Examining frailty phenotypes of community-dwelling older adults in Taiwan using the falls risk for older people in the community – Taiwan version (Tw-FROP-Com)

Introduction

Falls are the second leading cause of death from unintentional injuries worldwide [1] and among adults aged 65 and older in Taiwan. Globally, over 80% of the 684,000 annual fall-related deaths occur in low- and middle-income countries. Adults aged 60 and older have the highest rates of fatal falls, with approximately 37.3 million falls each year severe enough to require medical attention. In Taiwan, one in six adults aged 65 and older has experienced a fall, and one in twelve has sought medical care due to fall-related injuries [2].

Falls among older adults often lead to functional decline due to disability, increasing their dependence on others and raising the risk of becoming a burden on caregivers or facing premature admission to nursing homes [3]. Additionally, falls can significantly impact physical and mental health, social functioning, and overall quality of life [4]. In response to these challenges, the World Health Organization (WHO) emphasizes the need for comprehensive fall prevention strategies, including research, education, training, safer environments, and policy development aimed at reducing fall risk [5].

Many factors contribute to falls suffered by older people, such as frailty [6], balance dysfunction, mobility problems, impaired vision, lack of vitamin D, fear of falling, depression, side effects of specific medications, and the presence of home hazards [7]. Frailty is a clinical condition resulting from age-related declines in physiological reserves. Consequently, frail older adults, including persons who have entered the prefrail stage, are likely to experience recurrent falls [7, 8]. Recently, frailty in the older population has attracted much attention because of studies showing its link to negative outcomes such as falls, institutionalization, hospitalization, incident disability, and death [9, 10]. Notably, the frailty phenotype has been associated with a greater risk of falls in women aged 55 and older [11]. A prospective cohort study of older men indicated that a simple frailty index—comprising weight loss, the inability to rise from a chair, and poor energy (Study of Osteoporotic Fractures index)—predicted the risk of falls, disability, fractures, and mortality as effectively as the more complex Cardiovascular Health Study (CHS) index. Components of the CHS include unintentional weight loss, poor grip strength, poor energy, slowness, and a low level of physical activity [12]. Another longitudinal cohort study used the CHS index and the Women’s Health Initiative (WHI) phenotypes to predict fall risk in older women. Higher fall rates (Hazard Ratio (HR) = 1.48, p = 0.003) were observed in the WHI frailty phenotype group compared with the CHS group [13]. The WHI phenotype, which does not require direct physical performance measurements, may be useful for the general older population [13]. In Japan, a cross-sectional study examined the prevalence and associated factors of cognitive frailty and cognitive frailty-related falls in community-dwelling older adults. Age, chronic disease, the Timed Up and Go Test, and the Council on Nutrition Appetite Questionnaire were significantly associated with cognitive frailty and cognitive frailty-related falls [14].

Numeric frailty assessment tools have been developed and have demonstrated their ability to predict the risk of falling. However, there is no gold standard for assessing frailty [15]. Currently, a frailty index that does not require a physical examination is needed to predict fall risk and guide preventive interventions for the broader older population. Simple measurement tools can assess frailty as accurately as complex ones. Fried’s frailty phenotypes meet these criteria and have been widely adopted in several studies [16–18].

According to Fried, frailty comprises five elements: weakness, slowness, exhaustion, low physical activity, and unintentional weight loss [19]. In most studies on frailty and fall risk, frailty has been treated as a categorical scale, such as non-frail/frail or non-frail/pre-frail/frail [20–23]. However, few studies have examined the correlation between individual frailty phenotypes and fall risk in Taiwan. Moreover, it remains unclear which frailty phenotype contributes most to the fall risk among community-dwelling older adults in Taiwan.

