Abstract
Background
Frail geriatric patients have low adherence to physical activity, which is a critical concern as exercise intervention is a key method for managing frailty. The Capability Opportunity Motivation-Behaviour (COM-B) model provides a framework for understanding factors related to health behaviour. However, few studies have applied this model to investigate physical activity in hospitalized frail patients.
Methods
From February to June 2025, a questionnaire survey was conducted on 432 frail geriatric patients. Physical activity was assessed using the International Physical Activity Questionnaire-Short Form. Drawing on the COM-B model as an organizing framework, data on potential correlates were collected using a general information questionnaire, the Geriatric Locomotive Function Scale, the Family APGAR index, and the Revised Behavioural Regulation in Exercise Questionnaire-3.
Results
The total metabolic expenditure during physical activity for frail geriatric patients was 1521 (IQR: 495, 3786) MET-min/week, with an average daily sedentary time of 8.7 ± 1.8 h. Only 81 participants (18.8%) met the guidelines for physical activity. Hierarchical multiple linear regression analysis identified six significant factors associated with physical activity(R2=0.300): family support and behavioural regulation were positive correlates, whereas geriatric locomotive function, increasing age, number of chronic diseases, and medication status were negative correlates. Structural equation modelling explained 29.3% of the variance in physical activity. Mediation analysis showed that geriatric locomotive function and family support were directly associated with physical activity (direct effects of -0.263 and 0.159, respectively; P < 0.001) and also indirectly associated via behavioural regulation (indirect effects of -0.041 and 0.035; P < 0.05).
Conclusions
This study suggests that physical inactivity among frail geriatric inpatients is associated with geriatric locomotive function, family support, and motivation, alongside physiological constraints including age, comorbidities, and polypharmacy. Behavioural regulation emerged as a key mediating factor in these relationships. The findings suggest that interventions referencing the COM-B framework might benefit from targeting these specific functional and motivational correlates.
Clinical trial number
Not applicable.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12877-026-07074-w.
Keywords: Frailty, Geriatric patients, Physical activity, COM-B model, Correlates
Introduction
Frailty is a highly prevalent and clinically significant condition in geriatric populations [1]. The Integral Conceptual Model of Frailty (ICMF) provides a key theoretical framework, characterizing it not merely as a physical decline but as a dynamic, multidimensional state arising from the interplay of deficits in physical, psychological, and social domains [2]. Directly based on this model, the Tilburg Frailty Indicator (TFI) was developed as a 15-item self-report questionnaire to operationalize this concept, with a score of 5 or higher identifying frailty [3]. Importantly, the TFI is not only a screening tool but also a robust predictor of adverse outcomes. A comprehensive scoping review of 37 longitudinal studies confirmed that the TFI effectively predicts hospitalization, mortality, disability, decline in quality of life, and falls in older adults, underscoring the severe personal and societal burden of frailty [4]. The prevalence of this high-risk condition varies considerably by assessment tool and setting, as evidenced by a 42% rate in community studies using multidimensional tools like the TFI [5], compared to rates of 29.8% in hospitals and 51.5% in communities when using the Multidimensional Prognostic Index (MPI) [6]. Given this significant and variable burden, identifying modifiable factors, such as physical activity, that may alter the trajectory of frailty becomes an urgent priority.
Building on this priority, contemporary clinical guidelines endorse a holistic model of frailty management, emphasizing the integration of exercise and nutrition interventions as most effective for improving functional outcomes [7–9]. Within this integrated approach, physical activity constitutes a fundamental component. It is broadly defined as any bodily movement produced by skeletal muscles that results in energy expenditure, encompassing all daily activities [10]. Exercise represents a structured and planned subcategory of physical activity, performed repetitively with the specific aim of maintaining or improving physical fitness components such as strength, endurance, balance, or flexibility [11]. Substantial evidence confirms that regular physical activity, including structured exercise, improves physical function, muscle strength, and can mitigate frailty progression [10]. Accordingly, global and specific guidelines provide clear recommendations: the World Health Organization advises at least 150 min of moderate-intensity or 75 min of vigorous-intensity aerobic physical activity weekly, plus muscle-strengthening activities [12]. Frailty-specific guidelines further recommend tailored, multicomponent exercise programs [8]. Despite these clear recommendations and established benefits, achieving sustained engagement in physical activity among frail older adults remains a significant and prevalent challenge.
However, a significant gap persists between guideline recommendations and actual behavior. Frail older adults consistently demonstrate insufficient physical activity and high levels of sedentary behavior, falling far below recommended levels [10]. Recent reports indicate that only about 29.1% of frail geriatric patients adhere to regular structured exercise [13]. This pattern is powerfully corroborated by objective data, such as a national cohort study showing the most frail individuals averaged 9.57 h of daily sedentary time with only 8.3% meeting activity guidelines [14]. Moreover, physical inactivity is highly prevalent among institutionalized older adults in China, with one study reporting a rate as high as 88.46% among nursing home residents, highlighting the severity of this issue across care settings and underscoring the particular urgency of addressing it within related healthcare environments such as hospitals [15]. Critically, this widespread inactivity is not a normative aspect of aging but is directly attributable to the frailty syndrome, which is characterized by core components such as exhaustion and weakness that inherently reduce the capacity and motivation for activity [14, 16]. Since many existing interventions are implemented in controlled settings, they often ensure only short-term compliance, while long-term adherence remains unsatisfactory [17, 18]. Consequently, promoting sustained engagement is a major challenge, as traditional approaches frequently overlook the individual-level determinants that drive behavior. Therefore, to design effective and durable strategies, it is essential to systematically investigate the modifiable factors that influence physical activity within this vulnerable population.
To address this need for a more detailed understanding of behavioral determinants, this study draws on the Capability, Opportunity, Motivation-Behaviour (COM-B) model as an organizing theoretical framework. The COM-B Model, proposed by Michie et al., identifies key behavioural determinants through the interaction of capability, opportunity, and motivation [19]. It is widely used to synthesise evidence and inform clinical interventions. The model has been applied increasingly in physical activity research among older adults, including studies on nursing home residents [15], qualitative syntheses [20], and mixed-method reviews identifying barriers and facilitators for exercise adherence [18]. However, its use in quantitatively examining factors associated with physical activity among hospitalized frail geriatric patients remains limited.
Based on existing studies, this study uses the COM-B model as an organizing framework to find factors associated with physical activity (Fig. 1). In the specific clinical context of frail geriatric patients, we focused on three key correlates that theoretically align with the model’s core domains. As physical decline is a key part of frailty, geriatric locomotive function was selected to reflect the physical constraints relevant to Capability [21]. Given the pivotal role of family caregiving for hospitalized older adults in China, family support was chosen to reflect the social context relevant to Opportunity [22]. Behavioural regulation was identified to capture the psychological processes of Motivation [23]. Additionally, sociodemographic and clinical characteristics were included as background factors within this framework.
