The study protocol has been designed with reference to the CONSORT statement in order to ensure methodological validity [22
]. It is not possible to blind treating therapists nor patients to the treatment intervention so a single (assessor) blind, randomised controlled trial design will be used. It is also considered unethical to have a "no treatment" or a placebo control group as treatment for peripheral vestibular dysfunction has a robust evidence base [5
]. The aim of this study is to compare the outcome from conventional vestibular rehabilitation to virtual reality based vestibular rehabilitation in the treatment of peripheral vestibular impairment. A parallel design will be used. Figure shows the flow of diagram of the trial.
CONSORT (2010) flow diagram of the trial.
1. To compare the effect of conventional vestibular rehabilitation and virtual reality based rehabilitation on gait impairment.
2. To compare the effect of conventional vestibular rehabilitation and virtual reality based rehabilitation on balance.
3. To compare the effect of conventional vestibular rehabilitation and virtual reality based rehabilitation on dizziness and vertigo.
4. To compare the effect of conventional vestibular rehabilitation and virtual reality based vestibular rehabilitation on dynamic visual acuity.
5. To quantify and compare patient satisfaction and adherence with conventional and virtual reality based vestibular rehabilitation.
Participants and setting
Patients attending the otolaryngology and neurology outpatient clinics in two Dublin university teaching hospitals (Beaumont Hospital, Dublin and the Royal Victoria Eye and Ear Hospital, Dublin) for diagnosis and treatment of unilateral vestibular loss will be invited to participate in the trial. Participants will be provided with information leaflets and given a cooling off period to consider their involvement in the trial. Written informed consent will be obtained and patients will be free to withdraw from the study at any time. Non-consenting participants will be treated with usual care.
Participants will be included if they have a clinical diagnosis of peripheral vestibular dysfunction (confirmed where possible with vestibular function testing; canal paresis > 20%), and no other neurological deficit. Caloric testing is not available to all patients in the sites; where it is not available presence of a positive head thrust test, head shaking after nystagmus or spontaneous nystagmus (assessed with infrared oculomotor recording system) will be used to support the clinical diagnosis. Participants will also have one of the following subjective complaints indicating a failure of vestibular compensation; dysequilibrium, gait instability, vertigo/dizziness, or motion sensitivity.
Participants will be excluded if they have had previous vestibular rehabilitation or have bilateral peripheral vestibular pathology, CNS involvement, fluctuating symptoms Meniere's disease, migrainous vertigo), active benign paroxysmal positional vertigo, or other medical conditions in the acute phase (orthopaedic injury). As per Nintendo Wii® guidelines, participants will also be excluded if they have a pacemaker or epilepsy, or are unwilling or unable to use a NWFP. Pregnant patients will also be excluded.
A permuted blocked randomisation procedure will be used. A third party (HF) not involved in the day to day running of the trial will use an online randomisation program (http://www.randomization.com
) to assign patients to either conventional therapy or NWFP therapy in advance of recruitment. Block size will be chosen randomly from a block size of 4, 6 or 8 and block size will not be revealed to other study personnel. Group allocation will be performed by the third party and notified to the treating therapist by email after individual patients provide informed consent and undergo baseline assessments (by the blinded assessor), thus ensuring concealed allocation.
Consenting participants will complete baseline assessments prior to randomisation. They will be retested post treatment (at 8 weeks). The primary endpoint is gait velocity (m/sec) at 8 weeks post baseline.
Gait velocity will be assessed using three dimensional gait analysis (3DGA) in the Royal College of Surgeons in Ireland, School of Physiotherapy's Movement Laboratory. The laboratory has a Vicon 250 motion analysis system (Vicon Motion Systems Ltd., Oxford, UK) and a Kistler force plate (Kistler Instruments Ltd., Winterthur, Switzerland). Vicon consists of five infra red cameras that are able to localize reflective markers (placed on the participant) in a three-dimensional coordinate system within 1.5 mm error. Because of its high level of accuracy and reliability, it is considered a "gold standard" for motion analysis. The Kistler force plate, in synchronicity with the Vicon is able to compute kinetic forces during gait as well as centre of mass and centre of pressure. Gait velocity will be recorded during four conditions; self selected speed eyes open, self selected speed eyes closed (towards a target), cadence of 120 steps per minute eyes open (paced with a metronome), and walking with head turns.
Secondary gait variables will also be investigated; percentage of a gait cycle spent in double support, base of support width (cm), centre of mass/centre of pressure moment arm (mediolateral) and centre of mass velocity and displacement. These variables have been identified by Krebs et al (2003) [8
] as the most significant discriminators of success following vestibular rehabilitation The Dynamic Gait Index (DGI)[23
] will also be assessed.
Other secondary outcomes will be;
1. Computerised dynamic posturography (Equitest-NeuroCom) [24
]. This measures postural sway during six sensory conditions and returns sensory ratios that can be compared to normative data. The sensory organisation test (SOT) will be performed according to the protocol published by Equitest (Neurocom).
2. The Vestibular Rehabilitation Benefits Questionnaire (VRBQ) [25
] and the Activities Specific Balance Confidence Scale (ABC) [27
]. These are validated and reliable questionnaires for assessment of the effects of dizziness. Scores range from 0-100.
3. Dynamic Visual Acuity (DVA) [28
]. This will be measured using a computerised DVA test (Micromedical Technologies, Chatham, Illinois). The difference between static and dynamic visual acuity (at three head speeds) in LogMAR units will be measured using the protocol published by Micromedical Technologies.
