Neuroscience and Biobehavioral Reviews 47 (2014) 445–456

Contents lists available at ScienceDirect

Neuroscience and Biobehavioral Reviews

journal homepage: www.elsevier.com/locate/neubiorev

Review

Dance as an intervention for people with Parkinson’s disease:

A systematic review and meta-analysis
Kathryn Sharp ∗ , Jonathan Hewitt
Department of Geriatric Medicine, School of Medicine, Cardiff University, Academic Centre, Llandough Hospital, Penarth CW64 2XX, United Kingdom

a r t i c l e i n f o a b s t r a c t

Article history: Recent studies suggest dance may be able to improve motor and non-motor disabilities in Parkinson’s dis-
Received 6 May 2014 ease patients. A systematic review and meta-analysis of randomised controlled trials (RCT’s) regarding the
Received in revised form 3 September 2014 effectiveness of dance compared with no intervention and other exercise interventions was performed.
Accepted 9 September 2014
Five trials were included and methodological quality and mean or standardised mean differences were
Available online 28 September 2014
calculated. Dance significantly improved UPDRS motor scores (−10.73, CI −15.05 to −6.16; P = 0.004),
berg balance (0.72, CI 0.31 to 1.44; P = 0.0006) and gait speed (0.14 m/s CI 0.02 to 0.26; P = 0.02) when
Keywords:
compared with no intervention. When compared with other exercise interventions significant improve-
Parkinson’s disease
Dance
ments in berg balance (3.98, CI 1.52 to 6.44, P = 0.002) and quality of life (PDQ-39) (−4.00, CI −7.13 to
Exercise −0.87, P = 0.01) were found. Dance demonstrates short term clinically meaningful benefits in Parkinson’s
Meta-analysis disease. Future RCT’s should be well designed and determine the long term effects of dance, which dose
Systematic review and type of dance is most effective and how dance compares to other exercise therapies.
© 2014 Elsevier Ltd. All rights reserved.

Contents

1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446
2. Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446
2.1. Search methods for identification of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446
3. Data collection and analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
3.1. Selection of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
3.2. Data extraction and management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
3.3. Assessment of methodological quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
3.4. Data synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
4. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
4.1. Description of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
4.2. Included studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
4.3. Description of the included studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
4.4. Data available for analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
4.5. UPDRS MOTOR score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
4.5.1. Dance vs no intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
4.6. Dance vs exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
4.7. Berg balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
4.7.1. Dance vs no intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
4.8. Dance vs exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
4.9. Freezing of gait . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
4.9.1. Dance vs no intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
4.10. Dance vs exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449

∗ Corresponding author. Permanent address: Lily Farm, Garneddwen CH8 8JS, Rhes-y-cae, Flintshire, United Kingdom. Tel.: +44 7779592627.
E-mail addresses: katielsharp100@hotmail.com (K. Sharp), HewittJ2@cardiff.ac.uk (J. Hewitt).

http://dx.doi.org/10.1016/j.neubiorev.2014.09.009
0149-7634/© 2014 Elsevier Ltd. All rights reserved.
446 K. Sharp, J. Hewitt / Neuroscience and Biobehavioral Reviews 47 (2014) 445–456

4.11. Velocity (m/s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450

4.11.1. Dance vs no intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
4.12. Dance vs exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
4.13. 6MWT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
4.13.1. Dance vs no intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
4.14. PDQ-39 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
4.14.1. Dance vs no intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
4.14.2. Attrition and safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
5. Risk of bias in included studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
6. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451
7. Quality of evidence and the limitations of our data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452
8. Implications for practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
Appendix A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
Appendix B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455

1. Background Each dance form has different qualities and researchers have
hypothesised that certain qualities will target specific PD symp-
Parkinson’s disease (PD) is a chronic, progressive and disabling toms. Tango requires frequent movement initiation and cessation,
neurodegenerative disorder, with wide reaching implications for spontaneous directional changes and movement speeds, which
the people it affects. It impacts on physical, psychological, emo- may target movement initiation, turning and bradykinesia. In con-
tional, social and financial function and consequently has a trast ballet challenges strength and flexibility to emphasise posture,
profound effect on quality of life. body alignment, projection of eye focus and limb extension, as well
Historically physiotherapy and occupational therapy rehabili- as whole body coordination (Houston and McGill, 2012). Dance
tation models have utilised compensatory strategies as a basis for can also promote a range of cueing strategies. These may be audi-
therapeutic management. Emerging evidence however, suggests tory, when music directly initiates or facilitates movement; visual,
that rehabilitative exercise interventions can have a neuro- where an individual coordinates steps with a partner or uses a
protective effect and cause the brain to repair itself through neuro partner’s foot to step over; or cognitive, where imagery is used to
plasticity (Hirsh and Farley, 2009). In addition, animal PD stud- visualise sequences of movement.
ies have demonstrated behavioural recovery and an increase in A recent Cochrane review (Tomlinson et al., 2012a) included 2
dopamine synthesis and release, as a result of exercise (Tilerson dance studies in a large review of the effectiveness of physiother-
et al., 2003). apy interventions when compared with no intervention in patients
Research has suggested that there are five key principles of exer- with PD. Their results suggested dance improves UPDRS motor
cise and activity that enhance neuro plasticity in relation to PD: (a) scores (−8.48, CI −12.76 to −4.19), freezing of gait (FOG) (2.21, CI
intensive activity maximises synaptic plasticity; (b) complex activ- −4.63 to 0.22) and the six minute walk test (6MWT) (38.94, −3.18
ities promote greater structural adaptation; (c) rewarding activities to 81.06) when compared with no intervention. Via meta-analysis
increase dopamine levels and promote learning/relearning; (d) they indirectly compared dance RCT’s with other physiotherapy
dopaminergic neurons are highly responsive to exercise and inac- intervention RCT’s and concluded that there were no differences in
tivity; (e) introduced early in the disease process, progression can the effectiveness of the physiotherapy interventions.
be slowed (Fox et al., 2008). This systematic review and meta-analysis will directly evaluate
Exercise interventions are only beneficial if performed regu- the effectiveness of dance in comparison to other exercise inter-
larly. Studies usually report sustained gains from exercise sessions ventions and no intervention.
held 2–3 times per week over a 4–12 week period (Goodwin
et al., 2008; Keus et al., 2007). Unfortunately compliance and reg- 2. Method
ular participation is often insufficient to achieve or maintain any
improvements (Heiberger et al., 2011a) or lifestyle changes. The Randomised controlled trials comparing any form of dance with
“Parkfit” trial (van Nimwegen et al., 2013) delivered a multifaceted either no intervention or other exercise intervention were consid-
behavioural change programme incorporating motivational strate- ered for inclusion in the study
gies and ambulatory feedback with physiotherapy and exercise,
to a group of people with PD (n = 299) however, overall it did not 2.1. Search methods for identification of studies
increase overall physical activity. People with PD have previously
been shown to be 29% less active than controls (van Nimweggen A systematic search of the literature was undertaken up until
et al., 2011) therefore effective interventions that are enjoyable and the end of January 2014, without date or language limitations.
promote regular participation need to be offered. It is suggested This included searching electronic bibliographic databases, the
that dance maybe one such intervention that could target more Cochrane library, dissertation and theses databases, grey literature
sedentary people with PD as well as those who are more active. databases, relevant conference abstracts and proceedings, hand
Dance is a choreographed routine of movements usually per- searching relevant journals and searching the reference lists of
formed to music (Hui et al., 2009). It is a multi-dimensional activity retrieved papers. The full list of our reference search material can
offering auditory, visual and sensory stimulation, musical expe- be found in Appendix 1. We searched on the following base terms;
rience, social interaction, memory, motor learning and emotional Parkinson*, Danc* and Random*. The detailed search history is
perception, expression and interaction (Kattenstroth et al., 2010). given in Appendix 2.
Studies have demonstrated that people with PD are motivated to Our aim was to select well designed RCTs, hence we assumed
attend dance classes regularly, have a high rate of compliance with that their chosen endpoints would be appropriate. Therefore we
a low dropout rate and often continue with the activity after the did not specify exact criteria for outcome measures in order
study period (Hackney and Earhart, 2009a,b). to prevent limitation of potentially suitable RCTs. Within this
 K. Sharp, J. Hewitt / Neuroscience and Biobehavioral Reviews 47 (2014) 445–456 447

framework we anticipated our outcome measures would consider

motor disability, physical function, gait, balance and quality of
life.
Participants of any age or gender with a diagnosis (as defined
by the study author’s) of PD were included in the analysis regard-
less of duration of PD, type of drug or surgical therapy, duration of
treatment or level of impairment

