Background: Tai chi is considered a safe and low-cost treatment for improving balance ability among an older population. However, there is no existing evidence on the optimal exercise parameters of tai chi for improving balance in older adults.

Objectives: To investigate the optimal parameters of a tai chi intervention to improve balance performance of older adults.

Design: Systematic review and meta-analysis of randomized controlled trials (RCTs).

Setting: PubMed, Embase, Cochrane Library, Web of Science, Scopus, China National Knowledge Infrastructure, Wanfang, Chinese Science and Technology Periodical and China Biology Medicine were searched from inception until November 30, 2020.

Participants: Adults aged 60 years and over.

Measurements: Two reviewers independently extracted the data and assessed the quality of the included studies according to the Physiotherapy Evidence Database (PEDro) scale. Subgroup analyses and meta-regressions were conducted to elucidate the impact of tai chi training programs on balance measures.

Results: Twenty-six eligible RCTs were included in the meta-analysis. Pooled results showed that tai chi has moderate effects for improving proactive balance (weighted mean standardized mean differences [SMD(wm) ] = 0.61, 95% CI 0.33-0.89) and static steady-state balance (SMD(wm)  = 0.62, 95% CI 0.30-0.95) and small effects for improving dynamic steady-state balance (SMD(wm)  = 0.38, 95% CI 0.03-0.73) and balance test batteries (SMD(wm)  = 0.47, 95% CI 0.13-0.81) in adults over 60 years of age. The practice frequency could predict the effects of tai chi on static steady-state balance, and the 24-form simplified Yang style tai chi (45-60 min/session, more than four sessions per week and at least 8 weeks) was the most optimal.

Conclusions: Tai chi is effective at improving the balance ability of adults over 60 years of age. A medium duration and high frequency of 24-form tai chi may be the optimal program for improving balance, but this evidence should be recommended with caution due to limitations of the methodology and small sample sizes.