Professional Documents
Culture Documents
research-article2018
CRE0010.1177/0269215518791274Clinical RehabilitationKleffelgaard et al.
CLINICAL
Article REHABILITATION
Clinical Rehabilitation
Abstract
Objective: To investigate the effects of group-based vestibular rehabilitation in patients with traumatic
brain injury.
Design: A single-blind randomized controlled trial.
Setting: University Hospital (recruitment and baseline assessments) and Metropolitan University
(experimental intervention).
Subjects: A total of 65 patients (45 women) with mild-to-moderate traumatic brain injury (mean age
39.4 ± 13.0 years) were randomly assigned to intervention (n = 33) or control group (n = 32).
Intervention: Group-based vestibular rehabilitation for eight weeks. Participants were tested at baseline
(3.5 ± 2.1 months after injury) and at two post-intervention follow-ups (2.7 ± 0.8 and 4.4 ± 1.0 months after
baseline testing).
Main measures: Primary outcome: Dizziness Handicap Inventory. Secondary outcome: High-Level
Mobility Assessment Tool. Other outcomes: Vertigo Symptom Scale; Rivermead Post-concussion
Symptoms Questionnaire; Hospital Anxiety and Depression Scale; and Balance Error Scoring System.
Between-group differences were analyzed with a linear mixed-model analysis for repeated measurements.
Results: At baseline, no group differences were revealed (personal factors, clinical characteristics
and outcome measures). At the first follow-up, statistically significant mean differences in favor of the
intervention were found in the primary (−8.7, 95% confidence interval (CI): −16.6 to −0.9) and secondary
outcomes (3.7 points, 95% CI: 1.4–6.0). At the second follow-up, no significant between-group differences
were found. No significant between-group differences in the other outcomes were found at the two
follow-ups.
Conclusion: The intervention appeared to speed up recovery for patients with dizziness and balance
problems after traumatic brain injury. However, the benefits had dissipated two months after the end of
the intervention.
Keywords
Dizziness, neurological rehabilitation, physical therapy, postural balance, traumatic brain injury
or substance abuse reported in their medical record, blocks of four, prepared by a statistician not
insufficient command of Norwegian and/or cogni- involved in the trial. The allocation sequence was
tive dysfunction (unable to follow instructions and/ concealed from the interventionist. An uninvolved
or fill in forms), fractures, or other comorbidities research assistant stored the list and prepared
affecting mobility and independent gait, and scores sequentially numbered, opaque, sealed envelopes.
<15 points on the Dizziness Handicap Inventory.14 The envelopes were kept in a locked drawer and
The cut-off point at 15 points was selected because a drawn consecutively after consented inclusion in
total score of ⩾16 indicates a handicap for patients the study. The envelopes were given to the partici-
with peripheral or central pathology.15 pants by the interventionist after the baseline
Physiatrists at the outpatient clinic of the assessments. The participants opened the sealed
Department of Physical Medicine and envelopes and were subsequently enrolled in the
Rehabilitation at Oslo University Hospital referred intervention or control group.
eligible patients to the interventionist, who per- A blinded outcome assessor (physical educator)
formed clinical assessments of the patients and carried out the follow-up assessments. The partici-
provided the patients with oral and written infor- pants and the interventionist were not blinded for
mation about participation in the study. Written the group allocation. An uninvolved research assis-
informed consent was obtained from all partici- tant entered the data into the statistical database,
pants. Information regarding personal factors, and the interventionist was blinded for group allo-
cause of injury and severity of injury were recorded cation during the statistical analyses.
from the patients’ medical records. Personal factors
included age at the time of the injury, married/
cohabiting (yes/no), level of education (cut-off set
Intervention
at >12 years), preinjury employment/studies (yes/ Both groups were offered the usual multidiscipli-
no), and preinjury comorbidities (yes/no). Severity nary outpatient rehabilitation comprising clinical
of injury included the Glasgow Coma Scale score, examinations by a physiatrist and assessments and
loss of consciousness (yes/no or not reported), follow-ups by a multidisciplinary team if needed.
