The role of anxiety, depression, and illness perception in quality of life of patients with rheumatoid arthritis: A structural equation model approach

Abstract

This study aimed to construct and test a structural model of self-regulation for the relationship between health-related quality of life (HRQoL) and related factors in patients with rheumatoid arthritis. In this cross-sectional correlational study, 250 patients with rheumatoid arthritis who visited the outpatient clinic of a general hospital in Korea were selected through convenience sampling. Sociodemographic characteristics, clinical characteristics, HRQoL, awareness of disease outcomes, anxiety, depression, and stress coping strategies were evaluated using a questionnaire. Structural equation modeling was used to examine the proposed model. The final model for predicting HRQoL of patients had an explanatory power of 70.3% and demonstrated a good fit with the data. Anxiety, depression, and illness perception had direct negative effects on HRQoL. Self-efficacy and self-esteem showed indirect effects on HRQoL, fully mediated by illness perception. Social support, which was fully mediated by anxiety and depression, indirectly affected HRQoL. The self-regulation model tested in this study may serve as a framework for future intervention studies on improving HRQoL among patients.

1. Introduction

Rheumatoid arthritis (RA) is an autoimmune disease characterized by symptoms such as pain, joint swelling, stiffness, progressive joint destruction, and deformity, often leading to severe disability. Risk factors for RA include infection, smoking, hormones, genetics, and abnormal immune system activation.^[1] RA affects approximately 0.24% to 1% of the global population.^[2] Recent epidemiological data further indicate that, in 2019, approximately 18.5 million individuals worldwide were living with RA, with an estimated 1.07 million new cases reported annually. The age-standardized prevalence and incidence rates were 224.25 and 12.21 per 100,000 population, respectively, and these rates were consistently higher among females.^[3]

Drug therapies, including nonsteroidal anti-inflammatory drugs, steroids, disease-modifying antirheumatic drugs, and biological agents, have significantly reduced joint bone damage, improved joint function, and enhanced quality of life (QoL).^[4] However, sociodemographic characteristics, social roles, and environmental factors continue to influence patients’ QoL positively or negatively,^[5] necessitating continuous self-regulation management alongside drug therapy. Assessing health-related quality of life (HRQoL) in patients with RA is crucial for disease management and treatment planning, offering insights into their experiences living with the condition.^[6] According to several cross-sectional studies, factors influencing HRQoL in patients with RA include self-efficacy, depression, fatigue, functional disability, social support, employment status, disease activity, comorbidities, income, gender, age,^[7–10] sleep duration, body mass index, psychological counseling, and physiological indicators.^[11]

Patients with RA experience significant life limitations due to the negative impact of physical symptoms such as joint pain, inflammation, and stiffness^[12]; emotional challenges such as anxiety and depression; and psychosocial issues, including impaired social roles, reduced work ability, increased dependency, and loss of self-esteem, but few of these aspects are within their control.^[13] Therefore, identifying modifiable aspects of the self-regulation process through theoretical frameworks enables healthcare professionals to allocate resources effectively, which is essential for patient-centered disease management and ultimately improves QoL.

Self-regulation is a problem-solving process that involves regulating the body, emotions, and resources to achieve goals and control behaviors.^[14] It emphasizes specific behavioral strategies and active roles individuals can adopt to prevent illness and promote health in patients with chronic diseases.^[15] The self-regulation model focuses on self-regulation in health and disease, including patients’ responses to chronic conditions. It integrates personal and social factors influencing health behaviors from the patients’ perspectives.^[16] In this model, cognitive representation (beliefs about illness causes, consequences, timeline, and controllability)^[17] and emotional representation (e.g., anxiety, depression)^[12,18] shape coping strategies,^[19] which in turn influence health outcomes such as HRQoL. Social support, self-efficacy, and self-esteem are considered key resources^[9,20] that can buffer negative illness perceptions and improve adaptation, especially in RA patients who often face chronic pain, functional limitations, and psychological distress.^[12,18]

