in being able to speak fluently, in a range of everyday situations. Both adolescents and adults who stutter have both been found to have decreased self-efficacy for verbal fluency in comparison to their fluent peers (Manning 1994, Ornstein and Manning 1985). It consists of 100 everyday speaking situations which are scored on a scale of 1 to 10 for the degree to which self-efficacy for verbal fluency is anticipated, where 1 is low and 10 is high. Manning (2001) reported that the SEA-Scale differentiated adolescents who stutter and those who do not, reporting a mean score of 7.21 (SD = 1.8) for adolescents who stutter and a mean score of 8.64 (SD = 1.2) for fluent adolescents.
  • The Fear of Negative Evaluation (FNE) Scale (Watson and Friend 1969) measures fear of negative evaluation by others. It is widely used as a measure of this aspect of social anxiety and has been shown to perform well as a specific measure of social evaluative concern (Stopa and Clark 2001). It consists of 30 true or false statements, with a higher score out of 30 indicating greater FNE. It was originally standardized on a population of American college students, however Stopa and Clark (2001) have established norms based on a population of British university students. They propose a cut-off point for low FNE at 7–8 and below, and a cut-off point for high FNE at 20–22 and above. It has been shown to have satisfactory test–retest reliability and is routinely used in therapy and research (Musa et al. 2004).
  • The Wright and Ayre Stuttering Self-Rating Profile (WASSP) (Wright and Ayre 2000) measures five aspects of stuttering: stuttering behaviours, thoughts about stuttering, feelings about stuttering, avoidance due to stuttering and disadvantage due to stuttering. It is standardized on individuals who stutter aged 18 and over, although is also widely used for older adolescents who are not yet 18, and has been shown to have good internal consistency and high reliability. Norms are not established, however, and interpretation of scores is based on visual analysis. The authors report ‘extremely good’ test–retest reliability when the WASSP was repeated at 1 week, followed by a 2-week interval, with Pearson's Product Moment Correlation above 0.8 (p < 0.005) for all subtests, indicating that this measure can be used at frequent intervals and change interpreted as therapeutic.
  • The Locus of Control of Behaviour Scale (LCB) (Craig et al. 1984) measures the extent to which adults perceive themselves as active agents in influencing their life experience and by implication their stutter. Greater internality indicates that the individual sees themselves as being an active agent in influencing their speech. Andrews and Craig (1998) found that a shift towards greater internality, together with increased knowledge about fluency management skills and having a more positive attitude about speaking, was associated with long-term maintenance of therapy gains with adults who stutter. However, the evidence for this is equivocal and DeNil and Kroll (1995) did not find an association between LOC and long-term therapy outcomes in adults. The LCB scale consists of 17 items arranged on a six-point Likert-type scale; lower scores represent greater internality and a downward direction in scores is desirable.

Results

Participation

All three participants attended the full 10 days of the intensive course. P1 attended the second and third follow-up meetings, with attendance at the first and last meetings prohibited by disruptions to air travel. He attended a refresher day held for any teenagers who had attended a therapy group at the centre in the previous 3 years, which took place between the third and fourth follow-up for his group. He did not ask for any additional individual sessions. P2 attended the first and third follow-up days and also attended a series of individual sessions with his local speech and language therapist between these. P3 attended all follow-up days and did not ask for any additional individual sessions at the centre.

Analysis of percentage syllables stuttered

Participant 1

P1 made all requested video-recordings, although number 38 in phase D could not be analysed because of poor recording quality. Inter-rater reliability was 0.980 for reading and 0.971 for conversation. Visual inspection of the data, which are shown in figure 1, shows a marked reduction in frequency of stuttering early on in phase B, with this remaining stable during phase C. There is increased stuttering during phase D, at recordings 31–33 and 39, which corresponds to 2–4 and 9 months after the intensive course, respectively; however, overall the frequency of stuttering appears to be less than prior to treatment.

Figure 1.

P1 percentage stuttered syllables during reading and conversation.

This is supported by statistical analysis. There was no significant trend in phase A for either reading (z = 0.96, p = n.s.) or conversation (z = 0.46, p = n.s.), meaning that any subsequent trend can be associated with the effect of therapy. A significant downward trend was found when phases A–C were aggregated in both reading (z = 4.00, p < 0.01) and conversation (z = 4.64, p < 0.01). A significant trend was also found when data from phase D were included for reading (z = 3.80, p < 0.01) and conversation (z = 5.00, p < 0.01). Analysis therefore indicated that the frequency of stuttering reduced in response to treatment and that this was maintained.

