Psychometric properties of the Chinese version of the falls risk perception scale among community-dwelling elderly

Abstract

Background

Effective fall prevention requires accurate risk perception assessment. There remains a paucity of validated measurement tool adapted to China’s cultural context. This cross-cultural validation study aimed to evaluate the psychometric properties of the Chinese version of falls risk perception questionnaire (RPQ) among community-dwelling elderly.

Methods

Using a standard translation-backward method, the original English version of RPQ was translated into Chinese (Mandarin). Between March 2023 and October 2023, a convenient sampling of 300 elderly individuals was recruited from one community in Quzhou, China to test internal consistency, split-half reliability, content validity, structure validity, and the receiver operating characteristic (ROC) of this scale.

Results

A total of 272 patients (91.33%) completed questionnaires. The Chinese version of RPQ includes 16 items in total. Exploratory factor analysis extracted three factors, with a cumulative variance contribution rate of 63.32%. It achieved Cronbach’s alphas 0.874, McDonald’s omega coefficient 0.884 and the test-retest reliability 0.907, respectively, for the full scale. The scale-level content validity index was 0.875, and the item-level content validity index ranged from 0.833 to 1.000. Using the Falls Efficacy Scale-International (FES-I) as a benchmark, the correlation coefficient was 0.669. ROC analysis examined the correspondence between perceived risk and prior fall experience, with an area under the curve (AUC) of 0.562 (95% CI 0.487–0.637). However, its accuracy in predicting actual fall risk is limited as it is primarily intended for assessing perceived risk.

Conclusion

The Chinese version of RPQ demonstrated acceptable reliability and validity among community-dwelling elderly. It can be used to evaluate fall risk perception among this population. However, the scale should not be used to predict actual fall events, as ROC analysis indicated limited predictive accuracy.

Background

Falls represent a major public health concern among older adults [1]. According to the World Health Organization (WHO) [2], falls account for approximately 646,000 annual fatalities worldwide, making them the second-leading cause of death from unintentional injuries. In China, falls constitute the predominant cause of unintentional injury-related mortality among adults aged ≥ 65 years [3]. Annually, no fewer than 20 million older adults experience approximately 25 million fall incidents, resulting in direct medical expenditures surpassing 5 billion RMB and total economic burdens approaching 80 billion RMB [4]. Falls not only significantly compromise the health-related quality of life in elderly populations but also impose substantial pressures on both familial support systems and healthcare infrastructure. Given that community- and family-based care represent the primary care practice in China, the systematic identification and evaluation of fall-associated risk factors are critical for fall prevention strategies.

Current research has extensively examined risk factors for falls in older adults, with diverse discussions on their underlying determinants [5, 6]. However, increasing attention has been directed toward fall risk perception, a psychological construct reflecting an individual’s subjective appraisal of both the likelihood of falling and the potential severity of its consequences. Behavioral change theory posits that an individual’s subjective risk perception is a key driver of behavioral modification [7]. Specifically, theoretical frameworks such as the Health Belief Model conceptualize risk perception as a cognitive determinant that motivates individuals to adopt preventive health behaviors when perceived susceptibility and severity are high [8]. Empirical evidence indicates that heightened fall risk perception in older adults correlates with increased engagement in proactive fall prevention strategies [9]. Conversely, inaccurate risk perception—either underestimation or excessive overestimation—may lead to risk-taking behaviors or activity restriction, both of which can paradoxically increase fall risk.

Current fall risk assessment tools widely employed in clinical practice—such as the Morse Fall Scale, Thomas Fall Risk Assessment Tool, and Johns Hopkins Fall Risk Assessment Tool [10]—primarily emphasize clinician-administered evaluations while overlooking self-perceived risk assessment and failing to incorporate active patient engagement. Although several self-reported fall risk instruments exist for older adults, such as the Fall Risk Questionnaire [11], the STEADI Fall Risk Self-Assessment Tool [12], and the Falls Efficacy Scale-International [13], these instruments predominantly assess functional limitations, objective risk factors, or fear of falling rather than cognitive appraisal of fall risk. Importantly, fear of falling and fall risk perception represent related but distinct constructs: fear reflects an emotional response, whereas risk perception involves a cognitive judgment of susceptibility and potential harm.

