Validation of the Self-Assessment for Falls in the Elderly (SAFE) Questionnaire to Assess the Risk of Falls in People Aged 60 or Over.

Not Applicable

Not yet recruiting

• Conditions: Fall Prevention; Falls (Accidents) in Old Age

• Registration Number: NCT06961812
• Lead Sponsor: University Hospital, Montpellier

Brief Summary

This single-centre, cross-sectional diagnostic study will determine whether the new self-completed Self-Assessment for Falls in the Elderly (SAFE) questionnaire can classify fall-risk level (low, moderate, high) in adults ≥ 60 years as accurately as the current clinician-administered international algorithm that mixes yes/no questions with physical tests; to do so, about 300 participants will (1) sign consent, (2) answer the 5-minute SAFE, and (3) undergo guideline assessment and tests by a clinician in the same visit, after which researchers will compare SAFE and clinician results for sensitivity, specificity, and predictive values, examine agreement and correlations, and give immediate risk-based prevention advice:

* Eligibility: men or women ≥ 60 y, stable walking/balance ≥ 1 month, able to read French and perform brief tests; legal guardianship or refusal excludes.

* Participant tasks: complete SAFE; perform Timed Up-and-Go, 4 m gait speed, Five-Times-Sit-to-Stand and Short Physical Performance Battery (SPPB) under supervision

* Benefits/Risks: instant personalised fall-risk feedback; SAFE is risk-free and physical tests carry only minimal supervised exertion.

Falls are a leading cause of injury and death in older adults, and current assessments require trained staff; if SAFE proves equivalent, it could enable large-scale, low-cost self-screening and support future digital monitoring tools for fall prevention.

Detailed Description

This single centre diagnostic accuracy study validates the Self Assessment for Falls in the Elderly (SAFE), an auto-questionnaire that converts the binary items and performance tests of the 2022 World Guidelines for Falls Prevention and Management into visual analogue scales (100 mm each). SAFE is completed without assistance in ≈ 5 min and yields a continuous score that can be stratified into low, moderate, or high fall risk categories. The reference standard is the full guideline algorithm-Three Key Questions, detailed fall history characterisation, Timed Up and Go, Five Times Sit to Stand, 4 m gait speed, and Short Physical Performance Battery-administered by a trained clinician after the participant has returned the SAFE form to prevent incorporation bias.

The study adopts a cross sectional design and will recruit 300 community dwelling adults aged ≥ 60 years attending the Geriatrics Department of Montpellier University Hospital over 6 months. Sample size (300) was calculated for expected sensitivity and specificity of 0.85, 30 % prevalence of moderate to high fall risk, 5 % absolute precision, two sided α = 0.05, and 90 % power. After informed consent and eligibility verification, participants complete a single study visit in the following order: (1) SAFE self completion, (2) guideline interview plus physical tests. No investigational product, device, imaging, or biospecimen collection is involved.

Statistical plan. For every SAFE VAS (visual analogue scale) item, optimal cut points will be derived from non parametric ROC analysis versus its binary analogue. The primary analysis compares the three level SAFE classification with the reference standard using contingency tables and reports sensitivity, specificity, positive and negative predictive values, likelihood ratios, diagnostic odds ratio, and area under the summary ROC curve with 95 % confidence intervals. Secondary analyses will explore Pearson or point biserial correlations between individual SAFE scores and continuous physical test results, apply multivariable ordinal logistic regression to evaluate whether age, sex, or comorbidity modify SAFE performance, and perform decision curve analysis to estimate net clinical benefit across plausible threshold probabilities. Missing data will be handled by complete case analysis.

Data quality and monitoring. Source data will be captured in an electronic case report form with automated range and consistency checks. The CHU Montpellier sponsor will conduct risk based on site and/or remote monitoring and centralised statistical data surveillance. All study documentation will be archived for 15 years in line with French regulations.

Ethics, safety, and dissemination. The protocol has been submitted to a French Comité de Protection des Personnes and the Agence Nationale de Sécurité du Médicament (category 3). Physical testing is performed by certified staff in a dedicated fall prevention clinic with immediate assistance available; adverse events are therefore not expected but will be recorded and reported according to Good Clinical Practice. Results-positive, negative, or inconclusive-will be published in peer reviewed journals and communicated to participants on request. Should SAFE demonstrate acceptable diagnostic accuracy, its simplicity will enable large scale, low cost self screening, integration into mobile health applications, and targeted allocation of fall prevention resources for older adults.