The self-regulation model provides a comprehensive framework for coping with changing environments by managing illness-related thoughts, emotions, and behaviors to achieve health goals. Thus, it aids in predicting QoL in patients with RA. Shinan-Altman and Afuta-Goldstein^[15] examined the HRQoL structure in patients with RA using the self-regulation model. They verified the structure with illness representations as an exogenous variable and coping strategies, measured by resilience and social support, as an endogenous (mediator) variable. Other studies have reported only cross-sectional correlations between illness representations and HRQoL^[17,21] or coping strategies and HRQoL.^[19] Therefore, this study developed a hypothetical model to predict the QoL, focusing on the illness representations and coping strategies of the self-regulation model, including the self-system and sociocultural context that influence them, and aimed to verify this model.

2. Methods

2.1. Design

This structural model study established a hypothetical model based on self-regulation to explain factors affecting the QoL among patients with RA. It verified the model’s fit and hypotheses (Fig. 1).

Figure 1.

Hypothetical model of quality of life in rheumatoid arthritis patients.

2.2. Participants and data collection

This study collected data from April to August 2021. The participants in this study were patients with RA aged 20 or older who visited the outpatient clinic of a university hospital in South Korea and were conveniently recruited from an accessible population. The study was conducted with patients who had been diagnosed for over a year, considering the period when adaptation to RA treatment and self-regulation of the disease is expected to affect QoL.^[22] Those with a psychiatric history, disorders of consciousness, or a cancer diagnosis within the past 5 years were excluded from the study. The participants completed the questionnaire on their own, taking approximately 30 minutes, and if assistance was required, the researcher personally interviewed the participants to help complete the questionnaire.

A survey was conducted on 256 individuals according to the participant inclusion and exclusion criteria, and data from 250 participants were finally analyzed, excluding 6 individuals whose responses were almost missing. In structural equation models (SEMs), the sample size requires 10 to 20 times the number of samples per observed variable. Considering the rationale that 200 individuals are generally appropriate when using maximum likelihood estimation for parameter estimation and that the observed variables in this study are 21, the sample size met the requirement. Ethical approval for this study was obtained from the Eulji University Hospital Institutional Review Board to protect the study participants (EMC 202104018-HE001).

2.3. Measures

2.3.1. Sociodemographic and disease-related characteristics

The study investigated sociodemographic characteristics, including age, gender, spouse, education level, religion, job, and monthly income, and disease-related characteristics, including comorbidities, duration since diagnosis, and medication.

2.3.2. Self-efficacy

Self-efficacy was measured using the Korean version of the tool developed by Sherer et al.^[23] This tool was developed to measure self-efficacy in general situations, consisting of 7 items on a 4-point Likert scale ranging from 0 for “strongly disagree” to 3 for “strongly agree,” with higher scores indicating higher self-efficacy. Content validity was assessed by a panel consisting of 6 experts in nursing and psychological measurement. The item-level content validity index (I-CVI) values ranged from 0.83 to 0.97, and the scale-level content validity index, average (S-CVI/Ave) was 0.94, exceeding the recommended threshold of 0.90. Cronbach α in this study was 0.82.

2.3.3. Self-esteem

Self-esteem was measured utilizing the Korean version of the tool developed by Rosenberg.^[24] The items consist of positive and negative questions, using a 5-point Likert scale ranging from 1 for “strongly disagree” to 5 for “strongly agree,” where higher scores indicate higher self-esteem. Content validity was confirmed by a panel consisting of 6 experts in nursing and psychological measurement. The I-CVI values ranged from 0.85 to 1.00, and the S-CVI/Ave was 0.95. Cronbach α in this study was 0.88.

2.3.4. Social support

Social support was measured with the Korean version of the tool developed by Zimet et al.^[25] This tool includes 12 items, consisting of 3 sub-factors: friend support, family support, and support from healthcare professionals. Each item is scored on a 5-point Likert scale from 1 for “strongly disagree” to 5 for “strongly agree,” indicating that the higher the score, the higher the social support. The I-CVI values ranged from 0.86 to 0.98, and the S-CVI/Ave was 0.94. Cronbach α in this study was 0.89.