post-traumatic amnesia (yes/no or not reported) The main focus of the multidisciplinary rehabilita-
and the presence of intracranial abnormality on tion was to help the patient return to daily life
magnetic resonance imaging scans/computer activities and return to work. The multidisciplinary
tomography (yes/no or not reported). outpatient rehabilitation is described in a publica-
The clinical assessments of the vestibular system tion by Vikane et al.16
at baseline consisted of tests of the oculomotor sys- Patients in the intervention group received a
tem (smooth pursuit, saccadic eye movements) and group-based vestibular rehabilitation intervention
the vestibulo-ocular reflex (head-thrust test, clinical twice weekly for eight weeks. Attending all 16 ses-
test for dynamic visual acuity). Benign paroxysmal sions was considered 100% adherence to the inter-
positional vertigo was tested with the Dix-Hallpike vention. The intervention is described in detail in
and Roll test. All of these tests were performed and another publication17 and consisted of guidance,
evaluated according to descriptions in Herdman individually tailored exercises, a home exercise
and Clendaniel.7 In addition, the neck was screened program, and an exercise diary. Two Physical
for painful active range of movement. Therapists experienced in vestibular and traumatic
The interventionist assessed all included partici- brain injury rehabilitation were responsible for the
pants prior to randomization. Participants were intervention. The intervention was mainly based on
randomly assigned (1:1 allocation) to receive a principles from motor control theory for improving
group-based vestibular rehabilitation intervention balance18 and theory of positive psychology for
in addition to the usual multidisciplinary traumatic coping with symptom pressure and disease bur-
brain injury rehabilitation. Allocation followed a den.19 Elements from established vestibular reha-
computer-generated list of random numbers in bilitation interventions were included and followed
Kleffelgaard et al. 77
differences at the two follow-ups. As fixed effects, the approximately two months after the end of the
statistical model included the baseline value of the intervention at a mean of 4.4 ± 1.0 months after
outcome as a covariate, the main effects of treatment baseline testing. At baseline, the participants pre-
and follow-up, the interaction term between treatment sented with a mix of vestibular, visual, and muscu-
and follow-up, and the interaction term between the loskeletal clinical findings/characteristics (Table 1).
baseline value of the outcome and follow-up. The self-reported outcome measures indicated
Assumptions for all statistical tests were assessed that the included participants had moderate dizzi-
with adequate descriptive statistics. Results are pre- ness-related disability and severe symptoms of
sented as mean differences with 95% confidence dizziness.14,24 The participants also reported a con-
intervals (CIs) and P values at each follow-up time siderable burden of post-concussion symptoms and
point. IBM SPSS Statistics for Windows (v. 23; psychological distress.13,26 The performance-based
IBM Corp., Armonk, NY, USA) was used for the scores for balance and mobility were below popula-
statistical analysis. All statistical tests were two- tion norms, indicating reduced balance when stand-
sided and assumed a 5% significance level. ing with eyes closed and reduced functioning and
Sample size and power calculations were based tempo in usual mobility tasks (Table 2).29,30
on the primary outcome measure from a methodo- No adverse events of the intervention were reg-
logical study performed in Norway.21 Estimations istered. One patient had a fall during the mobility
were based on 80% power (β = 0.8), an alpha (α) of testing (High-level Mobility Assesment Tool for
0.05, an SD (δ) of 15 and a clinically relevant traumatic brain injury). No injuries were registered
between-group difference of approximately 11 after the fall, but he was offered an assessment by a
points and a 10% loss to follow-up. The trial thus physiatrist. For the 33 patients in the intervention
needed to include 35 patients in each group, a total group, 22 patients (67%) had ⩾50% adherence and
of 70 patients. 12 of these had ⩾80% adherence. Two patients
(6%) did not attend at all, and the remaining 9
(27%) patients attended between one and seven
Results sessions.
A total of 65 patients were included in the study. One At the first post-intervention follow-up, a statis-
patient withdrew after randomization and did not tically significant between-group mean difference
consent to the use of baseline data, leaving 64 was found in the respective primary and secondary
included patients. One participant (2%) was lost at outcome measures in favor of the intervention
the first follow-up. A total of nine (14%) were lost at group. At the second post-intervention follow-up,
the second follow-up (Figure 1). There were no sig- the between-group differences were no longer sta-
nificant differences between the groups in frequency tistically significant (Table 3). The intervention
or causes of drop-outs. Characteristics of both group maintained their level of improvement,
groups are presented in Table 1. There were no while the control group improved over time in diz-
major differences in baseline characteristics between ziness-related disability and mobility problems
the groups in terms of personal factors, clinical char- (Table 2). No significant differences between the
acteristics, and outcome measures or in participation groups were found in any of the other outcomes at
in the multidisciplinary rehabilitation and/or the the two post-intervention follow-ups (Table 3).
psychoeducational groups (Table 1). There were no
major differences between the groups in any of the
Discussion
outcome measures at baseline (Table 2).
