Sleep Disturbance as a Mediator Between Self-Reported Nighttime Noise Annoyance and Depression, Anxiety, and Suicide: A Mediation Analysis

Abstract

Objective:

To examine whether sleep disturbance mediates associations between self-reported nighttime noise annoyance and depression, anxiety, and suicide risk in Taiwanese adults.

Methods:

We analyzed cross-sectional online survey data from 776 Taiwanese adults collected between December 2024 and May 2025. Self-reported nighttime noise annoyance from road traffic, railways, aircraft, and residential building vibrations was assessed on a 5-point verbal scale. Sleep disturbance was assessed with the Pittsburgh Sleep Quality Index (PSQI), and depression, anxiety, and suicide risk were assessed with validated self-report scales. Mediation analyses estimated total, natural direct, and natural indirect effects using 5000 bootstrap resamples.

Results:

Self-reported nighttime noise annoyance from road traffic, aircraft, and residential building vibrations was associated with poorer subjective sleep quality. Self-reported nighttime noise annoyance from all four sources was associated with longer sleep latency, more sleep disturbances, greater use of sleeping medication, worse daytime dysfunction, and higher global PSQI scores (B = 0.652–0.788; all P < 0.001), but not with sleep duration or habitual sleep efficiency. Significant natural indirect effects through sleep disturbance were observed for anxiety, depression, and suicide risk for road traffic annoyance (B = 0.076–0.463) and residential building vibration annoyance (B = 0.082–0.503); 95% confidence intervals excluded 0. For residential building vibration annoyance, 40.1%–63.6% of total associations were mediated.

Conclusion:

Sleep disturbance may mediate associations between self-reported nighttime noise annoyance and depression, anxiety, and suicide risk. Populations reporting nighttime noise annoyance may benefit from sleep assessment and source-specific noise mitigation.

KEY MESSAGES

• (1)Self-reported nighttime noise annoyance from road traffic, railways, aircraft, and residential building vibrations is associated with poor sleep quality in Taiwanese adults.
• (2)Sleep disturbance significantly mediated the associations of self-reported nighttime noise annoyance from road traffic and residential building vibrations with depression, anxiety, and suicide risk.
• (3)Sleep disturbance mediated 40.1% to 63.6% of the total effects of self-reported nighttime noise annoyance from residential building vibrations on depression, anxiety, and suicidality.

INTRODUCTION

Self-reported nighttime noise annoyance can disrupt sleep. It has been associated with depression, anxiety, and suicide risk. Sleep disturbance is one of the most harmful nonauditory consequences of noise exposure. Even low maximum A-weighted sound pressure levels (33 dB) can elicit physiological arousals that fragment sleep, impair sleep continuity, and contribute to short-term and long-term health risks.[1] Longitudinal, cross-sectional, and population-based studies have linked residential road traffic noise with depressive symptoms and incident depression.[2,3,4,5] Both human and animal studies have associated noise annoyance with anxiety.[2,3,6] Residential noise exposure has been associated with anxiety and depression during late pregnancy.[7] Furthermore, nighttime or transportation noise, particularly from road traffic and railways, has been associated with suicide risk.[8,9] Whether sleep disturbance links self-reported nighttime noise annoyance to psychiatric symptoms and suicide risk remains unclear.

Although studies have reported associations between noise annoyance and mental health, several gaps remain in the literature. First, few studies have evaluated sleep disturbance as a mediator of the association between noise annoyance and mental health.[10,11,12] Second, most studies have focused on specific populations, such as young adults, university students, or individuals living near wind turbines, which limits generalizability. Third, whether sleep disturbance may mediate the association between noise annoyance and mental health in the general population remains unclear.[4] Two studies identified sleep disturbance as a mediator of the association between residential noise exposure and mental health in young adults or university students,[11,12] whereas a study of wind turbine sound modeled noise annoyance, rather than sleep disturbance, as the mediator linking sound exposure with sleep disturbance and psychological distress.[10] Finally, most studies have considered only a single noise source or general ambient noise, leaving uncertainty about whether mediation differs across source-specific self-reported nighttime noise annoyance and mental health outcomes.

The present study addressed the aforementioned gaps by examining whether sleep disturbance mediates the associations between self-reported nighttime noise annoyance from road traffic, railways, aircraft, and residential building vibrations and depression, anxiety, and suicide risk in Taiwanese adults. By comparing multiple noise sources, we aimed to clarify source-specific pathways linking self-reported nighttime noise annoyance with mental health.