2.3.5. Illness perception

Illness perception was measured using the Korean version of the Illness Perception Questionnaire-Revised, developed by Moss-Morris et al.^[26] The “perception of illness consequences” section (6 items) in the original tool was used to measure illness perception. Each item is scored on a 5-point Likert scale from 1 for “strongly disagree” to 5 for “strongly agree,” with higher scores indicating a higher negative perception of the illness consequences. Content validity was assessed by a panel consisting of 6 experts in nursing and psychological measurement. The I-CVI values ranged from 0.84 to 0.96, and the S-CVI/Ave was 0.93. Cronbach α in this study was 0.87.

2.3.6. Anxiety and depression

Anxiety and depression were measured using the Korean version of the Hospital Anxiety and Depression Scale (HADS) developed by Zigmond and Snaith.^[27] This tool contains a total of 14 items, consisting of 7 items each for anxiety (HADS-A) and depression (HADS-D). Each item is scored on a 4-point Likert scale ranging from 0 to 3, with higher scores meaning higher levels of anxiety or depression. The I-CVI values ranged from 0.83 to 0.95, and the S-CVI/Ave was 0.92. Cronbach α in this study was 0.91 and 0.84, respectively.

2.3.7. Stress coping strategies

Stress coping strategies were measured with the Korean version of the tool developed by Lazarus and Folkman.^[28] This study used 6 items for problem-focused coping and 6 for emotion-focused coping out of 40 items from the original tool. Each item is scored on a 4-point Likert scale from 1 for “never use” to 4 for “frequently use,” with higher scores indicating more use of the coping strategies in that area. The I-CVI values ranged from 0.85 to 0.97, and the S-CVI/Ave was 0.94. Cronbach α in this study was 0.81 and 0.87, respectively.

2.3.8. Health-related quality of life

HRQoL was measured using the Korean version of SF-36 developed by Ware and Sherbourne.^[29] The Korean version of the SF-36 was provided by QualityMetric and was used in this study. This tool consists of 36 items in 8 domains, divided into 2 domains: physical (physical functioning, role limitations due to physical health, bodily pain, and general health perception) and mental (mental health, role limitations due to emotional health, vitality/fatigue, and social role functioning) domains. Each item has a different response range; the total score ranges from 0 to 100, with higher scores representing a higher HRQoL. This study used the scores calculated by entering the raw data values into the software provided by QualityMetric and applying weights. Cronbach α in this study was 0.95.

2.4. Statistical analysis

Data were analyzed using SPSS version 23.0 (IBM Corp., Armonk) and AMOS version 22.0 (IBM Corp.). Descriptive analysis was conducted, utilizing frequency and percentages for categorical variables, and mean and standard deviation for continuous variables. The correlations among study variables were analyzed by Pearson correlation analysis. The validity of the study tools was verified with confirmatory factor analysis (CFA), with parameters of the model measured using maximum likelihood estimation. The model’s goodness-of-fit index (GFI) was analyzed using χ²/df, standardized root mean square residual, root mean square error of approximation (RMSEA), GFI, adjusted goodness-of-fit index (AGFI), normed fit index (NFI), Tucker-Lewis index (TLI), and comparative fit index (CFI). Adequate and acceptable model fit values were: 1 < chi-square minimum discrepancy/degrees of freedom < 5, NFI > 0.90, CFI > 0.90, TLI > 0.90, AGFI > 0.90, RMSEA < 0.08, and standardized root mean square residual <0.05.^[30] The significance of the study model’s direct, indirect, and total effects was verified using the bootstrapping method (200 resamples).

3. Results

3.1. Sociodemographic and disease-related characteristics of the study participants

The mean age of the participants was 53.1 ± 10.8 years, with 218 females (87.2%). The sociodemographic and disease-related characteristics are shown in Table 1.