3. Data collection and analysis

3.1. Selection of studies

Both authors independently completed the literature search,

removed duplicates and screened abstracts from the poten-
tially relevant studies. Full text articles were obtained, where
available. Where details of potentially eligible studies and con-
ference abstracts were unclear, attempts were made to contact
authors.

3.2. Data extraction and management

Both authors independently assessed each identified paper and

abstract for trial details and outcome data. The trial details were
recorded on a trial description form and included trial name (author
and year of publication), study design, blinding, use of intent to
treat analysis, number of group participants and characteristics,
eligibility criteria, intervention schedule (including type of inter-
vention, number of sessions and duration), outcome measures used
and assessment time points (baseline, 3 months and final outcome).
A data collection form was created to record the outcome data of
each trial.

3.3. Assessment of methodological quality

The Cochrane Collaboration’s risk of bias assessment tool

(Higgins and Green, 2011) was used to determine the method-
ological quality of each trial included in the review. The studies
were assessed for methodological quality in respect of method
of randomisation, allocation concealment (selection bias), blind-
ing of participants and personnel (performance bias), blinding Fig. 1. Study PRISMA diagram.
of subjective and objective outcome assessment (detection bias)
similarity of groups at baseline, co-interventions constant, and 4. Results
reporting of all outcome measures and variability (reporting
bias). 4.1. Description of studies

3.4. Data synthesis After the removal of duplicates, the search strategy identified 44
references for inclusion in the review. Of the references identified,
The primary analysis was a comparison of dance interventions the titles and abstracts were reviewed and 28 records were elimi-
versus no intervention (control) or exercise intervention. nated from the review. The reasons for exclusion are shown in the
Outcome variables were continuous data presented as means, PRISMA diagram (Fig. 1). For the remaining 16 references, full text
standard deviations (SD) and/or standard errors (SE). If standard articles were obtained.
deviations weren’t provided these were calculated for meta- A review of the references of these 16 articles revealed the
analyses purposes. Where a particular trial’s groups received potential eligibility of a further 5 articles, which were obtained but
similar interventions (i.e. two different types of dance) the found not to be eligible. A further 6 trials were then excluded (see
means and standard deviations of their outcome measures were Table 1). Therefore, 8 trials (10 articles) were identified as eligible
combined. for inclusion in the qualitative synthesis (see Fig. 1: PRISMA flow
The mean differences (MD) with 95% confidence intervals were diagram). Of these, three were ongoing trials (Duncan and Earhart,
calculated for all variables using the same outcome measure. Where 2013; Postuma, 2012; Shanhan, 2012) (see Table 2) and were not
trials used different outcome measures that were deemed compa- included in the meta-analyses, which ultimately included 5 trials.
rable, standardised mean differences (SMD) with 95% confidence
intervals were calculated (Deeks et al., 2002). The fixed effects 4.2. Included studies
inverse variance model was used throughout this study. Hetero-
geneity was assessed using the I2 value. Review Manager Software, There were two trials (Hackney and Earhart, 2009a,b; Duncan
version 5.2 was used for the analysis. and Earhart, 2012; Foster et al., 2013) comparing a dance
448 K. Sharp, J. Hewitt / Neuroscience and Biobehavioral Reviews 47 (2014) 445–456

Table 1 paper (Hackney and Earhart, 2009b). For the purposes of this review
Characteristics of excluded studies.
and meta-analysis the Tango and Foxtrot intervention groups’
Study ID Reason for exclusion results were combined, as these are both types of partnered dance
Hackney and Earhart (2010) Not an RCT, no control group. and the aim of this review was to determine the effects of dance
Observational study comparing two rather than compare types of dance. Both classes were also led by
types of dance the same instructor which may have controlled for distinctions
Heiberger et al. (2011) Uncontrolled observational pilot study, in teaching pedagogy (style of instruction). The tai chi arm was
Houston and McGill (2012) Mixed methods observational study,
excluded from the dance versus no intervention comparison as tai
no control group
Murray et al. (2010) Uncontrolled observational pilot study chi is considered to be a form of exercise (British Council for Chinese
Westbrook and McKibben (1989) Convenience sample of existing Martial Arts, 2013) and the no intervention arm was excluded in the
dance/exercise group users dance versus exercise comparison.
Lee et al. (2012) Further information regarding
The mean age of the participants in the dance versus no inter-
intervention was sought but was not
received. Hence we were unable to vention trials was 68 years, 66% were male, the mean Hoehn and
determine if intervention was dance or Yahr stage was 2.3, the mean baseline UPDRS motor score was 34.9
Chinese martial art and the mean disease duration was 7 years. The number of drop
outs before first assessment ranged from 10 to 13.
The trials in the dance versus exercise comparison were either
intervention with no intervention and four trials (Hackney and
two arm (Hackney et al., 2007a,b; Volpe et al., 2013) or three arm
Earhart, 2009a,b; Hackney et al., 2007a,b; Volpe et al., 2013) com-
(Hackney and Earhart, 2009a,b) parallel trials with the exercise
paring a dance intervention with another exercise intervention.
groups incorporating sitting and standing, breathing, stretching,
Their characteristics are summarised in Table 1.
strength and balance and dexterity exercises.
Outcome measures were obtained for the Unified Parkinson’s
The characteristics of all trials included in this review are sum-
Disease Rating Scale (UPDRS) (Fahn and Elton, 1987) motor scores,
marised in Table 3.
the Berg Balance Scale (BBS) (Berg et al., 1992), Freezing of
One trial (Hackney et al., 2007a) did not provide details of the
Gait Questionnaire (FOGQ) (Giladi et al., 2000), velocity (m/s)
characteristics of their sample. For the remaining participants the
(Tomlinson, 2012b), six minute walk test (6MWT) (Enright, 2003)
mean age of the participants was 66.8 years, 67% were men, the
and quality of life, assessed using the Parkinson’s Disease Question-
mean Hoehn and Yahr stage was 2.1, the mean baseline UPDRS
naire 39 (PDQ 39) (Jenkinson et al., 1997).
motor score was 26.7 and the mean disease duration 7 years.
The number of drop outs before first assessment ranged from 0
4.3. Description of the included studies
to 13.

One of the dance versus no intervention trials was presented

in two papers (Hackney and Earhart, 2009a,b). This was a multi- 4.4. Data available for analysis
arm trial, with two experimental intervention groups (Tango and
Foxtrot) in the first paper (Hackney and Earhart, 2009a) and three Adequate data was available for meta-analysis in all the included
experimental groups (Tango, Foxtrot and Tai Chi) in the second trials (Hackney and Earhart, 2009a,b; Duncan and Earhart, 2012;

Table 2
Characteristics of ongoing studies.