METHODS

Between 1 December 2024 and 31 May 2025, an online survey was conducted among adults (≥18 years) residing in Taiwan. Recruitment advertisements were disseminated through Facebook (Meta Platforms Inc., USA), Instagram (Meta Platforms Inc., USA), Dcard (Dcard Holdings Ltd., Taiwan), and the PTT Bulletin Board System (National Taiwan University, Taiwan). The study protocol was approved by the Central Regional Research Ethics Committee of China Medical University (approval number: CRREC-113-004). Written informed consent was obtained from all participants.

Measures

Sociodemographic Characteristics

Data on participants’ sex, age, education level, and yearly income were collected. These variables were treated as potential confounders.

Self-Reported Nighttime Noise Annoyance

Self-reported nighttime noise annoyance from road traffic, railways, aircraft, and residential building vibrations (of floors, ceilings, or walls) was assessed using the International Commission on Biological Effects of Noise verbal scale. Responses were rated using a 5-point scale with end points ranging from 1 (not at all) to 5 (extremely).[13] Participants were asked the following question: “During the past 12 months, when you wanted to sleep at home at night, how much did noise from [noise source] bother, disturb, or annoy you”?.[14]

Sleep Disturbance

Sleep disturbance was assessed using the Pittsburgh Sleep Quality Index (PSQI), a validated self-report questionnaire that evaluates seven components: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication, and daytime dysfunction. Each component is scored on a scale with end points ranging from 0 to 3. The total global PSQI score ranges from 0 to 21, with higher scores indicating poorer sleep quality. The PSQI demonstrated strong internal reliability (Cronbach’s α = 0.83).[15] We additionally assessed sleep duration (h) and habitual sleep efficiency (%) as continuous measures to capture variability beyond that reflected by the component scores.

Depression

Depressive symptoms were assessed using the Patient Health Questionnaire-9 (PHQ-9), a nine-item self-report instrument based on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria for major depressive disorder.[16] Respondents rated symptom frequency over the previous 2 weeks on a scale with end points ranging from 0 (not at all) to 3 (nearly every day). Higher total scores indicated more severe depressive symptoms. The PHQ-9 exhibited high internal consistency (Cronbach’s α = 0.89).[17]

Anxiety

Anxiety symptoms were assessed using the Generalized Anxiety Disorder-7 (GAD-7), a seven-item self-report instrument. Respondents rated symptom frequency over the previous 2 weeks on a 4-point Likert scale. Higher total scores indicated more severe anxiety symptoms. The GAD-7 exhibited high internal consistency (Cronbach’s α = 0.92).[18]

Suicide

Suicidal ideation and behavior were assessed using the six-item Columbia Suicide Severity Rating Scale (C-SSRS) Self-Report Recent instrument. This scale uses yes/no items to assess suicidal ideation and behavior. Responses range from a wish to die to active suicidal ideation with a specific plan and intent. The scores for all six items were summed. Higher scores indicated higher suicide risk. The C-SSRS exhibited strong internal consistency (Cronbach’s α = 0.937–0.946).[19]

Patient and Public Involvement

Patients, service users, caregivers, and members of the public were not involved in the design, conduct, analysis, or interpretation of this study. The study protocol—including the research questions, measurement selection, and the mediation analysis plan—was developed by the principal investigator based on disciplinary expertise and prior research experience.

Statistical Analysis

Descriptive statistics were used to summarize participants’ sociodemographic characteristics and scores for self-reported nighttime noise annoyance, sleep disturbance, and mental health outcomes. Categorical variables are presented as numbers and percentages. Continuous variables are presented as mean ± standard deviation or median (interquartile range), as appropriate. All statistical analyses were conducted using R version 4.6.1 (R Foundation for Statistical Computing, Vienna, Austria).[20]

We first examined the distributional characteristics of the psychometric mediator (sleep disturbance) and outcome variables (anxiety, depression, and suicide risk) using skewness and excess kurtosis. Substantial nonnormality was defined as an absolute skewness value of >2 or an excess kurtosis value of >4.[21] The global PSQI (skewness = 0.547; excess kurtosis = 0.031), PHQ-9 (skewness = 0.833; excess kurtosis = 0.616), and GAD-7 (skewness = 0.840; excess kurtosis = 0.202) scores did not exhibit substantial skewness or excess kurtosis. By contrast, the C-SSRS score exhibited substantial positive skewness and excess kurtosis (skewness = 2.268; excess kurtosis = 4.503).