For fall risk assessment, we used the Falls Risk for Older People in the Community (FROP-Com), which has been widely applied in various older populations worldwide [24–27]. FROP-Com is simple to use and does not require physical measurements. As the first research team to translate and validate the FROP-Com into the Taiwan version (Tw-FROP-Com), our study effectively evaluated fall risk among Taiwanese older adults [28]. Therefore, this study aimed to explore the relationship between Fried’s frailty phenotype and fall risk among community-dwelling older adults in Taiwan using the Tw-FROP-Com.

Results

Of the 425 subjects screened for a high risk of falling, only 375 participants who completed the frailty and fall risk assessments were included in the analysis. The sociodemographic characteristics of the participants and their fall risk are presented in Table 1. Participants had a mean age of 75.4 ± 6.8 years, were predominantly women (76.0%), and the majority had a low level of education (23.7% were illiterate with no formal education, and 46.5% had only an elementary school education, totaling 70.2%). Additionally, most participants were from the Hokkien ethnic group (85.1%), lived with a spouse (55.7%), lived with others (83.9%), were cared for by others (59.6%), and had no history of falls (67.3%). Regarding frailty components, most participants did not exhibit weakness (52.4%), slowness (61.1%), exhaustion (74.3%), or low physical activity (20.0%).

We used the Tw-FROP-Com to evaluate fall risk. There was no significant difference in Tw-FROP-Com scores between males and females (p = 0.388). Participants over 75 years old had a significantly higher fall risk compared to those aged 65–74 years (p = 0.001). Illiterate individuals had significantly higher Tw-FROP-Com scores than the other two groups (p = 0.022). Married participants had lower Tw-FROP-Com scores than unmarried participants (p = 0.007). Participants with a history of falling in the past year had significantly higher Tw-FROP-Com scores than those without a history of falling (p < 0.001). However, no significant difference was observed between participants living with others and those being cared for by others.

Among the frailty phenotypes, all five components—weakness, slowness, exhaustion, weight loss, and low physical activity—were significantly and positively associated with Tw-FROP-Com scores. We found that 18.7% of the 375 participants were classified as frail. Additionally, the comparison of characteristics between participants with and without a history of falls is presented in Table 2. Participants with a history of falls were significantly more likely to exhibit weakness (p = 0.014), slowness (p = 0.003), and a tendency toward frailty (p = 0.018).

Using multivariable linear regression, we evaluated the association between fall risk scores (Tw-FROP-Com scores) and a history of falling, along with all components of frailty, after adjusting for age, education, and marital status. The results are presented in Table 3. Older adults who exhibited weakness (β = 0.644, 95% CI: 0.082–1.206), slowness (β = 0.213, 95% CI: 0.151–0.274), exhaustion (β = 1.280, 95% CI: 0.666–1.895), unintentional weight loss (β = 3.992, 95% CI: 2.781–5.203), and low physical activity (β = 0.878, 95% CI: 0.194–1.563) had significantly higher Tw-FROP-Com scores. Based on the B values, unintentional weight loss was the most influential factor for fall risk scores, followed by exhaustion, low physical activity, weakness, and slowness. The distribution of fall risk scores predicted by individual frailty components is illustrated in Figure 1.

Frailty was treated as a dichotomous variable (Yes/No) in Table 4, whereas the model in Table 3 included five individual components. The results of the multivariable linear regression analysis for fall risk are presented in Table 4. Older adults with a history of falling had significantly higher fall risk scores than those without a history of falling (β = 3.886, 95% CI: 3.281–4.491). Additionally, frail older adults had higher fall risk scores compared to non-frail older adults (β = 3.886, 95% CI: 3.158–4.615).

The only difference between the models in Tables 3 and 4 was how frailty was measured. The model in Table 3 assessed frailty as five separate components, whereas the model in Table 4 treated it as a dichotomous Yes/No variable. Based on the R² values of the multivariable linear regression models, the R² values for Tables 3 and 4 were 0.552 and 0.484, respectively. This finding suggests that evaluating individual frailty components predicts fall risk more effectively than using a dichotomous frailty measure, with a 6.8% increase in predictive accuracy.