Study ID Earhart et al. (2013) Postuma (2012) Shanhan (2012)

Study name Exercise and Parkinson’s: comparing Tango for Treatment of Motor and A Randomized Controlled Feasibility
interventions and exploring neural Non-motor Manifestations in Trial to Determine the Effectiveness of
mechanisms Parkinson’s Disease Set Dancing for People With
Parkinson’s Disease
Methods RCT, single blind RCT, open label masking RCT–Investigator and assessors blinded
Participants 120 with PD diagnosis PD diagnosis PD diagnosis
Interventions Tango, treadmill training or stretching Tango, exercise leaflet Irish dancing, usual care
Intervention durations 24 h/12 wks 24 h/12 weeks 12 h/8 weeks
Outcome measures Walking velocity UPDRS Berg Balance Scale
Mini Best Total “off” time (h) UPDRS motor score
PDQ-39 Presence of Dyskinesia PDQ-39
UPDRS motor score Mini-best 6MWT
Falls Questionnaire from Canadian Zarit Care Giver Burden Interview
Longitudinal study of Aging
The Purdue Pegboard
FOGQ
Montreal Cognitive Assessment
Beck Depression Inventory
Apathy Evaluation Scale
Krupp Fatigue Severity Scale
PDQ-39
Adherence to treatment
Exit Questionnaire
Clinical Global Impression of Change
Adverse Events
Contact info Ryan P Duncan, PT, DPT;Gammon M Ronald Postuma -McGill University Joanne Shanahan, University of
Earhart, PhD, PT—Washington Health Centre Limerick
University School of Medicine
Start date February 2013 April 2012 December 2012
Estimated completion date June 2016 June 2013 June 2013
Expected publication date Not stated Not stated Not stated
 K. Sharp, J. Hewitt / Neuroscience and Biobehavioral Reviews 47 (2014) 445–456 449

Foster et al., 2013; Hackney et al., 2007a,b; Volpe et al., 2013). Some

Baseline and 6
of the results for two of the trials (Duncan and Earhart, 2012; Foster

(Week 10-13)

(Week 10-13)

(Week 10-13)
final outcome

final outcome

final outcome

final outcome
Base line and

Baseline, 3, 6
Baseline and

Baseline and

(12 months)
months and
Assessment
et al., 2013; Hackney et al., 2007a) were only published in graph

months
form. Requests for further information from were successful for one
points
of the trials (Duncan and Earhart, 2012; Foster et al., 2013) hence
medication included in the meta-analysis.
Assessed

4.5. UPDRS MOTOR score

on/off

Off
On

On

On

On
4.5.1. Dance vs no intervention
Intervention

104 h/52 wk
We studied 100 participants from two studies (Hackney and

39 h/26 wk
20 h/13 wk

20 h/13 wk

20 h/13 wk
Duration

Earhart, 2009a,b; Duncan and Earhart, 2012; Foster et al., 2013).

The results are given after 3 months of follow up and at the end of
the intervention. Both results were positive in favour of the inter-
vention, without heterogeneity; 3 months (−6.4 points, CI −10.73
Length of each

to −2.07; P = 0.004) and final (−10.6 points, CI −15.05 to −6.16;

intervention

P = 0.00001 (see Fig. 2).

1 h/twice

1 h/twice

1 h/twice
session

weekly

weekly

weekly

1.5 h
1h

4.6. Dance vs exercise

Dance (Tango)/Exercise

Dance (Tango)/Exercise

For the dance versus exercise comparison there were 43 par-

dance/physiotherapy
Tango)/exercise (Tai
chi)/no intervention
Dance (Foxtrot and

ticipants from 2 studies (Hackney et al., 2007b; Volpe et al., 2013).

Dance (Tango)/no

There was no difference between the two intervention arms (−1.39

Intervention

intervention

points, CI −3.56 to −0.78; P = 0.21). There was evidence of hetero-

exercise

geneity (P = 0.01, I2 = 83%) (see Fig. 3).

Irish

4.7. Berg balance

Mean Baseline
UPDRS motor

4.7.1. Dance vs no intervention

Not stated

Data were available for comparison for 100 participants from 2

score

29.3

46.3

24.3

trials (Hackney and Earhart, 2009a,b; Duncan and Earhart, 2012;

27

Foster et al., 2013). The outcome measures were combined using

standardised mean difference. When comparing the trials after 3
Not stated

months of intervention (0.5, CI 0.09 to 0.91, P = 0.02) and by final

% Male

outcomes (0.72, CI 0.31 to 1.14, P = 0.0006) balance was significantly

63

74

56

54

improved in the groups receiving dance. There was no evidence of

heterogeneity (see Fig. 2).
with PD (yrs)
Mean time

Not stated

4.8. Dance vs exercise

8.95
4.9

7.7

6.1

Berg balance data from 2 trials (Hackney et al., 2007b; Volpe

et al., 2013) and 43 participants were available for the dance
Mean H&Y

Not stated

versus exercise comparison. Berg balance scores were significantly

improved with dance when compared to exercise (3.98, CI 1.52 to
2.2

2.1

2.6

2.2

6.44, P = 0.002) (see Fig. 3).

Mean age (yrs)

4.9. Freezing of gait

Not stated

4.9.1. Dance vs no intervention

66.6

69.9

63.3
71

Data from the freezing of gait questionnaire were available in

100 participants from two trials (Hackney and Earhart, 2009a,b;
Duncan and Earhart, 2012; Foster et al., 2013). There was no dif-
No. randomised (% of
drop outs before 1st

ference between the intervention arms at 3 months (0, CI −2.56

to 2.56, P = 1.00) or at the final outcome (−0.22, CI −2.82 to 2.38,
assessment)

P = 0.87) (see Fig. 2).

75 (18.6)
Key characteristics of studies.

62 (16)
19 (0)

19 (0)

24 (0)

4.10. Dance vs exercise

Earhart (2012)
Earhart (2009)

Final outcome data were available from two trials (Hackney

Hackney et al.

Hackney et al.

Hackney and

Duncan and

Volpe et al.

et al., 2007b; Volpe et al., 2013) for the dance versus exercise
Study ID

(2007b)
(2007a)

comparison, involving 43 participants. The results favoured dance

(2013)
Table 3

(−0.52, CI −1.86–0.82, P = 0.45); however, the reduction in FOG

scores was not significant (see Fig. 3).
450 K. Sharp, J. Hewitt / Neuroscience and Biobehavioral Reviews 47 (2014) 445–456

Fig. 2. Dance versus no intervention.

4.11. Velocity (m/s) to the dance versus no intervention meta-analysis of the 6MWT.
Dance was not superior for these heterogeneous results which are
4.11.1. Dance vs no intervention given in Fig. 2.
Data on velocity were available from 2 trials (Hackney and
Earhart, 2009a,b; Duncan and Earhart, 2012; Foster et al., 2013) for 4.14. PDQ-39
100 participants in the dance versus no intervention analysis. The
results at three months and at the final results for the dance versus 4.14.1. Dance vs no intervention
no intervention analysis were (0.07 m/s CI −0.05 to 0.19, P = 0.25) Two studies (Hackney and Earhart, 2009a; Volpe et al., 2013)
and (0.14 m/s CI 0.02 to 0.26, P = 0.02), respectively. The improve- contributed 68 participants. There was an improvement in PDQ-39
ment at 3 months was not significant however the improvement at scores in favour of dance (−4.00, CI −7.13 to −0.87, P = 0.01) (see
final outcome was significant (see Fig. 2). Fig. 3).

