This study is a fully-powered randomised controlled trial and an essential follow-up study of our pilot work. The 21-style specific dance intervention to improve the balance and reduce the risks of falls (SDIIBRF) was suggested to be interesting, achievable, and compliable, according to the participants' feedback in the investigators' pilot work. Due to the expertly-designed and specifically-designed characteristics of the 21 dance styles in the SDIIBRF for improving balance and reducing risk of falls in older adults, the dance steps included in the SDIIBRF are unique and are largely different from the dance interventions that have already been reported in the literature. Based on the encouraging results of the preliminary efficacy in the investigators' pilot work and dance being a mind-body exercise, the investigators hypothesise that the specific dance programme will be efficacious in reducing the frequency of falls (primary outcome measure), improving the balance (secondary outcome measure for the main fall risk factor), and reducing the physical and psychological fall risk factors (secondary outcome measures) in Hong Kong's at-risk older adults. The findings could provide important new evidence for a feasible option for older adults as an innovative fall prevention exercise programme. The research question is "Can the 21-style SDIIBRF reduce the frequency of falls, improve the balance, and reduce the fall risks in Hong Kong's older adults at moderate to high risk of falling? Therefore, the aim of this proposed study is to examine the efficacy of the 21-style SDIIBRF in reducing the frequency of falls, improving the balance, and reducing the fall risks in Hong Kong's older adults by comparing it with a wait-list control.

One-hundred older adults will be recruited from different collaborating elderly community and rehabilitation centres in Hong Kong by convenience sampling. The participants (N=100) will be randomly assigned by an independent person using concealed block randomisation via a sealed and opaque envelope system to either the dance group (DANG; n=50) or the wait-list control group (WLCG; n=50). A computerised random-number generator will determine an allocation schedule. The older adults at moderate to high risk of falling living in Hong Kong, as indicated by a score of less than 25/28 in the Tinetti Performance Oriented Mobility Assessment (POMA) scale, will be randomly assigned to either the DANG or the wait-list control group (WLCG). The inclusion criteria will be (1) age 65 or above and (2) a total score of less than 25/28 in the POMA. The exclusion criteria will be (1) a total score of less than 24/30 in the Chinese version of the Mini-Mental State Examination (MMSE-C), (2) the existence of any untreated cerebral vascular disease, Parkinson's disease, or any other major neurological deficit, (3) the existence of any medical condition that can affect their safety during dance training sessions, (4) unable to stand independently for at least 15 minutes without a walking aid, (5) unable to walk independently indoor for at least 20 metres without a walking aid, (6) previous experience of dance training, and (7) currently participating in any other formal exercise programme.

The participants in DANG will participate in 12 dance training sessions (~120 minutes each with rest breaks) in one of the four groups of 12-13 participants once a week for 12 weeks. All training sessions will be conducted in Hong Kong by two experienced dance instructors under the supervision of a registered physiotherapist (e.g., principal applicant TWL Wong). Any adverse events will be tracked and reported. As an augmented safety measure for the dance training sessions in a group, we will ensure that the dance instructors are informed each participant's fall risk assessment so that close supervision of those with the highest fall risk can be guaranteed. The dance instructors will also gradually progress the balance challenge level as classes progress. The participants in the DANG will practise the 21-style specific dance exercises. The participants in the WLCG will be provided a fall prevention pamphlet with a set of conventional fall prevention exercises (e.g., including a set of balance, walking, and resistance exercises) at the start and will receive the training from the 21-style SDIIBRF after three scheduled assessment sessions with two waiting periods of 12 weeks and 3 months in between. The walking exercises will also include the balance training components (e.g., walking and turning around, sideways walking, heel walk, toe walk, and tandem walk). The participants in the WLCG will be asked to participate in 12 training sessions at home (~120 minutes each with rest breaks) once a week for 12 weeks. An exercise logbook will be provided to the participants in the WLCG to record their exercise compliance. If any of the participants is uncertain about the exercises, it is advised that the participant should contact the research staff immediately for further advice before the participant continues the exercises. The 21 dance styles (dance in a group), chosen and modified from seven dance styles, including cha-cha, jive, rumba, salsa, waltz, tango, and macarena, were developed by a group of experienced dance and rehabilitation experts and have been previously practised by a group of older adults in our pilot work, demonstrating good feasibility and positive participants' feedback. As a cautionary measure to maximize our participants' safety and restrict them from imprecisely practising the dance styles at home, all participants will be strongly advised not to practise without professional supervision. However, the participants will be encouraged to find the suitable opportunity to continue the dance activities under professional supervision after completion of the training sessions, so as to benefit from social interaction. For those older adults who can perform the dance styles excellently and stably under the assessment of the qualified dance instructors and practise at home becomes possible, additional safety advice will be provided, including the advice on appropriate warm up and cool down, sufficient and safe space, appropriate footwear, and dance at the appropriate level.

