Fatigue as a mediator between multidimensional influencing factors and lower limb function in older adults based on the theory of unpleasant symptoms

Abstract

Background

Fatigue is a highly prevalent and debilitating condition among older adults, significantly associated with adverse health outcomes and functional decline. Despite its clinical importance, the complex interplay between fatigue, its multidimensional influencing factors, and lower limb function remains to be fully elucidated. Guided by the Theory of Unpleasant Symptoms (TOUS), this study aimed to empirically test a conceptual model depicting these relationships.

Methods

A cross-sectional study was conducted in Grade A tertiary hospitals in China between September 2020 and March 2021. A sample of community-dwelling older adults was recruited and completed a battery of standardized questionnaires assessing bodily pain, psychological status, social support, fatigue, and lower limb function. The hypothesized model was tested using structural equation modeling (SEM).

Results

The structural equation model demonstrated an excellent fit to the data (χ²/df = 1.321, CFI = 0.983, RMSEA = 0.030). Bodily pain, depressed mood, anxious mood, and social support were found to have significant direct effects on fatigue. In turn, both depressed mood and fatigue exhibited significant direct effects on lower limb function. Furthermore, significant indirect effects on lower limb function, mediated through fatigue, were identified for bodily pain, depressed mood, anxious mood, and social support.

Conclusions

The findings confirm fatigue as a critical mediator in the pathway from physical, psychological, and social factors to lower limb function in older adults. This underscores the necessity for healthcare professionals to integrate routine fatigue assessment and management into geriatric care. Interventions should adopt a multifaceted approach, concurrently addressing psychological distress, optimizing social support, and managing bodily pain. Targeting fatigue represents a promising strategy for preserving mobility and functional independence in this population.

Background

As symptom research advances, a comprehensive and multidimensional understanding of symptom experiences becomes increasingly critical [1]. Fatigue represents a particularly prominent symptom in aging populations. Epidemiological studies have shown that while sleep disturbances, fatigue, and muscle tension are common among 70-year-olds, fatigue emerges as the most prevalent symptom among those aged 85 and older [2]. This symptom significantly limits the ability of older adults to maintain an active and fulfilling lifestyle and is associated with a wide range of etiological factors and functional outcomes. These observations align conceptually with the Theory of Unpleasant Symptoms (TOUS), a middle-range theory designed to integrate knowledge about multiple symptoms. The TOUS framework comprises three core components: influencing factors, symptoms, and performance outcomes, thereby providing a mechanism for a holistic understanding of symptoms [3]. In this study, we specifically focus on fatigue as the core symptom, lower limb function as the performance outcome, and bodily pain, depressed mood, anxious mood, and social support as key influencing factors. Based on this framework, we propose the following hypotheses:

• H1. Fatigue is significantly associated with lower limb function in older adults.

  A general overview of fatigue and its relationship to negative health outcomes indicates that fatigue-related declines in physical function often precede hospitalization, overt disease, and mortality [4]. This is particularly evident in lower limb function. For instance, one study accounting for familial kinship and covariates found that lower fatigue levels were significantly associated with faster gait speed in two generations, with effect sizes of 0.03 m/s (proband generation, P = 0.025) and 0.04 m/s (offspring generation, P < 0.001), respectively [5]. A subsequent follow-up study further confirmed that fatigue was negatively correlated with muscle endurance and gait speed among community-dwelling older adults [6].

• H2. Bodily pain is significantly associated with fatigue in older adults.

• H3. Depressed mood is significantly associated with fatigue in older adults.

• H4. Anxious mood is significantly associated with fatigue in older adults.

• H5. Social support is significantly associated with fatigue in older adults.

