Factors Associated with Adherence to the Atrial Fibrillation Better Care (ABC) Pathway Among Rural Patients with Atrial Fibrillation: A Path Analysis Based on the Integrated Theory of Health Behavior Change (ITHBC)

Introduction

Atrial fibrillation (AF) is the most common arrhythmia.¹ In China, there are approximately 7.9 million patients with AF.² AF can lead to a range of complications, the most serious of which is stroke.¹ Patients with AF have a 4–5 times higher risk of stroke than those without AF, and AF-related ischemic strokes are almost twice as fatal as those without AF.^1,3 The main symptoms of AF are palpitations, chest tightness, fatigue, and dyspnea, which impair the quality of life of more than 60% of patients.¹ Additionally, AF is often accompanied by cardiovascular diseases such as heart failure (HF), coronary artery disease (CAD), hypertension, diabetes, hyperlipidemia, and peripheral arterial disease (PAD).¹ In China, 73–90% of patients with AF have two or more cardiovascular diseases, and approximately 72% of deaths in patients with AF are attributed to cardiovascular disease.^4–7

Therefore, Lip⁸ proposed the atrial fibrillation better care (ABC) pathway in 2017 to guide the integrated management of AF. Moreover, the ABC pathway was also strongly recommended by the 2020 European Society of Cardiology (ESC) and 2021 Asia Pacific Heart Rhythm Society (APHRS) guidelines.^1,9 The ABC pathway refers to “A, anticoagulation or stroke avoidance”, “B, better control of symptoms through heart rate or rhythm control”, and “C, optimal management of cardiovascular comorbidities and risk factors”.^1,8,9 Adherence to the ABC pathway significantly reduces the risk of ischemic stroke, cardiovascular disease, and all-cause mortality (45%, 63%, and 58%, respectively).¹⁰ However, Since AF patients are primarily elderly and often have multiple comorbidities, this results in generally very low adherence to the ABC pathway. Improving adherence to the ABC pathway in AF patients is a challenging task worldwide. A study of 285,000 AF patients across developed countries showed that the adherence rate to the ABC pathway was 21%.¹⁰ In rural areas of China, where social economy and medical resources are relatively backward, only 5.75% of AF patients adhered to the ABC pathway.⁵ Improving adherence to the ABC pathway among rural patients with AF is an urgent need.

The integrated theory of health behavior change (ITHBC) has been widely used to guide health behavior change in patients with chronic diseases.¹¹ Based on multiple theories and a large body of empirical research, the ITHBC incorporates three constructs: knowledge and beliefs, self-regulation abilities, and social support.¹¹ (Figure 1). The ITHBC proposed that patients’ knowledge, beliefs, and social support can influence self-regulation ability and consequently change their health behavior.¹¹

Figure 1 The integrated theory of Health Behaviour Change (ITHBC).

Several studies have validated the relationship between some of the concepts in ITHBC and their effects on health behavior change in patients with AF. A study of patients with AF in high-altitude areas showed that patients’ knowledge and attitude could influence practice.¹² Ding et al ’s¹³ study of patients with AF within 3 months after radiofrequency catheter ablation (RFCA) showed that patients” knowledge of AF, self-management beliefs, and social support were associated with their self-management behavior of AF. A study by Li et al¹⁴ showed that knowledge of oral anticoagulants (OACs) in patients with AF was independently associated with attitude and practice. To date, no study has explored the variables of health behavior change that affect adherence to the ABC pathway among rural patients with AF.

Based on the ITHBC, this study aimed to explore the relationship between AF knowledge, AF health beliefs, AF self-management ability, and social support and adherence to the ABC pathway in rural patients with AF in China, to provide supportive evidence and identify intervention targets for improving adherence to the ABC pathway in rural patients with AF.

Methods

Design and Patients

This was a cross-sectional study involving a total of 940 rural patients with AF. The inclusion criteria were patients with an age ≥18 years, diagnosed with AF based on a 24-hour Holter, standard 12-lead ECG recording, or ≥30 second single-lead ECG tracing. All patients had to be able to communicate, understand the content and purpose of the study, and agree to participate in the study. The exclusion criteria were transient AF due to reversible causes (eg, temporary AF induced by surgery, hyperthyroidism, or pulmonary embolism) and terminal stage of malignant cancer or serious mental disorder. The patient inclusion flowchart was presented in Figure 2.

Figure 2 Flowchart of AF patients’ selection.