Study Timeline

• Status Verified: 2025-04
• Study First Submitted: 2025-04-29
• Study First Submitted QC: 2025-04-29
• Last Update Submitted: 2025-04-29
• Study Start: 2025-05-02
• Study First Posted: 2025-05-08
• Last Update Posted: 2025-05-08
• Primary Completion: 2026-05-02
• Study Completion: 2026-05-02

Recruitment & Eligibility

• Status: NOT_YET_RECRUITING
• Sex: All
• Target Recruitment: 300

Inclusion Criteria

• People aged 60 or over, judged by the investigator to be in stable clinical condition as regards walking and balance for at least one month,
• Able to perform the necessary physical assessments

Exclusion Criteria

• Unable, according to the investigator, to answer a self-questionnaire,
• Patient's refusal to take part in the study
• Absence of a caregiver if the participant is unable to answer the questionnaire alone
• Unable to understand the information note
• Opposition of the proxy or legal guardian if the participant is under legal protection (guardianship, curatorship, safeguard of justice)
• Subject deprived of liberty (art. L. 1121-6) (by judicial or administrative decision, or forced hospitalization)
• Not affiliated to a French social security scheme or beneficiary of such a scheme (L1121-8-1)
• Participation in other research involving a period of exclusion still in progress (Article L1121-12)