4.12. Dance vs exercise 4.14.2. Attrition and safety

Two trials (Hackney and Earhart, 2009a,b; Duncan and Earhart,
In the dance versus exercise analysis there were 2 trials 2012; Foster et al., 2013) reported dropout rates at three months
(Hackney et al., 2007a; Hackney et al., 2007b) with 38 participants (17% and 16%, respectively) at 3 months and one at one year (44%)
where data on velocity was available. There was no significant dif- (Duncan and Earhart, 2012; Foster et al., 2013). The reasons for
ference between the dance and exercise interventions (−0.03 m/s, dropping out varied but did not suggest it was due to problems
CU −0.08 to 0.02, P = 0.28). There was no evidence of heterogeneity related to the intervention.
(P = 1.00, I2 = 0%) (see Fig. 3).
5. Risk of bias in included studies
4.13. 6MWT
The methodological quality of the studies varied and trial
4.13.1. Dance vs no intervention reports did not always provide sufficient information to determine
Two studies (Hackney and Earhart, 2009a,b; Duncan and the risk of bias. Their methodological qualities are given in Fig. 4.
Earhart, 2012; Foster et al., 2013) contributed 100 participants It is not possible to blind participants and instructors/therapists to
 K. Sharp, J. Hewitt / Neuroscience and Biobehavioral Reviews 47 (2014) 445–456 451

Fig. 3. Dance versus exercise.

group allocation therefore all the included trials are subject to per-
formance bias. To mitigate a blinded assessor was used and deemed
low risk. The performance bias was addressed further in one trial
(Hackney and Earhart, 2009a,b) as participants were blinded to
study hypotheses. The primary bias relating to each of the studies
was a lack of description of the drug therapy participants received.
Further, it was not clear whether trials required stable medica-
tion regimes or whether variation was permitted. No trial detailed
whether participants were in receipt of other possible confound-
ing interventions or activities i.e. other movement based therapies
or PD education classes. Withdrawals were not described in two
studies (Hackney and Earhart, 2009a; Volpe et al., 2013).

6. Discussion

The aim of this systematic review and meta-analysis was to

Fig. 4. Risk of bias graph: review authors’ judgements about each risk of bias item investigate the effectiveness of dance as an intervention for people
presented as percentages across all included studies. with PD in comparison to either no intervention or other exercise
452 K. Sharp, J. Hewitt / Neuroscience and Biobehavioral Reviews 47 (2014) 445–456

interventions. We found that for a large number of our outcome Gait speed is an important predictor for a decline in over-
measures dance therapy showed good results compared to either all health and survival (Studentski et al., 2011). After 3 months
no intervention or other exercised based interventions, with simi- of dance intervention, comfortable walking pace velocity did not
lar completion rates. The outcome measures assessed were all well improve but had improved at the final outcome. This suggests that
validated measures and the large effect sizes recorded are likely a 3 month dose of exercise may not be sufficient to significantly
to reflect meaningful clinical improvement. Dance also compared affect outcomes whereas a much longer intervention period does.
favourably when compared to exercise in relation to balance and These improvements are less than the minimal clinically important
an improved quality of life. These improvements were obtained change in velocity of 0.18 m/s (Steffen and Senney, 2008) reported,
for an intervention which appeared safe and achieved good although the continued improvement over time is consistent with
compliance. the findings of a systematic review of physiotherapy versus no
Emerging evidence suggests that basal ganglia structures are intervention or placebo in PD (Tomlinson et al., 2012a).
involved in the control of dance movements. It has been noted that The lack of effect that we demonstrated in the 6MWT is in
there was increased activity in the basal ganglia and specifically the contrast to increases that have been demonstrated with treadmill
putamen via positron emission tomography, when tango move- training (Mirelman et al., 2001). Treadmill training forces people
ments were performed to a rhythmical beat (British Council for to user faster gait cycles and velocities whilst also increasing the
Chinese Martial Arts, 2013). Physical activity has also been found number of repetitions practiced (Merholz et al., 2010). It specifi-
to affect the regulation of neurotransmitters by enhancing the con- cally trains participants to cover a greater distance whereas dance
centration of serotonin, resulting in PD sufferers generally “feeling provides a more holistic intervention incorporating strength, bal-
better” (Heiberger et al., 2011b). The improvements seen from ance, aerobic and flexibility exercises. Superior outcomes have also
dance in our meta-analysis may reflect these pathological changes been achieved with specific task orientated training (Kwakkel et al.,
in practice. 2007); therefore, dance is unlikely to provide the distance required
Our results indicated that clinician rated (UPDRS motor scores) to influence this outcome.
disability was improved at 3 months (6.4) with further improve- Two of our outcome measures favoured dance over exercise
ments at the conclusion of the intervention (10.6). Estimates and none favoured exercise. In addition to balance which we have
suggest that a change of between 2.3 and 5 represents clinically already discussed, quality of life was improved for dance partici-
meaningful improvement (Schrag et al., 2006; Schulman et al., pants. This is further supported by comments from the trials we
2010) and 10.8 to be a large clinically important difference. assessed. For example, a number of dance participants continued
These results contrast the findings of a meta-analysis of the dancing after the trials ended whereas none continued with exer-
effects of music based movement therapy in patients with PD (de cise (Hackney and Earhart, 2009a; Hackney et al., 2007a,b). It was
Dreu et al., 2012). It found that interventions such as dance and also reported that those in the exercise group joined dance classes
walking programmes to music, did not significantly improve UPDRS after the trial ended (Hackney et al., 2007b).
motor scores. However, others have suggested that rehabilitative
therapies have a significant affect at an activity level as patients
learn to use compensatory strategies but they don’t affect overall 7. Quality of evidence and the limitations of our data
impairment, as measured by UPDRS motor scores (Kwakkel et al.,
2007). Our studies utilised different types and regimes of dance. Tango
Gait disorders and balance impairments are one of the most was widely used but not exclusively. It is likely that this dance
incapacitating symptoms of PD (Boonstra et al., 2008). With most style was chosen deliberately as it is believed to be well suited
falls in PD attributable to balance disorders (Bloem et al., 2004; Kerr to the symptoms encountered in PD (Houston and McGill, 2012).
et al., 2010), the need to identify interventions that can reduce the We would recommend that future studies compare different forms
risk of falls and subsequently the physical, psychological and finan- of dance to determine whether Tango itself conveys benefits over
cial costs of falls, is high on patient, carer and health care decision other forms of dance.
makers agenda’s. Our results showed an improvement in balance The improved final outcome scores imply that with a longer
at 3 months which was sustained at the end of the intervention. intervention period significantly greater clinical improvements in
These results are comparable to the findings of a meta-analysis UPDRS motor scores, balance and gait speed can be achieved. Oth-
into the effects of exercise on balance related activities (Allen et al., ers have also suggested that more superior results can be achieved
2011). They indicate that highly challenging balance training, such by people with PD when participating in dance classes over a longer
as dance, is superior to other exercise interventions in improv- duration (Earhart, 2009). There is however a lack of trials investi-
ing functional balance activities. Dancing requires participants to gating the longer term effects of dance and need to understand the
use postural control muscles to stabilise the body before muscles effects of “chronic” exercise has been highlighted (de Dreu et al.,
responsible for executing movement are activated. The patterns 2012).
of starts, stops, side steps and backwards walking, challenge the As none of our trials reported any follow up data from cessation
body’s ability to anticipate and respond to the different tasks and of the intervention it is not possible to comment on whether the
environmental constraints (Howe, 2011). Strength, posture and an benefits we demonstrated are continued long term. Future trials
ability to initiate corrective stepping strategies when balance is should incorporate a follow up period of at least 6 months into
challenged are essential to maintaining balance (Maki and McIiroy, their study designs but be mindful of the greater attrition rates
2005). with increased study lengths in this population (Keus et al., 2007;
Freezing of gait is a prevalent motor disturbance of PD that is Deane et al., 2001).
most commonly experienced in the advanced stages of the dis- The methodological quality of the included trials needs to be
ease (Bloem et al., 2004). There were no improvements seen in considered when interpreting these results. Of the 5 trials included
our review with FOG. This was unexpected as it was thought the in the review, only 2 provided evidence of a suitable random
auditory cues from the music and visual cues from the dance steps sequence generation method, however they did not go on to evi-
would assist patients to overcome FOG episodes as cueing training dence their procedures to conceal treatment group allocation. The
has previously been found to be effective (Nieuwboer et al., 2007). other trials stated only that “participants were randomised”. Four
However, our participants had low Hoehn and Yahr scores and a trials used blinded assessors, 1 trial blinded participants to the
low prevalence of FOG at baseline, is suggestive of early disease. study hypotheses and 1 used intent to treat analysis.
 K. Sharp, J. Hewitt / Neuroscience and Biobehavioral Reviews 47 (2014) 445–456 453