Participant 2

P2 submitted eight recordings for phase A, ten in phase B, of which one conversation sample could not be analysed due to poor quality, eight for phase C, and four in phase D. Inter-rater reliability was 0.957 for reading and 0.990 for conversation. Visual interpretation of the data, as shown in figure 2, and is less straightforward than with P1. A downward trend is suggested when phases A–C are examined, although clearly P2 has occasions of more marked stammering in conversational speech, notably at recordings 15, 18 and 33.

Figure 2.

P2 percentage stuttered syllables during reading and conversation.

Statistical analysis established that there was no significant trend during the baseline for either reading (z = 1.20, p = n.s.) or conversation (z = 1.32, p = n.s.). A significant downward trend was found for both reading (z = 2.32, p = < 0.01) and conversation (z = 1.80, p = < 0.05) when phases A–C were aggregated, indicating that the frequency of stuttering reduced in response to therapy during this time frame. P2 made four recordings in phase D which are included in figure 2 but were insufficient to allow an analysis of trend for this phase.

Participant 3

P3 completed 18 recordings throughout phases A–C, and two recordings in phase D. These are included in figure 3 but were insufficient to allow analysis of trend in this phase. P3 was atypical in that he did not stutter when reading, with both raters recording 0% stuttered syllables in each speech reading sample. Analysis of trend was therefore conducted on data from the conversation task alone. Inter-rater reliability was 0.978 and intra-rater reliability was 0.939. Visual inspection of the data, as shown in figure 3, suggests that while P3's stuttering continued to vary, the range reduced, and with one exception at recording 27, the%SS remained below 2.5% once he entered the therapy programme.

Figure 3.

P3 percentage stuttered syllables during conversation.

Statistical analysis established that there was no trend in the baseline phase (z = 0.01, p = n.s.). A significant downward trend (z = 2.77, p = < 0.01) was found when phases A–C were aggregated, indicating a response to treatment.

Self-report questionnaires

Participant 1

P1 scored below the mean for teenagers who stutter on the SEA-Scale before treatment. As can be seen in figure 4, his scores increased during the intensive course, continued to increase steadily and passed the mean for fluent adolescents 3 months after the course. P1 completed the FNE on six assessment occasions and submitted one incomplete form. As can be seen in table 1 his scores were above the cut-off point for high FNE range prior to the intensive course, however they fell to either below or near the cut-off point for low FNE during the next 10 months. There was a slight increase in FNE at 3 and 10 months although this was relatively slight. P1 completed all questionnaires for the WASSP. As can be seen in figure 5 his scores reduced sharply during the intensive course itself and continued to do so during the consolidation phase, remaining stable until 6 months after the course. There was an increase in scores at 6 and 10 months, however overall his WASSP scores remained lower following therapy than before. P1's LOC scores reduced during the intensive therapy phase, however, increased after this and showed no overall shift towards greater internality.

Table 1. Fear of Negative Evaluation (FNE) Scale scores for P1, P2 and P3
 5 weeksFirst dayLast day5 weeks3 months6 months10 months
Taskpre-courseof courseof coursepost-coursepost-coursepost-coursepost-course

Notes

  1. a

    Questionnaire incomplete.

  2. b

    Data not obtained.

P125a8812710
P213a1117bbb
P322241814151316
Figure 4.

SEA-Scale scores for P1, P2 and P3.

Figure 5.

WASSP scores for P1, P2 and P3.

Figure 6.

Locus of control of behaviour (LCB) scores for P1, P2 and P3.

Participant 2

P2 completed the SEA-Scale on the first four occasions only. He scored below the mean for teenagers who stutter before treatment in phase A and showed increased self-efficacy during the course itself, with his score at the end of the course approaching the mean for fluent adolescents. However P2's SEA-Scale score had reduced at the 5-week follow up and no further questionnaires were completed. P2 completed the FNE on three occasions, and one form was incomplete. His FNE scores were all in the mid-range and while there was a reduction during the course this was also not sustained. As before P2 completed the WASSP on the first four occasions only. His scores reduced sharply during the period of the course itself but this reduction was not sustained. P2's LOC scores dropped prior to starting the intensive course and increased during the course itself.