  Empirical evidence robustly supports the relationships posited in H2-H5. The high bodily pain group reported higher levels of fatigue (P < 0.01) and was more likely to meet the threshold for clinically significant self-reported fatigue (43.9%) compared to the low pain (15.9%) and no-pain control groups (12.9%) [7]. Another study, after controlling for confounders, found pain to be positively associated with fatigue (OR = 1.77, 95% CI = 1.11–2.81) [8]. Depressed and anxious mood, though often overlooked in later life, show strong links to fatigue. Analysis of data from 2,848 participants in the Cohorte Lausannoise study demonstrated a positive association between depressed mood and fatigue (OR = 3.26, 95% CI = 2.38–4.46) [9]. Similarly, a significant positive correlation has been observed between anxious mood and fatigue (r = 0.531, P < 0.01) [10]. Furthermore, social support plays a protective role, with one study identifying it as a significant predictor of lower fatigue levels (β=−0.13, 95% CI=−0.26- −0.02) [11].

• H6. Bodily pain is significantly associated with lower limb function in older adults.

• H7. Depressed mood is significantly associated with lower limb function in older adults.

• H8. Anxious mood is significantly associated with lower limb function in older adults.

• H9. Social support is significantly associated with lower limb function in older adults.

  The justification for H6-H9 is supported by a body of research linking these factors to physical performance. Evidence confirms an association between bodily pain and impaired lower limb function; older adults with pain demonstrated a significantly larger center of pressure movement area (indicating worse balance, P < 0.001), slower walking speed (P = 0.05), and longer times to complete the Timed Up and Go test (P = 0.004) [12]. Regarding psychological factors, a bidirectional relationship between depressed mood and lower limb function has been reported [13]. Older adults with more severe depressive symptoms score lower on measures of lower limb function and engage in less physical activity compared to their non-depressed counterparts [14]. Anxious mood also exerts a negative effect on functional mobility; both depressed and anxious moods have been significantly correlated with gait speed, functional mobility, and dynamic balance (P < 0.05) [15]. Emerging evidence directly supports the association between social support and lower limb function. A cross-sectional study demonstrated that a stronger social network was significantly associated with better performance on the Timed Up and Go test (β=−0.112, P = 0.044) [16]. Furthermore, longitudinal data indicate that tangible support—the availability of practical assistance—significantly attenuates gait speed decline, with low tangible support associated with 0.51 cm/s greater annual decline in simple gait (P = 0.008) and 0.58 cm/s in complex gait (P = 0.007) [17].

By elucidating the specific influencing factors and performance outcomes associated with fatigue in older adults, this study aims to deepen the understanding of their interrelationships, thereby informing the development of targeted health interventions. Our conceptual model, grounded in the Theory of Unpleasant Symptoms (TOUS) and prior research, illustrates these hypothesized relationships, as depicted in Fig. 1.

Fig. 1.

Conceptual model

Methods

Design and participants

Between September 2020 and March 2021, a convenience sample of older adults was recruited to complete this face-to-face survey in grade A tertiary hospitals in Beijing, China. Inclusion criteria were as follows (a) age 60 years or older, (b) stable vital signs, (c) ability to understand and complete the survey and (d) voluntary participation in this study. Participants with a diagnosis of mental illness, dementia, reading or communication disorders, or other forms of significant illness will be excluded. All participants were allowed to withdraw from the survey at any time.

Given that the chi-square value (χ2) of parameter estimation and fitness in Structural Equation Modeling (SEM) is very sensitive to the size of the sample size, if the sample size is small it will make the estimation lack of stability, and the statistical test power (Power) of the model will be reduced, but when the sample size is large, the χ2 value is easy to reach a statistically significant level (P < 0.05), and the chance of hypothesizing that the model does not fit the actual data increases, so it is difficult to strike a balance between the sample size and the overall model fitness. In general, in order to pursue more stable SEM results, it is best to choose a sample size of more than 200 cases for evaluation by the Maximum likelihood (ML) [18]. However, when the sample size is 400–500 cases, ML may be too sensitive to be suitable for the model. Therefore, considering the previous studies and the current clinical situation, the sample size of this study was estimated to be 200–400 cases.