Whilst 49% of the categories in the assessment of bias were viii. PsychINFO (1806–2013),
deemed low risk, 44% were of unclear risk as there was not suf- ix. Cochrane Central Register of Controlled Trials (last searched
ficient evidence in the trial reporting to make a clear assessment of Feb 2013),
the risk of bias. x. Web of Knowledge (Last Searched Feb 2013),
All trials were relatively small with only one (Duncan and xi. OTseeker (Last searched Feb 2013),
Earhart, 2012; Foster et al., 2013) including sample size calcula- xii. Physiotherapy Evidence Database (PEDro) (1929–2013),
tions in their methodology. Trials need to be adequately powered xiii. SpeechBITE (last Searched Feb 2013),
to reduce the risk of random error and consequently false posi- xiv. PsychBITE (last searched Feb 2013),
tive results and future trials should be larger than those already xv. Rehabdata (1995–2010).
conducted (Brok et al., 2008).
Further, it should be noted that none of the included trials con-
trolled for the attention and socialisation group interventions can The following English Language databases of foreign language
deliver. It is possible that the additional clinic visits, clinical assess- research and third world publications were also searched;
ments, contact with clinicians and researchers and the anticipation
of change may all influence outcomes in both the intervention and
control groups. i. LILACS (1982–2013),
The reporting of eligibility criteria was also variable amongst ii. IMEMR (1984–2013),
the included trials. PD diagnostic criteria were clearly defined in iii. MedCarib (17th century–2013).
three of the five trials (Hackney and Earhart, 2009a; Hackney et al.,
2007a,b) meaning the possible inclusion of patients with PD plus
syndromes was minimised. The baseline characteristics (sex, age, Dissertation and Grey literature data bases searched included:
Hoehn and Yahr stage, years with PD medications taken) of the
groups need to be described in detail, to enable an assessment to
be made of which PD patients the results are applicable. Baseline i. DART—Europe,
characteristics were sufficiently reported in three trials (Hackney ii. Electronic (Last searched Feb 2013),
and Earhart, 2009a; Volpe et al., 2013). iii. Electronic Theses online service (16th century-2010),
The risk of publication bias is low as we are not aware of any iv. Networked digital library of Theses (1900–2013).
completed unpublished trials. There are currently three ongoing
trials; see Table 2.
The trials in both analyses administered interventions of vary- Clinical trials databases were searched for ongoing and pro-
ing lengths. Further, the trials included used a range of different posed trials. The searches included:
outcomes and outcome measures to determine the effects of their
interventions. This limited the pooling of results. There needs to
be a consensus on a core set of outcome measures and length of i. The Cochrane controlled trials register, The Centre watch clinical
intervention, to enable the magnitude of effects reported in RCT’s trials listing service
to be combined and/or compared more consistently. ii. Clinical trials.gov
iii. National institute for health research
iv. Trials journal.com
8. Implications for practice

Dance is a feasible intervention for people with PD. Significant We also hand searched the conference abstracts from the Inter-
improvements were noted in the UPDRS motor scores and balance national Congress of Parkinson’s Disease and Movement Disorders
measures after 3 months of dance classes when compared with no (1990–2012) and reviewed extracted reference lists and contacted
intervention and with velocity using final outcomes. The long term trial authors for further information.
effects of these improvements are not yet known.
This review suggests that people with PD can benefit from
dance interventions however further evidence is required to con-
firm these results ideally from larger well conducted randomised Appendix B.
controlled studies. Further trials are also required to determine the
long term effects of dance, which dose and type of dance is most Search history:
effective and how dance rates against other exercise based thera-
pies in terms of functional improvements and cost effectiveness. 1 AMED; (Parkinson’s AND disease).ti,ab; 971 results.
2 AMED; parkinson’s.ti,ab; 993 results.
3 AMED; Parkins*.ti,ab; 1223 results.
Appendix A. 4 AMED; (Parkinsonian AND Disorders).ti,ab; 15 results.
5 AMED; Parkinsonism.ti,ab; 76 results.
We searched the following databases: 6 AMED; 1 OR 2 OR 3 OR 4 OR 5; 1223 results.
7 AMED; Dance.ti,ab; 273 results.
8 AMED; Dancing.ti,ab; 61 results.
i. Allied and Complementary Medicine (AMED) (1983–2013), 9 AMED; Danc*.ti,ab; 441 results.
ii. British Nursing Index (BNI) (1992–2013), 10 AMED; Tango.ti,ab; 10 results.
11 AMED; Ballroom.ti,ab; 3 results.
iii. Cumulative Index to Nursing and Allied Health Literature
12 AMED; Foxtrot.ti,ab; 2 results.
(CINAHL) (1980–2013), 13 AMED; Ballet.ti,ab; 90 results.
iv. Excerpta Medical Database (EMBASE) (1980–2013), 14 AMED; 7 OR 8 OR 9 OR 10 OR 11 OR 12 OR 13; 455 results.
v. Health Business Elite (HBE) (1922–2013), 15 AMED; 6 AND 14; 7 results.
vi. Health Management Information Consortium (HMIC)(1979– 16 AMED; Randomised.ti,ab; 1702 results.
17 AMED; Random*.ti,ab; 12807 results.
2013),
18 AMED; (Controlled AND clinical AND trial).ti,ab; 1226 results.
vii. MEDLINE (1950–2013), 19 AMED; 16 OR 17 OR 18; 12959 results.
454 K. Sharp, J. Hewitt / Neuroscience and Biobehavioral Reviews 47 (2014) 445–456

Appendix B. (Continued ) Appendix B. (Continued )