Fig. 1.
The hypothesized conceptual framework drawing on the COM-B model. Note: Validated instruments (GLFS-25, APGAR, BREQ-3) were used as proxy indicators for the COM-B domains. Solid arrows represent statistical associations tested via SEM rather than causal links. The dashed arrow indicates adjustment for sociodemographic and clinical covariates
Therefore, this study was designed with two specific aims: (1) to describe the physical activity levels of hospitalized frail geriatric patients, and (2) to examine the hypothesized relationships between geriatric locomotive function, family support, behavioural regulation, and physical activity, using structural equation modeling within a framework referencing the COM-B model.
Methods
Study design
We conducted a cross-sectional study to investigate the physical activity and associated factors of frail geriatric patients. This study used the STROBE statement for cross-sectional studies (Supplementary Table).
Setting and participants
This single-center, cross-sectional study was conducted at Hangzhou TCM Hospital Affiliated to Zhejiang Chinese Medical University from February to June 2025. A convenience sample of hospitalized patients aged ≥ 60 years was recruited from the inpatient wards of four clinical departments: Geriatrics, Orthopedics, Respiratory Medicine, and Medical Massage.
Frailty was assessed using the Chinese version of the Tilburg Frailty Indicator (TFI), a validated self-report questionnaire based on the integral conceptual model of frailty [24]. For the purpose of this study, only Part B of the TFI was administered, which focuses on the components of frailty itself. Part B comprises 15 items that assess three domains: physical frailty, assessed through 8 items including low physical activity, unexplained weight loss, and physical tiredness; psychological frailty, assessed through 4 items such as problems with memory and feeling anxious; and social frailty, assessed through 3 items including living alone and lack of social support. The total score ranges from 0 to 15, with a higher score indicating greater frailty severity. In accordance with the established standard, a cut-off score of 5 or higher was used to define frailty for participant inclusion in this study [3, 24].
The inclusion criteria were as follows: (1) age ≥ 60 years and (2) frailty status, defined as a score of ≥ 5 of the TFI; (3) Clear consciousness and good language expression and communication skills. The exclusion criteria were as follows: (1) severe impairment in basic mobility (inability to stand or walk independently); (2) severe systemic diseases or tumours involving the heart, brain, lungs, liver, kidneys, etc.; (3) severe psychiatric disorders, Alzheimer’s disease, or other types of dementia; and (4) unstable or critical phase of illness.
Trained research staff screened potential participants identified by their treating physicians. Eligible individuals who provided written informed consent were included. Of the 450 patients approached, 432 (96.0%) completed the assessment and formed the final sample.
Sample size calculation
On the basis of the sampling principles of observational studies, the sample size was calculated using G*Power 3.1 software. Multiple linear regression analysis of variance was employed, with a effect size f2 at 0.15 and a statistical power of 0.95; this study included 24 independent variables (general information and scale dimension levels), resulting in a required sample size of at least 238 cases. Considering a 10%–20% attrition rate, the recommended additional sample size is at least 265–298 cases. To ensure adequate power and account for potential screening failures, we planned to enroll approximately 450 participants.
Measures
Demographic data
The demographic data included gender, age, body mass index (BMI), educational attainment, economic income, smoking and drinking habits, and other general demographic characteristics, as well as the comorbidities (number of diseases), and medication status, for a total of 9 independent variables.
Physical activity assessment
Given the clinical context of frail geriatric inpatients, the target behaviour was defined broadly as “maintaining overall physical activity and minimizing sedentary behaviour within the hospital setting.” Unlike structured exercise interventions, the primary goal for this population is general mobilization. Therefore, the International Physical Activity Questionnaire-Short Form (IPAQ-SF) was selected to capture the total volume of physical activity.
The IPAQ-SF, developed by the International Physical Activity Group [25] and adapted to China by Qu and Li [26], was used to assess physical activity behaviour. The Chinese version (IPAQ-C) has demonstrated good reliability and validity in older adult populations, including a test-retest reliability coefficient of 0.84 over 9 days and a moderate correlation (Spearman and partial r = 0.33, P < 0.001) with pedometer-measured steps [27]. In our study, the test-retest reliability over an 8-day interval was also assessed in a subsample (n = 32), yielding an intraclass correlation coefficient of 0.82. Participants reported their physical activity over the past seven days. For our sample of hospitalized patients, this period corresponded to their current inpatient stay, as most had been hospitalized for more than seven days at the time of assessment.
The questionnaire includes four categories: vigorous physical activity, moderate physical activity, walking, and sedentary behaviour. Weekly physical activity volume: Weekly physical activity level (MET-min/w) = MET value for each physical activity×daily duration (min/day)×number of days per week (days/week). The MET values for walking, moderate-intensity activity, and vigorous-intensity activity are 3.3, 4.0, and 8.0 MET, respectively. Higher scores are positively correlated with physical activity levels. On the basis of the classification criteria in the manual, individuals’ total physical activity energy expenditure was categorized into three levels: physically active (≥ 3000 MET-min/week), moderately active (600 ~ 3000 MET-min/week), and physically inactive (< 600 MET-min/week). Additionally, we determined the proportion meeting the World Health Organization recommendation of at least 150 min of moderate-intensity or 75 min of vigorous-intensity activity per week [12].
Geriatric locomotive function(GLFS-25)
The Geriatric Locomotive Function Scale (GLFS-25) was developed by Seichi et al. [28] and adapted to China by Zhang et al. [29] for assessing the locomotive function of older people. It consists of 4 dimensions and 25 items: pain (4 items), activities of daily living (16 items), social activities (3 items), and mental health status (2 items). The scale uses a 5-point scoring system, with scores ranging from 0 to 4 points for “no difficulty,” “mild difficulty,” “moderate difficulty,” “considerable difficulty,” and “severe difficulty,”, respectively. The total score ranges from 0 to 100 points, with higher scores indicating poorer locomotor function in older people. A score of ≥ 16 points is considered to indicate locomotor dysfunction syndrome. In this study, the Cronbach’s α coefficient was 0.938, and the split-half reliability was 0.871.
Family care(APGAR)
The Family Care Index Scale (APGAR), designed by Smilkstein [22] and adapted to China by Lv et al. [30], consists of five items: adaptability and cooperation, growth, emotional connection, and intimacy. Each item is scored using a 3-point scale: “often” (2 points), “sometimes” (1 point), and “rarely” (0 points). The total score ranges from 0 to 10, with 0–3, 4–6, and 7–10 representing severe, moderate, and good functioning, respectively. In this study, the Cronbach’s α coefficient was 0.874, and the split-half reliability was 0.787.