To examine the associations between source-specific self-reported nighttime noise annoyance and PSQI component-level and item-level measures, we used linear regression models adjusted for sex, age, education level, and yearly income. Regression coefficients, 95% confidence intervals (CIs), and P-values were calculated. Statistical significance was defined as a two-sided α level of 0.05.

To determine whether sleep disturbance mediated the associations between self-reported nighttime noise annoyance with mental health outcomes, we conducted mediation analyses using the mediation package in R.[22] The models were adjusted for sex, age, education level, yearly income, and self-reported nighttime noise annoyance from sources other than that being assessed. We used 5000 nonparametric bootstrap samples to calculate 95% percentile CIs. This approach accommodates the potentially asymmetric sampling distribution of indirect effects. The total effect, natural direct effect, natural indirect effect (NIE), and proportion of association mediated are reported. Notably, the proportion of association mediated is not reported when the NIE exceeded the total effect in absolute magnitude or when the NIE and total effect had opposite directions. Significance was defined as the 95% percentile CI not including 0.

RESULTS

Table 1 presents the characteristics of the study cohort (N = 776; men: 63.9%; mean age: 34.3 ± 10.4 years). Because yearly income and the C-SSRS suicidal ideation score were right-skewed, both are presented as medians (interquartile ranges). Yearly income was 50.0 (24.0, 70.0) in units of 10,000 New Taiwan dollars, and the C-SSRS score was 0.0 (0.0, 1.0). The mean global PSQI score was 6.4 ± 3.3, the mean PHQ-9 score was 7.1 ± 5.4, and the mean GAD-7 score was 5.9 ± 5.0.

Table 1.

Sociodemographic, sleep-related, and mental health-related characteristics of the study cohort.

Variable	N = 776
Sex
Male	496 (63.9)
Female	280 (36.1)
Education level
Senior high school or below	88 (11.3)
College/University	522 (67.3)
Graduate school	166 (21.4)
Age (yr)	34.3 ± 10.4
Yearly income (10,000 NTD)	50.0 (24.0, 70.0)
Self-reported nighttime noise annoyance
Road traffic noise annoyance	2.2 ± 1.0
Railway noise annoyance	1.2 ± 0.7
Aircraft noise annoyance	1.3 ± 0.7
Residential building vibration noise annoyance	2.1 ± 1.0
Pittsburgh Sleep Quality Index (PSQI)
Subjective sleep quality	1.4 ± 0.8
Sleep latency	1.3 ± 0.9
Sleep duration	0.5 ± 0.9
Sleep hours	7.4 ± 1.5
Habitual sleep efficiency	0.4 ± 0.8
Habitual sleep efficiency percent	0.9 ± 0.1
Sleep disturbances	1.4 ± 0.7
Use of sleeping medication	0.4 ± 0.8
Daytime dysfunction	1.0 ± 0.8
Global PSQI	6.4 ± 3.3
Patient Health Questionnaire-9 (depression score)	7.1 ± 5.4
Generalized Anxiety Disorder-7 (anxiety score)	5.9 ± 5.0
C-SSRS (suicidal ideation score)	0.0 (0.0, 1.0)

Notes: Abbreviations: C-SSRS, Columbia-Suicide Severity Rating Scale; IQR, interquartile range; n, number; NTD, New Taiwan dollars; SD, standard deviation. Values are n (%), mean ± SD, or median (IQR), as appropriate. Yearly income and the C-SSRS suicidal ideation score are presented as median (IQR).

Adjusted linear regression models indicated that greater self-reported nighttime noise annoyance from road traffic, aircraft, and residential building vibrations was associated with poorer subjective sleep quality. However, the association between self-reported nighttime noise annoyance from railways and sleep quality was nonsignificant. Greater self-reported nighttime noise annoyance from all four sources was associated with longer sleep latency, more sleep disturbances, greater use of sleeping medication, worse daytime dysfunction, and higher global PSQI scores [Table 2]. No noise source was associated with sleep duration or habitual sleep efficiency, regardless of whether these measures were analyzed as continuous item-level measures (hours and percentages) or component scores (0–3).

Table 2.

Associations between self-reported nighttime noise annoyance and sleep-related variables.