In 2019, Gravesande and colleagues developed a Falls Risk Perception Questionnaire (RPQ) which is a multifaceted assessment tool for evaluating fall risk perception among community-dwelling older adults [6]. Although the RPQ was initially validated in an older adult population with diabetes, diabetes served as a high-risk exemplar rather than a defining characteristic of the construct. The RPQ is not disease-specific, and its items reflect experiences common to community-dwelling older adults. The underlying conceptual framework of the RPQ is theoretically applicable to general older adult populations.

From a cultural perspective, fall risk perception among Chinese older adults may be shaped by collectivist values, family-centered decision-making, and cultural norms emphasizing endurance and modesty in reporting vulnerability. These sociocultural characteristics may influence how older adults perceive, interpret, and report fall-related risk. As a result, direct application of Western-developed instruments without cultural adaptation may compromise measurement validity. To date, no validated Chinese-language instrument specifically targets fall risk perception as a cognitive construct distinct from fear of falling or functional impairment.

Therefore, we conducted a cross-cultural adaptation of the RPQ through rigorous translation into Chinese, followed by evaluation of its psychometric properties. This study aims to establish a culturally appropriate and psychometrically sound instrument for assessing fall risk perception among Chinese community-dwelling older adults, thereby supporting theory-driven and patient-centered fall prevention strategies.

Methods

Study design and participants

We conducted a cross-sectional study between March 2023and October 2023 in a urban community (residential compound) in Quzhou City, Zhejiang Province, China, which is classified as a third-tier city (there are four tiers in China with the first-tier as most developed), with economic development, population size, and urban infrastructure below provincial averages. In China, a residential compound refers to a planned and enclosed housing area that typically includes several apartment buildings, shared green spaces, and various community facilities. It’s a common and dominant form of urban housing in Chinese cities. Using convenience sampling, we enrolled 300 community-dwelling older adults who met the following inclusion criteria: (1) age ≥ 60 years; (2) fluent verbal communication in Mandarin; and (3) willing to participate voluntarily. Exclusion criteria were: (1) post-stroke hemiplegia; (2) acute or life-threatening medical conditions; and (3) psychosis. The first author obtained written informed consent after explaining the study objectives, potential benefits, and risks. The convenience sampling strategy and single-community recruitment may result in a sample that is not representative of the broader population of community-dwelling older adults, particularly those who are less socially active, have poorer health status, or reside in different geographic or socioeconomic contexts.

Instrumentation

Demographic characteristics

The general information was self-designed based on the literature review including gender, age, education level, whether they had fallen in the last one year, whether they used crutches and whether they lived alone in the community.

Falls risk perception questionnaire (RPQ)

The RPQ was developed by Janelle et al. through systematic integration of the Health Belief Model with established risk perception frameworks [6]. This 20-item instrument assesses fall risk perception in community-dwelling older adults across five validated dimensions: risk-perception of falling (item 1), risk factors (items 2–6), internal/external factors (items 7–11), individual perceptions (items 12a/b-16) and self-efficacy (items 17–19). Responses are recorded using a 7-point Likert scale (1="strongly disagree” to 7="strongly agree”), yielding total scores ranging from 20 to 140, where higher scores indicate greater perceived fall risk. The initial reliability of the RPQ was assessed by internal consistency (Cronbach’s alphas = 0.78), with retest reliabilities ranging from 0.78 to 0.82.

Falls efficacy Scale-International (FES-I)

The FES-I [13], originally developed by Kempen et al., is a 16-item instrument assessing concern about falling during both indoor and outdoor activities. Each item is scored on a 4-point Likert scale (1="not at all concerned” to 4="very concerned”), yielding a total score range of 16–64, with higher scores indicating greater fear of falling. Chan et al. [14] conducted cultural adaptation and validation of the Chinese version. The Chinese version of the FES-I has a Cronbach’s alpha coefficient of 0.921, and a retest reliability of 0.906. We employed the Chinese FES-I as the standard for evaluating convergent validity of the RPQ through scale-to-scale correlation analysis.