Behavioural regulation in exercise(BREQ-3)
This study utilized the Revised Behavioural Regulation in Exercise Questionnaire-3 (BREQ-3), revised by Markland and Tobin [31] and adapted to China by Luo et al. [23], to assess the types and levels of exercise motivation, including six dimensions and 24 items: intrinsic motivation, integrated regulation, identified regulation, introjected regulation, external regulation, and amotivation. Each dimension consists of four items, scored using a 5-point Likert scale, with each item rated from ‘strongly disagree’ to ‘strongly agree’ and assigned scores ranging from 0 to 4. Following previous measurement practices, the Relative Autonomy Index (RAI) was used as an observational indicator for measuring individual autonomous motivation and was calculated as follows: Autonomous Motivation = 3×Intrinsic motivation + 2×Introjected regulation + Identified regulation - Internal regulation − 2×External regulation − 3×Amotivation. Higher scores indicate stronger autonomy motivation. In this study, the Cronbach’s α coefficient was 0.833, and the split-half reliability was 0.843.
Variable conceptualization for analysis
To construct the structural equation model based on the COM-B framework, the measured variables were conceptualized as proxy indicators. Specifically, the GLFS-25 was used to represent the Physical Capability domain, acting as a marker for the functional limitations prevalent in this population. The Family APGAR Index was used as a proxy for Social Opportunity, and the BREQ-3 was mapped to Motivation. It is explicitly acknowledged that these clinical measures are partial indicators selected for their relevance to the inpatient context and do not fully capture the multidimensionality of the original COM-B constructs.
Data collection
Before the data were collected, the authors conducted three pilot surveys to optimise the wording and content of the questionnaire. The questionnaire was administered in the ward by two authors who had received training. The authors screened clinical records to obtain information about the diseases of the participants. Owing to the special nature of frail geriatric patients, the questionnaire was administered face-to-face. After obtaining the consent of the research subjects, the researchers asked the questions, the patients answered, and the authors filled out the questionnaire on the basis of their answers. The authors then checked the completeness and logical consistency of the questionnaire. Any questions with missing or incorrect responses were returned to the patients for completion and confirmation.
Data analysis
All statistical analyses were conducted using IBM SPSS Statistics for Mac (version 27.0) and R Studio 4.3.3. Quantitative data are expressed as mean (standard deviation, SD), and categorical data as frequency (percentage).
To examine factors associated with physical activity and ensure a theory-driven approach, a hierarchical multiple linear regression analysis (Enter method) was performed. Based on previous literature and theoretical relevance [18–20, 32, 33], sociodemographic and clinical covariates (age, economic income, alcohol consumption, and number of comorbidities) were entered in Block 1 to control for their confounding effects. The core predictor variables representing the COM-B domains (Geriatric Locomotive Function, Family Support, and Behavioural Regulation) were entered in Block 2. Model assumptions were rigorously verified: multicollinearity was assessed using the Variance Inflation Factor (VIF < 5); independence of residuals was evaluated using the Durbin-Watson statistic (values near 2 were considered acceptable); and the normality and homoscedasticity of residuals were checked using P-P plots (Probability-Probability plots) and scatterplots, respectively.
Structural equation modeling (SEM) was applied to test the hypothesized pathways mapped to the COM-B framework. The a priori selected covariates (consistent with the regression analysis) were included as exogenous variables controlling for the outcome. Parameter estimation used the Maximum Likelihood (ML) method. To rigorously test the significance of indirect effects (mediation), bias-corrected bootstrapping with 5,000 resamples was employed. Model fit was evaluated using established indices: Chi-square/df ratio (< 3), Comparative Fit Index (CFI > 0.90), Tucker-Lewis Index (TLI > 0.90), Root Mean Square Error of Approximation (RMSEA < 0.08), and Standardized Root Mean Square Residual (SRMR < 0.08). A two-tailed significance level of P < 0.05 was applied for all tests [34].
Ethical approval and consent
This study was reviewed and approved by the hospital ethics committee (No. 2025KLL027). All the participants voluntarily participated and signed a paper-based informed consent form.
Results
Demographic characteristics of participants
Table 1 showed that the mean age of the participants was 69.5 years (SD = 7.5, range 60–98), and 53.0% of whom were less than 70 years. The average BMI was 23.2 kg/m² (SD = 3.5, range 12.2–44.4). A total of 198 (45.8%) were male, and 234 (54.2%) were female. Other sociodemographic statistics, disease-related characteristics, and lifestyle characteristics are shown in Table 1.
Table 1.
Demographic characteristics of the participants (n = 432)
| Variable | Category | Number | Percentage (%) |
|---|---|---|---|
| Gender | Male | 198 | 45.8 |
| Female | 234 | 54.2 | |
| Age (years) | mean (SD) | 69.5(7.5) | |
| 60ཞ69 | 229 | 53.0 | |
| 70ཞ79 | 154 | 35.7 | |
| ≥ 80 | 49 | 11.3 | |
| BMI (kg/m2) | Underweight | 37 | 8.6 |
| Normal weight | 286 | 66.2 | |
| Overweight | 96 | 22.2 | |
| Obesity | 13 | 3.0 | |
| Educational attainment | Illiterate | 141 | 32.6 |
| Junior | 130 | 30.1 | |
| Senior | 86 | 19.9 | |
| Bachelor’s degree, associate degree, and above | 75 | 17.4 | |
| Economic income (yuan/month) | < 3000 | 17 | 3.9 |
| 3000–4999 | 87 | 20.1 | |
| 5000ཞ9999 | 228 | 52.8 | |
| ≥ 10,000 | 100 | 23.2 | |
| Smoking | Yes | 46 | 10.6 |
| None | 345 | 79.9 | |
| Quit smoking | 41 | 9.5 | |
| Drinking | Yes | 54 | 12.5 |
| None | 341 | 78.9 | |
| Quit drinking | 37 | 8.6 | |
| Medication status | 0 type | 142 | 32.9 |
| 1ཞ2 types | 171 | 39.6 | |
| 3ཞ4 types | 102 | 23.6 | |
| ≥ 5 types | 17 | 3.9 | |
|
Comorbidities (Number of diseases) |
0 type | 159 | 36.8 |
| 1 type | 181 | 41.9 | |
| ≥ 2 types | 92 | 21.3 | |
| Tilburg Frailty Indicator scale scores | Physical frailty, mean (SD) | 5.1 (1.7) | |
| Mental frailty, mean (SD) | 2.2 (0.7) | ||
| Social frailty, mean (SD) | 1.2 (0.5) | ||
| Total score, mean (SD) | 8.5 (1.9) | ||
Note. BMI Body mass index
Levels of physical activity among participants
The total metabolic equivalent (MET-min) for physical activity among participants was 1,521 (495, 3,786) MET-min/week. As shown in Fig. 2, Panel A, most participants reported no vigorous activity, reflecting a highly right-skewed distribution. Median moderate-intensity activity was 90 min/week, below the WHO recommendation of 150 min/week. Light activity averaged 210 min/week, and average daily sedentary time was 8.7 ± 1.8 h. According to Panel B, only 18.8% (n = 81) of frail geriatric patients met physical activity guidelines, while 34.3% were somewhat active and 38.2% had insufficient activity. Overall, these results indicate widespread physical inactivity and prolonged sedentary behaviour in this population.