Participant 3

P3 completed all questionnaires. His SEA-scale score was initially below the mean for adolescents who stutter, but increased during the course itself. While his scores decreased after the group this was a short-term trend and his score approached the mean for fluent adolescents by the end of the study. P3 completed the FNE on each assessment occasion. His scores before treatment placed him at, or above, the cut-off point for high FNE, and while his scores did not reach the cut-off point for low FNE there was a steady decrease throughout the period of the study. P3 completed all questionnaires for the WASSP. These scores also reduced sharply during the intensive period and while there was an increase between the end of the course and the three month follow up, overall there appears to be a stable downward trend from 3 months post-course onward. P3's LOC scores increased prior to the intensive group, readjusted during the group and overall showed no apparent shift towards greater internality.

Conclusions and implications

The aim of this study was to investigate whether this therapy programme was effective in reducing overt and covert aspects of stuttering for the participants involved. Two participants withdrew from the study during the baseline phase, and while not required to give a reason, one volunteered that making the recordings increased his self-consciousness about stuttering. It is assumed that the other reconsidered the commitment involved in the research methodology. Both withdrawees continued with therapy and there was no evidence that their decision was linked to their engagement with the therapy process. Analysis on the recordings received for the participants who continued with the study was deemed to have good inter- and intra-rater reliability.

P1 provided the most complete data of the three participants and his data showed the most unequivocal, stable and durable reduction in overt stuttering behaviours and the most evidence of change in self-reported experience of stuttering. Both visual analysis and mean percentage stuttered syllables showed some occasions of increased stuttering during phase D, however this was not statistically significant in terms of overall trend. It interesting to note that the higher frequency of stuttering found at recordings 31–33 was accompanied by only a slight increase in FNE, while SEA-Scale and WASSP scores were unaffected. While P1's speech may have been somewhat less fluent at these points in time, it is possible that he gained broader skills from the therapy programme than speech management skills alone, and as a result was more psychologically resilient in dealing with short-term variations in his fluency. Recording 39 also showed a higher frequency of stuttering, however as it was the final recording this cannot be placed in context. Fluency is expected to fluctuate with chronic stuttering and may be influenced by environmental factors, and it is notable that this period coincides with the beginning of the spring academic year when pressures with coursework and mock examinations are often described as high by clients within secondary education. As found by DeNil and Kroll (1995) there is not an apparent correlation between P1's LOC scores and maintenance of fluency gains, which invites consideration of the usefulness of this assessment. Overall, given P1's fluency gains, increased sense of self-efficacy, reduced FNE and his reduced WASSP scores overall it can be concluded that this 1-year therapy and follow-up programme was effective for him. Of the three participants, P1 engaged with the routine of completing homework tasks and recordings the most consistently. His commitment to this process may have been a reflection of his overall commitment to therapy, which could be expected to be a contributing factor to the outcome of therapy.

It was particularly important that a stable baseline trend was established for P2 as he, in common with many young people who stutter, entered the study with extensive experience of therapy. He withdrew from the study half-way through phase D and disengaged from therapy, so progress beyond phase C is unknown and the interpretation of his results needs to reflect that. His stutter reduced significantly in frequency in the short-term and appeared to stabilize well during phase C, suggesting that an accumulation of skills and practise was helpful for him in the short-term. However, data obtained in phase D point to some destabilization of skills at the time that he withdrew from therapy. P2 experienced increased frequency of stuttering on occasion, as seen in figure 2 at recordings 15, 32 and 33. This coupled with data from the other participants suggests that fluency gains are particularly vulnerable approximately 6 months after intensive therapy, which could be an artefact of the length of time after therapy or related to the environmental demands experienced at that time in the academic year. P2 completed questionnaires on the first four occasions only. In all four domains measured his scores showed a positive shift during the intensive course itself, however this was not sustained once the group had finished. Overall, his results indicate that intensive group therapy in itself was beneficial in the short-term, but that he had difficulty maintaining improvements beyond the context of the group, inviting a consideration of why this might be, and whether this might be related to internal or external factors. Of the three, P2 had the most severe stutter in terms of stuttering behaviours, and he had also received the most therapy in the past, however his FNE and SEA-Scale scores suggested that his stutter had less impact on his view of himself and his confidence in speaking than it did for P1 and P3. It may be that his stutter, while severe, had less impact on his social interaction and was a less central concern for him than might be anticipated, and that this in turn impacted on the degree to which he engaged in therapy in the long-term. It is also possible that he needed more in terms of on-going follow up and support after the course, although this is available on request and he did not request further support.