Translation of the RPQ

The scale was translated according to the Brislin [15] translation model after authorization from the original authors via email, including forward-backward translation, harmonization meetings, expert panel review, and pre-testing with the target population. Similar rigorous procedures have been applied in other cross-cultural validation studies [16, 17]. The detailed process was as follows:

• Stage 1: Translation: Two nursing graduate students independently translated the scale into the Chinese versions (T1 and T2), and the integrated Chinese version (T3) was formed after discussion by three research team members (one with master's degree in nursing, one with doctoral degree in nursing in the US, and one nurse with further study in the US) during harmonization meetings.

• Stage 2: Back-translation: Two English graduate students were selected to back-translate the Chinese version of T3 into the English versions of BT1 and BT2, and the research team compared the two back-translated versions with the original scale during harmonization meetings, and iteratively revised and back-translated the Chinese version of T3 until it was consistent with the meaning of original scale (T4). None of the back-translators had any knowledge of the original scale. The BT1, BT2, T4, and the translation issues were sent to the original authors for confirmation.

• Stage 3: Cultural adaptation：We conducted a rigorous cultural adaptation of the translated scale (T4) through expert panel review involving six multidisciplinary specialists (clinical nursing, community healthcare, gerontological nursing, nursing research, injury prevention, and health education) with 29.5±8.5 years of professional experience. Using Delphi methodology, the panel evaluated semantic clarity, content relevance, cultural appropriateness, and conceptual equivalence, achieving > 90% inter-rater agreement through iterative reviews to produce the culturally adapted version T5.

• Stage 4: Pre-test: We conducted a pre-test in one community in Quzhou - an aging neighborhood with approximately 6,000-7,000 residents including over 30% aged ≥65 years - where we recruited 30 elderly participants through convenience sampling. During face-to-face interviews, participants were asked to describe their understanding of each item in their own words, and interviewers probed for clarity, interpretation, and cultural relevance. Feedback was collected and used to refine items, ensuring comprehension and conceptual equivalence in the Chinese context.

Data collection

The first author (investigator) reached out to an elderly activity centre in the community and got approval for investigation. The elderly activity centre regularly organized indoor activities such as playing chess which could attract a lot of elderly people to visit. The investigator explained the study objectives, obtained written informed consent and distributed paper questionnaires using uniform instructions. The investigator also went to the community public green space to recruit more participants. Participants self-completed the questionnaires with the help of the investigator if they had questions. The investigator assisted with completion for those with visual impairment, illiteracy, or other limitations, but did not provide leading information. We did not conduct cognition assessments when recruiting participants. However, all participants were required to be able to communicate and provide informed consent. Face-to-face interviews were conducted to ensure understanding of the survey items. The investigator conducted immediate quality checks for missing responses. From the initial 300 questionnaires, we excluded invalid responses based on incomplete surveys and those who were not willing to finish questionnaire, resulting in 272 valid responses for final analysis.

Data analysis

Statistical analyses were conducted using SPSS 24.0 (IBM Corp.). Descriptive statistics for categorical variables were presented as frequencies and percentages (n, %). All statistical tests employed a two-tailed a level of 0.05 for significance. We conducted psychometric analyses as follows.

Item analysis

We calculated the Pearson correlation coefficient between each item and the total scale score. Items with correlation coefficients below 0.40 were removed [18]. The total scores of the 272 Chinese RPQ questionnaires were rank-ordered, and were stratified into high- and low-scoring subgroups (top and bottom 27%, respectively). Independent two-sample t-tests were performed to compute the CR values for each item, comparing responses between the high- and low-scoring subgroups. Items were retained only if they met the following thresholds: CR > 3.00 with a statistically significant between-group difference (P < 0.05) [18].

