Chronic Heart failure (CHF) is a clinical syndrome resulting from structural and/or functional abnormalities of the heart. It is clinically characterized by recurrent worsening of symptoms and signs and represents the advanced and often terminal stage of various cardiovascular diseases (Khan et al., 2024). Globally, CHF affects more than 64 million individuals. In China, data from the national urban employee medical insurance database indicate that approximately 12.1 million individuals living with CHF, with 3.0 million new cases diagnosed annually (Wang et al., 2021). With an ageing population, improved survival from cardiovascular diseases, and the increasing prevalence of chronic conditions such as coronary artery disease, hypertension, and diabetes, the burden of CHF is expected to rise steadily. Consequently, mitigating the social and economic impact of CHF has become an urgent public health priority in China (Savarese et al., 2023).

Exercise-based cardiac rehabilitation (ExCR) has been shown to improve the prognosis of patients with CHF by reducing morbidity, mortality, and hospital readmissions (Davies et al., 2010). The efficacy of ExCR in combination with pharmacological therapies and device-based interventions is recognized in multiple clinical guidelines, which endorse ExCR with Class Ι, Level A evidence (Taylor et al., 2023). Notably, the benefits of ExCR are highly dependent on sustained adherence (O'connor et al., 2009). However, studies have shown that 40 to 91% of individuals with CHF do not engage in regular exercise, and that adherence rates reported in clinical ExCR trials remain suboptimal (Pozehl et al., 2018). A randomized controlled trial of cardiac rehabilitation among patients with CHF in China found that the implementation rates of exercise rehabilitation behavior in the control group were only 28.99 and 21.59% at 8 and 12 weeks, respectively (Huang et al., 2022). Similarly, Yang et al. demonstrated that exercise adherence among CHF patients remained low across multiple dimensions of the adherence scale (Yang et al., 2024). These findings highlight that their exercise adherence is unsatisfactory both in terms of subjective psychological engagement and objective implementation of prescribed exercise regimens. Moreover, adherence to exercise was consistently lower than that observed for other CHF-related self-care behaviors (Van Der Wal et al., 2010). Despite widespread recognition of the importance of exercise as a health behavior among CHF patients, many exhibit poor adherence, thereby increasing the risk of unfavorable clinical outcomes. It is therefore essential to elucidate the factors influencing exercise adherence in CHF patients and the mechanisms through which they exert their effects.

Drawing on the HF-ACTION study (O'connor et al., 2009), Collins and Cooper investigated the determinants affecting exercise adherence among patients with CHF. The findings indicate that clinical and demographic variables are ineffective predictors of exercise adherence. Poor social support and high barriers to exercise were associated with lower exercise time (Collins et al., 2023; Cooper et al., 2015). Alonso et al. found that attitude, self-efficacy and relapse management were the effective components of exercise adherence intervention in patients with CHF (Alonso et al., 2021). Warehime et al. suggest that perceived health status may be key to the exercise adherence. Knowledge, social support, and motivation are critical influencing factors for long-term adherence to exercise (Warehime et al., 2020). Although increasing attention has been given to factors influencing exercise adherence among patients with CHF, existing studies still present several limitations. First, many studies lack guidance from a systematic theoretical framework, resulting in a narrow focus on determinants. Second, most analyses adopt a marginal approach, assuming that independent variables exert effects in isolation, thereby overlooking the inter-dependence among factors and the potential for synergistic interactions—what could be described as “chemical reactions.” Fundamentally, this reflects a neglect of the systemic and causally complex nature of individual behavior. From a holistic perspective, exercise adherence among patients with CHF results from multiple, concurrently acting antecedent conditions. These conditions may interact in various combinations, and no single pathway is solely responsible for achieving high adherence. However, the underlying configurational mechanisms associated with exercise adherence in this population remain insufficiently explored and poorly understood.