Sample Size

Because the study explored factors affecting adherence to the ABC pathway based on ITHBC, which is a model with two independent variables and two mediating variables, the sample size requirements of a complex mediating model were referred to when determining the sample size.¹⁵ There are no reports on the effect size of this complex mediation model in patients with AF; therefore, we used the sample size of the partial mediation effect and the average effect size model to determine the sample size. The minimum required sample sizes for models with two independent variables and two mediating variables are 480 and 360, respectively.¹⁵ Considering a 20% non-response rate, a minimum of 600 patients with AF were required for this study.

Data Collection

Convenience sampling was used in this study. Since all ECG machines in Shangyou County, Ganzhou City, Jiangxi Province, China, were connected to a local general hospital, the medical record system of the general hospital had a list of all patients with a previous diagnosis of AF. After obtaining authorization from the local health management department, we retrieved the list of all patients with AF diagnosed from the general hospital from 2015 to 2022. Then, we selected those registered as permanent residents in local rural areas according to the official household registration system. After excluding patients who had died, were in the terminal stage of malignant tumors, or had disorder of consciousness, we sent a recruitment text message or placed a call to all surviving patients with AF. Patients were invited to attend a local general hospital to complete a data collection form. They were verified by specialist nurses in the chronic disease management center. Missing data were supplemented by interviewing patients on the spot or by reviewing the medical record system. For patients who could not be contacted or had mobility barriers, data were collected by village doctors at home.

Variables and Measurements

Demographic and Clinical Variables

Demographic characteristics included sex, age, education level, marital status, occupation, type of medical insurance, and monthly household income per capita. Clinical characteristics included the type of AF, the duration of AF, history of previous hospitalization for AF, history of previous hospitalization in a department of cardiology or neurology, the CHA₂DS₂-VASc stroke risk score, the HAS-BLED bleeding risk score, RFCA, left atrial appendage closure (LAAC), pacemaker implantation, the presence of chronic comorbidities, and the medications being used. Multimorbidity was defined as patients having ≥2 chronic conditions in addition to AF.¹⁶ Polypharmacy was defined as patients being administered ≥5 medications.¹⁷ The medication information was collected based on the patient’s self-reported current medications being taken.

AF Knowledge

AF knowledge was measured using the AF knowledge scale developed by Xu et al^18. The questionnaire has five dimensions and 25 items: knowledge of AF, anticoagulation, healthy lifestyle, RFCA, and self-monitoring. A score of 1 was assigned to a correct answer, or “known,” and 0 to an incorrect answer, or “unknown.” The total scores range from 0 to 25. The content validity index of the scale was 0.946. The Cronbach’s α value was 0.835 in this study.

AF Health Beliefs

According to the health belief model, health beliefs include six dimensions: perceived susceptibility, perceived severity, perceived benefits, perceived barriers, cues to action, and self-efficacy.¹⁹ AF health beliefs were measured using the Atrial Fibrillation Health Belief Scale (AFHBS) developed by Xu et al¹⁸ and the six-item Chinese version of the Chronic Disease Self-Efficacy Scale (CDSES) developed by Lorig²⁰ and translated by Chow et al.²¹ Because AFHBS and CDSES measured different dimensions of AF health beliefs, the scores of these two scales were summed to obtain the total score of AF health beliefs. The total scores range from 23 to 145.

The AFHBS has five dimensions and 17 items: perceived severity, perceived susceptibility, perceived benefits, perceived barriers, and motivation to maintain health. A 5-point Likert scale was used to score from strongly disagree to strongly agree. The total scores range from 17 to 85. The content validity index of the scale was 0.921.The Cronbach’s α value was 0.840 in this study.

The 6-item CDSES was used to measure self-efficacy. It reflects patients’ confidence in overcoming fatigue, pain management, emotion management, symptom management, activity management, and medication management. Each item was scored from 1 to 10 points. A score of 1 indicated no confidence, and a score of 10 indicated high confidence. The Cronbach’s α value was 0.791 in this study.

AF Self-Management Ability

AF self-management ability was measured using the Self-Management Ability Scale for patients with AF (SMAS-AF), developed by Wen.²² SMAS-AF has five dimensions and 22 items: compliance behavior management, bad habits management, emotional and social management, anticoagulant drug management, and disease prevention and monitoring management. A 5-point Likert scale was used, with each item ranging from “not done” (1 point) to “always done” (5 points). The total scores range from 22 to 110. The content validity index of the scale was 0.926. The Cronbach’s α value was 0.891 in this study.

Social Support

Social support was measured using the Social Support Rating Scale (SSRS) developed by Xiao.²³ The SSRS has three dimensions and 10 items: objective support, subjective support, and utilization of social support. The total score of the scale was the sum of the scores of each item, with higher scores indicating better social support. The total scores range from 12 to 66. The Cronbach’s α value was 0.784 in this study.