Fig. 2.
Distribution of exercise behaviour levels among participants (n = 432). A Violin plots depict the probability density of the data at different values, combined with a box plot (white) inside showing the median (center line), interquartile range (IQR, box bounds), and adjacent values. The red dashed line indicates the WHO recommended minimum of 150 min of moderate-intensity physical activity per week. Note the logarithmic scale on the Y-axis for Panel A. B Bar chart showing the percentage of participants classified into different physical activity levels based on their adherence to standard guidelines
Levels of core variables among participants
Table 2 presents the mean scores, standard deviations, and ranges for the core variables. Overall, geriatric locomotive function impairments and behavioural regulation among frail geriatric patients were moderate, while family support levels were relatively high. Specifically, 307 (71.06%) participants exhibited signs of locomotive function decline, and 319 (73.84%) reported good family support.
Table 2.
Levels of core variables among participants (n = 432)
| Variable | Mean ± SD | Score range | Kurtosis | Skewness |
|---|---|---|---|---|
| 1 Geriatric Locomotive Function, GLFS-25 | 23.8 ± 13.5 | 0ཞ89 | 4.855 | 1.591 |
| 1.1 Pain | 2.8 ± 2.1 | 0ཞ12 | 0.299 | 0.505 |
| 1.2 Activities of daily living | 13.1 ± 9.5 | 0ཞ64 | 6.158 | 1.958 |
| 1.3 Social activities | 4.1 ± 2.3 | 0ཞ12 | 1.705 | 0.865 |
| 1.4 Mental health status | 3.5 ± 2.3 | 0ཞ8 | 4.855 | 0.436 |
| 2 Family Care Index Scale, APGAR | 8.0 ± 2.2 | 0ཞ10 | 0.140 | -0.908 |
| 3 Behavioural Regulation in Exercise Questionnaire-3, BREQ-3 | 8.9 ± 28.0 | -80ཞ92 | 0.101 | -0.062 |
| 3.1 Amotivation | 6.8 ± 4.2 | 0ཞ16 | 0.234 | 0.455 |
| 3.2 External regulation | 8.0 ± 3.7 | 0ཞ16 | -0.979 | -0.347 |
| 3.3 Introjected regulation | 5.4 ± 3.2 | 0ཞ14 | -0.383 | 0.266 |
| 3.4 Identified regulation | 7.8 ± 3.0 | 0ཞ16 | 0.081 | -0.250 |
| 3.5 Intrinsic motivation | 9.0 ± 3.5 | 0ཞ16 | -0.181 | -0.518 |
| 3.6 Integrated regulation | 7.9 ± 3.3 | 0ཞ16 | -0.362 | -0.228 |
Correlations among core variables
In this study, the skewness values of the core variables range from − 0.908 to 1.958, and the kurtosis values range from − 0.979 to 6.158,indicating that the data are approximately normally distributed [35]. Therefore, Pearson correlation analysis was performed. As shown in Table 3, physical activity was significantly positively correlated with behavioural regulation and family support, and significantly negatively correlated with geriatric locomotive function. Family support was weakly positively correlated with behavioural regulation, while geriatric locomotive function was negatively correlated with behavioural regulation. There was no statistically significant association between family support and geriatric locomotive function.
Table 3.
The pearson correlation of the core variable (n = 432)
| Variable | 1 | 2 | 3 | 4 |
|---|---|---|---|---|
| 1.International Physical Activity Questionnaire-Short Form (IPAQ-SF) | 1 | |||
| 2.Geriatric Locomotive Function, GLFS-25 | -0.456* | 1 | ||
| 3.Family Care Index Scale, APGAR | 0.233* | -0.012 | 1 | |
| 4.Behavioural Regulation in Exercise Questionnaire-3, BREQ-3 | 0.404* | -0.176* | 0.176* | 1 |
Note. *p < 0.01; there were significant differences between the correlations
Hierarchical multiple linear regression analyses of factors associated with physical activity
As shown in Table 4, the hierarchical regression model met all diagnostic assumptions, including independence of errors (Durbin-Watson = 1.709), absence of multicollinearity (VIF < 1.7), and normality of residuals. The final model (Block 2) explained 30.0% of the variance in physical activity (F = 14.931, P < 0.001).
Table 4.
Hierarchical multiple linear regression analysis of factors associated with physical activity
| Variates | B | Std. Error | Beta (β) | t Value | P Value | VIF |
|---|---|---|---|---|---|---|
| Constant terms | 4675.121 | 1508.383 | 3.099 | 0.002 | - | |
| Gender | 259.202 | 284.592 | 0.047 | 0.911 | 0.363 | 1.578 |
| Age | -43.628 | 16.741 | -0.119 | -2.606 | 0.009 | 1.246 |
| BMI (kg/m2) | 52.968 | 151.992 | 0.015 | 0.348 | 0.728 | 1.039 |
| Educational attainment | 190.985 | 105.214 | 0.076 | 1.815 | 0.070 | 1.036 |
| Economic income | 248.129 | 203.156 | 0.052 | 1.221 | 0.223 | 1.089 |
| Smoking | 156.144 | 222.720 | 0.037 | 0.701 | 0.484 | 1.659 |
| Drinking | -112.770 | 210.871 | -0.028 | -0.535 | 0.593 | 1.651 |
| Medication status | -333.919 | 141.823 | -0.103 | -2.354 | 0.019 | 1.140 |
| Comorbidities (Number of diseases) | -289.482 | 94.888 | -0.131 | -3.051 | 0.002 | 1.105 |
| Geriatric Locomotive Function (GLFS-25) | -52.977 | 9.408 | -0.258 | -5.631 | < 0.001 | 1.254 |
| Family Support (APGAR) | 197.343 | 51.719 | 0.160 | 3.816 | < 0.001 | 1.049 |
| Behavioural Regulation (BREQ-3) | 20.719 | 4.242 | 0.210 | 4.884 | < 0.001 | 1.103 |
| Note. B: Unstandardized coefficient; Beta (β): Standardized coefficient | ||||||
R2=0.300, after adjustment R2༝0.279; F༝14.931, p < 0.001.Durbin-Watson = 1.709
Significant positive correlates of physical activity included family support (β = 0.160, P < 0.001) and behavioural regulation (β = 0.210, P < 0.001). Conversely, geriatric locomotive function (β = -0.258, P < 0.001), increasing age (β = -0.119, P = 0.009), number of comorbidities (β = -0.131, P = 0.002), and medication status (β = -0.103, P = 0.019) were negatively associated with physical activity. Other covariates, including economic income and alcohol consumption, were not statistically significant in the final model.
