Presence of lower limb joint arthritis moderates the interrelationship between vision impairment and fear of falling avoidance behavior in older adults Authors: Lin-Wei Hu, Merrill R. Landers, Lung-Chang Chien, Szu-Ping Lee, Fu-Lien Wu Published: 2026 ABSTRACT Falling is a common concern for older adults. Previous studies showed that mobility performance is significantly lower in those with age-related eye diseases or lower limb arthritis, contributing to their increased fall risk. However, it is currently unknown how the co-presence of these 2 age-related sensorimotor impairments interacts to affect fear of falling (FOF) and associated activity-avoidance behaviors. A total of 200 participants recruited from an outpatient regional medical center (127 women, 73 men, age = 65.5 ± 8.4 years) participated in this study. The presence of age-related, uncorrected eye diseases (i.e., cataracts, glaucoma, retinopathy, and macular degeneration) and arthritis in the lower limb joints were recorded. Geriatric Fear of Falling Measure, Fear of Falling Avoidance Behavior Questionnaire, and Timed Up-and-Go test were used to assess FOF, avoidance behavior, and physical mobility, respectively. Moderated mediation analyses, corrected for age and body mass index, were performed. With the Geriatric Fear of Falling Measure as mediator, lower limb arthritis significantly moderated the indirect effects of eye disease on Timed Up-and-Go test performance (B = 1.00; 95% confidence interval = 0.06–2.11) and the Fear of Falling Avoidance Behavior Questionnaire score (B = 5.68; 95% confidence interval = 0.36–11.11). Namely, participants with uncorrected age-related eye disease and lower limb arthritis exhibited significantly poorer mobility and greater avoidance behavior than those without arthritis. The presence of age-related eye diseases exacerbates FOF avoidance behavior and leads to diminished mobility in older adults. The severity of FOF is moderated by the presence of lower limb arthritis. These findings suggest that effective fall prevention strategies for older adults should incorporate comprehensive vision, musculoskeletal, and behavioral evaluations. 1. Introduction Fall-related injuries and other consequences are major public health concerns and a substantial economic burden for society.[1] Annually, around 28%–35% of people over the age of 65 years experience at least 1 fall.[2] Moreover, 1 in 2 persons with a fall history experiences recurring falls.[3] In addition to physical disabilities resulting from fall-related injuries, the psychological impact of falling, including fear of falling (FOF),[4] FOF-related activity avoidance behavior,[5] reduced balance confidence,[6] depressive symptoms,[7] and social isolation,[8] have been reported. These psychosocial consequences of falls can result in a vicious cycle that leads to future functional decline.[4,7,9,10] FOF has been identified as a significant predictor for falls in previous studies.[11–13] Increased levels of FOF may lead to a decline in autonomy and mobility. This, in turn, may result in decreased levels of physical activity.[14,15] The construct commonly known as “Fear of Falling Avoidance Behavior” (FOFAB) has been identified, wherein individuals refrain from engaging in physical activities due to their FOF. FOFAB is a significant psychological risk factor for future falls, leading to a vicious cycle of deconditioning and functional decline.[5,10,16] However, a recent study showed that FOF alone may not predict lower physical activity levels in community-dwelling older adults.[17] Therefore, investigations on how psychological and physical factors interact to affect FOF and attitude toward activities are needed to understand the complex interrelationships underlying fall risk in older adults. Integration of visual, vestibular, and somatosensory inputs is key to stability and mobility during locomotor tasks.[18] Current evidence suggests that vision impairment, either central or peripheral, may be associated with an increased risk for falls in older adults.[19–21] Visual input provides crucial sensory information for maintaining balance, navigating obstacles when walking, and preventing falls.[22,23] Different from vestibular and somatosensory systems, the visual system plays a key role in anticipatory postural adjustments to destabilizing situations.[24,25] Age-related vision impairments have been shown to be associated with depression, lower activity levels, and limited capabilities to perform daily tasks and social functions.[18,26,27] A number of studies demonstrated that age-related vision impairment was significantly associated with FOF, avoidance behavior due to FOF, and diminished mobility.