The nature of P3's stutter differed in that he was fluent when reading and stuttered mildly in conversational speech. He had not had therapy before and his SSI-3 score might preclude him from some clinical programmes. However, while having an overtly mild stutter it is notable that his SEA-Scale score prior to treatment was the lowest of all three participants, while his FNE scores were high, suggesting that stuttering had a more substantial impact for him than could be ascertained by examining speech measures alone. This underlines the importance of basing clinical decisions, and of measuring change, in terms of client self-report as well as measures of stuttered syllables, and of having sufficient flexibility within a treatment programme to respond to individual needs. Given the reduction in overt stuttering behaviours, reduced FNE and increased self-efficacy and the overall direction of WASSP scores it seems reasonable to conclude that the year's treatment programme was effective for P3. P3 commented that he had learnt to slow down his rate of speech. While he had been aware of this technique before he had not internalized it as a concept. It may be that he was more ready for therapy, as he approached transition to university, or that the group context with its emphasis on personal responsibility for change helped him to explore this idea in a way that he had not done previously.

Both P1 and P3 both showed increases in stuttering approximately 6 months after the intensive component of the course and this is where P2 discontinued therapy. This corresponds to a time of increased coursework and examination demands within the UK academic year and for this reason additional follow-up days could be experienced as an unwanted pressure. It is possible that distance sessions using tele-health would be seen as a more positive option at this time as this would minimize disruption to college schedules while providing additional support if required.

One of the challenges in studying this age group has been a lack of robust self-report measures although this more recently been addressed as new measures specific to stuttering in adolescence have been developed. While initially developed with an adult population, the FNE is theoretically relevant and has apparent clinical value in charting change. Two out of the three participants reached the cut-off point for high FNE before therapy, which is consistent with studies finding high scores on measures of social anxiety amongst adults who stutter (Kraaimaat et al. 2002, Stein et al. 1996). Both participants then either reached, or approached, the cut-off point for low FNE after therapy, supporting the use of CBT and social anxiety measures in stuttering therapy. There was no clear association between locus of control and therapy outcomes in this study and the usefulness of the LCB could be questioned. Finally, the importance of measuring covert as well as overt features of stuttering is highlighted by P3, who had very mild outward stuttering but high anxiety about it.

The use of beyond-clinic measures demonstrated that a medium-term treatment effect occurred independently of the treatment setting for the participants studied. Limitations of this study include the need for more substantive long-term follow-up although the mobility of this age group and changing circumstances and demands of adolescents’ lives makes this challenging as demonstrated in the lack of complete data for P2 and P3. The authors are mindful of this limitation and the importance of interpreting the data with this in mind.

This study aimed to replicate a study described by Fry et al. (2009), which used the same measures and statistical analysis to investigate the impact of the same intensive therapy programme on one participant. Similarly to this study, Fry et al. found that the participant studied exhibited a trend of reduced frequency of stuttering across treatment phases and consolidation phases, consistently increasing SEA-Scale scores, as seen in this study with P1. There was a similar picture when examining LOC scores in that they moved in an unexpected direction, with reduction in scores prior to treatment and some increase in the B and C phases. The participant in this initial study did not complete the FNE and this could not be compared.

This therapy programme aims to equip clients with skills that will enable them to manage both overt and covert aspects of stammering more effectively. All three participants showed a significant reduction in overt stuttering behaviours in the short-term, replicating the findings of Fry et al. (2009). Participants who maintained contact further through the period of the study continued to show reduced stuttering and positive changes in those aspects of stuttering measured by self-report. On-going engagement with the therapy process may be a factor influencing long-term gains. While the interpretation of results needs to be cautious due to the term of data collection and the missing data for two participants, this study adds to a previous single-subject study showing evidence of the efficacy of this therapy programme. As more data become available, it becomes possible to examine the variability in responses to this therapy programme and to begin to hypothesize about that factors that may have an impact on outcome. With further consideration it is hoped that in future it may be possible to identify which young people are most likely to benefit from this therapy programme, and what alternatives or modifications should be considered for others.

Acknowledgements

The authors would like to acknowledge the contribution made by Stephanie Ryder, Speech and Language Therapist for analysis of data for P1 as part of her postgraduate degree, and the support of Whittington Health and Action for Stammering Children. We would also like to thank the participants for their efforts in contributing to the study. This study was partially supported by funding from the Rayne Foundation. The Rayne Foundation had no role in the design of the study, collection, analysis or interpretation of data, writing up or submission of this paper. Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Ancillary