Adherence to the ABC Pathway

Adherence to the ABC pathway in patients with AF was assessed based on the 2020 ESC guideline criteria.¹ A - adherence was defined as patients at a high risk of stroke (CHA₂DS₂-VASc scores ≥ 2 in males or ≥3 in females) being administered an OAC or LAAC or not at high risk of stroke. Otherwise, patients were classified as A-non-adherence. B - adherence was defined as having European Heart Rhythm Association (EHRA) scores of I (asymptomatic) or II (mild symptoms not affecting daily activities). Otherwise, patients were classified as B-non-adherence. C - adherence was defined as cardiovascular risk or other comorbidities associated with AF that have been treated with appropriate medications or have achieved control standards as follows: (1) Effective control of hypertension defined as baseline blood pressure ≤140/90 mmHg; (2) The treatment for HF included the use of Angiotensin Receptor-Neprilysin Inhibitors (ARNI)/angiotensin converting enzyme inhibitors (ACEI)/angiotensin receptor blockers (ARB) and β-blockers; (3) Treatment for CAD involved the use of ACEI/ARB, β-blockers, and lipid-lowering drugs (LLD); (4) PAD treated with LLD; (5) Stroke/TIA treated with LLD; (6) Diabetes treated with insulin or oral hypoglycaemic drugs; (7) Hyperlipemia treated with LLDs; and (8) A body mass index (BMI)<24.0 kg/m². Individuals were classified into the C-adherent group when all these diseases were properly medicated or well-controlled, or when individuals were not accompanied by any of these diseases. Otherwise, patients were classified as C-non-adherence. Adherence to the ABC pathway was defined as meeting all three components of the guideline criteria.

Data Analysis

Data analysis was performed using SPSS version 25.0 (SPSS Inc., Chicago, IL, USA) and Mplus version 8.0.²⁴ Categorical variables are reported as frequencies (percentages), while continuous variables are reported as means (standard deviations [SD]). Chi-square tests, t-test and analyses of variance were used to compare sociodemographic characteristics, clinical characteristics, and variables in the ITHBC between ABC-adherent and ABC-non-adherent patients with AF. A correlation analysis was used to analyze the correlation between variables. After controlling for sociodemographic and clinical characteristics that were significant in the univariate analysis, path analysis using bootstrapping with 5000 sample draws was used to explore the mediating hypothesis of the model. P-values < 0.05 (two-sided) were considered statistically significant.

Results

Sample Description

Among the 940 rural patients with AF, 50.7% were male, 79.3% were ≥65 years of age, 56.3% had a primary or lower level of education, 51.9% had a family per capita income of <1000 Renminbi Yuan (RMB), 77.2% were married, 44.5% were farmers, and 30.1% were unemployed. In terms of clinical characteristics, 78.2% had paroxysmal AF, 89.7% had a CHA₂DS₂-VASc score ≥2 in males or ≥3 in females, 93.1% had a HAS-BLED score ≤2, and 47.8% had a disease duration of ≤1 year. Regarding multimorbidity and polypharmacy, 89.1% patients had ≥2 chronic diseases, and 37.8% were being administered ≥5 medications (Table 1).

Table 1 Sociodemographic and Clinical Characteristics in Rural Patients with AF (N=940)

Adherence to the ABC Pathway in Rural Patients with AF

In this study, only 5.7% of rural patients with AF adhered to the ABC pathway. Of the three components, component C (management of cardiovascular comorbidities and risk factors) had the lowest adherence rate (19.8%), followed by component A (anticoagulation or stroke avoidance, 32.2%). Approximately half (54.1%) of the patients adhered to component B (control of symptoms) (Table 2).

Table 2 Adherence to ABC Pathway in Rural Patients with AF. (N=940)

Descriptions and Correlation Between Study Variables

The level of AF knowledge, AF health beliefs, AF self-management ability and social support of rural patients with AF were extremely low. The total scores were 4.15 ± 3.13, 73.79 ± 14.56, 53.05 ± 10.94 and 27.7 ± 4.80 respectively, which were less than or close to half of the total scores on the scale. There were moderately or highly significant positive correlations between these variables, with correlation coefficients ranging from 0.611 to 0.843. (Table 3).

Table 3 Descriptions and Correlation Between Study Variables. (N=940)

Univariate Analysis of Adherence to the ABC Pathway in Rural Patients with AF

Univariate analysis showed that patients with older age, lower education level, no employment or farming as their occupation, low income, persistent AF, and multimorbidity were more likely to be in the ABC-non-adherent group. Patients in the ABC-non-adherent group had lower levels of AF knowledge, AF health beliefs, AF self-management ability, and social support than those in the ABC-adherent group. (Table 4).