Both the DANG and WLCG will undertake three assessment sessions: before training at baseline (T0), just after the completion of the 12 dance training sessions (T1), and 3 months after the completion of the training sessions (T2) to examine the programme's efficacy. An independent assessor who is blinded to the participants' group allocation will perform the assessments. During T0, a questionnaire will be used to collect participants' demographic information, fall history (past 12 months and others by self-recall), medical history, current medications, and social history. The participants will be asked to prospectively record the number of falls they have experienced each day between T0, T1, and T2, using a structural calendar, and their frequency of falls (primary outcome measure) will be collected at T1 and T2. A fall will be counted if it will cause our participant(s) unintentionally coming to rest on the ground or other lower level. Afterwards, a series of assessments will be used to evaluate the participants' balance ability, through the Berg Balance Scale (BBS - secondary outcome measure for the main fall risk factor), and other physical and psychological fall risk factors (secondary outcome measures) at T0, T1, and T2. Since all eligible participants will be at moderate to high risk of falling, no ceiling effect of the BBS is anticipated; that is, all will begin treatment with relatively low scores. Additionally, the participants will be asked to record their ongoing compliance with dance and other physical activities during the follow-up period (from T1 to T2) using a daily diary. As an experienced registered physiotherapist in Hong Kong, the principal applicant TWL Wong will train all independent assessors and ensure that they perform the assessments accurately.

BBS will assess the balance ability (secondary outcome measure for the main fall risk factor) of the participants. BBS is a well-recognised and widely used functional balance assessment scale with good validity and reliability. The scale comprises 14 daily balance tasks in which participants have to maintain balance while completing them. The maximum total score is 56, and each of the 14 items could be scored between 0 and 4. Lower scores indicate poorer balance ability and a higher risk of falling (i.e., 0-20 = high risk; 21-40 = moderate risk; 41-56 = low risk). The BBS is a comprehensive tool to assess balance with the advantages of clinically relevant, standardised scoring, sensitive to change, and can assist in clinical decision-making and intervention planning. Concerning the other physical fall risk factors (secondary outcome measures), a functional gait and balance assessment will be conducted using the POMA and Timed Up and Go (TUG) tests. These two tests have been extensively used and widely recognised as clinical assessment tools for evaluating functional gait and balance with good validity and reliability. The POMA assesses the components of gait (12 points with seven items) and balance (16 points with nine items) with a total score of 28 points. A lower total score suggests a higher risk of falling (i.e., ≤18 = high risk; 19-24 = moderate risk; ≥25 = low risk). For example, the item 'Nudged' is useful in measuring the ability of an older adult to withstand a perturbation that could cause a fall. The TUG test has both discriminative and predictive properties related to older adult falls. The test requires participants to stand up from a chair, walk 3 metres, turn around, walk back 3 metres, turn around, and sit down on the chair at a comfortable pace. The total time taken to complete this task will be recorded. A longer completion time illustrates a higher risk of falling (e.g., taking more than 13.5 seconds to complete the TUG test indicates a high risk of falling in older community-dwelling adults). Flexibility will be measured via the "Chair Sit-and-Reach" (CSR) test. A shorter reaching distance represents poorer flexibility and a higher risk of falling. Muscle strength of the major lower limb muscles groups (e.g., knee and hip extensors) will be measured using the well-recognised Lafayette manual muscle testing system (e.g., Model 01163 or equivalent). A lower score represents poorer muscle strength and a higher risk of falling. Agility will be assessed by the 8 Foot Up-and-Go Test (8FUG). A longer completion time indicates poorer agility and a higher risk of falling.

In addition, various valid and reliable assessment tools will assess different psychological fall risk factors (secondary outcome measures). The Chinese version of the Falls Efficacy Scale International [FES-I (Ch)] will quantify the fear of falling. A higher score indicates a higher fear of falling, indicating a higher risk of falling. The Chinese version of the Activities-specific Balance Confidence Scale [ABC (Ch)] will examine the balance confidence. A lower score indicates poorer balance confidence and, thus, a higher risk of falling. Depression levels will be evaluated by the Chinese version of the four-item Geriatric Depression Scale (GDS-4C). GDS-4C was developed as a simple but valid and reliable alternative to the 15-item and 30-item versions of the GDS and displays good sensitivity (60-76%) and specificity (65-81%), using the cut-off point of 2 or more in older population. A higher score represents a higher depression level, suggesting a higher risk of falling. Anxiety levels will be evaluated by the Chinese version of the State-Trait Anxiety Inventory (STAT-SC & STAI-TC). A higher score suggests a higher anxiety level and, thus, a higher risk of falling. The MMSE-C will only be used as a screening tool to inspect the preliminary cognitive function at T0. A structured questionnaire will be used to record participants' feedback. Participants' satisfaction will be measured by a Participant Satisfaction Questionnaire. The adherence of the participants in the dance programme will be assessed by recording the attendance to dance training sessions and compliance with the dance styles. The treatment fidelity will also be assessed by comparing all the dance training sessions with the established dance training protocol. Participants' ongoing compliance with dance and other physical activities from T1 to T2 will be recorded.

Inclusion Criteria:

1. age 65 or above;
2. a total score of less than 25/28 in the Tinetti Performance Oriented Mobility Assessment (POMA) scale.

Exclusion Criteria:

1. a total score of less than 24/30 in the Chinese version of the Mini-Mental State Examination (MMSE-C);
2. the existence of any untreated cerebral vascular disease, Parkinson's disease, or any other major neurological deficit;
3. the existence of any medical condition that can affect their safety during dance training sessions;
4. unable to stand independently for at least 15 minutes without a walking aid;
5. unable to walk independently indoor for at least 20 metres without a walking aid;
6. previous experience of dance training;
7. currently participating in any other formal exercise programme.