Instruments

The Short Form 36 (SF-36) [19] is a multipurpose self-administered questionnaire that covers generalized measures of physical and mental health. There are a total of 36 items categorized into 8 dimensions, including physical functioning (PF), role functioning (RP), bodily pain (BP), general health (GH), vitality (VT), social functioning (SF), role-emotional (RE), mental health (MH). This study utilized the vitality (VT) subscale to measure fatigue [20] and the bodily pain (BP) subscale to assess pain levels. Each dimension is scored on a 0 to 100 scale, with lower scores representing poorer health. The selected subscales demonstrated good internal consistency with Cronbach’s alpha of 0.86 for VT and 0.90 for BP, respectively.

The Geriatric Depression Scale-15 (GDS-15) [21] is an instrument specifically designed to assess depressive symptoms in older adults with a total of 15 items. It was developed by Yesavage and Sheikh based on the GDS [22]. Items require yes/no responses, with total scores ranging from 0 to 15. Scores above 5 indicate risk for depression. The scale showed good internal consistency among Chinese older adults with Cronbach’s α of 0.793 [23]. GDS-15 was used in this study to measure depressed mood.

The Self-rating Anxiety Scale (SAS) [24] is a self-report scale that covers a wide range of anxiety symptoms, both psychological and physical. The 20 items are scored on a 4-point scale ranging from 1 (none or occasionally) to 4 (most or all of the time), with some items requiring reverse scoring. The SAS Index is calculated by converting raw scores to a 0–100 scale, with higher scores indicating more severe anxiety. Cronbach’s alpha was 0.69 for the normal sample and 0.81 for the outpatient sample [25]. SAS was employed to assess anxious mood.

The Social Support Rating Scale (SSRS) [26] is the most commonly used instrument to measure the level of social support in the Chinese population. The scale contains 10 items divided into three dimensions: objective support, subjective support, and support utilization. Total scores range from 7 to 56, with higher scores indicating better social support. The scale demonstrated excellent reliability with Cronbach’s alpha of 0.896 for the total scale and 0.825–0.894 for the three dimensions [27]. SSRS was used to measure social support.

The Short Physical Performance Battery (SPPB) [28] was employed as an objective performance-based measure to assess lower limb function. The SPPB consists of three components: standing balance, walking speed, and chair stand tests, which collectively evaluate lower extremity function. The standing balance test assesses the ability to maintain three standing positions (side-by-side, semi-tandem, and full tandem) for 10 s each. The walking speed test measures the time required to complete an 8-foot walking course. The chair stand test evaluates the time needed to complete five repeated rises from a seated position. Total scores range from 0 to 12, with higher scores indicating better lower limb function. The Chinese version demonstrated good reliability with Cronbach’s alpha of 0.76.

Data collection

A structured questionnaire as described above was used to survey older adults who met the inclusion criteria. In addition, general characteristics included age, gender, marital status, education level, financial burden and exercise habits. Prior to the survey, uniform training of investigators was conducted to ensure the quality of the study. After the questionnaires were filled out, the investigators checked for completeness. If there were omitted items, the elderly were immediately asked to add them. After data collection, each participant was thanked and given a small gift as an incentive. The study involving human participants was reviewed and approved by the Medical Ethics Committee (S2021-077-01).

Data analyses

Descriptive statistics were used to characterize the sample and variables. The sample size of this study was less than 5000 and normality was tested using the Shapiro-Wilk test (SW test). Maximum likelihood (ML) can be chosen if the absolute values of skewness coefficient and kurtosis coefficient are less than 2 and 7, respectively [29]. Maximum likelihood (ML) was used for parameter estimation. Pearson correlation analysis assessed the correlation between variables (IBM SPSS Statistics version 26.0). Structural equation modeling (SEM) was used to test the fit of the overall model because it allows for testing the relationship between measured (observed) and latent variables in the hypothetical model, taking into account measurement error (AMOS version 24.0). The model fit was judged using a variety of statistics and was considered good if the following conditions were met: (1) chi-square (χ2) with a correlation of P >0.05 (2), goodness-of-fit index (GFI) of 0.9 or above (3), adjusted goodness-of-fit index (AGFI) of 0.9 or above (4), comparative fit index (CFI) of 0.9 or above (5), normed fit index (NFI) of 0.9 or above (6), root-mean-square error of approximation (RMSEA) of 0.08 or lower (7), standardized root mean residual (SRMR) of 0.05 or lower [30, 31]. All analyses were considered statistically significant at P < 0.05.

