Stage 1: development
Theory selection.
The first step in the identification of a theory and a theoretical model was a comprehensive review of the literature of explanatory processes and models of MUS in general [9]. This process identified a general explanatory model of functional symptoms as having predisposing, precipitating and perpetuating factors. The perpetuating factors included physical, behavioural, cognitive, affective and social components. This expanded CBT model of MUS [9] formed the theoretical basis of the current study of FD patients.
Model building
Next, the literature on predisposing, precipitating and perpetuating factors in functional dysphonia was reviewed in the context of the expanded CBT model of MUS. Evidence was found for predisposing vulnerabilities (gender, personality, occupation, other MUS); precipitating triggers (life events, viruses) and perpetuating factors (dysregulation of the laryngeal and paralaryngeal muscles, anxiety and depression) (for full review see [1].
Model development—patient interviews
Eight patients with functional dysphonia were interviewed by one member of the team (VD, a cognitive behavioural therapist) using an assessment structure based on the expanded CBT model identified above. This was used to elicit as wide a range of predisposing, precipitating and perpetuating factors of patients’ problems as possible. Individualised formulations, describing the interaction of multiple factors in the onset and maintenance of functional dysphonia and its associated distress, were worked out in session in collaboration with patients.
Initial development and delivery of the intervention.
Common factors, themes and patterns of interaction identified from the patient interviews and formulations as being important in the causation and maintenance of functional dysphonia were used alongside the theory and modelling data to develop an individually adaptable, expanded CBT model of functional dysphonia. This formed the basis of the speech and language therapist training package. This training was initially piloted in a small consecutive cohort study [18].
Refinement of the functional dysphonia model
Two notable further insights were gained through clinical supervision during this small cohort study and through discussion with speech and language therapists. Functional dysphonia patients routinely reported exhaustion, with disturbed patterns of activity, rest and sleep. They also tended to report perfectionist tendencies. This leads to a questionnaire-based case-control study into these aspects of dysphonia, which confirmed that this group was significantly more fatigued and perfectionist than matched normal controls [19]. This further shaped the CBT model and training.
Assessing and maximising the likelihood of intervention uptake
The prior consecutive cohort study [18] provided evidence that, for at least one speech and language therapist, delivering the CBT was feasible and acceptable. To further tailor the training for the generic and specific context in which it was to be implemented, VD carried out interviews with individual therapists, clinical teams and a voice special interest group. This process gave an understanding of how presenting problems other than voice issues (particularly anxiety, depression and personality problems) were routinely dealt with by SLTs in their management of functional dysphonia patients. Next, the SLT to be trained to deliver the CBT intervention (TM) was interviewed, to establish training needs, learning style and normal practice in the management of functional dysphonia and to establish a protocol for how emotional issues were to be dealt with in the usual care randomised controlled trial (RCT) arm.
Stage 2: feasibility—an external pilot randomised controlled trial
As a preliminary, a further set of patient interviews were conducted with greater attention to factors highlighted in the initial intervention development stages. These interviews confirmed the findings on perfectionism, fatigue, disordered activity, rest and sleep. Furthermore, the interviews supported the feasibility of addressing these with patients through developing a shared multi-factorial understanding of their condition. The insights gained from the above stages were used to further refine the model, the training of the therapist and the design and conduct of the pilot RCT. The objectives of this trial were as follows:
Acceptability and feasibility objectives
-
To assess the feasibility and acceptability of procedures and methods for trial participant identification, recruitment and data collection.
-
To assess the feasibility and acceptability of the CBT intervention to a SLT by evaluating their ability to integrate CBT training into clinical practice and by evaluating the amount and nature of supervision required to embed the CB intervention in usual care.
-
To assess the fidelity of delivery, acceptability and clinical utility of a CBT intervention to functional dysphonia patients.