The Capacity, Opportunity, Motivation – Behavior (COM-B) model (Michie et al., 2011), proposed by Michie et al., posits that the enactment of a behavior requires the presence of three essential conditions. Capability and opportunity can directly influence behavior or indirectly affect it through their impact on motivation. The strength of the COM-B model lies in its provision of a multidimensional analytical framework for understanding the determinants of individual behavior. In our previous research, we conceptualized exercise adherence as a behavior characterized by subjective intention (Wang et al., 2022). From the perspective of facilitating behavior change, we employed the COM-B model to comprehensively identify the influencing factors of exercise adherence in patients with CHF and constructed a corresponding measurement instrument (Gao et al., 2021; Gao et al., 2023). Specifically, the capability dimension included exercise rehabilitation literacy; the opportunity dimension comprised exercise prescription and multiple social support; and the motivation dimension encompassed perceived benefits, kinesiophobia, and exercise self-efficacy (Figure 1).

Qualitative Comparative Analysis (QCA), which was introduced by Charles Ragin in 1987, has since become a widely used method for examining causal complexity in management research (Ragin, 1987). This approach is rooted in the theoretical principles of set theory, Boolean operations, and configuration ideology. For this reason, in all QCA based methods independent and dependent variables are referred to as “conditions” and “outcomes,” respectively (Ragin, 2000, 2008, 2009). A configuration refers to a specific set of interacting conditions that, in combination, are sufficient to produce a particular outcome. QCA integrates the advantages of both qualitative and quantitative analyses, identifies conditions and outcomes from empirical data, and clarifies the underlying causal mechanisms and pathways to realization. This methodology is especially well-suited for examining the intricate interactions between conditions and outcomes they generate. The incorporation of fuzzy-set enables QCA to assess the degree of membership of each case within the set, while simultaneously maintaining the essential qualitative states of full membership and full non-membership (Ragin, 2000). In recent years, an increasing number of studies adopting fuzzy-set qualitative comparative analysis (fsQCA) have been published in the field of health. Chen et al. employed fsQCA to identify causal pathways leading to depression (Chen et al., 2024). Li et al. revealed four distinct configurations of conditions associated with physical activity behavior, demonstrating that such behavior among older adults with subjective cognitive decline results from multifactorial synergies (Li et al., 2024). Palas et al. applied fsQCA based on the extended UTAUT2 model to examine the key determinants influencing older adults’ intention and behavior regarding m-health use (Palas et al., 2022). These studies collectively illustrate the growing application of fsQCA in the health domain, particularly for uncovering complex, configurational mechanisms underlying health behaviors. Consequently, this study examined the mechanisms of exercise adherence among patients with CHF grounded in the COM-B model through the application of fsQCA. The following are the three main research questions addressed in this study: 1. Are there distinct adherence profiles of exercise adherence among patients with CHF? 2. What combination of capability, opportunity, and motivation lead to exercise adherence? 3. Does a complex and asymmetric causal relationship underlie exercise adherence in CHF patients?

This study employed a mixed-method approach integrating LPA and fsQCA to explore the mechanism of exercise adherence among CHF patients. LPA identified three distinct adherence profiles highlighting substantial interindividual variability. Building on this classification, fsQCA revealed four configurations associated with high adherence and two configurations associated with low adherence. Notably, the perception of exercise benefits consistently emerged as a core condition across all high-adherence configurations, while exercise prescription was present in both low-adherence configurations despite poor behavioral outcomes. These findings underscore the configurational and asymmetric nature of exercise adherence determinants and demonstrate that similar outcomes can arise from multiple distinct pathways.