Results

General characteristics

A total of 380 older adults were initially enrolled in this study. During the investigation period (September 2020 to March 2021), 12 participants withdrew for personal reasons, resulting in a final sample of 368 participants (140 males, 228 females) with a response rate of 96.8% (Fig. 2). Participants’ ages ranged from 60 to 95 years (mean = 74.2 ± 7.47). Detailed demographic and clinical characteristics are presented in Table 1.

Fig. 2.

Flow chart of survey process

Table 1.

Demographics of the sample (N = 368)

Variables	Categories	N	%
Age (years)	60∼69	110	29.9
	70∼79	163	44.3
	≥ 80	95	25.8
Sex	Male	140	38.0
	Female	228	62.0
Marital status	Married	313	85.1
	Unmarried, divorced or widowed	55	14.9
Education level	Elementary school or less	21	5.7
	Middle school	55	14.9
	High school	96	26.1
	University and above	196	53.3
Financial burden	No burden	206	56.0
	Mild	94	25.5
	Moderate	47	12.8
	Heavy	21	5.70
Exercise habit	Regular a	284	77.2
	Occasional b	44	12.0
	No exercise c	40	10.9

^aRegular refers to at least three times a week, and each time at least 30 min

^bOccasional refers to less than three times a week, or less than 30 min each time

^cNo exercise refers to never or rarely

Descriptive statistics and correlation analysis

Descriptive statistics and normality testing indicated that all variables demonstrated acceptable univariate normality, with skewness absolute values < 2 and kurtosis absolute values < 7 (Table 2). Correlation analyses revealed significant relationships among all study variables (correlation coefficients ranging from − 0.647 to 0.550), with no evidence of multicollinearity (Table 3).

Table 2.

Convergent validity of variables in respondents (N = 368)

Variables	Scale Range	M (SD)	Skewness	Kurtosis	SW test P-value
Fatigue	0∼100	72.3 (18.50)	− 0.818	0.378	0.000
Bodily Pain	0∼100	66.8 (24.39)	− 0.192	− 0.664	0.000
Depressed Mood	0∼15	4.4 (3.13)	0.794	− 0.015	0.000
Anxious Mood	4∼80	40.6 (9.56)	0.884	1.265	0.000
Social Support	12∼66	38.2 (7.60)	− 0.239	− 0.109	0.008
Lower Limb Function	0∼12	9.7 (1.92)	− 0.726	0.064	0.000

M mean, SD standard deviation, SW test, Shapiro-Wilk test

Table 3.

Correlation coefficients between variables in respondents (N = 368)

	1	2	3	4	5	6
1 Bodily Pain	1
2 Depressed Mood	−0.295**	1
3 Anxious Mood	−0.387**	0.550**	1
4 Social Support	0.176**	−0.291**	−0.165**	1
5 Fatigue	0.369**	−0.637**	−0.518**	0.312**	1
6 Lower Limb Function	−0.180**	−0.316**	−0.248**	0.140**	0.332**	1

*P < 0.05, **P < 0.01

Hypothesized model evaluation

The overall fit of the structural model was good: χ2 [11] = 14.894 (P = 0.187) >0.05, χ2/df = 1.354, GFI = 0.990 >0.9, AGFI = 0.968 >0.9, CFI = 0.994 >0.9, NFI = 0.979 >0.9, RMSEA = 0.031 < 0.08, SRMR = 0.023 < 0.05. Consistent with standard SEM practice, covariances among all exogenous variables (bodily pain, depressed mood, anxious mood, and social support) were freely estimated in the model to account for their interrelationships. These indices collectively indicate the model adequately represents the observed data. Figure 3 showed the standardized path estimates for the conceptual model of this study. As shown in Fig. 3, six of the nine hypothesized paths were statistically significant. The supported pathways revealed that: (1) fatigue was directly influenced by bodily pain, depressed mood, anxious mood, and social support; and (2) lower limb function was directly affected by both depressed mood and fatigue.