Content validity

Six experts (who also served as cultural adaptation expert panel) evaluated the relevance of each item to the fall risk perception in older adults using a 4-point Likert scale (1 = not relevant, 2 = weakly relevant, 3 = strongly relevant, 4 = very relevant). Item-level (I-CVI) and scale-level (S-CVI) content validity indices were computed. For each item, the number of experts with a score of 3 or 4 divided by the total number of experts participating in the evaluation is I-CVI. The number of items rated as 3 or 4 by all experts divided by the total number of items yields S-CVI. The scale was deemed to exhibit good content validity if it met the following thresholds: I-CVI > 0.78 and S-CVI > 0.80 [19].

Structural validity

Exploratory factor analysis (EFA) was conducted using principal component analysis with varimax rotation, when sampling adequacy was confirmed (KMO > 0.8) and Bartlett’s test of sphericity reached significance (χ², p < 0.001). The scale demonstrated good structural validity when all items exhibited primary factor loadings > 0.4, cross-loadings were < 0.4, and cumulative variance exceeded 50% [19]. Then we did confirmatory factor analysis (CFA) based on the structure from EFA. The model fit were evaluated according to the following standards: standardized root mean square residual (SRMR) < 0.08, comparative fit index (CFI) > 0.90, Tucker-Lewis index (TLI) > 0.90 [20].

Convergent validity

Convergent validity indicates the degree to which an instrument correlates with other measures assessing related constructs. The relation between the Chinese version RPQ total score and FES-I total score were calculated.

Reliability analysis

Reliability was assessed through internal consistency (Cronbach’s α and McDonald’s omega) and test-retest reliability, with coefficients exceeding 0.80 and 0.70, respectively, demonstrating excellent measurement stability and reproducibility [21].

Receiver operating characteristic (ROC) analysis

Although the RPQ is a perception-based instrument, ROC analysis was conducted exploratorily to examine the correspondence between perceived fall risk and history of falls during the past 12 months. This analysis does not imply diagnostic or predictive validity. The area under the ROC curve (AUC) was calculated with AUC > 0.7 as good performance.

Results

Translation and cultural adaptation

During the cultural adaptation stage, the modifications and reasons are provided in Table 1. During the pre-test stage, some elderly people mentioned that item 11 (“I feel that getting older increases my risk of falling”) and item 12a (“I feel that people my age are more likely to fall than people who are younger”) were similar in meaning and they felt reluctant to answer twice. After discussion with the research team, item 12a was deleted. The original scale was based on a 7-point Likert scale, ranging from “strongly disagree” to “strongly agree”. However, some older adults found it too granulated and were unable to express themselves accurately. After discussion with research team, it was changed to a 5-point Likert scale, with anchors of 1 (strongly disagree), 2 (disagree), 3 (neutral/uncertain), 4 (agree), and 5 (strongly agree). This is supported by prior methodology research which suggested that 5-point and 7-point Likert scales often yield comparable reliability and construct validity [22].

Table 1.

Revisions in cultural adaptation stage and item analysis

Original item	Revised item	Reason
Item 9 “Fall prevention messages from doctors, physical therapists, or health care workers will affect the incidence of falls”	“Fall prevention messages from community nurses or health care workers will affect the incidence of falls”	Generally, the community nurses who give fall education to the elderly in China
Item 10 “My cultural or religious beliefs affect the incidence of falls”	“Taboos against mentioning falls affect the incidence of falls”	Some elderly people are superstitious and they believe that people who often mention “falling” are more likely to fall in Chinese culture
Item 19 “I feel confident that I will not fall when moving around in my house”	“I feel confident that I will not fall when moving around at home”	Differences in building structure in China
Item 1 “I feel that I am at risk for having a fall”	Deleted due to low item–total correlations	This item assessed a general, overall personal risk. Its deletion does not compromise the content validity as other retained items assess specific risk factors. Also, admitting personal risk may be seen as pessimistic or culturally undesirable
Item 5 “Using a cane or walker may increase risk for falling”	Deleted due to low item–total correlations	Less relevant in Chinese community-dwelling elderly. Older adults who require mobility aids are less likely to live independently and engage in outdoor activities
Item 18 “I feel confident that I will not fall when walking outside in poor weather”	Deleted due to low item–total correlations	In Chinese residential habits, elderly participants may avoid going outside in poor weather, so they rarely experience this situation. Responses may cluster around neutral options, reducing variability

Characteristics of participants

A total of 300 elderly people were recruited. Among them, 272 completed the questionnaire, resulting in a response rate of 91.33%. The majority of community- dwelling older adults were aged 60–70 years (65.4%), females (64.3%), and had an educational of middle school or lower (65.3%). Further details are presented in Table 2.