Structural equation model of physical activity in frail geriatric patients
On the basis of the hypothesized framework referencing the COM-B model, we constructed a structural equation model (Fig. 3). The IPAQ-SF score was used as the manifest variable; geriatric locomotive function, family support, and behavioural regulation were incorporated as latent variables. To ensure a robust and parsimonious model specification, age, number of comorbidities, and medication status were included as covariates. These variables were identified as the primary physiological and clinical confounders in the hierarchical regression analysis.
Fig. 3.
Structural equation model of physical activity in frail geriatric patients. Note.*P < 0.001.χ² = 1.952, df = 3, P = 0.582, CFI = 1.000, TLI = 1.003, RMSEA = 0.000, SRMR = 0.014,AIC = 12,020.4, and BIC = 12,061.0
The model demonstrated satisfactory fit indices (χ² = 1.952, df = 3, P = 0.582, CFI = 1.000, TLI = 1.003, RMSEA = 0.000, SRMR = 0.014, AIC = 12,020.4, and BIC = 12,061.0). The model explained 29.3% of the variance in physical activity. The results indicated that geriatric locomotive function had a significant negative direct association with physical activity (β = -0.263, P < 0.001) and behavioural regulation (β = -0.187, P = 0.002). Family support was positively associated with behavioural regulation (β = 0.157, P = 0.001) and physical activity (β = 0.159, P < 0.001). Furthermore, behavioural regulation had a strong positive association with physical activity (β = 0.222, P < 0.001). Among the covariates, age (β = -0.128, P < 0.001), comorbidities (β = -0.133, P < 0.001), and medication status (β = -0.098, P = 0.013) were all significantly negatively associated with physical activity.
Table 5 shows the decomposition of effects based on bias-corrected bootstrapping. The 95% confidence intervals for all indirect pathways did not include 0, confirming significant mediating roles of behavioural regulation. Specifically, geriatric locomotive function exerted a significant indirect effect on physical activity via behavioural regulation (Effect = -0.041, 95% CI: -0.073 to -0.009), accounting for 13.5% of the total effect. Similarly, family support had a significant indirect effect via behavioural regulation (Effect = 0.035, 95% CI: 0.010 to 0.060), accounting for 18.0% of the total effect.
Table 5.
Direct effects, indirect effects, and total effects of locomotive function and family support on physical activity
| Pathways | Point Estimate (β) | 95% CI (Bootstrap) | P Value |
|---|---|---|---|
| Total Effects | |||
| GLFS ◊ Physical Activity | -0.304 | [-0.372, -0.237] | < 0.001 |
| Family Support◊ Physical Activity | 0.194 | [0.121, 0.266] | < 0.001 |
| Direct Effects | |||
| GLFS ◊ Physical Activity | -0.263 | — | < 0.001 |
| Family Support◊ Physical Activity | 0.159 | — | 0.001 |
| Behavioural Regulation ◊ Physical Activity | 0.222 | — | < 0.001 |
| Indirect Effects (via Regulation) | |||
| GLFS ◊ Behavioural Regulation ◊Physical Activity | -0.041 | [-0.073, -0.009] | 0.011 |
| Family Support ◊ Behavioural Regulation ◊Physical Activity | 0.035 | [0.010, 0.060] | 0.007 |
Note. GLFS Geriatric Locomotive Function. CI Confidence Interval
Discussion
Despite broad recognition of the importance of exercise for managing frailty, this study found that physical activity levels among hospitalized frail older adults in China remain insufficient. As one of the first studies to examine the factors associated with physical activity in this population drawing on the COM-B model, this research addresses a significant gap in understanding the behavioural correlates of activity in geriatric frailty. The results show that physical activity was primarily associated with geriatric locomotive function, family support, and behavioural regulation, alongside physiological constraints including age, comorbidities, and medication status. Furthermore, behavioural regulation was found to partially mediate the relationships between locomotive function/family support and physical activity, accounting for mediation proportions of 13.5% and 18.0%, respectively.
Consistent with most reports [17, 18, 36], our study indicated that frail geriatric patients in China engage in insufficient physical activity. Specifically, the total metabolic equivalent of physical activity (MET-min/week) was 1521 (495, 3786), with 27.55% of patients classified as having insufficient physical activity. This finding aligns with national data, which reported a prevalence of 28.82% among Chinese adults aged 59 and older [36]. Notably, only 18.75% of frail geriatric patients met the physical activity guidelines, with most participants engaging in low-intensity exercise and few participating in moderate-to-vigorous physical activity, which is lower than that reported in developed countries (21.2%) [12]. This disparity may be related to the concept of “yangsheng” (nourishing life) in Chinese culture, which emphasizes achieving a balance between dynamic and static activities [15]. Consequently, moderate-to-vigorous physical activity may be perceived as physically taxing and disruptive to this balance. Furthermore, patients’ functional limitations likely restricts intense engagement in activity. Conversely, the high average daily sedentary time( 8.71 h) reflects environmental constraints within the hospital setting, highlighting the need for culturally tailored and setting-specific physical activity interventions to slow frailty progression and improve health outcomes.
The results of this study indicated that 71.06% of participants exhibited signs of geriatric locomotive function decline, a prevalence similar to that reported in Japanese older cohort studies [37], suggesting that motor dysfunction is widespread among frail individuals. Additionally, we found that the severity of locomotive dysfunction was negatively associated with physical activity in frail geriatric patients. Interpreted within the COM-B framework, this finding aligns with the Physical Capability domain, suggesting that declines in the musculoskeletal system and overall physiological function may restrict physical activity. This observation is consistent with findings in nursing homes [15] and community populations [38]. Previous research has similarly identified functional capacity as a key correlate of physical activity among older adults [15, 38, 39]. A possible explanation is that most participants, owing to postural frailty and cognitive decline, tend to underestimate their physical capabilities, perceive their functional limitations as “insurmountable,” and develop excessive concerns about the risks of physical activity, which is consistent with the “fear-avoidance” model. These findings underscore the necessity of incorporating resistance training and functional rehabilitation into routine older care to mitigate functional loss.