Fig. 3.

Structural model

Effect decomposition

Table 4 presents the comprehensive decomposition of direct and indirect effects. Depressed mood demonstrated the strongest direct effect on fatigue (β = −0.440, 95% CI [−0.540, −0.338]), followed by anxious mood (β = −0.195, 95% CI [−0.297, −0.095]), social support (β = 0.151, 95% CI [0.058, 0.243]), and bodily pain (β = 0.130, 95% CI [0.038, 0.222]). For lower limb function, both fatigue (β = 0.194, 95% CI [0.074, 0.318]) and depressed mood (β = −0.155, 95% CI [−0.280, −0.019]) showed significant direct effects.

Table 4.

Decomposition of effects of variables for the structural model (N = 368)

	Effect	Estimate	SE	95% CI
Lower Limb Function ←Fatigue	Direct	0.194	0.063	[0.074, 0.318]
	Indirect	0.000	0.000	[0.000, 0.000]
	Total	0.194	0.063	[0.074, 0.318]
Fatigue ←Bodily Pain	Direct	0.130	0.047	[0.038, 0.222]
	Indirect	0.000	0.000	[0.000, 0.000]
	Total	0.130	0.047	[0.038, 0.222]
Lower Limb Function ←Bodily Pain	Direct	0.046	0.056	[−0.061, 0.158]
	Indirect	0.025	0.013	[0.006, 0.057]
	Total	0.071	0.054	[−0.032, 0.178]
Fatigue ←Depressed Mood	Direct	−0.440	0.052	[−0.540, −0.338]
	Indirect	0.000	0.000	[0.000, 0.000]
	Total	−0.440	0.052	[−0.540, −0.338]
Lower Limb Function ←Depressed Mood	Direct	−0.155	0.068	[−0.280, −0.019]
	Indirect	−0.085	0.030	[−0.152, −0.034]
	Total	−0.240	0.060	[−0.353, −0.121]
Fatigue ←Anxious Mood	Direct	−0.195	0.050	[−0.297, −0.095]
	Indirect	0.000	0.000	[0.000, 0.000]
	Total	−0.195	0.050	[−0.297, −0.095]
Lower Limb Function ←Anxious Mood	Direct	−0.045	0.059	[−0.163, 0.072]
	Indirect	−0.038	0.016	[−0.080, −0.014]
	Total	−0.083	0.060	[−0.202, 0.035]
Fatigue ←Social Support	Direct	0.151	0.047	[0.058, 0.243]
	Indirect	0.000	0.000	[0.000, 0.000]
	Total	0.151	0.047	[0.058, 0.243]
Lower Limb Function ←Social Support	Direct	0.000	0.000	[0.000, 0.000]
	Indirect	0.029	0.013	[0.009, 0.063]
	Total	0.029	0.013	[0.009, 0.063]

Mediation analysis demonstrated that all influencing factors exerted significant indirect effects on lower limb function through fatigue, with effects mirroring the direct effect pattern: depressed mood (β = −0.085, 95% CI [−0.152, −0.034]), anxious mood (β = −0.038, 95% CI [−0.080, −0.014]), social support (β = 0.029, 95% CI [0.009, 0.063]), and bodily pain (β = 0.025, 95% CI [0.006, 0.057]). Combining the results of the direct and indirect effects, fatigue fully mediated three pathways: lower limb function ← bodily pain (H6), lower limb function ← anxiety mood (H8), and lower limb function ← social support (H9). For the lower limb function pathway ← depressed mood (H7), proportion analysis showed that fatigue partially mediated 35.4% of the total effect.