Table 2.

The characteristics of the participants (N = 272)

Characteristic	Frequency	%
Sex
Male	97	35.7%
Female	175	64.3%
Age
60–70 years old	178	65.4%
71–80 years old	67	24.6%
Above 80 years old	27	9.9%
Education
Middle school and below	179	65.3%
High school or equivalent	67	24.5%
College or above	28	10.2%
Whether fall in the last 1 year
Yes	77	28.3%
No	195	71.7%
Use of crutches
Yes	9	3.3%
No	263	96.7%
Live alone
Yes	10	3.6%
No	262	96.4%

Item analysis

Pearson correlation analyses were conducted between each item of the Chinese version RPQ and the total score. The correlation coefficients for items 1, 5, and 18 with the total score were 0.189, 0.342, and 0.241 respectively (< 0.400), so these three items were removed from further analysis (Table 1). The correlation coefficients of the remaining items with the total score ranged from 0.419 to 0.743 (all P < 0.01). Participants were divided into a high-scoring group (≥ 79 points) and a low-scoring group (≤ 66 points) on the Chinese version RPQ. The CR values of each item ranged from 7.020 to 18.488 (all above 3.00 and all P < 0.01).

Content validity

The results indicated that the scale - level content validity index (S - CVI) reached 0.875. Item - level content validity indices (I - CVIs) ranged from 0.833 to 1.00. Collectively, these findings suggest that the scale demonstrates excellent content validity.

Construct validity

The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy reached 0.880, exceeding the recommended threshold of 0.800. The Bartlett’s test of sphericity yielded a χ² value of 2386.761 (P < 0.001), indicating that the data were suitable for EFA. Principal component analysis with varimax orthogonal rotation was employed, with a criterion of eigenvalue greater than 1. EFA extracted three distinct factors. Cumulatively, these factors accounted for 63.320% of the total variance, surpassing the 50% benchmark. Specifically, items 2–4, 6, 11, 12b, 14–16 loaded onto factor 1; items 7–10 loaded onto factor 2; and items 13, 17, and 19 loaded onto common factor 3. Factor loadings for each item on its respective factor ranged from 0.512 to 0.908, all exceeding 0.400, with no evidence of cross-loading. Based on these findings, the scale retained three factors: individual fall risk perception (9 items), extrinsic fall risk influence (4 items), and fall risk self-efficacy (3 items). In total, 16 items were retained (Table 3). In the CFA, we built a three-factor model. This model showed an acceptable fit: SRMR = 0.056, CFI = 0.902 and TLI = 0.905.

Table 3.

Factor - item matrix of the Chinese version RPQ (n = 272)

Items	Factor1	Factor2	Factor3
14. If I had a fall, it would result in serious consequences	0.789	0.198	0.101
6. Being physically active may increase your risk for falling	0.745	0.197	-0.068
11. I feel that getting older increases my risk of falling	0.683	0.097	0.318
2. A fall in the past year may increase your risk for falling.	0.672	0.203	-0.127
4. Difficulty moving around the house or community may increase your risk for falling.	0.651	0.243	0.285
3. Weakness in your legs may increase your risk for falling.	0.650	0.088	-0.126
16. Avoiding a fall is just as important as my other health concerns	0.574	0.359	0.342
12b. I am less likely to fall when compared to other people my age.	0.523	0.194	0.171
15. My attitude about aging influences my risk of falling	0.512	0.384	0.507
8. I believe that my risk of falling is influenced by Information given to me by the media	0.205	0.908	0.102
9. Fall prevention messages from community nurses or health care workers will affect the incidence of falls	0.256	0.889	0.067
7. I believe that my risk of falling is influenced by Information given to me by my friends and family	0.239	0.866	0.104
10. Taboos against mentioning falls affect the incidence of falls	0.340	0.592	0.347
19.I feel confident that I will not fall when moving around at hone	0.044	0.080	0.868
17. I feel confident that I can maintain my balance when I walk outside in the community	0.009	0.040	0.841
13. I feel that I am able to protect myself from falling	0.053	0.172	0.732