-
To test the sensitivity to change of a selection of candidate outcome measures
These objectives were measured and monitored through recruitment and retention rates (patient acceptability of intervention and trial procedures), through observation of the training process (therapist acceptability) and through monitoring the process and content of the CBT participants’ treatment, in clinical supervision, case recordings, case notes and case summaries (fidelity, therapist and patient acceptability and feasibility, clinical utility). The main outcomes were the feasibility and acceptability estimates, assessed as described above. In addition, measures were taken of voice (Voice Performance Questionnaire, Carding et al. 1999), general health (General Health Questionnaire, Goldberg and Williams 1988) and psychological distress (Hospital Anxiety and Depression Scale Zigmond and Snaith, 1983, at baseline, at discharge from treatment (usually 6–8 sessions happening every other week) and 6 months after the end of treatment. These outcomes were assessed with regard to their acceptability and responsiveness to change, to assist in powering a future trial and will be the subject of a separate paper.
Design
The trial was a single-centre external pilot, patient randomised controlled trial with two arms: standard voice therapy versus voice therapy plus CBT, both delivered by a single SLT (TM), experienced in treating functional dysphonia.
The study was conducted at the Speech Voice and Swallowing Clinic of the Freeman Hospital, Newcastle upon Tyne, UK, between October 2007 and August 2010. Participants who remained in treatment or follow-up after this point continued to receive fully supervised treatment. Ethical permission was sought and obtained from Newcastle and North Tyneside Research Ethics Committee 1 (ethics reference number: 07/H0906/118). The trial was registered with the ISRCTN under the title: Training a Speech and Language Therapist in Cognitive Behavioural Therapy to treat Functional Dysphonia - A Randomised Controlled Trial. Trial Identifier: ISRCTN20582523.
Inclusion/exclusion criteria
Study participants were patients who had been referred to the Speech and Voice Clinic, for assessment of their dysphonia. Patients were screened by endoscopy, which excluded the presence of an injury, a lesion or a movement disorder in the patient’s voice box. Patients who had been thus diagnosed as having functional dysphonia were approached regarding entry into the trial. They were given participant information sheets describing the study in detail and at least 24 h to consider participation.
For the sake of generalisability, inclusion criteria were as broad as possible. Patients were considered eligible for randomisation if they were aged 18 or over and presented with an alteration or loss of voice where there was no evidence of a non-functional reason for vocal impairment (other than vocal nodules), a score of ≥ 1 on the overall Grade component of the clinician-rated voice quality Grade Roughness Breathiness Asthenia Strain Scale (GRBAS) [20] and a score of ≥ 20 on the self-rated Vocal Performance Questionnaire (VPQ) [21] (a self-report measure of voice quality and voice related disability rated 12–60 with 12 being normal). Patients were excluded from the trial if they had any of the following: previous experience of CBT for their voice problem; an acute or ongoing serious medical illness or severe mental health problem which was likely to interfere with their ability to comprehend, engage and/or comply with treatment; a learning disability; and a mild vocal condition which did not merit a full course of treatment.
Interventions—usual care: voice therapy
The control condition of “usual care” aimed to be as close to standard voice therapy practice as possible. Patients were offered an average of six to eight sessions every 2 weeks of approximately 1 h of voice therapy, although length and number of sessions were allowed to vary as needed. The content typically had the following elements: vocal hygiene and education (such as maintaining adequate vocal hydration); elimination of voice misuse and abuse (such as excessive throat clearing or shouting); breath control and coordination with phonation; and in-session and between-session exercises to promote vocal flexibility and resonance. When emotional issues arose in the course of therapy sessions, the SLT employed non-directive counselling skills, whereby patients were encouraged to speak about difficult issues, reflecting TM’s normal pre-CBT practice.
Interventions—the CBT intervention
In the experimental CBT arm, in addition to the standard voice therapy, patients also received the following CBT elements. As with usual care, treatment sessions lasted approximately 1 h.
Assessment and formulation
The CBT assessment identified the predisposing, precipitating and perpetuating factors. It was derived from the CBT model of functional dysphonia previously described. This information formed the basis of a formulation that attempted to explain how current factors might be interacting to maintain both poor voice and general distress, and how these had developed through the interaction of predisposing and precipitating factors. This formulation formed the basis for both treatment delivery (by TM) and treatment supervision (by VD). The ability to reach an agreed formulation with the patient was also a key measure of the acceptability of CBT for the patient group.