20 AMED; 15 AND 19; 2 results. 94 HEALTH BUSINESS ELITE; 87 OR 88 OR 89 OR 90 OR 91 OR 92 OR
21 BNI; (Parkinson’s AND disease).ti,ab; 333 results. 93; 3939 results.
22 BNI; parkinson’s.ti,ab; 361 results. 95 HEALTH BUSINESS ELITE; 86 AND 94; 2 results.
23 BNI; Parkins*.ti,ab; 451 results. 96 HEALTH BUSINESS ELITE; Randomised.ti,ab; 470 results.
24 BNI; (Parkinsonian AND Disorders).ti,ab; 1 results. 97 HEALTH BUSINESS ELITE; Random*.ti,ab; 8904 results.
25 BNI; Parkinsonism.ti,ab; 5 results. 98 HEALTH BUSINESS ELITE; (Controlled AND clinical AND trial).ti,ab;
26 BNI; 21 OR 22 OR 23 OR 24 OR 25; 451 results. 628 results.
27 BNI; Dance.ti,ab; 84 results. 99 HEALTH BUSINESS ELITE; 96 OR 97 OR 98; 9148 results.
28 BNI; Dancing.ti,ab; 26 results. 100 HEALTH BUSINESS ELITE; 95 AND 99; 0 results.
29 BNI; Danc*.ti,ab; 109 results. 101 HMIC; (Parkinson’s AND disease).ti,ab; 178 results.
30 BNI; Tango.ti,ab; 5 results. 102 HMIC; parkinson’s.ti,ab; 196 results.
31 BNI; Ballroom.ti,ab; 1 results. 103 HMIC; Parkins*.ti,ab; 223 results.
32 BNI; Foxtrot.ti,ab; 0 results. 104 HMIC; (Parkinsonian AND Disorders).ti,ab; 1 results.
33 BNI; Ballet.ti,ab; 5 results. 105 HMIC; Parkinsonism.ti,ab; 15 results.
34 BNI; 27 OR 28 OR 29 OR 30 OR 31 OR 32 OR 33; 115 results. 106 HMIC; 101 OR 102 OR 103 OR 104 OR 105; 223 results.
35 BNI; 26 AND 34; 1 results. 107 HMIC; Dance.ti,ab; 76 results.
36 BNI; Randomised.ti,ab; 2180 results. 108 HMIC; Dancing.ti,ab; 27 results.
37 BNI; Random*.ti,ab; 3211 results. 109 HMIC; Danc*.ti,ab; 109 results.
38 BNI; (Controlled AND clinical AND trial).ti,ab; 142 results. 110 HMIC; Tango.ti,ab; 12 results.
39 BNI; 36 OR 37 OR 38; 3236 results. 111 HMIC; Ballroom.ti,ab; 1 results.
40 BNI; 35 AND 39; 0 results. 112 HMIC; Foxtrot.ti,ab; 0 results.
41 CINAHL; (Parkinson’s AND disease).ti,ab; 3947 results. 113 HMIC; Ballet.ti,ab; 2 results.
42 CINAHL; parkinson’s.ti,ab; 4298 results. 114 HMIC; 107 OR 108 OR 109 OR 110 OR 111 OR 112 OR 113; 123
43 CINAHL; Parkins*.ti,ab; 6277 results. results.
44 CINAHL; (Parkinsonian AND Disorders).ti,ab; 59 results. 115 HMIC; 106 AND 114; 0 results.
45 CINAHL; Parkinsonism.ti,ab; 581 results. 116 HMIC; Randomised.ti,ab; 4809 results.
46 CINAHL; 41 OR 42 OR 43 OR 44 OR 45; 6277 results. 117 HMIC; Random*.ti,ab; 9355 results.
47 CINAHL; Dance.ti,ab; 1223 results. 118 HMIC; (Controlled AND clinical AND trial).ti,ab; 841 results.
48 CINAHL; Dancing.ti,ab; 421 results. 119 HMIC; 116 OR 117 OR 118; 9409 results.
49 CINAHL; Danc*.ti,ab; 1915 results. 120 HMIC; 115 AND 119; 0 results.
50 CINAHL; Tango.ti,ab; 59 results. 121 MEDLINE; (Parkinson’s AND disease).ti,ab; 45544 results.
51 CINAHL; Ballroom.ti,ab; 18 results. 122 MEDLINE; parkinson’s.ti,ab; 47016 results.
52 CINAHL; Foxtrot.ti,ab; 4 results. 123 MEDLINE; Parkins*.ti,ab; 67291 results.
53 CINAHL; Ballet.ti,ab; 293 results. 124 MEDLINE; (Parkinsonian AND Disorders).ti,ab; 1164 results.
54 CINAHL; 47 OR 48 OR 49 OR 50 OR 51 OR 52 OR 53; 2014 results. 125 MEDLINE; Parkinsonism.ti,ab; 12087 results.
55 CINAHL; 46 AND 54; 19 results. 126 MEDLINE; 121 OR 122 OR 123 OR 124 OR 125; 67291 results.
56 CINAHL; Randomised.ti,ab; 15859 results. 127 MEDLINE; Dance.ti,ab; 2066 results.
57 CINAHL; Random*.ti,ab; 99038 results. 128 MEDLINE; Dancing.ti,ab; 896 results.
58 CINAHL; (Controlled AND clinical AND trial).ti,ab; 7467 results. 129 MEDLINE; Danc*.ti,ab; 3674 results.
59 CINAHL; 56 OR 57 OR 58; 99802 results. 130 MEDLINE; Tango.ti,ab; 342 results.
60 CINAHL; 55 AND 59; 4 results. 131 MEDLINE; Ballroom.ti,ab; 31 results.
61 EMBASE; (Parkinson’s AND disease).ti,ab; 60906 results. 132 MEDLINE; Foxtrot.ti,ab; 10 results.
62 EMBASE; parkinson’s.ti,ab; 62817 results. 133 MEDLINE; Ballet.ti,ab; 608 results.
63 EMBASE; Parkins*.ti,ab; 87235 results. 134 MEDLINE; 127 OR 128 OR 129 OR 130 OR 131 OR 132 OR 133;
64 EMBASE; (Parkinsonian AND Disorders).ti,ab; 1649 results. 4149 results.
65 EMBASE; Parkinsonism.ti,ab; 14790 results. 135 MEDLINE; 126 AND 134; 29 results.
66 EMBASE; 61 OR 62 OR 63 OR 64 OR 65; 87235 results. 136 MEDLINE; Randomised.ti,ab; 54517 results.
67 EMBASE; Dance.ti,ab; 2893 results. 137 MEDLINE; Random*.ti,ab; 628358 results.
68 EMBASE; Dancing.ti,ab; 1072 results. 138 MEDLINE; (Controlled AND clinical AND trial).ti,ab; 40109 results.
69 EMBASE; Danc*.ti,ab; 4848 results. 139 MEDLINE; 136 OR 137 OR 138; 635816 results.
70 EMBASE; Tango.ti,ab; 431 results. 140 MEDLINE; 135 AND 139; 9 results.
71 EMBASE; Ballroom.ti,ab; 54 results. 167 MEDLINE; Dance.ti,ab; 2066 results.
72 EMBASE; Foxtrot.ti,ab; 21 results. 231 MEDLINE; Ballroom.ti,ab; 31 results.
73 EMBASE; Ballet.ti,ab; 767 results. 236 MEDLINE; Randomised.ti,ab; 54517 results.
74 EMBASE; 67 OR 68 OR 69 OR 70 OR 71 OR 72 OR 73; 5459 results. 239 MEDLINE; 236 OR 237 OR 238; 635816 results.
75 EMBASE; 66 AND 74; 63 results. 240 MEDLINE; 235 AND 239; 9 results.
76 EMBASE; Randomised.ti,ab; 74169 results. 241 PsycINFO; (Parkinson’s AND disease).ti,ab; 13463 results.
77 EMBASE; Random*.ti,ab; 783672 results. 242 PsycINFO; parkinson’s.ti,ab; 13723 results.
78 EMBASE; (Controlled AND clinical AND trial).ti,ab; 51943 results. 243 PsycINFO; Parkins*.ti,ab; 18491 results.
79 EMBASE; 76 OR 77 OR 78; 793303 results. 244 PsycINFO; (Parkinsonian AND Disorders).ti,ab; 374 results.
80 EMBASE; 75 AND 79; 18 results. 245 PsycINFO; Parkinsonism.ti,ab; 3387 results.
81 HEALTH BUSINESS ELITE; (Parkinson’s AND disease).ti,ab; 1370 246 PsycINFO; 241 OR 242 OR 243 OR 244 OR 245; 18491 results.
results. 247 PsycINFO; Dance.ti,ab; 3533 results.
82 HEALTH BUSINESS ELITE; parkinson’s.ti,ab; 1478 results. 248 PsycINFO; Dancing.ti,ab; 1072 results.
83 HEALTH BUSINESS ELITE; Parkins*.ti,ab; 1745 results. 249 PsycINFO; Danc*.ti,ab; 4920 results.
84 HEALTH BUSINESS ELITE; (Parkinsonian AND Disorders).ti,ab; 5 250 PsycINFO; Tango.ti,ab; 113 results.
results. 251 PsycINFO; Ballroom.ti,ab; 53 results.
85 HEALTH BUSINESS ELITE; Parkinsonism.ti,ab; 57 results. 252 PsycINFO; Foxtrot.ti,ab; 5 results.
86 HEALTH BUSINESS ELITE; 81 OR 82 OR 83 OR 84 OR 85; 1745 253 PsycINFO; Ballet.ti,ab; 296 results.
results. 254 PsycINFO; 247 OR 248 OR 249 OR 250 OR 251 OR 252 OR 253;
87 HEALTH BUSINESS ELITE; Dance.ti,ab; 2041 results. 5103 results.
88 HEALTH BUSINESS ELITE; Dancing.ti,ab; 868 results. 255 PsycINFO; 246 AND 254; 21 results.
89 HEALTH BUSINESS ELITE; Danc*.ti,ab; 3209 results. 256 PsycINFO; Randomised.ti,ab; 4500 results.
90 HEALTH BUSINESS ELITE; Tango.ti,ab; 215 results. 257 PsycINFO; Random*.ti,ab; 116029 results.
91 HEALTH BUSINESS ELITE; Ballroom.ti,ab; 294 results. 258 PsycINFO; (Controlled AND clinical AND trial).ti,ab; 5668 results.
92 HEALTH BUSINESS ELITE; Foxtrot.ti,ab; 9 results. 259 PsycINFO; 256 OR 257 OR 258; 117225 results.
93 HEALTH BUSINESS ELITE; Ballet.ti,ab; 493 results. 260 PsycINFO; 255 AND 259; 4 results.
 K. Sharp, J. Hewitt / Neuroscience and Biobehavioral Reviews 47 (2014) 445–456 455