Structural equation modelling indicated that family support was significantly positively associated with physical activity through both direct and indirect relationships. Interpreted within the COM-B framework, this highlights the role of Social Opportunity as a critical enabling factor. Family members act as key promoters of healthy behaviour, playing a vital role in the health and well-being of geriatric patients [40, 41]. These findings are similar to those of Wen and Zhang [42], who suggested that intergenerational support from family members can serve as an important facilitator for healthy behaviours among geriatric people. Such support includes providing companionship, encouragement, and supervision, which may help mitigate loneliness and is linked to enhanced self-efficacy. Notably, our study found that family support functions independently of the patient’s physical limitations. Particularly for frail patients, improving attitudes alone is insufficient; active family involvement appears crucial to sustaining engagement in exercise [43],. Additionally, exercising with a spouse has been linked to improve long-term adherence [44]. Therefore, healthcare providers should integrate family support into personalised exercise interventions and health education strategies for this population.
In this study, frail geriatric patients scored low on the Behavioural Regulation in Exercise Questionnaire, indicating a general lack of internal drive that was significantly associated with insufficient physical activity. Interpreted through the COM-B lens, this highlights the critical role of the Motivation domain. A potential explanation involves physiological aging; the prefrontal cortex, which is responsible for self-regulatory behaviour, typically deteriorates with age [45]. Furthermore, our multivariate analysis explicitly identified advanced age, comorbidities, and medication status as significant constraints on physical activity. These factors reflect a high clinical burden and are linked to impaired bodily functions [15]. This functional decline may challenge their ability to effectively regulate physical activity [17]. Qualitative research further contextualizes the lack of motivation, noting that some frail geriatric patients lack confidence in overcoming exercise barriers, possess limited health beliefs, and lack initiative and exercise habits [46]. Therefore, health care providers should proactively assess the physical function and medication burden of frail geriatric patients. When physical function declines, interventions should aim to reinforce reflective motivation and autonomy, activate behavioural compensation strategies (e.g., using a walker for seated exercises), and mitigate the constraints imposed by functional decline on exercise self-regulation.
Multivariate analysis and SEM results indicated that age, the number of comorbidities, and medication status were significant sociodemographic and clinical correlates of physical activity in frail geriatric patients. Advanced age is associated with accumulated chronic inflammation, reduced muscle strength, and functional decline [47], which may collectively limit functional ability and contribute to the tendency for patients to engage primarily in low-intensity activities such as walking or mind-body exercises [18]. Additionally, the presence of multiple chronic diseases and polypharmacy were identified as factors negatively associated with activity. Patients with comorbidities often adhere to traditional health beliefs favouring rest over exercise [15]. Furthermore, taking multiple medications often reflects a higher burden of disease and may introduce side effects such as fatigue or dizziness, which further reduce the capacity for physical activity [48]. These findings help identify priority groups for intervention, especially those with high clinical and medication burdens, and support the development of stratified physical activity strategies for frail older adults.
SEM results supported the applicability of referencing the COM-B framework to understand physical activity in frail geriatric patients, consistent with previous studies [32, 49, 50]. More importantly, behavioural regulation served as a central mediating factor, linking locomotive function and family support to physical activity, which aligns with existing systematic reviews [51]. This suggests that strategies addressing functional limitations, social opportunity, and motivation may synergistically support activity [15, 39]. Our correlation results indicated no significant association between family support and locomotive function, reinforcing that social resources operate distinctively from physical capabilities. Consequently, family support exerted a significant positive influence on physical activity independent of locomotive function. This underscores the vital role of family involvement in compensating for physical decline. Thus, healthcare professionals should collaborate with families to foster intrinsic motivation, support continued exercise despite functional limitations, and aim to improve overall physical and mental well-being [20].
This study has limitations. First, the single-center sampling limits generalizability, suggesting the need for future multi-center studies. Additionally, self-reported physical activity data may be biased; objective measures are recommended for future research. Second, the cross-sectional design precludes the determination of causal relationships or temporal ordering between the variables. The pathways presented in the SEM represent statistical associations based on the hypothesized framework. Longitudinal studies are needed to verify causality and further unravel behavioral mechanisms. Third, regarding the theoretical application, the target behaviour was operationalized as overall physical activity time (IPAQ-SF) rather than a specific self-directed exercise behaviour. Consequently, the measures used served as proxy indicators and did not fully capture the multidimensional subcomponents of the COM-B model. For instance, while the GLFS assesses physical capability, it may not reflect the specific psychological capability (e.g., knowledge, skills) required for targeted exercise. Similarly, the Family APGAR captures social opportunity but does not encompass the physical environment (e.g., hospital corridor space) and support from medical staff. Furthermore, the assessment of motivation focused on behavioural regulation, potentially overlooking automatic motivational processes (e.g., habits, emotional impulses). Future research should utilize comprehensive tools designed to evaluate the full spectrum of COM-B subdomains in relation to clearly defined rehabilitation behaviours. Finally, by excluding individuals with severe impairment in basic mobility to ensure safety and assessment feasibility, our findings may not generalize to the most functionally impaired segment of the frail population.
Conclusion
This study indicates that physical activity levels among frail geriatric patients in China are insufficient to meet WHO recommendations. Drawing on the COM-B framework, key factors associated with physical activity were identified, including geriatric locomotive function, family support, and behavioural regulation, as well as physiological constraints such as age, number of comorbidities, and medication status. Structural equation modelling demonstrated statistical associations linking these variables to physical activity, characterized by direct associations involving locomotive function, family support, and behavioural regulation, as well as indirect relationships mediated by behavioural regulation. These findings offer practical insights for identifying target populations with high clinical burden and informing the design of future interventions that integrate functional training with motivational and family-based strategies referencing the COM-B framework.
Supplementary Information
Acknowledgements
The authors wish to thank all patients who participated in this study.
Abbreviations
- COM-B
The Capability Opportunity Motivation-Behaviour (COM-B) model
- ICMF
Integral Conceptual Model of Frailty
- TFI
Tilburg Frailty Indicator
- IPAQ-SF
the International Physical Activity Questionnaire-Short Form
- GLFS
the Geriatric Locomotive Function Scale
- APGAR
the Family Care Index Scale
- BREQ-3
the Revised Behavioural Regulation in Exercise Questionnaire-3
- IQR
Interquartile Range
- VIF
Variance Inflation Factor
- SEM
Structural equation modeling
- ML
Maximum Likelihood
- CFI
Comparative Fit Index
- TLI
Tucker-Lewis Index
- RMSEA
Root Mean Square Error of Approximation
- SRMR
Standardized Root Mean Square Residual
Authors’ contributions
Xueyan Huang: Research design; Data analysis; Paper writing. Na Zhang : Data curation; Software. Haifang Zhou: Resources; Supervision. Rui Wang: Data curation; Editing. Jing Cui: Investigation; Organization. Bo Xin: Investigation; data collection. Mengchi Li: Review ; Editing; Paper revision. Wenhui Jiang: Conceptualization; Funding acquisition; Paper revision.