Furthermore, we conducted a logistic regression analysis by categorizing the Tw-FROP-Com scores into two groups: the mild fall risk group (0–11) and the moderate fall risk group (12–60). We performed both crude odds ratio analyses and analyses adjusted for age, education, and marital status. Table 5 presents the association between fall history and each individual frailty component, evaluating their relationship with the fall risk groups. Older adults who exhibited weakness (aOR = 9.285, 95% CI: 2.672–32.262), slowness (aOR = 0.551, 95% CI: 0.163–1.860), exhaustion (aOR = 3.368, 95% CI: 1.063–10.673), unintentional weight loss (aOR = 23.843, 95% CI: 5.172–109.904), and low physical activity (aOR = 3.935, 95% CI: 1.263–12.261) had a higher likelihood of being in the moderate fall risk group.

The results of the multivariable logistic regression model for fall risk grading are shown in Table 6. Older adults with a history of falling had a significantly higher fall risk than those without a history of falling (aOR = 13.097, 95% CI: 4.117–41.667). Additionally, frail older adults had a significantly higher fall risk compared to non-frail older adults (aOR = 15.510, 95% CI: 4.526–53.153). According to the Nagelkerke R², the values for Tables 5 and 6 were 0.459 and 0.335, respectively. This finding suggests that frailty components predict fall risk grading more effectively than the dichotomous frailty measure (Yes/No), with a 12.4% improvement in predictive accuracy.

Discussion

To the best of our knowledge, this study is the first to evaluate the association between frailty components and fall risk scores, as measured by the Tw-FROP-Com assessment tool, among community-dwelling older adults in Taiwan. Our findings indicate that frailty components account for more than 50% of the variation in fall risk scores among individuals with a history of falls. Older adults experiencing unintentional weight loss, weakness, exhaustion, slowness, and low physical activity had significantly higher Tw-FROP-Com scores after adjusting for sociodemographic factors and fall history. As expected, frail older adults exhibited significantly higher fall risk scores than non-frail individuals. Among the five frailty components, unintentional weight loss emerged as the most influential factor affecting fall risk. Treating frailty as five distinct components provided a more precise prediction of fall risk than using a dichotomous frailty measure (Yes/No). Unintentional weight loss was identified as a critical intrinsic factor, highlighting the need for a differentiated approach to the frailty phenotype among the oldest old, taking into account the specific influence of its components [15]. Each individual indicator can be used to screen older adults at high risk of falls, enabling early detection and timely intervention. Based on the specific scores of each frailty component, tailored fall prevention strategies can be implemented, such as weight management, dietary recommendations, and exercises to improve muscular strength and endurance. These targeted interventions aim to effectively reduce the risk of falls among the elderly population.

Since unintentional weight loss emerged as a key factor in frailty associated with fall risk among older adults, effective weight management and regular monitoring are critical components of fall prevention strategies. Despite the various criteria used to define frailty, weight loss has consistently been a significant predictor of future falls among both community-dwelling [29] and hospitalized older adults [30]. However, previous studies primarily used Fried frailty status (Yes/No) as an outcome variable rather than examining frailty through its five distinct phenotypes. Moreover, no existing studies have assessed fall risk among community-dwelling older adults using both the Fried frailty index and the Tw-FROP-Com. In contrast, our study distinguished the five frailty phenotypes and assessed the overall fall risk. Additionally, the Tw-FROP-Com proved to be a practical and efficient tool, requiring no supplementary instruments to comprehensively assess the 13 fall risk domains. Its simplicity makes it highly applicable in both clinical and community settings for the early identification of older adults at risk of falling. Numerous studies have demonstrated a connection between frailty status and a history of falls [14, 21, 31]. Our findings offer novel insights compared to previous studies. For example, a study conducted in Brazil interviewed 1,413 older adults to assess the association between frailty and falls. The frailty components associated with fall risk included: (1) reduced grip strength (no falls: 21.8%; falls: 31.5%; relative risk (RR) = 1.44; p = 0.003) and (2) exhaustion (no falls: 7.6%; falls: 14.7%; RR = 1.93; p = 0.003) [32]. In our study, unintentional weight loss was the most powerful predictor of fall risk among the five frailty components. This finding suggests that the contribution of frailty components to fall risk may vary by race, age, and other physical conditions.