Table 1.

Sociodemographic and disease-related characteristics of the participants (N = 250).

Characteristics	Categories	n (%)	M ± SD (range)
Sociodemographic characteristics
Age (yr)	<40	24 (9.6)	53.1 ± 10.8 (21–82)
	40–49	62 (24.8)
	50–59	88 (35.2)
	≥60	76 (30.4)
Gender	Male	32 (12.8)
	Female	218 (87.2)
Spouse	Yes	202 (80.8)
	No	48 (19.2)
Education level	≤Middle school	42 (16.8)
	High school	104 (41.6)
	≥University	104 (41.6)
Religion	Yes	124 (49.6)
	No	126 (50.4)
Job	Yes	151 (60.4)
	No	99 (39.6)
Monthly income (10,000 won)	<100	57 (22.8)
	100–199	59 (23.6)
	200–299	61 (24.4)
	300–399	32 (12.8)
	≥400	41 (16.4)
Disease-related characteristics
Comorbidities*
Hypertension	Yes	41 (16.8)
	No	208 (83.2)
Diabetes	Yes	15 (6.0)
	No	235 (94.0)
Cardiopulmonary disease	Yes	9 (3.6)
	No	241 (96.4)
Duration since diagnosis (yr)	<5	66 (26.4)	14.32 ± 8.25 (1–40)
	5–9	50 (20.0)
	10–14	61 (24.4)
	15–19	25 (10.0)
	≥20	48 (19.2)
Medication*
Biologics	Yes	38 (15.2)
	No	212 (84.8)
Glucocorticoids	Yes	121 (48.4)
	No	129 (51.6)
DMARD	Yes	190 (76.0)
	No	60 (24.0)
Methotrexate	Yes	177 (70.8)
	No	73 (29.2)

DMARD = disease-modifying antirheumatic drug, M = mean, SD = standard deviation.

^*Multiple responses.

3.2. Descriptive statistics, normality tests, and correlations among study variables

Table 2 presents the mean and standard deviation of each study variable. Skewness (−0.81 to 0.87) and kurtosis (−0.87 to 0.74) were calculated to assess normality, and since both fell within ±2, the measured variables were assumed to follow a normal distribution. As a result of analyzing the correlation between variables, HRQoL had a significant positive correlation with the self-system (r = 0.36, P < .001), sociocultural context (r = 0.23, P < .001), and stress coping strategies (r = 0.27, P < .001), whereas a significant negative correlation was observed with cognitive representations (r = −0.46, P < .001) and emotional representations (r = −0.71, P < .001; Table 2).

Table 2.

Descriptive statistics, confirmatory factor analysis, and correlations among variables (N = 250).