	Road traffic noise annoyance		Railway noise annoyance		Aircraft noise annoyance		Residential building vibration noise annoyance
Outcome	B (95% CI)	P-value	B (95% CI)	P-value	B (95% CI)	P-value	B (95% CI)	P-value
PSQI component level
Subjective sleep quality	0.148 (0.091, 0.204)	<0.001	0.066 (−0.017, 0.149)	0.117	0.103 (0.022, 0.184)	0.013	0.154 (0.099, 0.210)	<0.001
Sleep latency	0.125 (0.064, 0.186)	<0.001	0.122 (0.034, 0.210)	0.007	0.122 (0.035, 0.209)	0.006	0.162 (0.102, 0.221)	<0.001
Sleep duration	0.060 (−0.002, 0.121)	0.058	0.023 (−0.066, 0.112)	0.614	0.063 (−0.025, 0.150)	0.161	0.034 (−0.027, 0.094)	0.276
Habitual sleep efficiency	0.054 (−0.005, 0.114)	0.072	0.047 (−0.039, 0.132)	0.285	0.005 (−0.079, 0.089)	0.909	0.020 (−0.038, 0.078)	0.500
Sleep disturbances	0.134 (0.085, 0.183)	<0.001	0.101 (0.030, 0.173)	0.005	0.152 (0.082, 0.222)	<0.001	0.162 (0.114, 0.209)	<0.001
Use of sleeping medication	0.108 (0.052, 0.164)	<0.001	0.176 (0.096, 0.256)	<0.001	0.151 (0.072, 0.230)	<0.001	0.089 (0.034, 0.144)	0.001
Daytime dysfunction	0.148 (0.088, 0.207)	<0.001	0.120 (0.033, 0.207)	0.007	0.101 (0.016, 0.187)	0.021	0.167 (0.109, 0.225)	<0.001
PSQI Global score
Global PSQI	0.777 (0.547, 1.007)	<0.001	0.652 (0.315, 0.989)	<0.001	0.696 (0.364, 1.028)	<0.001	0.788 (0.563, 1.013)	<0.001
PSQI Item level
Sleep hours	−0.032 (−0.141, 0.078)	0.572	−0.011 (−0.169, 0.146)	0.890	−0.121 (−0.270, 0.034)	0.126	0.001 (−0.106, 0.109)	0.980
Habitual sleep efficiency percent	−0.010 (−0.020, 0.001)	0.070	−0.012 (−0.027, 0.004)	0.133	−0.001 (−0.016, 0.014)	0.879	−0.006 (−0.017, 0.004)	0.235

Note: Abbreviations: B, regression coefficient; CI, confidence interval; PSQI, Pittsburgh Sleep Quality Index. All analyses were adjusted for sex, education level, age, and yearly income. Significant P-values (<0.05) are boldfaced.

Mediation analyses were conducted to determine whether sleep disturbance mediated the associations between self-reported nighttime noise annoyance and mental health outcomes [Table 3]. The NIEs of self-reported nighttime noise annoyance from road traffic and residential building vibrations on anxiety, depression, and suicide risk were significant (95% percentile CIs did not include 0). However, the NIEs of self-reported nighttime noise annoyance from railways and aircraft on mental health outcomes were nonsignificant. Notably, sleep disturbance mediated 45.7%–52.3% of the total effects of self-reported nighttime noise annoyance from road traffic on anxiety and depression and 40.1%–63.6% of the total effects of self-reported nighttime noise annoyance from residential building vibrations on anxiety, depression, and suicide risk.

Table 3.

Mediation of the associations between self-reported nighttime noise annoyance and mental health outcomes by sleep disturbance.