[28–31] Clearly, the current evidence indicates deficits in visual function are a major contributing factor to an increased risk of falling in older adults. What is unknown is how age-related vision impairment interacts with other forms of age-related sensory deficits to impact the risk of falling and associated behavioral changes in this population.[32] One of the most common conditions that affect somatosensory and physical mobility in the older adult population is degenerative arthritis of the lower extremity joints.[33] It is generally agreed that people with arthritis exhibit poorer proprioception acuity than age-matched controls.[34] A 12-month longitudinal study by Doré et al revealed that people with symptomatic osteoarthritis had 53% higher odds of falling than those without.[35] These findings imply that impaired proprioception, weakness, and/or pain due to arthritis may increase the risk of falling by disrupting the sensorimotor functions required for postural control, leading to increased fall risk and potentially FOF and activity avoidance behaviors.[10] Furthermore, individuals with knee arthritis have been shown to exhibit reduced physical activity levels,[36] which in turn could lead to negative downstream consequences.[37] Therefore, the presence of arthritis in older adults may moderate (magnify) the effects of age-related vision impairment on physical mobility and activity avoidance behavior. This theory, while plausible, has not been investigated previously. The purpose of this study was to examine the moderating role of lower limb arthritis in the interrelationship between the presence of age-related eye diseases, physical mobility, and FOF-related psychological measures in community-dwelling older adults. We hypothesized that the presence of arthritis significantly moderates the effects of vision impairment in exacerbating FOF, FOF-related activity avoidance behavior, and physical mobility. 2. Materials and methods 2.1. Participants A cross-sectional study was conducted in an outpatient hospital setting. Participants were recruited from the rehabilitation department of Far Eastern Memorial Hospital, New Taipei City, Taiwan. The inclusion criteria were age >50 years, and able to independently transfer from sitting to standing. Participants were excluded if they could not understand and follow simple instructions or had any conditions that prevented them from walking independently for 50 meters with or without using an assistive device. The procedure for this study was verbally explained to each potential participant prior to enrollment. The study protocol was approved by the Far Eastern Memorial Hospital Institutional Review Board. 2.2. Procedures A licensed physical therapist performed all the data collection after obtaining written informed consent from the participants. Demographic data (i.e., age, gender, height, and weight) and fall incidence frequency in the past year were collected from all participants. In addition, participants were classified into the eye disease subgroup if they did not receive corrective treatments (e.g., lens implant, laser surgery, intravitreal injections) for any of the following diagnoses of age-related eye diseases in at least 1 eye: cataracts, glaucoma, retinopathy, and macular degeneration. Presbyopia was not considered an eye disease in this study.[38,39] People who had more than 1 eye disease in both eyes but were only receiving interventions for only 1 eye or 1 disease at the time of the study were still considered as having eye diseases. Corrective glasses were allowed during the study if the participant wears them regularly to conduct daily activities. The definition of lower limb arthritis was that participants had a medical diagnosis of arthritis in any of the major lower extremity joints (i.e., ankle, knee, or hip) at any severity. 2.3. Outcome measures Geriatric Fear of Falling Measure (GFFM) was used to quantify the severity of FOF experienced by the participant.[40] The GFFM is a self-administered questionnaire consisting of 15 items in 3 aspects: Risk Prevention Attitude, Modifying Behavior, and Psychosomatic Symptoms.[40,41] Evidence for the reliability and validity of the GFFM has been established in Taiwanese older adults.[40] Scoring for each item ranges from 1 to 5, with higher scores indicating greater FOF. Fear of Falling Avoidance Behavior Questionnaire (FFABQ) is an established assessment for evaluating how much a participant’s activity and participation were restricted by his or her FOF.