Funding
The study was supported by the Medical Scientific Research Foundation of Zhejiang Province (No. 2025KY160), the National Social Science Fund of China (No. 22BGL253), the Major Program of Medical Scientific Research Foundation of Hangzhou(No.Z20250286), the Construction Fund of Key Medical Disciplines of Hangzhou(No.2025HZPY06), and Shaanxi Provincial Department of Science and Technology - Key R&D Program - Social Development Field (No. 2025SF-YBXM-236).
Data availability
The datasets generated and analysed during the current study are not publicly available but are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
This study was approved by the Ethics Board of Hangzhou TCM Hospital Affiliated to Zhejiang Chinese Medical University (NO. 2025KLL027). All participants gave written informed consent. The research protocol was established according to the ethical guidelines of the Declaration of Helsinki.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Mengchi Li, Email: mengchili@mail.xjtu.edu.cn.
Wenhui Jiang, Email: jiangwenhui@mail.xjtu.edu.cn.
References
- 1.Doody P, Asamane EA, Aunger JA, et al. The prevalence of frailty and pre-frailty among geriatric hospital inpatients and its association with economic prosperity and healthcare expenditure: A systematic review and meta-analysis of 467,779 geriatric hospital inpatients. Ageing Res Rev. 2022;80:101666. [DOI] [PubMed] [Google Scholar]
- 2.Gobbens RJ, Luijkx KG, Wijnen-Sponselee MT, Schols JM. Towards an integral conceptual model of frailty. J Nutr Health Aging. 2010;14(3):175–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Gobbens RJJ, van Assen MALM, Luijkx KG, Wijnen-Sponselee MT, Schols JMGA. The Tilburg frailty indicator: psychometric properties. J Am Med Dir Assoc. 2010;11:344–55. [DOI] [PubMed] [Google Scholar]
- 4.Lee J, Kim GS, Kim S, Park J, Lee H, Shim MS, et al. Use of the Tilburg frailty indicator in longitudinal studies with older adults: A scoping review. J Adv Nurs. 2023;79(7):2429–43. [DOI] [PubMed] [Google Scholar]
- 5.Qiu Y, Li G, Wang X, et al. Prevalence of multidimensional frailty among community-dwelling older adults: A systematic review and meta-analysis. Int J Nurs Stud. 2024;154:104755. [DOI] [PubMed] [Google Scholar]
- 6.Veronese N, Custodero C, Cella A, Demurtas J, Zora S, Maggi S, et al. Prevalence of multidimensional frailty and pre-frailty in older people in different settings: A systematic review and meta-analysis. Ageing Res Rev. 2021;72:101498. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Dent E, Morley JE, Cruz-Jentoft AJ, et al. Physical frailty: ICFSR international clinical practice guidelines for identification and management. J Nutr Health Aging. 2019;23(9):771–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dent E, Martin FC, Bergman H, Woo J, Romero-Ortuno R, Walston JD. Management of frailty: Opportunities, challenges, and future directions. Lancet. 2019;394:1376–86. [DOI] [PubMed] [Google Scholar]
- 9.Lorbergs AL, Prorok JC, Holroyd-Leduc J, et al. Nutrition and physical activity clinical practice guidelines for older adults living with frailty. J Frailty Aging. 2022;11:3–11. [DOI] [PubMed] [Google Scholar]
- 10.Angulo J, El Assar M, Álvarez-Bustos A, Rodríguez-Mañas L. Physical activity and exercise: strategies to manage frailty. Redox Biol. 2020;35:101513. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Caspersen CJ, Powell KE, Christenson GM. Physical activity, exercise, and physical fitness: definitions and distinctions for health-related research. Public Health Rep. 1985;100(2):126–31. [PMC free article] [PubMed] [Google Scholar]
- 12.Bull FC, Al-Ansari SS, Biddle S, et al. World health organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med. 2020;54:1451–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Yu S, Lin J, Song S, Huang S, Liu F, Xiao M. Understanding regular exercise behavior in frail older adults: A structural equation model based on social-cognitive variables. BMC Geriatr. 2025;25:73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Blodgett J, Theou O, Kirkland S, Andreou P, Rockwood K. The association between sedentary behaviour, moderate–vigorous physical activity and frailty in NHANES cohorts. Maturitas. 2015;80:187–91. [DOI] [PubMed] [Google Scholar]
- 15.Shi Y, Xie XY, Lao AD, Shao L, Wang ZA, Zhang J. Prevalence of physical inactivity and its determinants among older adults living in nursing homes: A cross-sectional study based on COM‐B model. J Clin Nurs. 2025;34:204–17. [DOI] [PubMed] [Google Scholar]
- 16.Kehler DS, Theou O. The impact of physical activity and sedentary behaviors on frailty levels. Mech Ageing Dev. 2019;180:29–41. [DOI] [PubMed] [Google Scholar]
- 17.Izquierdo M, Merchant RA, Morley JE, et al. International exercise recommendations in older adults (ICFSR): expert consensus guidelines. J Nutr Health Aging. 2021;25:824–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Yang Y, Gao Y, An R, Wan Q. Barriers and facilitators to exercise adherence in community-dwelling older adults: A mixed-methods systematic review using the COM-B model and theoretical domains framework. Int J Nurs Stud. 2024;157:104808. [DOI] [PubMed] [Google Scholar]
- 19.Michie S, van Stralen MM, West R. The behaviour change wheel: A new method for characterising and designing behaviour change interventions. Implement Sci. 2011;6:42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Meredith SJ, Cox NJ, Ibrahim K, et al. Factors that influence older adults’ participation in physical activity: A systematic review of qualitative studies. Age Ageing. 2023;52:afad145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kobayashi T, Morimoto T, Shimanoe C, Ono R, Otani K, Mawatari M. Clinical characteristics of locomotive syndrome categorised by the 25-question geriatric locomotive function scale: a systematic review. BMJ Open. 2023;13(5):e068645. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Smilkstein G. The family APGAR: A proposal for a family function test and its use by physicians. J Fam Pract. 1978;6:1231–9. [PubMed] [Google Scholar]
- 23.Luo Y, Mullin EM, Mellano KT, Sha Y, Wang C. Examining the psychometric properties of the Chinese behavioral regulation in exercise Questionnaire-3: A bi-factor approach. PLoS ONE. 2022;17:e0265004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Xi X, Guo G, Sun J. Reliability and validity of Chinese version of Tilburg frailty indicator. J Nurs(China). 2013;20:1–5. [In Chinese]. [Google Scholar]