Source of self-reported nighttime noise annoyance	Outcome	Total Effect	NIE	NDE	Proportion mediated
		B (95% CI)	B (95% CI)	B (95% CI)	Estimate (95% CI)
Road traffic noise annoyance	Anxiety	0.742 (0.300, 1.217)	0.388 (0.157, 0.626)	0.354 (-0.007, 0.727)	0.523 (0.273, 1.000)
	Depression	1.013 (0.520, 1.532)	0.463 (0.179, 0.758)	0.550 (0.203, 0.925)	0.457 (0.245, 0.708)
	Suicide	0.042 (−0.071, 0.158)	0.076 (0.030, 0.130)	−0.033 (−0.141, 0.075)	N/A
Railway noise annoyance	Anxiety	−0.076 (−0.701, 0.560)	0.072 (−0.302, 0.447)	−0.147 (−0.700, 0.421)	N/A
	Depression	0.163 (−0.524, 0.891)	0.086 (−0.344, 0.533)	0.077 (−0.527, 0.710)	0.526 (0.000, 1.000)
	Suicide	0.037 (−0.179, 0.278)	0.014 (−0.063, 0.091)	0.023 (−0.168, 0.236)	0.377 (0.000, 1.000)
Aircraft noise annoyance	Anxiety	0.657 (0.107, 1.236)	0.287 (−0.025, 0.623)	0.369 (−0.089, 0.844)	0.438 (0.000, 1.000)
	Depression	0.826 (0.141, 1.506)	0.343 (−0.018, 0.728)	0.483 (−0.083, 1.030)	0.415 (0.000, 1.000)
	Suicide	0.346 (0.109, 0.581)	0.056 (−0.006, 0.123)	0.290 (0.078, 0.498)	0.162 (0.000, 1.000)
Residential building vibration noise annoyance	Anxiety	1.052 (0.617, 1.501)	0.421 (0.211, 0.634)	0.631 (0.259, 1.034)	0.401 (0.227, 0.635)
	Depression	0.931 (0.482, 1.386)	0.503 (0.258, 0.756)	0.428 (0.066, 0.788)	0.540 (0.331, 0.888)
	Suicide	0.129 (0.018, 0.247)	0.082 (0.041, 0.129)	0.047 (−0.062, 0.160)	0.636 (0.240, 1.000)

Note: Abbreviations: B, regression coefficient; CI, confidence interval; NDE, natural direct effect; NIE, natural indirect effect. All analyses were adjusted for sex, education level, age, yearly income, and self-reported nighttime noise annoyance from sources other than that being assessed. The bootstrap percentile method was used to calculate 95% CIs. Significant total effects, NIEs, or NDEs are boldfaced. N/A indicates that the proportion of association mediated was not calculated because the NIE exceeded the total effect in absolute magnitude or because the total effect and NIE had opposite directions, rendering the proportion mediated uninterpretable.

DISCUSSION

We examined whether sleep disturbance mediated the associations between self-reported nighttime noise annoyance and mental health outcomes in adults. Greater self-reported nighttime noise annoyance from road traffic, aircraft, and residential building vibrations was associated with poorer subjective sleep quality. Furthermore, greater self-reported nighttime noise annoyance from road traffic, aircraft, railways, and residential building vibrations was associated with longer sleep latency, more sleep disturbances, greater use of sleeping medication, worse daytime dysfunction, and higher global PSQI scores. None of the noise sources was associated with sleep duration or habitual sleep efficiency. Mediation analyses revealed that the NIEs of self-reported nighttime noise annoyance from road traffic and residential building vibrations on depression, anxiety, and suicide risk were significant. These findings suggest that sleep disturbance links self-reported nighttime noise annoyance to psychiatric symptoms and suicide risk.

The aforementioned associations may be explained by a noise-induced arousal mechanism. Discrete noise events can cause autonomic and cortical microarousals (e.g., tachycardia, movement, and awakening), even at indoor maximum A-weighted sound pressure levels of approximately 30–35 dB.[23] Repeated microarousals may increase presleep vigilance, delay sleep initiation, and fragment sleep. The lack of association between self-reported nighttime noise annoyance and sleep duration or habitual sleep efficiency may reflect discrepancies between objective and subjective sleep. In the study conducted by Lauderdale et al.,[24] participants who reported sleeping for 5 and 7 h overestimated their sleep duration by 1.2 and 0.4 h (average values), respectively, compared with actigraphy-derived measurements. The overall correlation between reported and measured sleep duration was moderate (r = 0.47). These findings are consistent with evidence that self-reported sleep duration and efficiency agree less with objective measures than do other sleep indicators.[25]

