Huang 2010.

Methods	Study design: Cluster RCT Number of study arms: 2 Number of clusters: 4 (2 clusters included in this review) Length of follow‐up: 5 months
Participants	Setting: Taipei, Taiwan Number of participants: 115 Number analysed: 78 Number lost to follow‐up: 37 Sample: people registered as living in 4 randomly‐selected villages Age (years): mean 71.5 (SD 0.6) in people not lost to follow‐up Sex: 30% female Inclusion criteria: aged > 65 years; living in a non‐organised community of Taiwan Exclusion criteria: immobile; living outside registered living area
Interventions	Randomised into 4 groups: 3 intervention groups (1 group‐based Tai Chi, 1 education group, 1 Tai Chi plus education group) and 1 control group. Only group‐based Tai Chi and control groups included in this review 1. Group‐based Tai Chi: 13 simple movements, 40‐minute sessions, 3 a week for 20 weeks 2. Control group: usual care
Outcomes	1. Number of people who experienced 1 or more falls (risk of falling) 2. Health‐related quality of life
Duration of the study	20‐72 weeks
Adherence	Not reported
Notes	Source of funding: The National Science Council, Taiwan Economic information: not reported Reported results not adjusted for clustering. Raw data at 5 months used in the review and adjusted for clustering. No raw data for 18 months so not possible to adjust for clustering. Number of clusters allocated to intervention: 1; number of clusters allocated to control: 1; number of clusters analysed (intervention): 1; number of clusters analysed (control): 1 Email communication regarding fall data, response received, data not included in review
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The three intervention groups and one control group were then assigned randomly to one each of the four selected villages."
Allocation concealment (selection bias)	High risk	Individual participant recruitment was undertaken after group allocation of the 4 villages. There was no mention of active blinding of research team members recruiting participants
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Participants and personnel unblinded but impact of unblinding unknown
Blinding of outcome assessment (detection bias) Falls	Unclear risk	Insufficient information to determine how falls were monitored in each group or whether assessors were blinded
Blinding of outcome assessment (detection bias) Fractures	Unclear risk	Not applicable
Blinding of outcome assessment (detection bias) Hospital admission, medical attention and adverse events	Unclear risk	Not applicable
Blinding of outcome assessment (detection bias) Health related quality of life (self report)	High risk	Participants were not blinded to group allocation
Incomplete outcome data (attrition bias) Falls and fallers	High risk	More than 20% of fall outcome data are missing (32%). Loss of fall data was unbalanced in the Tai Chi (n = 34) and control (n = 3) groups, with the reasons for withdrawal not clear
Selective reporting (reporting bias)	High risk	Falls measured, but number of falls not reported. Adverse events not reported
Method of ascertaining falls (recall bias)	Unclear risk	No mention of how falls were monitored Quote: "The fall or non‐fall situation was checked at preintervention, postintervention and at one and half year later with the aim of examining the effectiveness of the interventions"
Cluster‐randomised trials	Unclear risk	Individuals were recruited to the trial after the clusters were randomised and personnel recruiting participants were not blind to cluster; clusters were not comparable at baseline for gender or education level; missing outcomes for clusters or within clusters were not reported; did not account for clustering in analysis; results comparable with individually randomised trials