Variables	Range	M ± SD	Skewness	Kurtosis	r (P)					Confirmative factor analysis
					1	2	3	4	5	β	AVE	CR
Self-system
Self-efficacy	0–21	12.98 ± 3.21	−0.59	0.64							0.77	0.81
Self-esteem	10–50	35.62 ± 6.60	−0.60	0.48						0.55
Social context	1–5	3.53 ± 0.69	−0.14	0.22	0.38*					0.85	0.50	0.74
Support from healthcare professionals	1–5	3.10 ± 0.97	0.11	−0.54						0.54
Family support	1–5	3.99 ± 0.82	−0.70	0.22						0.70
Friend support	1–5	3.55 ± 0.85	−0.32	−0.28						0.78
Cognitive representation	1–5	3.07 ± 0.86	−0.05	−0.69	−0.17*	−0.12					0.52	0.81
Illness perception 1	1–5									0.77
Illness perception 2	1–5									0.74
Illness perception 3	1–5									0.82
Illness perception 4	1–5									0.73
Emotional representation					−0.48*	−0.40*	0.41*				0.87	0.93
Anxiety	0–21	6.07 ± 4.17	0.87	0.74						0.78
Depression	0–21	7.01 ± 4.12	0.61	0.12						0.88
Coping strategies	1–4	2.63 ± 0.49	−0.41	0.41	0.56*	0.35*	−0.02	−0.33*			0.84	0.91
Problem focus coping	1–4	2.61 ± 0.62	−0.38	0.14						0.91
Emotion-focused coping	1–4	2.67 ± 0.49	−0.27	0.20						0.56
HRQoL	0–100	58.23 ± 17.07	−0.53	0.06	0.36*	0.23*	−0.46*	−0.71*	0.27*
General health	0–100	37.70 ± 18.62	0.26	0.27						0.71	0.57	0.91
Physical function	0–100	63.32 ± 23.51	−0.59	−0.27						0.64
Role physical	0–100	63.93 ± 26.05	−0.52	−0.38						0.78
Role emotional	0–100	71.17 ± 28.87	−0.81	−0.36						0.67
Social function	0–100	73.45 ± 24.04	−0.75	−0.12						0.72
Pain	0–100	54.38 ± 26.80	−0.15	−0.87						0.77
Mental health	0–100	64.24 ± 19.74	−0.69	0.43						0.73
Vitality/fatigue	0–100	45.70 ± 16.53	−0.27	0.11						0.73

β = standardized estimate, AVE = average variance extracted, CR = construct reliability, HRQoL = health-related quality of life, M = mean, SD = standard deviation.

^*P value < .001.

3.3. Structural equation model

Before constructing the SEM, the measurement model was first assessed using CFA to ensure validity and reliability. The CFA demonstrated acceptable model fit (χ²/df = 2.23, GFI = 0.90, AGFI = 0.86, CFI = 0.94, RMSEA = 0.07, NFI = 0.90, TLI = 0.91, and RMR = 0.04), meeting the recommended thresholds. All measurement properties were examined, and necessary refinements were applied to optimize model fit. Once the measurement model satisfied the acceptable goodness-of-fit criteria (Table 2), the SEM was subsequently developed and refined through multiple modifications. The final SEM demonstrated acceptable model fit, with indices as follows: χ²/df = 2.21, GFI = 0.90, AGFI = 0.86, CFI = 0.94, RMSEA = 0.070, NFI = 0.90, TLI = 0.91, and RMR = 0.04. These values indicate that the model satisfies the recommended fit criteria and supports its scientific validity (Fig. 2).

Figure 2.

Path diagram of the model for quality of life in rheumatoid arthritis patients.

In this study, 5 out of a total of 17 pathways were statistically significant (Fig. 2, Table 3). The self-system in patients with RA showed significant direct (β = −0.34, P < .001) effects on cognitive representations, and the sociocultural context showed significant direct (β = −0.77, P < .001) effects on emotional representations. No significant direct effects of coping strategies were found on the self-system (β = 0.02, P = .92), cognitive representations (β = −0.06, P = .34), and emotional representations (β = −0.15, P = .14), whereas significant direct (β = 0.81, P < .001) and total (β = 0.62, P = .001) effects were found on the sociocultural context.

Table 3.

Standardized direct, indirect, and total effects of the model (N = 250).