Table 4 Univariate Analysis of Adherence to ABC Pathway in Rural Patients with AF. (N=940)

Test of the Hypothesized Path Models

ITHBC includes three constructs: knowledge and beliefs, self-regulation abilities, and social support. Therefore, path model tests were performed step by step for each construct and the overall construct of ITHBC, respectively (Figure 3). The fit measures for path Models 1 and 2 indicated a good fit (Model 1: χ²/dƒ = 58.603/14=4.186 < 5, RMSEA = 0.058 < 0.08, CFI = 0.984 > 0.9, TLI = 0.970 > 0.9, SRMR = 0.027 < 0.05. Model 2: χ²/dƒ = 34.540/7 = 4.934 < 5, RMSEA = 0.065 < 0.08, CFI = 0.979 > 0.9, TLI = 0.948 > 0.9, SRMR = 0.017 < 0.05). The fit measure for path Model 3 indicated an acceptable overall fit (Model 3: χ²/dƒ = 210.896/15=14.060 > 5, RMSEA = 0.118 > 0.08, CFI = 0.937 > 0.9, TLI = 0.878 > 0.8, SRMR = 0.034 < 0.05). This may mainly be caused by the complexity of the model paths. The R² values for Models 1, 2 and 3 were 40.7%, 48.9% and 48.0% respectively.

Figure 3 Test of the hypothesized model. *:The symbol indicates that the path coefficient has statistical significance (P < 0.05).

Stepwise path model tests showed stable and consistent results. (Figure 3) Regarding the overall path structure of ITHBC (Model 3), AF self-management ability (β = 0.449, P < 0.001) and social support (β = 0.429, P < 0.001) directly and positively affected adherence to the ABC pathway. Social support also indirectly positively affected adherence to the ABC pathway through AF self-management ability (β = 0.085, P < 0.001, relative mediating effect = 16.5%). AF health beliefs had no direct effect on adherence to the ABC pathway, but they indirectly positively affected adherence to the ABC pathway through AF self-management ability. Possibly due to a masking effect, AF knowledge had a direct negative effect on adherence to the ABC pathway. However, it indirectly positively affected adherence to the ABC pathway through AF self-management ability (β = 0.211, P < 0.001, relative mediating effect = 41.5%). AF knowledge also indirectly positively affected adherence to the ABC pathway through a series of mediations via AF health beliefs and AF self-management ability (β = 0.121, P < 0.001, relative mediating effect = 23.8%). (Table 5).

Table 5 Mediation Effects Between Study Variables. (N=940)

Discussion

The study shows that the ITHBC-based pathway model can effectively predict adherence to the ABC pathway in rural patients with AF. Four findings were explored and validated in the study: (1) Adherence to the ABC pathway in rural patients with AF was extremely low (5.7%). (2) AF self-management ability and social support can directly and positively affect adherence to the ABC pathway. (3) AF knowledge, AF health beliefs, and social support can indirectly positively affect adherence to the ABC pathway through AF self-management ability. (4) Although AF health beliefs have no direct effect on adherence to the ABC pathway, the serial mediating effects of AF health beliefs and AF self-management ability also played an important role in the relationship between AF knowledge and adherence to the ABC pathway.

It seems that education and economic status in rural areas are associated with few patients (5.7%) being treated according the ABC pathway in this study, which was much lower than that in developed countries (21%) and urban areas in China (22.8%).^10,25,26 Furthermore, considering that 39.5% (688/1762) of the patients died and 11.3% (120/1060) of the patients were lost to follow-up in the medical registration system, the rate of adherence to the ABC pathway for rural patients with AF in the real world might be even lower. Similarly, rural patients also had relatively lower levels of AF knowledge, AF health beliefs, AF self-management ability, and social support. AF knowledge (4.15 ± 3.13 vs 11.79 ± 5.58), AF health beliefs (AFHBS: 59.63 ± 15.67 vs 71.23 ± 6.64; CDSES: 14.16 ± 5.67 vs 32.13 ± 1.50), AF self-management ability (53.05 ± 10.94 vs 76.73 ± 17.54) and social support [27.7 ± 4.80 (total score 66) vs 63.84 ± 10.01, (total score 85)] scores of rural patients with AF were much lower than those of urban patients in China.^18,27–29 Furthermore, compared with urban areas, medical resources and quality in rural areas are relatively insufficient.³⁰ It lead to a decrease in the utilization of health care resources in rural patients with AF.³¹ Therefore, clarify the prediction of adherence to the ABC pathway by AF knowledge, AF health beliefs, social support, and AF self-management ability and elucidating the mechanisms involved based on the ITHBC, and maximizing the accessibility and utilization of medical resources, is of crucial importance for the construction of a targeted and effective integrated management program for rural patients with AF.