4. The Hospital Anxiety and Depression Scale (HAD) [29
]. This validated scale has been used previously in studies in vestibular rehabilitation and assesses non-somatic symptoms of anxiety and depression. Scores range between 0 and 21, scores between 8 and 10 are considered borderline and those above ten indicate clinical depression or anxiety.
5. Usability of virtual reality technology will be measured by the System Usability Questionnaire [21
]. This is a 10 item questionnaire using Likert scales. Adherence will be recorded in the exercise booklets (and where allocated, on the NWFP) and satisfaction with treatment will be measured with a questionnaire developed by the researchers.
A six month follow up will be carried out. This will be confined to the primary outcome measure of gait velocity (m/sec), and secondary outcomes of computerised posturography, the VRBQ, ABC and DGI. Outcome measurement at each time point is shown in Table .
Time frame and outcome measures in the study
Sample size and power calculation
Based on previously published work the estimated sample sizes for each outcome variable have been calculated (Table ). The calculations were performed to detect what would be a clinically important difference between the means, with a significance level of 0.05 using the power calculation by Pocock [30
] given as;
Sample size and power calculations for the trial
Where σ is the standard deviation, μ is the mean difference expected between the two groups using a two sample t-test, α is the significance level (P < 0.05) and β is the lower (0.80) and f (α,β) is 7.9. Allowing for a 5% drop-out rate a minimum of 37 patients will be required pergroup. Current referral rates of patients with unilateral vestibular loss to physiotherapy approximate 40-50 participants per year over the two sites, thus a two year recruitment period is required.
Following baseline assessments (by the blinded assessor), participants will be randomised (as described above) to one of two groups;
1. Conventional vestibular rehabilitation
2. Virtual reality vestibular rehabilitation
Based on best available evidence [5
] 6 treatments will be provided over an 8 week timeframe. Treatment will only be provided by physiotherapists who are at senior level and have experience in vestibular rehabilitation and all treating physiotherapists will attend periodic protocol training sessions. On their first attendance participants will undergo a standardised physiotherapy assessment which will elucidate their main problems. The interventions for both groups have been developed and will be based on six identified core elements of vestibular rehabilitation used in current clinical practice- education, relaxation, adaptation exercises, habituation exercises, balance and gait retraining and re-conditioning [33
]. Programs are customised to symptoms and impairments, and are progressive. All participants will be asked to perform a home exercise program daily for 30-40 minutes. Each participant will receive a standardised weekly booklet detailing their exercises and containing an adherence diary. They will return the booklets at each treatment session and receive a new one. The booklets have been designed to look similar. Each is tripartite containing adaptation exercises, balance exercises and a walking program. The booklets, provided to both groups differ only in the balance exercises, the conventional group receiving conventional balance exercises [2
] and the NWFP group receiving exercises developed in pilot work [20
]. The NWFP records type, duration and frequency of exercises and this will be used as an additional record in the NWFP group.
Participants in the virtual reality group will be instructed in the use of the NWFP® and will be given one on loan for use at home. Participants in the conventional group will be provided with a standard therapeutic foam balance mat (Sissel Balance Fit Pad). All patients will be reviewed a minimum of 4 times over 8 weeks for re-assessment, progression of exercises and advice by the treating therapist.
To address the risk of falling, participants will be provided with the necessary written and verbal instructions to prevent a fall (e.g. support nearby). Any adverse events will be reported to the participants consultant physician and an adverse event form filled out. Participants experiencing adverse events will be followed up and assessed by their treating physiotherapist and physician.
Data will be coded and entered in Microsoft Excel and statistical analysis will be performed using PASW 18 (SPSS Inc. Chicago) and Stata Statistical Software (Release 11, StataCorp LP, TX:). Intention to treat analysis and per protocol analyses will be performed [34
]. Baseline data will be examined for comparability. Where there are missing data, the last observation will be carried forward. Data will be summarised using means, standard deviations and 95% confidence intervals for continuous variables, median and interquartile ranges for non-normal continuous or ordinal data and percentages for categorical data. Per-protocol analyses will be performed excluding patients with major deviations from the treatment protocol (defined as less than 50% adherence with treatment). Data will be examined for normality and t-tests used for data with two time points. A mixed ANOVA model will be used for data with three time points with the factors of group (NWFP or Conventional Group) and time (baseline, 8 weeks, six months). Within and between group differences and interaction effects will be calculated and post hoc analyses performed where indicated. Comparison of outcome measures between groups will be performed controlling of the baseline distribution of known predictor variables and the results will be reported as differences in means/medians and their confidence intervals [32
]. The non-parametric equivalent will be used where data are not normally distributed or are non-parametric. A significance level of p < 0.05 will be set. Effect sizes for within and between groups comparison will also be calculated.
Ethical approval has been obtained from the Beaumont Hospital (Ref 10/75) and the Royal Victoria Eye and Ear Hospital medical ethics committees (Ref 24/10/2011). Participants will receive a verbal explanation of the study from the principle investigator (DM), will be given an opportunity to ask questions and will receive an information sheet. A cooling off period of a time designated by the participant will be given. Written informed consent will be obtained. Participants will be informed that they may leave the study at any time without a need to give an explanation. The study will be carried out in accordance with the Declaration of Helsinki Guidelines.