Preprint · Feb 26, 2026
The Role of Physical Activity and Psychological Factors in Fall Prevention: Insights from a Mediation Analysis of an RCT (CHAnGE trial)
By Hamdani N, Saragiotto BT, Rizzo RRN et al.
Research question Why might a physical activity coaching intervention fail to reduce fall risk in older adults? Rationale The CHAnGE trial was a pragmatic, assessor-blinded, cluster-randomised controlled trial with two parallel arms, where each community-based group served as a cluster. The primary aim of this trial was to evaluate the effect of the physical activity and fall prevention programme on objectively measured physical activity and self-reported falls compared with a healthy eating programme in people aged 60 years and over. The authors hypothesised that the CHAnGE program would lead to an increase in physical activity (i.e., activity counts per minute) and a reduced fall rate compared with the healthy eating control. The rationale for the intervention to reduce falls was that walking and preventive behaviour changes would improve strength and balance, reduce fear of falling, increase mobility-related confidence, and reduce risk-taking behaviors. However, there was no significant effect on the overall fall rate at 12 months. The intervention group had an average of 0.71 falls per person per year, compared to 0.87 in the control group, with an incidence rate ratio of 0.86 (95% CI 0.65 to 1.14), indicating a non-significant 14% reduction in falls. The intervention significantly improved daily steps (MD: +649 steps per day, 95% CI: 283 to 1015) at 6-month follow-up. No significant differences were found in secondary outcomes, such as fear of falling (7 to 28 Falls Efficacy Scale MD: −0.26, 95% CI −0.58 to 0.06, point estimate favouring less fear in the intervention group), mobility confidence (10 to 100 Modified Gait Efficacy Scale MD: 1.56, 95% CI −0.04 to 3.17, point estimate favouring greater mobility confidence in the intervention group), and risk-taking behaviour (5−20 Risk-taking behaviour: MD 0.21, 95% CI −0.08 to 0.50, point estimate favouring greater risk taking in the intervention group) at 6-months. Hypothesis This exploratory analysis aims to understand potential mechanisms underlying the lack of an apparent intervention effect on fall rates. Insights from this mediation evaluation will guide the optimisation of future fall-prevention programs. Methods Study registration and Ethical approval The CHAnGE Trial registration number is ACTRN12615001190594. The University of Sydney Human Research Ethics Committee approved the trial (number 2015/517). The analysis plan for this mediation analysis will be registered on Open Science Framework. This research protocol follows the AGReMA guidelines (A Guideline for Reporting Mediation Analyses). Study Design and Data Source This study is a mediation analysis of the CHAnGE Trial, a cluster-randomised controlled trial. The trial examined the effectiveness of a telephone-based health-coaching intervention, combined with a fall-prevention program, compared with an active control group among older adults. A total of 72 clusters (605 participants) were randomised: 37 clusters (290 participants) in the intervention group and 35 clusters (315 participants) in the control group. Follow-up data on physical activity, fall rate, psychological measures and well-being were collected over 12 months (i.e., baseline, 3 months, 6 months, and 12 months). Participants Participants were recruited from community-living individuals in metropolitan Sydney and regional Orange, NSW, through established community-based organisations for older adults, using the following criteria: only older adults aged 60+ years who held meetings or events at least once every 2 months were selected. Group members were potentially eligible for the trial if they were 60+ years old, were living in a private dwelling or a retirement village and regularly attended meetings (at least once every 2 months) or other activities at the participating community group. Intervention and control groups Intervention group The intervention group received written information on increasing physical activity and preventing falls, a 2-hour home visit by a research physiotherapist trained as a health coach to increase walking activity, and telephone-based health coaching contacts. Participants in the intervention clusters received the CHAnGE intervention, which included a fall-prevention booklet, a two-hour home visit from a trained physiotherapist, and telephone health coaching. Participants completed a fall risk assessment (QuickScreen) at the first visit and established a fall-prevention and physical-activity plan with their health coach. The treatment plan included recommendations for strengthening and balancing exercises as well as advice based on the fall-risk assessment, with a focus on vision (e.g., regular eye examination, avoiding bifocals when walking outdoors, improving lighting at home), peripheral sensation (e.g., checking for reduced sensations on the feet, appropriate footwear), and medication (e.g., identify and adjust dose of drugs that increase fall risks), and home adaptations (e.g., remove trip hazards, preference for non-slip mats). In addition, the program encouraged participants to increase their walking activities. Participants received a daily step monitor (pedometer or Fitbit) to be used during waking hours to record their steps and provide feedback and motivation on the amount of walking activity achieved. The health coach (physiotherapist) provided follow-up telephone coaching using motivational interviewing, problem-solving, goal setting, and action planning to increase physical activity and prevent falls throughout the program. Health coaches also assisted with Fitbit issues, local exercise options, and fall risk management. Calls were made every two weeks for six months, then monthly for the next six months. Health coaches with health coaching experience, extensive training in empowerment-focused coaching, and expertise in behaviour change science and self-determination theory conducted home visits and telephone coaching sessions. Control group Participants in the control clusters received a 12-month nutrition program with a healthy nutrition pamphlet and telephone-based health coaching from The Get Healthy Information and Coaching Service®, provided by the NSW Ministry of Health. They were contacted via telephone every two weeks for six months, then monthly for the next six months.The health coach who was a qualified dietitian helped the participants to set healthy eating goals and discuss strategies for achieving those goals. Participants had the option for continued access to the publicly available Get Healthy service at the conclusion of the study if they wish.Measurements Timepoints Patient characteristics, outcome measures, mediators, and potential confounders were assessed at baseline. Mediators were assessed again at 6 months. The outcome was measured monthly over the final 6 months of the 12-month trial period. Outcomes We will consider cumulative fall rates, recorded with monthly postal calendars over the final 6 months of the trial, to ensure a temporal precedence in the mediation model (i.e., the intervention and changes in the mediators precede the outcome). Mediators We will investigate the following four potential mediators measured at 6 months follow-up: (i) Objective walking measured by daily steps assessed over 7 days using a hip-worn activity monitor (ActiGraph GT3X+ accelerometer). A higher step count indicates higher physical activity levels. (ii) Fear of falling (Falls Efficacy Scale International- [FES-I] Short Form) consists of 7 items that assess an individual’s concern about falling during various daily activities. The scale is designed to be quick and easy to administer while still providing reliable and valid results. Each item is rated on a 4-point scale, from 1 (not at all concerned) to 4 (very concerned), with higher scores indicating greater fear of falling. The Short FES-I is particularly useful in both clinical and research settings where time is limited and helps identify individuals who may benefit from interventions to reduce their fear of falling. (iii) Mobility-related confidence (Modified Gait Efficacy Scale) is a 10-item questionnaire designed to measure walking confidence in older adults, especially in challenging situations. It is rated on a 10-point Likert scale. The scale ranges from 1 (no confidence) to 10 (complete confidence), which means that for each of the 10 items, individuals rate their confidence in performing specific walking tasks under various conditions, with higher scores indicating greater confidence. (iv) Risk-taking behaviour (risk-taking behaviour scale); assesses how frequently individuals engage in or avoid activities that could impact their safety, such as checking shoe soles for slipperiness, hurrying to answer the phone, using handrails on stairs, crossing streets at lights and avoiding crowded areas. Responses are rated on a scale from “Always” to “Never,” with some items reverse-scored to indicate higher risk-taking behaviour. The scores were calculated from a scale with a minimum possible score of 5 and a maximum of 20. A higher score indicates a lower risk-taking behaviour related to falls. This means that a person with a score of 20 is a high-risk taker (not concerned about falling), while a person with a score of 5 is a low-risk taker (very concerned about falling). Confounders We assume that treatment–mediator and treatment–outcome relationships are not confounded by random assignment of treatments. Using the disjunctive cause criterion, we identified potential confounders of the mediator–outcome association. This entailed identifying measured baseline covariates that were thought to be a cause of the outcome, the mediator, or both. Included in the minimally necessary adjustment set are age, sex, living situation, language, total medication use, total co-morbidities, walking aid use, fallen in the past year, self-rated balance, self-rated fear of falling, physical activity at baseline (self-reported physica