In this study, AF self-management ability and social support had a direct positive effect on adherence to the ABC pathway. Although the path analysis showed a negative direct effect of AF knowledge on adherence to the ABC pathway, this may be due to a masking effect. The indirect effect of AF knowledge on adherence to the ABC pathway through AF health beliefs and AF self-management ability was positive. This result is consistent with theories on health behavior change, such as cognitive behavioral theory, the theory of reasoned action, and social cognitive theory. The effect of improving the level of AF knowledge, AF self-management ability, and social support on improving adherence to “A” (ie, anticoagulation) has been widely verified.^32,33

AF knowledge, AF health beliefs, and social support also had indirect positive effects on adherence to the ABC pathway through AF self-management ability in the study. The same result was shown in a study predicting lymphedema self-management behavior in patients with breast cancer that was based on the ITHBC.³⁴ AF self-management ability is an important mediating variable for improving adherence to the ABC pathway. Model 3 showed that AF self-management ability had the largest positive direct effect on adherence to the ABC pathway. The indirect positive effects of AF knowledge and social support on adherence to the ABC pathway through AF self-management ability accounted for 41.5% and 16.5% of the total effect, respectively. The serial mediating effect of AF health beliefs and AF self-management ability between AF knowledge and adherence to the ABC pathway accounted for 23.8% of the total effect. A study showed that the activation of self-management ability was positively associated with health outcomes in patients with AF.³⁵ It suggests that future intervention strategies should focus more on strengthening AF self-management ability and consider AF knowledge, AF health beliefs, and social support as important components of self-management interventions.

Notably, there was no direct effect of AF health beliefs on adherence to the ABC pathway in the current study. This is inconsistent with the findings in other studies, which showed that health beliefs directly influence health behaviors.¹⁴ In the current study, the effect of AF health beliefs on adherence to the ABC pathway was fully mediated by AF self-management ability. This may be mainly due to the poor AF self-management abilities of rural patients with AF. Consequently, patients do not have the corresponding skills and abilities for self-management, even if they are confident enough to actively change their health behaviors. However, this does not mean that AF health beliefs are not important for adherence to the ABC pathway in rural patients with AF. The serial mediation of AF health beliefs and AF self-management ability between AF knowledge and adherence to the ABC pathway was the second largest mediating effect. AF health belief is also a key component of intervention strategies for rural patients with AF, and it should be closely integrated with the components of AF knowledge and AF self-management ability.

Limitations

This study has some limitations. Firstly, this is a cross-sectional study. Although path analysis can explain complex relationships between variables, its application to inferring causal relationships is limited. In the future, prospective longitudinal studies or empirical studies are needed to further verify the effects and mechanism of the variable model based on ITHBC on adherence to the ABC pathway in rural patients with AF. Secondly, the risk factors assessed in the C component of this study were not comprehensive enough, and future research should further consider lifestyle-related factors. Thirdly, the integrated management of patients with AF requires the cooperation of patients and medical multidisciplinary teams. The study only explored patient-related factors. Further exploration of the influencing factors and perceptions of other stakeholders on patients’ adherence to the ABC pathway is required. Fourth, although Models 1 and 2 have good fits, some of the fit indicators of Model 3 were only close to the fit criteria. Therefore, Model 3 was mainly used to test the path relationships and conclusion reliability based on the complete ITHBC model in the study, rather than as an accurate effect estimation model. Finally, the study was conducted among rural patients with AF from only one county in China. Therefore, the conclusions of the study need to be extrapolated with caution.

Conclusions

Adherence to the ABC pathway by rural patients with AF in China is extremely low. In the current study, the path analysis based on ITHBC elucidated the effects and mechanisms of AF knowledge, AF health beliefs, AF self-management ability, and social support on adherence to the ABC pathway among rural patients with AF. AF self-management ability has the greatest direct positive effect and an important mediating effect on adherence to the ABC pathway. AF knowledge, AF health beliefs, and social support have indirect positive effects on adherence to the ABC pathway through AF self-management ability. We suggest that effective interventions to strengthen AF self-management ability should be explored among rural patients with AF in the future, and that AF knowledge, AF health beliefs, and social support should be closely combined with AF self-management ability as important intervention components.