[42,43] This questionnaire consists of 14 activity scenarios. The participant had to rate each item on a scale between 0 (no avoidance) to 4 (complete avoidance) to express their tendency to avoid the activity due to FOF, where a higher total score suggests greater avoidance behavior. The main difference between GFFM and FFABQ is that FFABQ is centered around avoidance behaviors. For example, each FFABQ item asks the participant if they agree with the statement, “Due to my fear of falling, I avoid…(activity).” Standardized physical mobility performance was assessed using the Timed Up-and-Go (TUG) test. The TUG test has been found to predict fall risk in older adults and is widely used for assessing functional mobility.[44,45] Participants were instructed to stand up from a standardized armchair with a seat height of 45 cm and then walk toward a target 3 meters away.[46] After arriving at the target, participants would turn around, walk back to the chair, and sit back down. Participants were instructed to complete the TUG test at a fast but safe pace.[47] They were allowed to use assistive devices if necessary. The average time of 3 trials of the TUG test was recorded for analysis. 2.4. Statistical analysis Demographic data of the study participants were analyzed with descriptive statistics. Independent t tests were used to determine if the presence of eye disease (yes or no) and lower limb arthritis (yes or no) influences GFFM, FFABQ, and TUG. Chi-square tests were used to analyze the effects of the presence of eye diseases and lower limb arthritis on fall incidence frequency and gender distribution. Pearson correlation analyses were used to examine the associations among the continuous variables. We adopted the moderated mediation analysis to investigate the relationship between TUG or FFABQ and the presence of eye diseases (Fig. 1).[48] Models were developed to predict TUG time or FFABQ based on the presence of eye diseases, with GFFM as a mediator and lower limb arthritis as a moderator. Participants’ age and body mass index (BMI) were controlled in the model as potential confounding variables. These relationships were formulated as the following model equations. Statistical models. BMI = body mass index, FFABQ = Fear of Falling Avoidance Behavior Questionnaire, GFFM = Geriatric Fear of Falling Measure, TUG = Timed Up-and-Go. where M represents GFFM as the mediator, X represents eye disease as the main predictor, Y represents either TUG or FFABQ as the main outcome, W represents lower limb arthritis as the moderator, and C1 and C2 age and BMI as confounders. Moreover, α and θ are regression intercepts, β is the regression slope of W to predict M. In particular, the regression slope c was used to estimate the direct effect, and (a+dW)b was used to estimate the conditional indirect effects of X on Y. Lastly, γ1, γ2, δ1, and δ2 are regression slopes of confounders, aneM and eY are error terms. All parameters in the models were estimated by the ordinary least squares method. Conditional direct and indirect effects were computed with and without lower limb arthritis. The 95% confidence interval (CI) of each effect comprised the 2.5th and 97.5th percentiles from 5000 bootstrapping samples to determine the effect significance. The index of moderated mediation is the difference between the parameter estimates for the indirect effect at 2 levels (vs W=0, W=1) of the moderator, which tests whether the moderator changes the mediating effect. Sensitivity analyses were performed by entering/removing covariates to evaluate model robustness. All statistical analyses were performed with SPSS version 22.0 (IBM Corp.). Mediation analyses were performed through the PROCESS macro.[49] A post hoc power analysis using G*Power 3.1 was performed to estimate the power of the observed effects in the study sample. The significance level was set as 0.05. 3. Results Data from 200 participants were collected from September to November of 2020 (127 women, 73 men, age = 65.5 ± 8.4 years, BMI = 24.5 ± 3.0 kg/m2). Excluding missing data (n = 6), the most common eye disease was cataracts (n = 53, 27.3%), followed by glaucoma (n = 6, 3.1%), macular degeneration (n = 6, 3.1%), retinopathy (n = 2, 1.0%), and more than 1 disease (n = 19, 9.8%). Overall, 46.0% (n = 92) of the participants reported having age-related eye disease, and 19.6% (n = 39) reported having lower limb arthritis. Stratifying by the presence of age-related conditions, participants with eye diseases were older (68.5 years vs 62.9 years; P < .0001) with significantly higher proportions of having lower limb arthritis (30.4% vs 10.1%; P = .0011; Table 1) as well as a higher likelihood of experiencing a fall in the past year (32.6% vs 22.0%; P = .041). Participants with lower limb arthritis were older (69.8 years vs 64.4 years; P = .0003) with a significantly higher BMI (25.45 vs 24.27; P = .0293) and a significantly higher GFFM (36.4 vs 28.5; P = .0003) when compared to those without arthritis. Stratifying by fall experience, participants who had fallen in the past year (n = 54, 27.0%) had significantly higher FFABQ (18.24 vs 8.58; P = .0001) and GFFM scores (36.65 vs 27.88; P = .001), as well as longer TUG test time (11.88 vs 9.72; P = .005). The correlation analyses between all continuous variables showed that GFFM was moderately correlated to FFABQ (R = 0.63, 95% CI = 0.52–0.74) and TUG (r = 0.53, 95% CI = 0.41–0.65). FFABQ was correlated to TUG (r = 0.38, 95% CI = 0.25–0.51). Table 1The demographic data and fall history in both groups.Eye diseases (n = 92; % = 46.0)No eye diseases (n = 108; % = 54.0)P valueAge (yr)68.5 ± 7.662.9 ± 8.4<.0001*Male, n (%)30 (32.6)43 (39.5).2914†Body mass index (kg/m2)24.7 ± 2.924.4 ± 3.1.4206*Had lower limb arthritis (missing data n = 1), n (%)28 (30.4)11 (10.1).0011†Had at least 1 fall in the past year, n (%)30 (32.6)24 (22.0).0410†GFFM total score32.0 ± 13.128.8 ± 11.9.0687* RP11.6 ± 5.010.5 ± 4.7.1118* MB12.1 ± 5.811.0 ± 5.5.1932* PS8.4 ± 4.67.2 ± 3.7.0369*FFABQ12.8 ± 13.59.8 ± 12.9.1068*TUG (s)10.5 ± 2.810.1 ± 4.1.4910*Lower limb arthritis (n = 39; % = 19.5)No lower limb arthritis (n = 160; % = 80.0)P valueAge (yr)69.8 ± 7.364.4 ± 8.4.0003*Male, n (%)11 (28.2)62 (38.8%).3531†Body mass index (kg/m2)25.5 ± 2.824.3 ± 3.1.0293*Had at least 1 fall in the past year, n (%)13 (33.3)41 (25.6).5176†GFFM total score36.4 ± 12.028.5 ± 11.8.0003* RP13.1 ± 4.310.4 ± 4.7.0016* MB13.5 ± 5.410.9 ± 5.6.0118* PS9.9 ± 5.07.1 ± 3.5.0025*FFABQ13.5 ± 13.910.6 ± 13.1.2144*TUG (s)11.1 ± 2.410.0 ± 3.7.0980*Data were presented as mean ± standard deviation for continuous variables.FFABQ = Fear of Falling Avoidance Behavior Questionnaire, GFFM = Geriatric Fear of Falling Measure, MB = Modifying Behavior, PS = Psychosomatic Symptoms, RP = Risk Prevention Attitude, TUG = Timed Up-and-Go test.*Independent t test.†Chi-square test. The demographic data and fall history in both groups. Data were presented as mean ± standard deviation for continuous variables. FFABQ = Fear of Falling Avoidance Behavior Questionnaire, GFFM = Geriatric Fear of Falling Measure, MB = Modifying Behavior, PS = Psychosomatic Symptoms, RP = Risk Prevention Attitude, TUG = Timed Up-and-Go test. Independent t test. Chi-square test. In our moderated mediation analyses, data from 199 participants (one missing) were analyzed. Because age and BMI were found to be significantly different between arthritis and eye disease groups, they were included in the models as covariates. Table 2 shows that the interaction between eye disease and lower limb arthritis on GFFM was statistically significant (B = 9.43; 95% CI = 0.58–18.28; P = .0369). Moreover, significant effects of GFFM on FFABQ (B = 0.69; 95% CI = 0.57–0.82; P < .0001) and TUG (B = 0.12; 95% CI = 0.09–0.16; P < .0001) were detected. A second moderated mediation model examined the moderated mediation effect of lower limb arthritis (Fig. 1). The indirect effect of eye disease on FFABQ (5.68; 95% CI = 0.39–11.34) and TUG (1.00; 95% CI = 0.08–2.11) through the mediation of GFFM was significant when participants also had lower limb arthritis (Table 3). However, the direct effect of eye disease on FFABQ (0.93; 95% CI = −2.15 to 4.01) and TUG (−0.72; 95% CI = −1.58 to 0.14) was not significant. The index of moderated mediation tests if the moderator (arthritis) significantly moderates the indirect effect of the predictor (eye disease) on the dependent variable (FFABQ or TUG). The results revealed that lower limb arthritis was a significant moderator, while GFFM was only a significant mediator when the moderator (i.e., arthritis) was included in the model (Fig. 1). The post hoc power analysis demonstrated that the achieved power of the 2 mediation models was both >0.99. Sensitivity analyses (see Supplementary data, Table S1, Supplemental Digital Content 1) showed the index of moderated mediation in the unadjusted model on FFABQ and TUG was 5.52 (95% CI = −0.60 to 11.62) and 1.18 (95% CI = −0.12 to 2.76), respectively, similar to the original (adjusted) model (B = 6.62; 95% CI = 0.67–12.70; 1.16; 95% CI = 0.14–2.51). Table 2Testing the moderated mediation effect of the presence of eye disease on FFABQ and TUG through the mediator (GFFM) conditionally at the level of the moderator (lower limb arthritis; N = 199).Dependent variablePredictorAdj-B*95% CIP valueModel I: eye disease on GFFM (R2 = 0.1506) GFFM (M)Eye disease (X)−1.36(−5.17 to 2.45).4812 GFFM (M)Lower limb arthritis (W)−0.90(−8.23 to 6.43).8085 GFFM (M)Eye disease × lower limb arthritis (XW)9.54(0.70 to 18.38).0346Model II: eye disease on FFABQ (R2 = 0.4046) FFABQ (Y)GFFM (M)0.69(0.57 to 0.82)<.0001 FFABQ (Y)Eye disease (X)0.93(−2.15 to 4.01).5525Model II: eye disease on TUG (R2 = 0.3361) TUG (Y)GFFM (M)0.12(0.09 to 0.16)<.0001 TUG (Y)Eye disease (X)−0.72(−1.58 to 0.14).0983CI = confidence interval, FFABQ = Fear of Falling Avoidance Behavior Questionnaire, GFFM = Geriatric Fear of Falling Measure, TUG = Timed Up-and-Go test.