Study & Design

• Study Type: INTERVENTIONAL
• Study Design: SINGLE_GROUP

Primary Outcome Measures

Name	Time	Method
Performance of the SAFE Self-Assessment Questionnaire compared with the global recommendations algorithm (Hetero-administered questionnaire and physical tests). sensitivity	At inclusion	Based on predefined VAS thresholds in the self-questionnaire, we apply the global recommendations algorithm to classify patients into low, moderate, and high risk of falling: High Risk: At least one of the following: Two or more falls in the last year. One fall with injury in the last year. One fall with inability to get up in the last year. Unexplained fall in the last year. Frailty: At least three of the five Fried factors (fatigue, \>5 seconds for 4 meters, \>15 seconds for 5 chair lifts, \>4.5 kg weight loss in one year, sedentary lifestyle).; Moderate Risk: Non-severe fall, fear of falling, or instability while walking/standing, associated with: TUG ≥ 15 seconds, or ≥ 5 seconds to walk 4 meters.; Low Risk: One non-severe fall, fear of falling, or instability, but no frailty and normal TUG (Timed Up and Go) and walking speed.; We compare the statistical performance of SAFE with the reference standard using contingency tables, reporting sensitivity
Performance of the SAFE Self-Assessment Questionnaire compared with the global recommendations algorithm (Hetero-administered questionnaire and physical tests). specificity	At inclusion	Based on predefined VAS thresholds in the self-questionnaire, we apply the global recommendations algorithm to classify patients into low, moderate, and high risk of falling: High Risk: At least one of the following: Two or more falls in the last year. One fall with injury in the last year. One fall with inability to get up in the last year. Unexplained fall in the last year. Frailty: At least three of the five Fried factors (fatigue, \>5 seconds for 4 meters, \>15 seconds for 5 chair lifts, \>4.5 kg weight loss in one year, sedentary lifestyle).; Moderate Risk: Non-severe fall, fear of falling, or instability while walking/standing, associated with: TUG ≥ 15 seconds, or ≥ 5 seconds to walk 4 meters.; Low Risk: One non-severe fall, fear of falling, or instability, but no frailty and normal TUG and walking speed.; We compare the statistical performance of SAFE with the reference standard using contingency tables, reporting specificity
Performance of the SAFE Self-Assessment Questionnaire compared with the global recommendations algorithm (Hetero-administered questionnaire and physical tests). positive predictive values	At inclusion	Based on predefined VAS thresholds in the self-questionnaire, we apply the global recommendations algorithm to classify patients into low, moderate, and high risk of falling: High Risk: At least one of the following: Two or more falls in the last year. One fall with injury in the last year. One fall with inability to get up in the last year. Unexplained fall in the last year. Frailty: At least three of the five Fried factors (fatigue, \>5 seconds for 4 meters, \>15 seconds for 5 chair lifts, \>4.5 kg weight loss in one year, sedentary lifestyle).; Moderate Risk: Non-severe fall, fear of falling, or instability while walking/standing, associated with: TUG ≥ 15 seconds, or ≥ 5 seconds to walk 4 meters.; Low Risk: One non-severe fall, fear of falling, or instability, but no frailty and normal TUG and walking speed.; We compare the statistical performance of SAFE with the reference standard using contingency tables, reporting positive predictive values,
Performance of the SAFE Self-Assessment Questionnaire compared with the global recommendations algorithm (Hetero-administered questionnaire and physical tests). negative predictive values	At inclusion	Based on predefined VAS thresholds in the self-questionnaire, we apply the global recommendations algorithm to classify patients into low, moderate, and high risk of falling: High Risk: At least one of the following: Two or more falls in the last year. One fall with injury in the last year. One fall with inability to get up in the last year. Unexplained fall in the last year. Frailty: At least three of the five Fried factors (fatigue, \>5 seconds for 4 meters, \>15 seconds for 5 chair lifts, \>4.5 kg weight loss in one year, sedentary lifestyle).; Moderate Risk: Non-severe fall, fear of falling, or instability while walking/standing, associated with: TUG ≥ 15 seconds, or ≥ 5 seconds to walk 4 meters.; Low Risk: One non-severe fall, fear of falling, or instability, but no frailty and normal TUG and walking speed.; We compare the statistical performance of SAFE with the reference standard using contingency tables, reporting negative predictive values
Performance of the SAFE Self-Assessment Questionnaire compared with the global recommendations algorithm (Hetero-administered questionnaire and physical tests). likelihood ratios	At inclusion	Based on predefined VAS thresholds in the self-questionnaire, we apply the global recommendations algorithm to classify patients into low, moderate, and high risk of falling: High Risk: At least one of the following: Two or more falls in the last year. One fall with injury in the last year. One fall with inability to get up in the last year. Unexplained fall in the last year. Frailty: At least three of the five Fried factors (fatigue, \>5 seconds for 4 meters, \>15 seconds for 5 chair lifts, \>4.5 kg weight loss in one year, sedentary lifestyle).; Moderate Risk: Non-severe fall, fear of falling, or instability while walking/standing, associated with: TUG ≥ 15 seconds, or ≥ 5 seconds to walk 4 meters.; Low Risk: One non-severe fall, fear of falling, or instability, but no frailty and normal TUG and walking speed.; We compare the statistical performance of SAFE with the reference standard using contingency tables, reporting likelihood ratios
Performance of the SAFE Self-Assessment Questionnaire compared with the global recommendations algorithm (Hetero-administered questionnaire and physical tests). diagnostic odds ratio	At inclusion	Based on predefined VAS thresholds in the self-questionnaire, we apply the global recommendations algorithm to classify patients into low, moderate, and high risk of falling: High Risk: At least one of the following: Two or more falls in the last year. One fall with injury in the last year. One fall with inability to get up in the last year. Unexplained fall in the last year. Frailty: At least three of the five Fried factors (fatigue, \>5 seconds for 4 meters, \>15 seconds for 5 chair lifts, \>4.5 kg weight loss in one year, sedentary lifestyle).; Moderate Risk: Non-severe fall, fear of falling, or instability while walking/standing, associated with: TUG ≥ 15 seconds, or ≥ 5 seconds to walk 4 meters.; Low Risk: One non-severe fall, fear of falling, or instability, but no frailty and normal TUG and walking speed.; We compare the statistical performance of SAFE with the reference standard using contingency tables, reporting diagnostic odds ratio

Secondary Outcome Measures

Name	Time	Method
VAS thresholds for the self-questionnaire	At inclusion	In this section, for each binary question on the hetero-questionnaire or involving a physical test, we aim to determine a VAS threshold from the self-questionnaire for the corresponding question. The hetero-questionnaire measure is used as a reference measure: fear of falling, unsteadiness when standing up, unsteadiness when walking, getting up easily from a chair without using hands, walking speed, fatigue, reduction in physical activity.
Concordance analyses	At inclusion	We wish to evaluate the concordance between the SAFE self-questionnaire responses and the global recommendations algorithm hetero-assessment via the Gwet coefficient for qualitative variables and the intra-class correlation coefficient (ICC) for quantitative variables.
Point-biserial correlations	At inclusion	We wish to evaluate the correlation between the VAS of the self-questionnaire and the dichotomous variables of the hetero-questionnaire (point-biserial correlations).
Pearson correlation analysis	At inclusion	We wish to evaluate the correlation between the VAS of the self-questionnaire and the physical tests of the hetero-questionnaire (Pearson correlations).

Trial Locations

Locations (1)

University Hospital; 🇫🇷 Montpellier, Hérault, France