Endogenous variable	Exogenous variables	β	NSRW (SE)	CR (P)	Direct effect (P)	Indirect effect (P)	Total effect (P)	SMC
Cognitive representation	Self-system	−0.34	−0.41 (0.09)	−4.33 (<.001)	−0.34 (<.001)		−0.34 (<.001)	0.085
Emotional representation	Social context	−0.77	−0.82 (0.12)	−6.86 (<.001)	−0.77 (<.001)		−0.77 (<.001)	0.474
Coping strategies	Self-system	0.02	0.02 (0.21)	0.10 (.922)	0.02 (.922)	0.02 (.291)	0.04 (.606)	0.640
	Social context	0.81	0.62 (0.15)	4.15 (<.001)	0.81 (<.001)	−0.19 (.150)	0.62 (.001)
	Cognitive representation	0.06	0.05 (0.05)	0.95 (.340)	−0.06 (.344)		−0.06 (.344)
	Emotional representation	0.25	0.15 (0.10)	1.48 (.132)	−0.15 (.138)		−0.15 (.138)
HRQoL	Self-system	0.21	0.27 (0.41)	0.65 (.514)	0.21 (.514)	0.13 (.001)	0.34 (.060)	0.703
	Social context	−0.91	−0.90 (0.80)	−1.23 (.220)	−0.31 (.056)	0.59 (.003)	0.28 (.117)
	Cognitive representation	−0.37	−0.023 (0.04)	−5.75 (<.001)	−0.38 (<.001)	0.00 (.438)	−0.38 (<.001)
	Emotional representation	−0.96	−0.99 (0.22)	−4.44 (<.001)	−0.83 (<.001)	0.07 (.202)	−0.76 (<.001)
	Coping strategies	0.27	0.46 (0.46)	1.28 (.202)	0.27 (.202)		0.27 (.202)

CR = critical ratio, HRQoL = health-related quality of life, NSRW = non-standardized regression weight, SE = standard error, SMC = squared multiple correlation.

There were no significant direct effects of HRQoL on the self-system (β = 0.21, P = .51), sociocultural context (β = 0.31, P = .06), and coping strategies (β = 0.27, P = .20). However, HRQoL had a significant direct effect (β = −0.38, P < .001) on cognitive representations and also showed significant direct (β = −0.83, P < .001) and total (β = −0.76, P < .001) effects on emotional representations. The pathway from the self-system to HRQoL revealed that the direct (β = 0.21, P = .51) and total (β = 0.34, P = .06) effects were not significant, but the indirect effect (β = 0.13, P = .001) was significant. These results indicate that cognitive representations completely mediate the relationship between the self-system and QoL. In addition, in the pathway from sociocultural context to HRQoL, the direct (β = −0.31, P = .06) and total (β = 0.28, P = .12) effects were not significant. However, the indirect effect (β = 0.59, P = .003) was significant, suggesting that emotional representations serve as a complete mediator between sociocultural context and HRQoL. These variables explained 70.3% of the outcome variable, HRQoL.

4. Discussion

Based on the self-regulation model, this study constructed the relationship between related factors and verified the validity of predicting HRQoL in patients with RA. The study results showed that emotional representations (anxiety and depression) were the most influential factors in the HRQoL of patients with RA, and cognitive representations (illness perception) also had a significant effect. Moreover, emotional representations and cognitive representations fully mediated the relationship where sociocultural context and self-system affect HRQoL.

According to the results of this study, the most substantial direct factor affecting HRQoL in patients with RA was anxiety and depression. It was reported that anxiety and depression occur more than twice as often in patients with RA than in the general population, leading to worse long-term outcomes in patients with RA.^[12,18,31] Although anxiety and depression in patients with chronic diseases who require repeated treatment throughout their lives can lead to fatal problems such as suicide, they can be controlled by implementing active interventions, leading to changes in patients’ lives.^[12] Therefore, anxiety and depression should be assessed regularly from the treatment plan for the early stage of the disease, and active support measures, including psychological counseling requests, depending on the severity, should be prepared. Furthermore, healthcare professionals should consider anxiety and depression when evaluating treatment response and HRQoL in patients with RA.