- 25.Craig CL, Marshall AL, Sjöström M, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003;35:1381–95. [DOI] [PubMed] [Google Scholar]
- 26.Qu N, Li K. Study on the reliability and validity of international physical activity questionnaire (Chinese vision, IPAQ). Chin J Epidemiol. 2004;25(3):265–8. [In Chinese]. [PubMed] [Google Scholar]
- 27.Deng HB, Macfarlane DJ, Thomas GN, Lao XQ, Jiang CQ, Cheng KK, Lam TH. Reliability and validity of the IPAQ-Chinese: the Guangzhou biobank cohort study. Med Sci Sports Exerc. 2008;40(2):303–7. [DOI] [PubMed] [Google Scholar]
- 28.Seichi A, Hoshino Y, Doi T, Akai M, Tobimatsu Y, Iwaya T. Development of a screening tool for risk of locomotive syndrome in the elderly: the 25-question geriatric locomotive function scale. J Orthop Sci. 2012;17:163–72. [DOI] [PubMed] [Google Scholar]
- 29.Zhang N, Zhang RL, Li HJ. Validity and reliability of the Chinese version of geriatric locomotive function scale. Chin J Nurs. 2016;51:747–51. [In Chinese]. [Google Scholar]
- 30.Lv F, Zeng G, Liu S, Zhong T, Zhan Z. A study on validity and reliability of the family APGAR. Chin J Public Health. 1999;15(11):987–8. [In Chinese]. [Google Scholar]
- 31.Markland D, Tobin V. A modification to the behavioural regulation in exercise questionnaire to include an assessment of amotivation. J Sport Exerc Psychol. 2004;26:191–6. [Google Scholar]
- 32.Makizako H, Akaida S, Shiratsuchi D, Tateishi M. Are Middle-Aged and older adult users of physical activity monitoring systems more physically active and at a lower risk of locomotive syndrome? A Cross-Sectional Web-Based online survey. Ann Geriatr Med Res. 2024;28(3):323–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Chen D, Zhang H, Shao J, et al. Determinants of adherence to diet and exercise behaviours among individuals with metabolic syndrome based on the capability, opportunity, motivation, and behaviour model: A cross-sectional study. Eur J Cardiovasc Nurs. 2023;22:193–200. [DOI] [PubMed] [Google Scholar]
- 34.Kline RB. Principles and practice of structural equation modeling. 4th ed. New York: Guilford Press; 2016. [Google Scholar]
- 35.Kim HY. Statistical notes for clinical researchers: assessing normal distribution(2) using skewness and kurtosis. Restor Dent Endod. 2013;38(1):52–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Whiting S, Mendes R, Abu-Omar K, et al. Physical inactivity in nine European and central Asian countries: an analysis of National population-based survey results. Eur J Public Health. 2021;31:846–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Ohyama S, Hoshino M, Takahashi S, et al. Predictors of dropout from cohort study due to deterioration in health status, with focus on sarcopenia, locomotive syndrome, and frailty: from the Shiraniwa elderly cohort (Shiraniwa) study. J Orthop Sci. 2021;26:167–72. [DOI] [PubMed] [Google Scholar]
- 38.Chen X, Yang S, Zhao H, Li R, Luo W, Zhang X. Self-efficacy, exercise anticipation and physical activity in elderly: using bayesian networks to elucidate complex relationships. Patient Prefer Adherence. 2022;16:1819–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Nilstomt A, Gustavsson J, Beckman L, et al. Physical activity from the perspective of older adults: A convergent mixed-method study. BMC Geriatr. 2024;24:768. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Shamsikhani S, Ahmadi F, Kazemnejad A, Vaismoradi M. Meaning of respect for older people in family relationships. Geriatrics. 2022;7:57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Shulyaev K, Spielberg Y, Gur-Yaish N, Zisberg A. Family support during hospitalization buffers depressive symptoms among independent older adults. Clin Gerontol. 2024;47:341–51. [DOI] [PubMed] [Google Scholar]
- 42.Wen C, Zhang Q. The role of intergenerational support in shaping oral healthcare-seeking behavior among older adults in China. Front Public Health. 2023;11:1234539. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Huang WY, Wu CE. Predict the exercise behavior intention of the older adults in Taipei City to promote exercise behavior. Sci Prog. 2021;104:368504211042468. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Osuka Y, Jung S, Kim T, Okubo Y, Kim E, Tanaka K. Does attending an exercise class with a spouse improve long-term exercise adherence among people aged 65 years and older: A 6-month prospective follow-up study. BMC Geriatr. 2017;17:170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Lerner RM, Lerner JV, Bowers EP, Lewin-Bizan S, Gestsdottir S, Urban JB. Self-regulation processes and thriving in childhood and adolescence: A view of the issues. New Dir Child Adolesc Dev. 2011;2011:1–9. [DOI] [PubMed] [Google Scholar]
- 46.Zhou Y, Lu HH, Sun X, Dai YW, Gu CY, Gu ZJ. Influencing factors of exercise behavior in older patients with type 2 diabetes mellitus combined with frailty: A qualitative study. Geriatr Nurs. 2025;62:138–44. [DOI] [PubMed] [Google Scholar]
- 47.Hoseinpour AN, Bassami M, Ahmadizad S, et al. The influence of resistance training on inflammatory markers, body composition and functional capacity in healthy older adults: A systematic review and meta-analysis. Arch Gerontol Geriatr. 2025;130:105731. [DOI] [PubMed] [Google Scholar]
- 48.Katsimpris A, Linseisen J, Meisinger C, Volaklis K. The association between polypharmacy and physical function in older adults: a systematic review. J Gen Intern Med. 2019;34(9):1865–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Li W, Hong S, Chen Y, Zhao Y, Wang L. Factors influencing physical activity behavior in older adults with subjective cognitive decline: an empirical study using SEM and FsQCA methods. Front Public Health. 2024;12:1409614. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Silveira SL, Riemann-Lorenz K, Heesen C, Motl RW. Current and long-term physical activity among adults with multiple sclerosis in the united states: COM-B variables as explanatory factors. Int J Behav Med. 2021;28:561–74. [DOI] [PubMed] [Google Scholar]
- 51.Rhodes RE, Cox A, Sayar R. What predicts the physical activity intention-behavior gap? A systematic review. Ann Behav Med. 2022;56:1–20. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated and analysed during the current study are not publicly available but are available from the corresponding author on reasonable request.