Baseline assessments were conducted 3.5 ± 2.1 The results from this randomized controlled trial
(mean ± SD) months after injury. The first follow- showed that an eight-week modified group-based
up assessments were conducted at a mean of vestibular rehabilitation intervention, in addition to
2.7 ± 0.8 months after the baseline testing. The sec- the usual multidisciplinary traumatic brain injury
ond follow-up assessments were conducted rehabilitation, improved dizziness-related disability
Kleffelgaard et al. 79
and mobility in favor of the intervention group at intervention group. No differences between the
the first post-intervention follow-up. The effect of groups were found on the severity of vertigo symp-
the intervention was maintained. However, after the toms, post-concussion symptoms, psychological
end of the intervention, the control group continued distress, and balance at the two follow-ups.
to improve and at the second post-intervention fol- The results from this study support findings from
low-up, they were at the same level as the other studies that indicate benefits of vestibular
80 Clinical Rehabilitation 33(1)
Table 1. Baseline characteristics of participants in personal factors, clinical characteristics, and attendance in
multidisciplinary rehabilitation per treatment group.
BPPV, benign paroxysmal positional vertigo; CT, computed tomography; DVAT, dynamic visual acuity test; GCS, Glasgow Coma
Scale; HIT, head impulse test; IQR, interquartile range; LOC, loss of consciousness; MRI, magnetic resonance imaging; PTA, post-
traumatic amnesia; SD, standard deviation.
Values are shown as numbers (percentages within group), medians (IQR), or mean ± SD.
rehabilitation after traumatic brain injury.8,9 The stability in relation to contextual factors” were sup-
result on the main outcome measure favoring the ported by the secondary outcome measure. In con-
intervention group at the first post-intervention fol- trast to the multidimensionality of the Dizziness
low-up indicated that the intervention speeded up Handicap Inventory, the High Mobility Assessment
the recovery and as such had a positive effect on Tool is a unidimensional performance-based meas-
dizziness-related disability. The findings reflect dif- ure of mobility and dynamic balance that captures
ferent factors of functioning covered in the Dizziness the physical dimension of motor performance rather
Handicap Inventory like “emotional functioning and than cognitive or behavioral dimension that may
participation in social life,” “specific activities/ limit motor performance.22 Thus, the intervention
movements that provoke dizziness and/or unsteadi- seemed to have a positive effect on both motor per-
ness,” and “self-perceived walking ability and the formance and the patients’ self-perception of their
feeling of postural stability in relation to contextual disability.
factors.”31 The reported findings in favor of the The beneficial effects of the vestibular rehabili-
intervention group regarding aspects of “self-per- tation intervention were maintained in this study
ceived walking ability and the feeling of postural although no between-group differences were found
Kleffelgaard et al. 81
Table 2. Scores on all outcome measures at baseline, first, and second follow-up per treatment group.