Treatment techniques
As each patient had an individualised formulation, no two treatments were identical, but they typically consisted of a mixture of the following treatment techniques. For low energy and low mood, patients were advised gradually to do more, in a structured planned way, and gradually to resume activities that used to be done for enjoyment and achievement. These evidence-based methods [22] are relatively simple and hence easily transmissible from trainer to therapist and from therapist to patient. Graded exposure was used to address anxiety-based issues [23]. People who are anxious tend to avoid what they are anxious of (and thus become more anxious) encouraging people to gradually confront difficult situations in a planned and structured way and at their own pace is the best evidence-based treatment [23]. This work also incorporated simple cognitive techniques, such as helping the patient identify what kind of anxious thoughts they might be having about avoided situations and helping them to test out the reality of these thoughts by confronting the situation in a safe, planned manner. Cognitive work was thus conceptualised as being an adjunct and aide to behavioural change. In addition, the therapist was trained in specific cognitive techniques for the negative aspects of perfectionism such as very high self-standards and self-criticism [24]. Other common unhelpful beliefs concerned the best way to manage voice and other physical symptoms, with patients often interpreting symptoms as harmful and as a cue to stop activity and to socially withdraw, thus keeping going a cycle of physical dis-use, low energy and low mood. Cognitive techniques, such as guided questioning during therapy sessions and thought diaries in between sessions, helped patients to identify their unhelpful beliefs and test them out by looking for evidence both for and against them.
These assessment and treatment techniques were taught to TM over a total of 7 days, over a 2-week period, by VD, with extensive use of skills rehearsal and supported by a full training manual. Patient implementation was supported with a patients’ manual, outlining each of the above approaches, with examples.
Monitoring and supervision
All treatment sessions were recorded using digital audio equipment. All patients in the CBT arm were reviewed weekly using a CBT supervision framework where the individualised formulation formed the basis for case discussions between TM and VD. TM’s management of those randomised to usual care (voice therapy alone) was supervised as usual by PC (another experienced speech and language therapist). To monitor for the intrusion of CBT techniques into usual care, TM identified patients in this cohort where distress was an issue and VD listened to recorded sessions and asked her to describe her treatment approach. Detailed clinical notes were also kept by TM of session content for both arms of the trial. These provided content for supervision and allowed the performance of fidelity, feasibility and acceptability analyses of the intervention for both participants and therapist.
Sample size
As this was an external pilot RCT, the primary determinant of the sample size was pragmatic, i.e. the throughput of the unit over the period of the trial. No formal power calculation was used but the guideline of Lancaster et al. [25] was followed which suggests that in an external pilot trial, 30 patients per arm may be sufficient to estimate the parameters of interest for a larger trial. Factoring in estimates of likely non-suitability, we anticipated being able to recruit this number over 18 months. We continued to recruit up to this latter time-point, resulting in 74 patients randomised to the study.
Randomisation, allocation concealment and blinding
Once a patient consented, they were randomly allocated either to voice therapy alone or voice therapy plus CBT. An independent researcher at Newcastle Clinical Trials Unit prepared a randomisation list, using permuted blocks of random sizes to reduce risk of breach of concealment of allocation. Sequentially numbered, opaque envelopes were prepared and provided to the clinic. The randomisation envelope was opened by the speech and language therapist in the presence of the patient following receipt of written consent. Patient details, date and time of randomisation were recorded, to maintain an audit trail of randomisation. Neither participant nor the therapist (TM), who was also responsible for data collection, nor the researcher responsible for data analysis (VD) was blind to treatment allocation.
Data analysis and use
Data were analysed in the following stages. Rates of recruitment, retention and protocol violations were analysed with respect to the feasibility and acceptability of the interventions from patient and therapist standpoints. Success of randomisation was assessed by informally comparing both groups’ baseline characteristics. A tally of problem formulations, treatment targets and treatment techniques was done in the CBT group to assess the feasibility of transmitting CBT techniques to an SLT, and the acceptability of CBT techniques to patients. A tally was also done of the content of any voice therapy only sessions where emotional issues formed a significant part of the work.
Full trial protocol is available [26].