*All estimates were adjusted by age and body mass index. Testing the moderated mediation effect of the presence of eye disease on FFABQ and TUG through the mediator (GFFM) conditionally at the level of the moderator (lower limb arthritis; N = 199). CI = confidence interval, FFABQ = Fear of Falling Avoidance Behavior Questionnaire, GFFM = Geriatric Fear of Falling Measure, TUG = Timed Up-and-Go test. All estimates were adjusted by age and body mass index. Table 3The direct and indirect effect of eye disease on FFABQ or TUG.B95% CIDependent variable: FFABQ (index of moderated mediation = 6.62; 95% CI = 0.67–12.70) Direct effect0.93(−2.15 to 4.01) Indirect effect(with arthritis) 5.68(0.36 to 11.11)(without arthritis) −0.95(−3.71 to 1.74)Dependent variable: TUG (index of moderated mediation = 1.16; 95% CI = 0.14–2.51) Direct effect−0.72(−1.58 to 0.14) Indirect effect(with arthritis) 1.00(0.06 to 2.11)(without arthritis) −0.17(−0.79 to 0.29)CI = confidence interval, FFABQ = Fear of Falling Avoidance Behavior Questionnaire, TUG = Timed Up-and-Go test. The direct and indirect effect of eye disease on FFABQ or TUG. CI = confidence interval, FFABQ = Fear of Falling Avoidance Behavior Questionnaire, TUG = Timed Up-and-Go test. 4. Discussion The findings supported our hypotheses that greater FOF as measured by GFFM is significantly related to greater FOF avoidance behavior and diminished physical mobility. The presence of lower limb arthritis significantly moderates the effects of age-related eye disease on mobility and avoidance behavior. Specifically, the presence of lower limb arthritis magnified the effects of the presence of eye disease, resulting in lower physical mobility and more severe activity avoidance behavior. This effect was only significant when both eye disease and arthritis were considered together. To our knowledge, the current study is the first to describe the interrelationship between 2 age-related sensory impairments (i.e., vision and somatosensory) and how they interact to affect physical functioning and psychological behaviors in older adults. FOFAB and TUG have previously been identified as predictors of future falls in older adults.[5] Lower limb arthritis, particularly in the knee, has been confirmed to reduce balance confidence[50,51] and increase the risk of falling[52,53] and FOF.[50,54,55] These adverse effects were potentially caused by pain or stiffness,[35] weakness,[56,57] impaired proprioception,[56] and reduced mobility.[57] Therefore, it is reasonable to believe that concurrently impaired somatosensory and visual inputs due to arthritis and eye disease may result in worse performance on TUG and excessive FOFAB. This result was in agreement with Rosadi et al, who showed that poor proprioception and FOF are significantly related to falls in Taiwanese older adults.[58] Our findings highlight the importance of considering the cumulative effects of sensorimotor impairments due to age-related degeneration in managing the risks of falling in this population. Assessing FOF and the associated behavioral changes such as FOFAB is key to understanding the complex psycho-physical aftermath of the fall experience in older adults.[5,59] For example, individuals with physical activity avoidance behavior, a common sequela after experiencing falls, are more likely to exhibit greater catastrophization,[60] depression,[61] and anxiety.[61] Depression is significantly related to social isolation,[62] which has been considered a public health emergency for the aging population.[63] Although previous studies have shown that FOF plays a significant role in predicting future falls and decreased mobility, understanding what factors contribute to the decline of function and participation is an ongoing field of investigation.[64,65] The presence of uncorrected age-related eye diseases and impaired vision has previously been identified as a key risk factor for falling in older adults.[66] The prevalence of age-related eye diseases in Taiwanese older adults has been reported to be around 17%, leading to disrupted quality of life, particularly in physical and social functions.[67,68] Correcting age-related vision impairments (i.e., cataract) has been shown to significantly reduce the frequency of falling.