Hackney 2009.

Methods	Parallel group design. Randomisation was conducted by one author by selecting 1 of the 4 groups from a hat. Method of analysis not described. Treated as outpatients for 20 hours within 13 weeks (1‐hour sessions). Assessed at baseline and within one week of completing 20 sessions. Assessors were blinded.
Participants	19 participants in the tango group, 19 in the waltz/foxtrot group, 17 in the Tai Chi group, and 20 in the control group. 5, 2, 4 and 3 dropouts from the tango, waltz/foxtrot, Tai Chi, and control group respectively. Participants' mean age, 68.2 years (tango), 66.8 years (waltz/foxtrot), 64.9 years (Tai Chi), 66.5 years (control); male/female 11/3 (tango), 11/6 (waltz/foxtrot), 11/2 (Tai Chi), 12/5 (control). Mean Hoehn and Yahr 2.1 (tango), 2.0 (waltz/foxtrot), 2.0 (Tai Chi), and 2.2 (control). Mean duration of PD 6.9 years (tango), 9.2 years (waltz/foxtrot), 8.7 years (Tai Chi), 5.9 years (control). No baseline characteristics were given for dropouts. Inclusion criteria: Hoehn and Yahr stages 1‐3, at least 40 years of age, could stand for at least 30 minutes, walk independently 3 or more metres with or without assistive device, diagnosis of idiopathic Parkinson's disease using diagnostic criteria for clinically defined 'definite PD' based on published standards, participants demonstrated clear benefit from levodopa, cognitively intact. Exclusion criteria: history of neurological deficit other than Parkinson's disease, dementia, another measure of cognitive function, and a separate part of the study not reported where all participants were required to perform a subtraction task while walking (all completed with 85% accuracy), considered cognitively intact.
Interventions	Dance: experienced professional ballroom dancer taught progressive tango or waltz/foxtrot lessons for 1 hour twice weekly. Instructor equally versed in both dances attempted to give all students equal attention. Both genders spent equal time leading and following dance roles. All steps done in closed practice position where participants maintain contact through upper extremities and face one another. Martial arts: Received progressive lessons for 1 hour twice weekly on Tai Chi's first and second circles including 37 postures of the Yang Short Style of Cheng Manching from an experienced instructor. Control: No intervention. Drug therapy was kept constant during the trial.
Outcomes	PDQ‐39. UPDRS III. Berg Balance Scale. Timed Up and Go. 6‐Minute walk test. Freezing of gait questionnaire. Forward and backward gait. Gait speed. Stride length. Single support time. Exit questionnaire. Tandem Stance Test (TS). One‐Leg Stance test (OLS). Assessments took place at a standardised time, when participants were in the 'on' state.
Notes	1 participant was excluded from the study as the result of medication change. Participants were instructed not to change their habitual exercise routines. Data taken from all three publications. The tango and waltz/foxtrot arms assessed were suitably similar and were therefore combined to give one comparison of dance.
Risk of bias
Bias	Authors' judgement	Support for judgement
Eligibility Criteria	Low risk
Randomisation Method	High risk	Conducted by one author by selecting 1 of the 4 groups from a hat.
Concealment of Allocation	High risk	Conducted by one author by selecting 1 of the 4 groups from a hat.
Similarity at Baseline	Low risk
Withdrawals Described	High risk	19% withdrawals.
Intention To Treat Analysis	Unclear risk	Method of analysis not described.
Cointerventions Constant	Low risk	Drug therapy was kept constant during the trial.
Blinded Assessors	Low risk	Assessors were blinded to group allocation.