The results of this study showed that social support was an antecedent factor of anxiety and depression, indirectly affecting the QoL by completely mediating anxiety and depression. That is, the higher the social support, the lower the perception of anxiety and depression, which had a positive effect on improving HRQoL. These results are similar to previous correlation studies that found that lower anxiety and depression^[8,12] and higher social support^[7–9] were associated with higher QoL. Shinan-Altman and Afuta-Goldstein^[15] have shown that patients with RA perceived social support as high when emotional representations were low. In contrast, social support did not have a significant effect on the QoL in the structural model of the QoL based on the self-regulation model, suggesting the possibility of a potential factor between these variables. It was reported that social support is negatively correlated with depression and stress,^[32,33] plays a role in increasing resilience in stressful situations,^[34] is a predictor of disease activity in long-term disease duration,^[35] and has a positive effect on overall health condition and QoL.^[3] The results of this study showed that it is difficult to improve the HRQoL of patients with RA if anxiety and depression are not resolved, even though external resources such as support from healthcare professionals, family, and friends are sufficient. However, it also means that if anxiety and depression are aggravated in patients with low social support, their QoL may further decline. When caring for patients with RA with low social support, healthcare professionals should examine anxiety and depression to understand their QoL. In addition, it should be noted that active interventions to alleviate anxiety and depression can further increase the positive impact of social support on QoL.^[9] When considering the results of the study showing that the size of the social support network and number of social roles reported by patients with RA were not factors affecting QoL, but the quality of social support was an affecting factor,^[36] efforts should be made to provide high-quality social support.

The next factor that directly affected HRQoL was illness perception. It is likely that patients, due to the uncertainty of the disease, accumulated pain and physical damage, and functional restrictions throughout the illness, develop a negative perception of illness consequences,^[12] leading to lower HRQoL. Previous studies have also reported that the perception of illness consequences in patients with RA directly and negatively impacts HRQoL^[15] and that illness perception accounted for 51% and 45% of the variance in physical and mental HRQoL, respectively,^[21] supporting the results of this study. Patients’ perception of their illness may change depending on the duration of illness and time, and a negative perception can negatively impact treatment consequences and QoL.^[37] It was reported that the physical HRQoL among patients with RA who were sensitive to the chronicity of the disease and its negative impact on daily life was lower than that of patients who were not.^[21] Thus, healthcare professionals should help patients with RA develop positive illness perceptions and optimistic thoughts about illness consequences by paying close attention to the unmet needs of patients with RA who manage their health at home and providing counseling and education that have been reported to help develop problem-solving processes and long-term health management in chronically ill patients. According to a previous study, patients with RA sometimes feel guilty because they believe they are responsible for the disease.^[38] Hence, healthcare professionals should not overlook the inner feelings of these patients to improve the HRQoL among patients with RA. Further studies will find it interesting to explore whether alleviating the perceived illness threats and enhancing positive disease beliefs can positively affect HRQoL.

According to the results of this study, lower self-efficacy and self-esteem strengthened negative illness perception, leading to a decrease in HRQoL. This study found that illness perception was a crucial mediator factor in the pathway from self-efficacy and self-esteem to HRQoL. Previous studies have reported that self-efficacy^[7,8,39] and self-esteem^[40] positively influence HRQoL. The findings of this study indicate that negative illness perception can directly lead to negative changes in HRQoL and interfere with the positive effects of psychological resources. Previous studies have shown that patients with higher negative perceptions of the illness consequences were more sensitive to pain and had lower physical function^[19,21] and that the higher their fear-avoidance beliefs, the higher their perceived physical disability, leading to a lower QoL,^[10] which supported the results of this study. These results suggest the need to consider illness perception when designing QoL interventions based on improving self-efficacy and self-esteem.

Although only social support directly affected stress coping strategies in this study,^[41] no significant relationship with HRQoL was found. These results are similar to a study that found that patient-perceived social support positively affected resilience.^[35] However, these results differed from those of a study that found low social support was associated with higher avoidance coping after 3 and 5 years^[36] and that problem-focused coping positively affected HRQoL.^[19] This is believed because the correlation between coping strategies and QoL has a relatively small influence compared with other variables, including representations. Therefore, further replication studies with different measured variables and models are required. This discrepancy may be attributable to the specific characteristics of the study sample, including disease stage and demographic profile, or to differences in measurement timing and tools compared with prior studies. These factors should be considered when designing interventions, ensuring that social support strategies are tailored to patient needs and contextual factors to maximize their potential impact on HRQoL.