[69,70] While visual input is important for fall prevention, it is important to understand which quality of vision contributes to this effect. For example, a study by Lee et al showed that the presence of age-related eye disease predicted older adult participants’ history of falling, while visual acuity as commonly administered in general eye exams did not.[31] Similarly, Mehta et al found that contrast sensitivity and stereoacuity but not visual acuity were associated with fall history.[66] Depth perception and contrast sensitivity have been shown to be more important than visual acuity for maintaining balance and avoiding falls.[21] Dynamic visual acuity testing has also been proposed as a better screening tool for fall risks in older adults.[71] These seem to suggest that a paradigm shift to replace the currently most common vision test (i.e., static visual acuity using a Snellen chart) is needed to effectively predict fall risk in older adults. The current study provides insights to address the public health concern that older adults with glaucoma or cataracts, 2 of the most common age-related eye diseases, had an increase in the rate of falls when compared to age-similar controls without eye diseases.[72–75] It is reasonable to postulate that fall incidences due to vision-related reasons may be higher in populations where uncorrected age-related eye diseases are more prevalent. For example, the prevalence of cataracts in this study (27.3%) is similar to the 2017 Taiwan national report (28.4%),[76] and was significantly higher than the global average (17.2%).[77] This may have contributed to the significantly higher prevalence of falls in Taiwanese older adults when compared to other regions of the world.[78] Comprehensive evaluation of risk factors associated with falls, particularly vision impairment, may be particularly important in certain societies with an aging population. Our findings further confirm that visual impairment can adversely affect balance and physical mobility in individuals with lower limb arthritis. Good postural/balance control relies on the effective integration of visual, vestibular, and somatosensory inputs. Therefore, impaired vision may compromise postural stability and increase the risk of falls.[79] Individuals with visual impairment may also develop greater concerns about falling and avoid activities perceived as challenging or unsafe. Such avoidance would reduce physical activity, leading to deconditioning and further functional decline, subsequently increasing the risk of falls and reinforcing a vicious cycle of fear and activity restriction.[10] Accordingly, fall prevention strategies should consider the interplay among visual impairment, FOF, avoidance behavior, and coexisting neurological or neurodegenerative conditions.[80,81] In addition to visual impairment, other factors, including cognitive decline and physical activity levels, may also lead to balance and mobility limitations in individuals with lower limb arthritis. 5. Study limitations Limitations of the study included the following: first, we did not evaluate all potential factors that might be related to falls, such as psychological factors outside of FOF,[82,83] medications,[84] cognitive function,[85] and extremity muscle strength.[86] Second, the severity of arthritis, pain severity, number of arthritic joints, severity of eye disease, vestibular function, and other pathological conditions that may affect somatosensation were not assessed. Lastly, the causal relationship cannot be established due to the cross-sectional nature of the study. Prospective research in the future is needed to confirm the interrelationship we discovered and to examine how it responds to interventions. 6. Conclusion The presence of lower limb arthritis aggravates the impact of age-related eye disease on FOF and physical mobility, subsequently resulting in greater activity avoidance behaviors in older adults. We recommend that fall prevention and intervention strategies should incorporate physical performance, psychological factors (FOF or other related factors), and other age-related diseases affecting sensorimotor functions. Author contributions Conceptualization: Szu-Ping Lee. Data curation: Lin-Wei Hu, Szu-Ping Lee. Formal analysis: Lung-Chang Chien, Szu-Ping Lee, Fu-Lien Wu. Funding acquisition: Lin-Wei Hu, Szu-Ping Lee. Investigation: Lin-Wei Hu, Szu-Ping Lee. Methodology: Lung-Chang Chien, Szu-Ping Lee, Fu-Lien Wu. Project administration: Szu-Ping Lee. Resources: Szu-Ping Lee. Software: Lung-Chang Chien, Szu-Ping Lee, Fu-Lien Wu. Supervision: Merrill R. Landers, Lung-Chang Chien, Szu-Ping Lee. Validation: Lung-Chang Chien. Visualization: Lung-Chang Chien, Fu-Lien Wu. Writing – original draft: Szu-Ping Lee, Fu-Lien Wu. Writing – review & editing: Merrill R. Landers, Lung-Chang Chien, Szu-Ping Lee, Fu-Lien Wu.