Baseline First follow-up Second follow-up Baseline First follow-up Second follow-up
n = 33 n = 32 n = 27 n = 31 n = 31 n = 28
DHI (0–100 47.9 ± 16.6 32.9 ± 21.3 32.1 ± 20.7 41.4 ± 19.2 36.4 ± 22.7 30.0 ± 24.3
best–worst)
HiMAT (0–54 40.9 ± 9.3 47.6 ± 7.0, n = 29 47.3 ± 8.2, n = 25 39.1 ± 11.9 41.2 ± 12.3, n = 23 44.3 ± 9.6, n = 26
worst–best)
VSSv (0–32 10.9 ± 6.0 6.7 ± 6.0 6.9 ± 4.6 10.2 ± 6.6 8.4 ± 6.6 6.1 ± 5.5
best–worst)
VSSa (0–28 8.3 ± 5.3 7.4 ± 5.7 6.6 ± 5.0 7.7 ± 5.7 6.6 ± 5.1 5.5 ± 5.2
best–worst)
RPQ3 (0–12 6.6 ± 2.8 4.9 ± 3.3 5.3 ± 3.3 5.5 ± 2.4 4.7 ± 2.7 3.6 ± 2.9
best–worst)
RPQ13 (0–52 25.6 ± 8.4 21.1 ± 11.3 20.4 ± 12.0 25.4 ± 10.2 19.6 ± 11.7 17.0 ± 10.9
best–worst)
HADSa (0–21 9.1 ± 4.7 7.9 ± 5.8 7.6 ± 5.0 8.4 ± 3.8 7.9 ± 4.6 6.9 ± 3.9
best–worst)
HADSd (0–21 7.1 ± 4.4 6.1 ± 4.3 6.7 ± 5.2 6.5 ± 5.1 6.2 ± 5.4 5.1 ± 4.8
best–worst)
BESS (0–60 29.7 ± 11.6 19.1 ± 10.6, n = 31 17.5 ± 10.4, n = 26 29.0 ± 9.6 23 ± 9.1, n = 26 20.8 ± 9.0, n = 28
best–worst)
BESS, Balance Error Scoring System; DHI, Dizziness Handicap Inventory; HiMAT, High-level Mobility Assessment Tool for trau-
matic brain injury; HADSa, Hospital Anxiety and Depression Scale, anxiety subscale; HADSd, HADS depression subscale; RPQ,
Rivermead Post-Concussion Symptoms Questionnaire; RPQ3, RPQ Physiological subscale; RPQ13, RPQ Psychological subscale;
SD, standard deviation; VSSv, Vertigo Symptom Scale, vertigo subscale; VSSa, VSS anxiety subscale.
Values are in mean ± SD.
Table 3. Mean difference for each outcome between the groups (intervention compared with control) at the first
and second post-intervention follow-ups controlled for baseline values using a linear mixed model for repeated
measurements.
BESS, Balance Error Scoring System; CI, confidence interval; DHI, Dizziness Handicap Inventory; HADSa, Hospital Anxiety and
Depression Scale, anxiety subscale; HADSd, HADS depression subscale; HiMAT, High-level Mobility Assessment Tool for trau-
matic brain injury; RPQ, Rivermead Post-Concussion Symptoms Questionnaire; RPQ3, RPQ Physiological subscale; RPQ13, RPQ
Psychological subscale; VSSa, VSS anxiety subscale; VSSv, Vertigo Symptom Scale, vertigo subscale.
rehabilitation, which may be beneficial after mild– and pain might be additional barriers.33
moderate traumatic brain injury.16 How to cope with Furthermore, post-concussion symptoms, psycho-
post-concussion symptoms and psychological dis- logical distress, and neck pain, which were reported
tress was specifically addressed by the multidiscipli- by the included participants, might have affected
nary rehabilitation team, and this may explain that the adherence to the intervention.
there were no difference between the groups in the
Post-Concussion Symptoms Questionnaire and the
Limitations
Hospital Anxiety and Depression Scale. Furthermore,
attendance in the psychoeducative group sessions Some limitations warrant caution when interpret-
supplying information on coping strategies and the ing the result of this study. Due to a slower than
importance of physical activity may have had a posi- expected recruitment rate, and limited time and
tive impact on the burden of symptoms and psycho- economic resources, the study stopped at 65
logical distress.16 The attendance in the usual included patients. Thus, the study included a rela-
multidisciplinary traumatic brain injury rehabilita- tively small number of patients with mild-to-mod-
tion and psychoeducative group sessions and the erate traumatic brain injury from a single hospital.
occurrence and treatment of benign paroxysmal Studies with an exercise intervention might appeal
positional vertigo was similar in both groups and did to healthier and more motivated individuals.
not compromise the internal validity of the study. Furthermore, the study recruited more women than
Adherence is important in clinical trials. In this men. This differs from epidemiological studies
study, 11 of 33 (33%) attended fewer than eight where the incidence of traumatic brain injury is
sessions. This relatively low adherence may have highest among male patients.1 Hence, the results
contributed to a smaller difference between the must be generalized with caution.
intervention and control groups. Several factors Another limitation is that vestibular diagnoses
may affect adherence after traumatic brain injury. were not determined by vestibular and computer-
Personal health, fatigue, and a lack of motivation ized assessments by ear–nose–throat specialists,
have been reported as important barriers to physi- but based on symptoms and clinical assessments
cal activity, and cognitive problems with memory made by a physiotherapist. Thus, the intervention
Kleffelgaard et al. 83
targeted deficits in function rather than specific traumatic brain injury rehabilitation team at Oslo
diagnoses. University Hospital and the colleagues at Oslo
Ascertainment bias was a potential threat to the Metropolitan University for collaboration and support
internal validity in this study because all partici- during the study. We would also like to thank Oslo
Metropolitan University for the funding of the study and
pants and the investigator were aware of the group
for providing necessary equipment and facilities for the
allocation. Ideally, the investigator should not be
intervention.
involved in the implementation of the intervention. Trial registration: Clinical Trials No. NCT01695577.