Items 1, 5, 12a, and 18 were removed due to redundancy or low item–total correlations (< 0.40). These items may have limited relevance or interpretability in the Chinese community-dwelling elderly population

Convergent validity

The total score of the Chinese version of the RPQ was positively correlated with the total score of the Chinese version of the FES-I (r = 0.669, p < 0.001) indicating adequate convergent validity.

Reliability and ROC curve analysis

The results showed that Cronbach’s alpha coefficient of the Chinese version RPQ was 0.874, with subscales achieving 0.863, 0.896 and 0.805 respectively. The McDonald’s omega coefficient was 0.884 for total scale, with subscales achieving 0.872, 0.903 and 0.818. To assess test-retest reliability, 30 elderly people were recruited. They re-administered the questionnaire via face-to-face interviews after two weeks of the initial administration. The test-retest reliability coefficient for the overall scale was 0.907. The area under the ROC curve (AUC) was 0.562 [95% confidence interval (CI): 0.487, 0.637] (Fig. 1).

Fig. 1.

Area under the ROC curve

Discussion

The Chinese version of RPQ serves as a reliable and valid instrument for evaluating fall risk perception in community-dwelling elderly. This scale enables elderly people to take a proactive stance in fall prevention and facilitates a comprehensive understanding of elderly people’s perceived fall risk levels. Equipped with this knowledge, community nurses and health workers can provide tailored fall prevention and intervention.

There are certain discrepancies between the factors extracted in this study and the original scale. These differences may be attributed to disparities in cultural backgrounds, religious beliefs, and aging situation between Asian and Western countries. The original scale consists of five dimensions: risk-perception of falling (item 1), risk factors (items 2–6), internal/external factors (items 7–11), individual perceptions (items 12a/b-16) and self-efficacy (items 17–19), with a total of 20 items. During the pre-testing phase and item analysis, items 12a, 1, 5 and 18 were removed, leaving a total of 16 items. Exploratory factor analysis was then conducted on these 16 items, resulting in three factors: individual fall risk perception (items 2–4, 6, 11, 12b, and 14–16), extrinsic fall risk influence (items 7–10), and fall risk self-efficacy (items 13, 17 and 19). The original second dimension (items 2–6) pertains to individual fall risk factors and the original fourth dimension (items 12b, 14–16), along with item 11, represents the individual’s perception of fall risk. These elements encompass elderly’s perceptions of falls and fall risk factors, and were thus combined into a single dimension. Items 7–10 incorporate influences from the media, healthcare professionals, family and friends, and certain beliefs. These factors are all extrinsic influences on fall risk and correspond to the second dimension. Items 13, 17, and 19 all relate to older adults’ judgments regarding their confidence levels in avoiding falls and are categorized under the third dimension.

In this study, the accuracy of the ROC curve analysis was employed to evaluate the fall risk in older adults, with a history of falls within the past one year serving as the criterion. The results demonstrated that the area under the ROC curve (AUC) of the Chinese version of the RPQ for assessing the all risk in community-dwelling older adults was 0.562. The low AUC confirms that perceived fall risk does not necessarily match objective fall experience. This reinforces the conceptual distinction between cognitive appraisal (risk perception) and actual fall outcomes. This finding suggested that the diagnostic utility of this scale as a tool for evaluating fall risk in elderly is relatively low. This might be explained that this scale can assess the fall risk perception of the elderly, but it cannot be used as an assessment of the actual fall risk of the elderly, and the risk perception does not represent the actual fall risk of the elderly. Our study revealed that the actual fall risk experienced by older adults often diverges from their perceived fall risk, a finding consistent with previous international research [23–25]. Various factors including demographic and sociological aspects, attitudes towards aging, physical activity function, cognitive status, and the presence of depression may lead to either overestimate or underestimate their fall risk [26, 27]. Community nurses and health workers should prioritize the early assessment of fall risk perception in elderly. By providing personalized health education and tailored interventions, it may be possible to reduce the incidence of falls among this population.

