Hwang 2016.

Methods	Study design: RCT Number of study arms: 2 Length of follow‐up: 18 months
Participants	Setting: Taipei, Taiwan Number of participants: 456 Number analysed: 334 Number lost to follow‐up: 122 Sample: community‐dwelling Age (years): mean 72 Sex: 67% female Inclusion criteria: aged ≥ 60 who received fall‐related medical attention ‐ an older person was presumed to have recovered from a fall injury within 6 months and who could walk independently were invited by telephone to enrol in the study and participate in the baseline assessment Exclusion criteria: major unstable cardiopulmonary disease (ischaemic chest pain or shortness of breath on mild exertion), cognitive impairment (MMSE score < 24), and contraindications to physical exercise (e.g. severe arthritis that limits exercise capability)
Interventions	1. Individually‐supervised Tai Chi: taught individually each week for 24 consecutive weeks, 60‐minute sessions, 1 a week for 6 months 2. Individually‐supervised balance and strength training: exercises at increasing difficulty levels using own body weight; 60‐minute sessions, 1 a week for 6 months
Outcomes	1. Rate of falls 2. Number of people who experienced 1 or more falls (risk of falling) 3. Number of people who died
Duration of the study	72 weeks
Adherence	Adherence measured as participation in sessions 1. Individually‐supervised Tai Chi group: 145 (78%) people participated in 20 or more sessions 2. Supervised balance and strength training group: 132 (72%) people participated in 20 or more sessions
Notes	Source of funding: National Health Research Institute, Ministry of Science Technology Economic information: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Statisticans using computer‐generated sequence; block‐randomised in groups of 8
Allocation concealment (selection bias)	Low risk	Using an automated secure website operated by an off‐site independent service
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	Low risk	Research assistants who conducted fall‐related phone calls were blinded to allocation
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)	Unclear risk	Not applicable
Incomplete outcome data (attrition bias) Falls and fallers	High risk	More than 20% of fall outcome data are missing (27%)
Selective reporting (reporting bias)	Unclear risk	Minimum set of expected outcomes not reported (adverse events not reported)
Method of ascertaining falls (recall bias)	Low risk	Quote: "Falls were prospectively monitored and recorded daily using a diary, and these records were mailed monthly to the study coordinator." "When a participant failed to return the diary or provided incomplete data, two research assistants blinded to the group assignment provided telephone reminders, making a maximum of five calls. Monthly follow‐up of fall records was continued in participants who were unavailable for certain periods".