This could have affected the treatment and the fol-
low-ups in favor of the intervention group and thus Declaration of conflicting interests
have led to exaggerated effect estimates. However,
The author(s) declared no potential conflicts of interest
the outcome assessor was blinded to the group
with respect to the research, authorship, and/or publica-
allocation.
tion of this article.
The intervention group received additional
rehabilitation and attention from the researchers
Funding
that was not equivalent to the control group. Thus,
the improvements might also be attributed to the The author(s) received no financial support for the
research, authorship, and/or publication of this article.
increased rehabilitation time or increased attention
from the researchers and not just from the modified
ORCID iDs
group-based vestibular rehabilitation intervention.
Nevertheless, the results from this study suggest Ingerid Kleffelgaard https://orcid.org/0000-0002-
4994-7292
that vestibular rehabilitation is a safe and beneficial
Birgitta Langhammer https://orcid.org/0000-0002-
intervention, speeding up recovery for patients
9639-0570
with dizziness and balance problems after trau-
matic brain injury. The findings imply that vestibu- References
lar rehabilitation safely can be included as a part of
1. Peeters W, van den Brande R, Polinder S, et al.
multidisciplinary rehabilitation. Further studies are Epidemiology of traumatic brain injury in Europe. Acta
needed to secure outcomes that are more robust, to Neurochir 2015; 157: 1683–1696.
examine the impact of dose and progression of 2. Maskell F, Chiarelli P and Isles R. Dizziness after trau-
exercises and duration of the intervention period on matic brain injury: overview and measurement in the
clinical setting. Brain Inj 2006; 20: 293–305.
rehabilitation outcomes.
3. Akin FW, Murnane OD, Hall CD, et al. Vestibular conse-
quences of mild traumatic brain injury and blast exposure:
a review. Brain Inj 2017; 31: 1188–1194.
Clinical message 4. Arshad Q, Roberts RE, Ahmad H, et al. Patients with
•• An eight-week modified, group-based chronic dizziness following traumatic head injury typi-
cally have multiple diagnoses involving combined
vestibular rehabilitation intervention, in
peripheral and central vestibular dysfunction. Clin Neurol
addition to the usual multidisciplinary Neurosurg 2017; 155: 17–19.
traumatic brain injury rehabilitation, can 5. Fife TD and Kalra D. Persistent vertigo and dizziness after
speed up recovery in dizziness-related mild traumatic brain injury. Ann N Y Acad Sci 2015; 1343:
disability and mobility after a mild-to- 97–105.
6. Ellis MJ, Leddy JJ and Willer B. Physiological, vestibulo-
moderate traumatic brain injury
ocular and cervicogenic post-concussion disorders: an
evidence-based classification system with directions for
treatment. Brain Inj 2015; 29: 238–248.
Acknowledgements 7. Herdman SJ and Clendaniel RA. Vestibular rehabilita-
tion. 4th ed. Philadelphia, PA: F. A. Davis Company,
We would like to thank all the participants who took part 2014.
in the study. We thank the physiatrists Nada Andelic, 8. Schneider KJ, Meeuwisse WH, Nettel-Aguirre A, et al.
Torgeir Hellstrøm and Fredrik Walter Smith who took Cervicovestibular rehabilitation in sport-related concus-
part in the recruitment of participants to this study. sion: a randomised controlled trial. Br J Sports Med 2014;
Furthermore, we are grateful to the multidisciplinary 48: 1294–1298.
84 Clinical Rehabilitation 33(1)
9. Alsalaheen BA, Mucha A, Morris LO, et al. Vestibular cross-sectional and longitudinal designs. Health Qual Life
rehabilitation for dizziness and balance disorders after Outcomes 2009; 7: 101.
concussion. J Neurol Phys Ther 2010; 34: 87–93. 22. Williams G, Robertson V, Greenwood K, et al. The high-
10. Gottshall K. Vestibular rehabilitation after mild traumatic level mobility assessment tool (HiMAT) for traumatic
brain injury with vestibular pathology. Neurorehabilitation brain injury. Part 1: item generation. Brain Inj 2005; 19:
2011; 29: 167–171. 925–932.