In our study, sleep disturbance significantly mediated the associations between self-reported nighttime noise annoyance from road traffic and residential building vibrations and mental health outcomes. However, the NIEs of self-reported nighttime noise annoyance from railways and aircraft on mental health outcomes were nonsignificant. These findings align with those of studies in which sleep disturbance mediated the associations between residential noise exposure and general mental health[11] and depressive symptoms.[12] In contrast, a study of wind turbine sound found that noise annoyance mediated associations between sound exposure and both sleep disturbance and psychological distress.[10] However, this study did not test sleep disturbance as a mediator between noise annoyance and mental health. Differences among source-specific noise annoyance pathways may reflect variations in acoustic characteristics, exposure patterns, and reported annoyance rates. In our cohort, self-reported nighttime noise annoyance from road traffic and residential building vibrations may have involved recurrent nighttime events that provided more opportunities for sleep disruption, whereas self-reported nighttime noise annoyance from railways and aircraft might have been less frequent or more spatially concentrated. Several mechanisms may link self-reported nighttime noise annoyance, sleep disturbance, and mental health. Repeated nocturnal arousal and circadian disruption may lead to chronic activation of the hypothalamic–pituitary–adrenal axis, with downstream increases in cortisol and catecholamines that may contribute to anxiety, depression, and suicide risk.[26]

Limitations

This study has several limitations. First, self-reported nighttime noise annoyance may have been subject to source misclassification. Moreover, the direction of any resulting bias is unclear because self-reported nighttime noise annoyance may have been influenced by sleep or mental health symptoms. Second, sleep quality was assessed using the PSQI components and was thus subject to recall and reporting biases. Future studies should incorporate actigraphy to reduce measurement error. Furthermore, mental health outcomes were measured using self-report questionnaires rather than diagnoses identified from structured clinical interviews or medical records. Therefore, the reported outcomes should be interpreted as indicators of psychological symptoms or distress rather than as evidence of clinically diagnosed psychiatric morbidity. Future studies should use objective acoustic monitoring, actigraphy, standardized diagnostic interviews, or medical records to replicate and extend the findings and to clarify their relevance to clinically defined mental disorders. Third, the generalizability of our findings may be limited to Taiwanese adults living in dense, mixed-use urban settings, which are characterized by heavy motor scooter traffic. These urban and housing characteristics differ from those in lower-density settings, where neighborhoods are more often characterized by detached housing and greater distances from roadways, which may alter environmental noise patterns and sleep environments. Fourth, because we used an online survey rather than a probability-based random sampling approach, selection bias may have influenced the results, limiting the representativeness of the study population. For example, our participants were disproportionately young and had high educational attainment, which may limit the generalizability of the findings to the broader Taiwanese population. Future studies should perform probability-based or population-based sampling across geographic regions, residential densities, housing types, and traffic environments to verify the robustness and external validity of our findings. Fifth, several potential confounders were not measured or included in the analysis. For example, data on chronic medical conditions, sleep medication use, shift work, and residential environmental characteristics were unavailable. These factors may be associated with self-reported nighttime noise annoyance, poor sleep quality, and adverse mental health outcomes. Therefore, residual confounding may have influenced the observed associations and the estimated mediation effects. Future studies should identify comprehensive individual-level, occupational-level, and residential-level covariates and incorporate them into causal mediation models to minimize confounding effects in the exposure–mediator, mediator–outcome, and exposure–outcome associations. Finally, this study had a cross-sectional design, which precluded the investigation of temporal and causal associations. Future studies should adopt prospective cohort or randomized intervention designs to establish temporal pathways and provide more robust evidence for causal mediation.

CONCLUSION

Sleep disturbance may mediate the associations between self-reported nighttime noise annoyance and depression, anxiety, and suicide risk. Longitudinal studies using objective acoustic and sleep measurements are required to clarify temporality and validate the proposed mediation pathways.

Availability of Data and Materials

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Author Contributions

Dian-Jeng Li: Conceptualization, methodology, and writing—review and editing.

Cheng-Fang Yen: Conceptualization, methodology, supervision, and writing—review and editing.

Cheng-Chia Yang: Data curation, formal analysis, visualization, and writing—review and editing.

Hung-En Liao: Data curation, investigation, validation, and writing—review and editing.

Yi-Lung Chen: Conceptualization, funding acquisition, project administration, supervision, and writing—original draft.

All authors reviewed and approved the final manuscript.

Ethics Approval and Consent to Participate

The study protocol was reviewed and approved by the Central Regional Research Ethics Committee of China Medical University (IRB No. CRREC-113-004). Asia University does not maintain an institutional review board for this study type, and the authors’ institution has a formal agreement with the Central Regional Research Ethics Committee of China Medical University for ethics review services. Informed consent was obtained from all participants of this study.

Conflicts of Interest

The authors declare that they have no competing interests.

Funding Statement

This work was supported by grants from the National Science and Technology Council, Taiwan (NSTC 113-2314-B-468-005).