Discussion

This study analyzed the relationship between symptoms (fatigue) and influencing factors (bodily pain, depressed mood, anxious mood, and social support) and performance outcomes (lower limb function) in older adults using a cross-sectional design guided by the TOUS framework. Our findings suggested that (1) depressed mood, anxious mood, social support and bodily pain all contribute directly to fatigue (2), both fatigue and depressed mood contribute directly to lower limb function, and (3) fatigue fully mediates the three pathways between bodily pain, anxious mood, and social support on lower limb function and partially mediates the pathway between depressed mood and lower limb function.

Given the significant increase in the aging population, further research on fatigue is urgently needed, as there is evidence that fatigue contributes to adverse health outcomes in older adults [4]. However, fatigue is often considered a natural process of aging, and therefore, most older adults may ignore and underreport fatigue, which may be one of the reasons why fatigue is underappreciated [32, 33]. Several studies have found multiple influences to be associated with fatigue [34]. Indeed, our findings confirmed that a variety of influencing factors work together to cause fatigue. Furthermore, our exploratory analysis revealed an ordinal pattern in the magnitude of effects on fatigue, with psychological factors exhibiting the strongest influence, followed by social and physical factors. To our knowledge, this hierarchical pattern represents a novel finding that warrants further investigation. Importantly, this observation should be considered exploratory rather than conclusive, as it may be influenced by methodological considerations, particularly the use of self-report instruments. The SPPB used in this study was an objective measure of lower limb function, but the remaining four questionnaires were in self-report form. Self-report data are subject to various biases that may affect data accuracy [35]. Cultural context represents another potential influence on self-reporting behaviors [36]. These methodological aspects should be considered when interpreting the study findings. Although the current research findings are exploratory in nature, they remain helpful for clinical practice.

Our findings demonstrated that both fatigue and depressed mood have a direct effect on lower limb function and that fatigue has a greater effect (β = 0.194) than depressed mood (β = −0.155). This may stem from changes in lower limb biometrics. When fatigued, both contractile and diastolic muscles are affected to varying degrees, resulting in the inability of the muscles to achieve effective self-relaxation. Excessive muscle stiffness decreases lower limb strength and increases body sway in older adults, leading to poor balance stability, and in turn, the gait rhythm begins to become unstable, and eventually the dynamic stability of the entire body is enhanced by slowing down the pace [37]. Unlike fatigue, the effects of depressed mood on lower limb function may rely more on pessimistic beliefs and negative cognitive tendencies. The core of depressive mood is known as “anhedonia”, which is mainly characterized by diminished interest and loss of pleasure in all or almost all activities [38], so there is less exercise with depressive mood [39]. This implies that lack of exercise may lead to decreased muscle strength and balance, which may adversely affect low-limb function.

Our findings suggested that bodily pain and anxious mood do not directly affect lower limb function, which may be inconsistent with previous studies [12, 15]. The reasons for this are: first, older adults often use words “uncomfortable” to describe the sensation of bodily pain, which may lead to lower reporting of pain [40]. Second, the older adults in this study had more years of education, which means that they may be more attentive to these “uncomfortable” physical conditions and treat bodily pain at an earlier stage. Third, although anxious mood is characterized by fear, irritability, and difficulty concentrating [41], anxious individuals still have comparable levels of daily function and emotional intimacy to non-anxious individuals [42], so anxious mood has less of a direct impact on lower limb functioning than depressed mood. However, both influences have an indirect effect on lower limb function through fatigue, which further suggests a mediating role of fatigue on lower limb function. Lower limb function is critical for independent daily activities in older adults and is an independent predictor of long-term survival [43]. Lower limb functional limitations can lead to health problems and affect later life abilities. Therefore, the mediation patterns underscore fatigue’s central role in translating multiple risk factors to functional outcomes, suggesting that targeted fatigue management could effectively mitigate the impact of bodily pain, psychological distress, and limited social support on physical function in older adults.