11. Murray DA, Meldrum D and Lennon O. Can vestibular 23. Kleffelgaard I, Roe C, Sandvik L, et al. Measurement
rehabilitation exercises help patients with concussion? A properties of the high-level mobility assessment tool
systematic review of efficacy, prescription and progres- for mild traumatic brain injury. Physther 2013; 93:
sion patterns. Br J Sports Med 2017; 51: 442–451. 900–910.
12. Quatman-Yates C, Cupp A, Gunsch C, et al. Physical 24. Wilhelmsen K, Strand LI, Nordahl SH, et al. Psychometric
rehabilitation interventions for post-mTBI symptoms last- properties of the Vertigo symptom scale—short form.
ing greater than 2 weeks: systematic review. Phys Ther BMC Ear Nose Throat Disord 2008; 8: 2.
2016; 96: 1753–1763. 25. Eyres S, Carey A, Gilworth G, et al. Construct validity and
13. King NS, Crawford S, Wenden FJ, et al. The Rivermead reliability of the Rivermead Post-Concussion Symptoms
Post-Concussion Symptoms Questionnaire: a measure of Questionnaire. Clin Rehabil 2005; 19: 878–887.
symptoms commonly experienced after head injury and 26. Bjelland I, Dahl AA, Haug TT, et al. The validity of the
its reliability. J Neurol 1995; 242: 587–592. hospital anxiety and depression scale. An updated litera-
14. Jacobson GP and Newman CW. The development of the ture review. J Psychosom Res 2002; 52: 69–77.
dizziness handicap inventory. Arch Otolaryngol Head 27. Kleffelgaard I, Langhammer B, Sandhaug M, et al.
Neck Surg 1990; 116: 424–427. Measurement properties of the modified and total balance
15. Yorke A, Ward I, Vora S, et al. Measurement characteris- ERror SCoring system—the BESS, in a healthy adult
tics and clinical utility of the dizziness handicap inventory sample. Eur J Physiother 2018; 20: 25–31.
among individuals with vestibular disorders. Arch Phys 28. Zigmond AS and Snaith RP. The hospital anxiety and
Med Rehabil 2013; 94: 2313–2314. depression scale. Acta Psychiatr Scand 1983; 67: 361–
16. Vikane E, Hellstrom T, Roe C, et al. Multidisciplinary 370.
outpatient treatment in patients with mild traumatic brain 29. Iverson GL and Koehle MS. Normative data for the bal-
injury: a randomised controlled intervention study. Brain ance error scoring system in adults. Rehabil Res Pract
Inj 2017; 31: 475–484. 2013; 2013: 846418.
17. Kleffelgaard I, Soberg HL, Bruusgaard KA, et al. 30. Williams G, Rosie J, Denisenko S, et al. Normative values
Vestibular rehabilitation after traumatic brain injury: case for the high-level mobility assessment tool (HiMAT). Int
series. Phys Ther 2016; 96: 839–849. J Ther Rehabil 2009; 16: 370–374.
18. Shumway-Cook A and Woollacott MH. Motor control: 31. Kurre A, Bastiaenen CH, van Gool CJ, et al. Exploratory
translating research into clinical practice. 5th ed. Alphen factor analysis of the Dizziness Handicap Inventory
aan den Rijn: Wolters Kluwer, 2017. (German version). BMC Ear Nose Throat Disord 2010;
19. Bandura A. Self-efficacy: the exercise of control. New 10: 3.
York: W.H. Freeman, 1997. 32. Gurley JM, Hujsak BD and Kelly JL. Vestibular
20. Hilton MP and Pinder DK. The Epley (canalith reposition- rehabilitation following mild traumatic brain injury.
ing) manoeuvre for benign paroxysmal positional vertigo. Neurorehabilitation 2013; 32: 519–528.
Cochrane Database Syst Rev 2014; 12: CD003162. 33. Reavenall S and Blake H. Determinants of physical activ-
21. Tamber AL, Wilhelmsen KT and Strand LI. Measurement ity participation following traumatic brain injury. Int J
properties of the Dizziness Handicap Inventory by Ther Rehabil 2010; 17: 360–369.