The limitations of this study are as follows. First, the participants were recruited using convenience sampling in some areas, and there is a possibility that severely ill patients were omitted. Hence, caution is required in generalizing the results of this study, and future research should use multicenter random sampling to enhance representativeness. Second, since this is a cross-sectional study, there is a limitation to judging the apparent causal relationship between variables over time, and longitudinal designs are recommended to examine dynamic changes and establish causality. Third, this study could not include the evaluation and adjustment process of the self-regulation model. Thus, longitudinal studies are further required to show how illness representations, coping strategies, and illness consequences interact and change as patients adapt to an unpredictable progressive rheumatoid disease, and future studies should incorporate the assessment and iterative adjustment of model components throughout the research process. Fourth, as the study includes a higher proportion of females, caution is required when applying the results to males; ensuring a balanced gender ratio in recruitment will strengthen the applicability of the findings. Therefore, it is proposed to include objective variables such as biological markers, the occurrence of complications, and medication adherence, in addition to subjective self-reports, in further evaluations of illness consequences. Moreover, some path coefficients were small or nonsignificant, which may be due to factors such as sample characteristics, measurement error, or complex interrelationships among variables, warranting further investigation. In addition, this study did not examine potential interaction effects among variables; hence, future research should incorporate interaction terms to better capture the complex relationships influencing HRQoL in patients with RA. Finally, this study did not account for generational factors that may influence HRQoL. Accordingly, future research should incorporate variables reflecting generational differences and expand the theoretical framework to capture these dimensions, thereby enhancing the generalizability and depth of understanding of QoL in patients with RA.

This study found that the self-regulation model, which includes cognitive, emotional, and behavioral responses of patients not addressed in conventional health behavior theories,^[16] is suitable for predicting the relationship between HRQoL and related factors in patients with RA. Thus, the self-regulation model can be used as a theoretical foundation for intervention studies promoting self-management to enhance HRQoL in patients with RA.

4.1. Implications for clinical practice

This study provides key implications for rheumatology nursing care, emphasizing the importance of integrating psychological and cognitive assessments into routine disease management. The findings suggest that effective management of RA requires a comprehensive approach beyond physical symptom control.

First, nurses should be equipped to recognize patients’ illness perceptions and emotional burdens early in the care process. By incorporating brief, validated tools into outpatient settings, healthcare providers can proactively identify those at risk for negative illness perceptions and psychological distress, facilitating timely interventions.

Second, the study underscores the potential impact of nurse-led self-management programs. Specifically, interventions focusing on cognitive reframing, emotional regulation, and social resource activation can enhance patient engagement and promote positive coping strategies.

Finally, embedding these elements into standard clinical protocols may improve treatment adherence and ultimately enhance HRQoL in individuals living with RA. By acknowledging the psychological and cognitive dimensions of RA management, healthcare professionals can deliver more holistic and patient-centered care.

5. Conclusion

Anxiety, depression, and illness perception among patients with RA in this study were found to be variables that directly and negatively affected HRQoL, as well as completely mediate the relationship between self-efficacy, self-esteem, social support, and HRQoL. Therefore, nurses must focus on the relationships between these factors to identify those that may hinder QoL in self-management among patients with RA. In particular, it is crucial to assess early those patients who are at increased risk of having a negative perception of their disease outcomes or experiencing severe depression or anxiety to provide timely interventions.

Author contributions

Conceptualization: Sanghyun Bae, Ok-Hee Cho.

Data curation: Sanghyun Bae, Ok-Hee Cho.

Formal analysis: Sanghyun Bae, Ok-Hee Cho.

Investigation: Sanghyun Bae.

Methodology: Sanghyun Bae, Ok-Hee Cho.

Supervision: Ok-Hee Cho.

Validation: Sanghyun Bae, Ok-Hee Cho.

Writing – original draft: Sanghyun Bae, Ok-Hee Cho.

Writing – review & editing: Sanghyun Bae, Ok-Hee Cho.