Building on these findings, future research should prioritize the development and evaluation of targeted interventions for fatigue management. Promising directions include multimodal approaches addressing physical, psychological, and social dimensions of fatigue, as well as adapting existing interventions to specifically target fatigue reduction. Such intervention studies would not only validate our exploratory findings but also provide practical strategies for clinical application.

The current study has some limitations. First, our study design was a cross-sectional survey, and even with SEM, causality could not be determined, so the results need to be interpreted with more caution. Nonetheless, the reported results provide useful inter-factor path relationships to guide fatigue management theoretically. Second, the present study was more concerned with whether there were interactions between symptoms, influencing factors, and performance outcomes guided by TOUS. In particular, our assessment of fatigue using the SF-36 Vitality subscale, while psychometrically sound, may not fully encompass the multidimensional nature of fatigue, representing a conceptual limitation in operationalizing this core symptom. Furthermore, specific assessments of other symptom characteristics, including symptom onset, intensity, duration, distress, and quality, were not exhaustive. More detailed information would help to further understand the underlying mechanisms. Finally, although this study has tried to balance the stability of the model as much as possible based on previous studies and the current clinical situation, SEM does not yet have a completely unified method of calculating the sample size, so there may still be some potential effects on the reliability and accuracy of the statistical results. Moreover, this study used convenience sampling. Convenience sampling is practical during the exploratory phase of research, saving resources and providing results quickly. However, due to the lack of randomization, there is a possibility of sampling bias, resulting in an inability to accurately represent the entire population, thus limiting the generalizability of findings. As a next step, we will focus on implementing longitudinal designs, expanding sample diversity, and developing fatigue management interventions to enhance the generalizability and clinical relevance of future findings.

Conclusion

For older adults, lower limb function is critical for independent daily activities and an independent predictor of long-term survival. According to the results of this study, fatigue not only directly affected lower limb function in older adults, but also played a mediating role in the indirect effects of depressed mood, anxious mood, social support, and bodily pain on lower limb function. Therefore, it is suggested that healthcare professionals should prioritize comprehensive fatigue assessment and integrate evidence-based fatigue management into routine geriatric care practices. Healthcare providers should consider implementing multimodal fatigue management strategies that address the identified influencing factors, while researchers should focus on developing and validating targeted interventions for fatigue reduction.

While our study focused on older adults, emerging evidence suggests that fatigue-related issues may be affecting younger populations as well, potentially due to changing lifestyle patterns, work-related stress, and other contemporary factors. However, this observation remains speculative and requires further investigation. Future research on fatigue should consider expanding to include younger age groups to better understand the evolving epidemiology of fatigue across the lifespan.

In addition, the direct effects of depressed mood, anxious mood, social support, and bodily pain on fatigue followed a hierarchical pattern (psychological → social → physical). This suggests that fatigue management strategies might benefit from considering this sequence when prioritizing intervention components. However, as noted in the limitations, this pattern may be influenced by methodological factors, particularly the reliance on self-report measures. Future studies should incorporate objective biomarkers and performance-based measures to validate and refine these preliminary findings. Ultimately, such research advancements will enable more precise and effective fatigue management, contributing to improved physical function and quality of life in aging populations.

Author contributions

YH conceptualized the idea and supervised the project. YH and QW drafted the manuscript. HZ revised the manuscript. All authors participated in data collection and data analysis. All authors read and approved the final manuscript.

Funding

This work was supported by the Shanghai Municipal Health Commission Health Industry Clinical Research Special Program (20234Y0165), Shanghai University of Traditional Chinese Medicine Science and Technology Development Program (23HLZX20), and Key Laboratory of Geriatric Long-term Care(Naval Medical University), Ministry of Education(LNYB-2023-09).

Data availability

The datasets generated and/or analysed during the current study are not publicly available due researchers have not yet completed planned or anticipated analyses for future publication but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The studies involving human participants were reviewed and approved by Chinese PLA General Hospital Medical Ethics Committee (S2021-077-01). This study was conducted in accordance with the ethical principles of the Declaration of Helsinki. The patients/participants provided their written informed consent to participate in this study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.