Appendix B. (Continued ) the functional mobility and on the quality of life of individuals with Parkinson’s
disease. Front. Aging Neurosci. 3 (14), 1–15.
368 PsycINFO; Dancing.ti,ab; 1072 results. Heiberger, L., Christoph, M., Amtage, F., Mendez Balbuena, I., Schulte-Monting, J.,
380 PsycINFO; 375 AND 379; 4 results. Heppo-Retmond, M.-C., Kristeva, R., 2011b. Impact of a weekly dance class on
381 AMED,CINAHL,EMBASE,MEDLINE,PsycINFO; Duplicate filtered: [15 the functional mobility and on the quality of life of individuals with Parkinson’s
AND 19], [55 AND 59], [75 AND 79], [235 AND 239], [375 AND disease. Front. Aging Neurosci. 3 (14), 1–15 (citations of Y. Jeong, S. Hong, M.
379]; 37 results Lee, M. Park, Y. Kim therein).
Higgins, J., Green, S., 2011. Cochrane Handbook for Systematic Reviews of Inter-
ventions, Version 5.1.0. The Cochrane Collaboration, London, www.cochrane-
handbook.org (accessed 13 April 2013).
Hirsh, M.A., Farley, B.G., 2009. Exercise and neuroplasticity in person’s living with
References Parkinson’s disease. Eur. J. Phys. Rehabil. Med. 45, 215–229.
Houston, S., McGill, A., 2012. A Mixed methods study into ballet for people living
Allen, N., Sherrington, C., Paul, S., Canning, C., 2011. Balance and falls in Parkinson’s with Parkinson’s. Arts Health, http://dx.doi.org/10.1080/17533015.2012.745580.
disease: a meta-analysis of the effect of exercise and motor training. Mov. Disord. Hui, E., Chui, B., Woo, J., 2009. Effects of dance on physical and psychological well
26 (9), 1605–1615. being in older persons. Arch. Gerontol. Geriatr. 49 (1), e45–e50.
Berg, K., Wood-Dauphinee, S., Williams, J., Maki, B., 1992. Measuring balance in Jenkinson, C., Fitzpatrick, R., Peto, V., Greenhall, R., Hyman, N., 1997. The Parkinson’s
the elderly: validation of an instrument. Can. J. Public Health 83 (Suppl 2), S7– Disease Questionnaire (PDQ-39): development and validation of a Parkinson’s
S11. disease summary index score. Age Aging 26 (5), 353–357.
Bloem, B., Hausdorff, J., Visser, J., 2004. Falls and freezing of gait in Parkinson’s dis- Howe, T., Rochester, L., Neil, F., Skelton, D., Ballinger, C., 2011. Exercise for improving
ease: a review of two interconnected episodic phenomena. Mov. Disord. 19 (8), balance in older people. Cochrane Database Syst. Rev. (11) (citations of Winter
871–884. therein).
Boonstra, T., Vdk, H., Munnenke, M., Bloem, B., 2008. Gait disorders and balance Kattenstroth, J.-C., Kolankowska, I., Kailish, T., Dinse, H., 2010. Superior sensory,
disturbances in Parkinson’s disease: clinical update and pathophysiology. Curr. motor, and cognitive performance in elderly individuals with multi-year danc-
Opin. Neurol. 21 (4), 461–471. ing activities. Front. Aging Neurosci. 2 (31), 1–9.
British Council for Chinese Martial Arts. Tai Chi Overview.http://www.bccma.com/ Kerr, G., Worringham, C., Cole, M., Lacherez, P., Wood, J., Silburn, P., 2010. Predictors
chinese-martial-arts.asp?style = 12013.(accessed 13/04/2013). of future falls in Parkinson’s disease. Neurology 75, 116–124.
Brok, J., Thorlund, K., Gluud, C., Wettersley, J., 2008. Trial sequential analysis reveals Keus, S., Bloem, B., Hendriks, E., Bredero-Cohen, A., Munneke, M., Practice Rec-
insufficient information size and potentially false positive results in many meta- ommendations Development Group, 2007. Evidence based analysis of physical
analyses. J. Clin. Epidemiol. 61 (8), 763–769. therapy in Parkinson’s disease with recommendations for practice and research.
de Dreu, M.J., Vdw, A.S.D., Poppe, E., Kwakkel, G., van Wegen, E.E.H., 2012. Reha- Mov. Disord. 22 (4), 451–460.
bilitation, exercise therapy and music in patients with parkinson’s disease: a Kwakkel, G., de Goede, C., van Wegen, E., 2007. Impact of physical therapy for Parkin-
meta-analysis of the effects of music based movement therapy on walking abil- son’s disease: a critical review of the literature. Parkinson. Relat. Disord. 13,
ity balance and quality of life. Parkinson. Relat. Disord. 18 (Supplement 1), S114- S478-S87.
S9. Lee, H., Kim, S., Chae, Y., Kim, M., Yin, C., Jung, W., Cho, K., Kim, S., Park, H., Lee, H.,
Deane, K., Jones, D., Ellis-Hill, C., Clarke, C., Playford, E., Ben-Schlomo, Y., 2001. A 2012. Qi dance in patient’s with Parkinson’s disease. Parkinson. Relat. Disord.
comparison of physiotherapy techniques for patient’s with Parkinson’s disease. 18, S157.
Cochrane Database Syst. Rev. 1. Maki, B., McIiroy, W., 2005. Change in support balance reactions in older per-
Deeks, J., Higgins, J., Riis, J., Silagy, C.V.P., Cumpston, M., Macdonald, S., sons: an emerging research area of clinical importance. Neurol. Clin. 23 (3),
Manukonga, C., 2002. The Cochrane Collaboration Open Learning Mate- 751–783.
rial, www.cochrane-net.org/openlearning/html/mod0.htm (accessed 13 April Merholz, J., Friis, R., Kugler, J., Twork, S., Storch, A., Pohl, M., 2010. Treadmill training
2013). for people with Parkinson’s disease. Cochrane Database Syst. Rev.
Duncan, R.P., Earhart, G.M., 2012. Randomized controlled trial of community-based Mirelman, A., Maiden, I., Herman, T., Deutsch, J., Giladi, G., Hausdorff, J., 2001. Vir-
dancing to modify disease progression in Parkinson disease. Neurorehabil. Neu- tual reality for gait training: can it induce motor learning to enhance complex
ral Repair 26 (2), 132–143. walking and reduce fall risk in patients with Parkinson’s disease. J. Gerontol.,
Duncan and Earhart Exercise and Parkinson’s: Comparing Interventions and Explor- Ser. A, Biol. Med. Sci. 66, 234–240.
ing Neural Mechanisms. http://clinicaltrials.gov/ct2/show/NCT01768832 Murray, E., Jung, C., Millar, J., Daniel, L., Talles, E., Dunlop, S., Solomon, D., Von Coella,
(accessed 08/01/2013). R., Mari, Z., Ying, S., 2010. Dance therapy: effect on gait measures and quality of