Our findings highlight a notable discrepancy between perceived fall risk, as measured by the RPQ, and actual fall history. This discrepancy has important implications when interpreted through behavioral change theories. According to the Health Belief Model, individuals are more likely to engage in preventive behaviors when they perceive themselves to be susceptible to a health threat and when they recognize the potential severity of its consequences. The observed mismatch between perceived and actual risk suggests that some older adults may underestimate their vulnerability, potentially reducing engagement in fall-prevention behaviors [28]. Conversely, those who overestimate risk may experience unnecessary fear, which can also influence behavior. These results underscore the importance of interventions that not only educate older adults about objective fall risk factors but also calibrate their subjective risk perception to support adaptive preventive actions [29]. By explicitly linking perceived risk to behavioral intentions, this study contributes to a theoretical understanding of how cognitive appraisal shapes health behaviors in aging populations. Furthermore, community nurses and healthcare providers can use the RPQ to assess perceived fall risk among older adults. Although the scale does not directly guide interventions, understanding individuals’ risk perception allows practitioners to tailor education and fall-prevention strategies to address underestimation or overestimation of risk, thereby promoting more effective, personalized care.

This study has several limitations. First, the sample was from one community which may limit the generalizability. Second, we did not conduct cognition assessments when recruiting participants. However, we recruited elderly with fluent verbal communication and conducted face-to-face investigation. They were able to ask questions for clarification and we did not find their difficulty in understanding the questions. We acknowledge that residual confounding may remain and future studies should include formal cognitive assessments to better account for this potential source of bias. Third, due to resource limitations, we did not invite native English speakers at the back-translation stage, but we invited two graduate students with professional level of English. This may have limited the detection of subtle semantic or idiomatic discrepancies between the Chinese and original English versions, potentially affecting linguistic and conceptual equivalence. However, this risk was partially mitigated through expert panel review and pilot testing among older adults. Fourth, we conducted the investigation in an elderly activity centre where we may reach out to healthier participants and lead to selection bias. Fifth, no statistical cross-cultural equivalence testing, such as differential item functioning (DIF) analysis, was performed in this study. We changed the Likert scale from 7-point to 5-point. Future research should incorporate DIF analysis to further ensure measurement equivalence across populations and consider testing measurement invariance across different response formats. Finally, we conducted EFA followed by CFA for structural validity analysis which may carry a risk of data-driven capitalization on chance. Future studies should validate the CFA using an independent sample. In addition, increasing the sample size and recruiting a more diverse and representative population would strengthen the generalizability of the findings and provide further evidence of the reliability and validity of the Chinese version of the RPQ.

Conclusion

The Chinese version of RPQ exhibits sound reliability and validity. Community nurses and healthcare workers can use this scale to enhance older adults’ awareness of fall risks, prompting them to take proactive steps in fall prevention. However, the scale should not be used to predict actual fall events, as ROC analysis indicated limited predictive accuracy. These findings underscore the distinction between perceived and objective risk, highlighting the RPQ’s role in assessing perceived risk rather than forecasting falls.

Authors’ contributions

**Guanjun Bao and Yiming An collected data. Guanjun Bao did analysis and drafted the manuscript. Yuanfei Liu and Ye Luo reviewed the manuscript. All authors agree to submit this paper.**.

Funding

There is no funding in this study.

Data availability

Data is available upon reasonable request from corresponding author.

Declarations

Ethics approval and consent to participate

All participants provided written informed consent prior to enrollment. The research adhered to the ethical principles of the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.