Unintentional weight loss is commonly associated with malignant diseases, psychiatric disorders, gastrointestinal conditions, endocrine disorders, and cardiovascular diseases [33]. Many older patients with unintentional weight loss also report experiencing concomitant malnutrition [34]. Numerous studies have shown that both unintentional and intentional weight loss are unfavorably associated with increased mortality, morbidity, in-hospital complications, and reduced functional capacity for independent living [33, 35–37]. Rapid unintentional weight loss in older adults often indicates underlying diseases and accelerates age-related muscle loss. Weight loss can result from various physical, psychological, and social conditions, as well as age-related physiological changes. However, up to one-fourth of patients may have no identifiable cause [33].

Malnutrition is a key contributing factor to both weight loss and frailty in older adults. Studies examining the effects of malnutrition and falls among elderly inpatients have identified several risk factors [38–40]. Poor nutrition reliably predicts falls in hospitalized older patients, with malnourished individuals exhibiting higher fall rates [39]. In clinical practice, interventions such as nutritional screening and assessment should be implemented to prevent these avoidable falls. Poorly nourished participants had an elevated risk of self-reported falls over six months (RR = 1.5, 95% CI: 1.0–2.5, p = 0.03) [40]. The relationship between malnutrition and body weight loss is inextricably linked [38]. Older adults experiencing body weight loss may have an increased risk of falls due to chronic malnutrition. Malnutrition is often a subtle underlying condition, whereas subsequent body weight loss is a visible clinical manifestation. Evaluating body weight loss is more accessible and intuitive for clinicians, enabling them to quickly identify patients at high risk of falling. However, clinicians often face challenges in diagnosing weight loss in older adults.

In most cases, a diagnosis can only be made after a thorough patient history, physical examination, and basic laboratory evaluation, followed by the prescription of nutritional supplements [41]. If the initial evaluation is unremarkable, malnutrition should be considered a contributing factor [33]. Malnutrition screening tools are effective in identifying older adults at risk, enabling early prevention. Additionally, several quantitative malnutrition screening tools have been developed and evaluated for their effectiveness [42, 43]. A review of 74 papers covering 119 validation studies on 34 malnutrition screening tools used in older adults across various settings was conducted. The validation results varied considerably across tools, studies, and settings [43]. Therefore, tools for assessing malnutrition should be carefully selected based on the age of the subjects, the target population, and the intended setting.

Identifying and addressing any underlying causes is the primary goal in managing weight loss [33]. Treatment for unintentional weight loss often requires improving access to adequate nutrition [33]. Therefore, several essential non-pharmacological strategies can be implemented to prevent or manage malnutrition and promote adequate food intake. Additionally, other factors contributing to poor diet quality should be considered, including poverty, poor dental health, difficulties with chewing or swallowing, stress, and emotional distress [33].

Conclusions

Unintentional weight loss has been identified as the most influential frailty component in predicting fall risk among older adults in Taiwanese community settings. The five individual frailty phenotypes predict fall risk more effectively than overall frailty status. However, the relationship between frailty, its components, and fall risk is complex and warrants further investigation. Our findings highlight the need for future studies to prioritize longer follow-up cohort designs to confirm the causal relationship between specific frailty phenotypes and fall risk among community-dwelling older adults. A deeper understanding of these associations could inform more targeted and effective fall prevention strategies for this population.

Materials and Methods

Supplementary Materials

Supplementary Table 1