Earhart, G.M., 2009. Dance as therapy for individuals with Parkinson’ disease. Eur. J. life in Parkinson’s disease patients. Mov. Disord. 25, S301.
Phys. Rehabil. Med. 45 (2), 231–238. Nieuwboer, Kwakkel, G., Rochester, L., Jones, D., van Wegen, A., Wiilams, A., Chavret,
Enright, P., 2003. The six minute walk test. Respir. Care 48 (8), 783–785 (citations of F., Hetherington, V., Baker, K., Lim, I., 2007. Cueing training in the home improves
B. Baulke therein). gait related mobility in the home—the RESCUE trial. J. Neurol., Neurosurg. Psy-
Fahn, S., Elton, R., 1987. In: Fahn, S., Marsden, C., Calne, D., Golstein, M. (Eds.), chiatry 78, 134–140.
Recent Developments n Parkinson’s Disease. Macmillan Healthcare Information, Postuma, 2012. Tango for Treatment of Motor and Non-motor Manifestations
Florham Park, NJ. in Parkinson’s Disease, http://clinicaltrials.gov/show/NCT01573260 (accessed
Foster, E., Golden, L., Ryan, P., Duncan, R., Earhart, M., 2013. A community-based 08/01 2013).
argentine Tango dance program is associated with increased activity participa- Schrag, A., Sampaio, C., Counsell, N., Poewe, P., 2006. Minimal clinically impor-
tion among individuals with Parkinson disease. Arch. Phys. Med. Rehab. 94 (2), tant change on the unified Parkinson’s disease rating scale. Mov. Disord. 21 (8),
240–249. 1200–1208.
Fox, C.M., Rammig, L.O., Ciucci, M.R., Sapir, S., McFarland, D.H., Farley, B.G., Good- Schulman, L., Gruber Baldini, A., Anderson, K., Fishman, P., Reich, S., Weiner, W., 2010.
win, V., Richards, S., Taylor, R., Taylor, A., Campbell, J., 2008. The effectiveness of The clinical important difference on the unified Parkinson’s Disease Question-
exercise interventions for people with Parkinson’s disease: a systematic review naire. J. Am. Med. Assoc.: Neurol. 67 (1), 64–70.
and meta-analysis. Mov. Disord. 23 (5), 632–710. Shanhan, 2012. Randomized Controlled Feasibility Trial to Determine the
Giladi, N., Shabati, H., Simon, E., Biran, S., Tal, J., Korczyn, A., 2000. Construction of Effectiveness of Set Dancing for People with Parkinson’s Disease, http://
freezing of gait questionnaire for people with Parkinsonism. Parkinson. Relat. clinicaltrials.gov/ct2/show/NCT01757509?term=parkinson%27s+dance&rank=5
Disord. 6 (3). (accessed 08/01/2013).
Goodwin, V., Richards, S., Taylor, R., Taylor, A., Campbell, J., 2008. The effectiveness of Steffen, T., Senney, S., 2008. Test–retest reliability and minimal detectable change
exercise interventions for people with Parkinson’s disease: a systematic review on balance and ambulation tests, the 36-item short-form health survey, and the
and meta-analysis. Mov. Disord. 23 (5), 631–640. unified Parkinson disease rating scale in people with Parkinsonism. Phys. Ther.
Hackney, M.E., Earhart, G.M., 2009a. Effects of dance on movement control in Parkin- 88 (6), 733–746.
son’s disease: a comparison of Argentine Tango and American ballroom. J. Studentski, S., Perera, S., Patel, K., Rosano, C., Faulkner, K., Inzitari, M., Brach, J.,
Rehabil. Med. 41 (6), 475–481. Chandler, J., Cawthorn, P., Barrett Connor, E., Nevitt, M., Visser, M., Kritchevsky,
Hackney, M.E., Earhart, G.M., 2009b. Health-related quality of life and alternative S., Bandinelli, S., Harris, T., Newman, A., Cauley, J., Ferrucci, L., Guralnik, J.,
forms of exercise in Parkinson disease. Parkinson. Relat. Disord. 15 (9), 644– 2011. Gait speed and survival in older adults. J. Am. Med. Assoc. 305 (1),
648. 50–58.
Hackney, M., Earhart, G., 2010. Effects of dance on gait and balance in Parkinson’s Tilerson, J., Caudle, W., Reveron, M., Miller, G., 2003. Exercise induced behavioural
disease: a comparison of partnered and non partnered dance movement. Neu- recovery and attenuates neurochemical deficits in rodent models of Parkinson’s
rorehabil. Neural Repair 24 (4), 384–392. disease. Neuroscience 119, 899–911.
Hackney, M., Krantorovich, S., Earhart, G., 2007a. A study on the effects of Argentine Tomlinson, C.L., Patel, S., Meek, C., Clarke, C.E., Stowe, R., Shah, L., et al., 2012a. Phys-
Tango as a form of partnered dance for those with Parkinson’s disease and the iotherapy versus placebo or no intervention in Parkinson’s disease. Cochrane
healthy elderly. Am. J. Dance Ther. 29 (2), 109–127. Database Syst. Rev. (8).
Hackney, M., Kantorovich, S., Levin, R., Earhart, G., 2007b. Effects of Tango on func- Tomlinson, C.L., Patel, S., Meek, C., Clarke, C.E., Stowe, R., Shah, L., et al., 2012b. Phys-
tional mobility in Parkinson’s disease: a preliminary study. J. Neurol. Phys. Ther. iotherapy versus placebo or no intervention in Parkinson’s disease. Cochrane
31 (4), 173–179. Database Syst. Rev. 8 (citations of C. Trew therein).
Heiberger, L., Christoph, M., Amtage, F., Mendez Balbuena, I., Schulte-Monting, J., van Nimwegen, M., Speelman, A., Overreem, S., van de Warrenburg, B., Smulders,
Heppo-Retmond, M.-C., Kristeva, R., 2011a. Impact of a weekly dance class on K., Dontje, M., Borm, G., Backx, F., Bloem, B., Munneke, M., 2013. Promotion
456 K. Sharp, J. Hewitt / Neuroscience and Biobehavioral Reviews 47 (2014) 445–456

of physical activity and fitness in sedentary patients with Parkinson’s disease: Volpe, D., Signorini, M., Marchetto, A., Lynch, T., Morris, M., 2013. A comparison of
randomised controlled trial. Br. Med. J. 346, 1–11. Irish set dancing and exercises for people with Parkinson’s disease: a phase II
van Nimweggen, M., Speelman, A., Hofman-van Rossum, E., Overreem, S., Deeg, D., feasibility study. Biomed. Cent. Geriatr. 13 (54).
Borm, G., Backx, F., Bloem, B., Munneke, M., 2011. Physical inactivity in Parkin- Westbrook, B., McKibben, H., 1989. Dance movement therapy with groups of out-
son’s disease. J. Neurol. 258, 2214–2221. patients with Parkinson’s-disease. Am. J. Dance Ther. 11 (1), 27–38.