The identification of perception of exercise benefits as a necessary condition across all high-adherence configurations underscores its fundamental role in initiating and sustaining exercise behavior among patients with CHF (Malone et al., 2012). Cardinal et al. have similarly noted that a pivotal moment in the behavioral change process occurs when individuals perceive that the benefits of exercise outweigh its barriers (Cardinal J et al., 2004). Within the framework of fsQCA, a necessary condition does not guarantee an outcome, but its absence renders the outcome nearly impossible. This is consistent with our finding that none of the patients exhibiting high adherence lacked benefit perception. While the perception of benefits alone may not be sufficient to achieve sustained adherence, it serves as a prerequisite motivational foundation, without which other facilitating conditions may not be effectively activated. This perspective aligns with most behavior change models, in which behavioral intention is considered a precursor to action. However, benefit perception is more closely aligned with the broader construct of motivation, and growing evidence points to a discrepancy between intention and actual behavior (Feil et al., 2023). Moreover, the predictive power of intention in explaining behavioral execution remains inconclusive. Translating motivation, shaped by benefit perception, into consistent behavioral engagement remains challenging—particularly in CHF populations where physical limitations, fatigue, dyspnea, and fear of movement often result in short-term rather than sustained participation in exercise rehabilitation (Nelson et al., 2023).

The four configurations associated with high exercise adherence exhibit strong theoretical alignment with the COM-B model, which posits that behavior arises from the dynamic interplay among capability, opportunity, and motivation. Although the COM-B model has been extensively applied to identify and categorize determinants of health behaviors (Yang et al., 2024; Young-Silva et al., 2024), few studies have explicitly examined how these three components interact configurationally to produce behavioral outcomes. Our findings illustrate that while comprehensive integration of capability, opportunity, and motivation clearly facilitates optimal adherence, certain partial combinations of these elements can also be sufficient to support adherence behavior.

Configuration H1 is primarily driven by reflective motivation, with opportunity and capability conditions working synergistically to support behavior. It showed the highest raw coverage among all high-adherence configurations, indicating that many patients achieve adherence through this comprehensive behavioral structure. This reflects an optimal scenario for behavior change, where full alignment of the COM-B components facilitates high adherence. Consistent with Paterson et al., who found that targeting at least two COM-B components is effective for improving physical activity (Paterson et al., 2024), this highlights the importance of comprehensive interventions. However, such ideal conditions are often difficult to achieve in practice, especially in resource-limited settings (Ferrel-Yui et al., 2024). While H1 offers a valuable benchmark for intervention design, flexible and adaptive strategies remain essential to address varying clinical realities. This is also supported by the relatively low unique coverage of Configuration H1, suggesting that its distinctiveness is limited and that it is not irreplaceable.

Both H2a and H2b represent motivation-dominant, opportunity-assisted configurations, where motivation serves as the core driver of exercise adherence, and external opportunities act as supportive rather than essential conditions. The key distinction lies in the level of opportunity provided: H2a reflects adherence behavior in a relatively well-supported context, whereas H2b represents a minimalist structure relying solely on a single opportunity element—namely, the exercise prescription. This contrast highlights a gradient in patients’ dependence on environmental resources, indicating that similar adherence outcomes can be achieved under varying degrees of structural support. A comparative analysis of the coverage shows that H2a and H2b exhibit similar raw coverage, suggesting that both configurations contribute comparably to explaining high exercise adherence within the study population. However, the significantly higher unique coverage of H2b indicates that this configuration accounts for a more distinct subset of patients whose behavior cannot be explained by other pathways. In contrast, H2a shares a large proportion of its explanatory power with other configurations, limiting its distinctiveness. This pattern suggests that while both configurations are common among adherent individuals, H2b captures a more irreplaceable behavioral mechanism, particularly relevant for patients whose adherence is sustained by motivation and minimal structural support alone.

Configuration H3 represents a capability-compensated pathway, where behavior is supported by both reflective and autonomous motivation, along with internal resources such as exercise rehabilitation literacy, despite the absence of external opportunity conditions. Although its raw coverage is relatively low, its unique coverage is notable, suggesting that this configuration accounts for a distinct subgroup of patients who rely on internal cognitive and motivational resources to sustain adherence. Notably, this finding aligns with prior research indicating that once individuals establish a habitual exercise routine, the influence of external factors tends to diminish, and behavior becomes increasingly self-regulated (Malone et al., 2012).

To examine causal asymmetry, this study conducted a sufficiency analysis of antecedent condition combinations associated with low exercise adherence among CHF patients, yielding two distinct configurations: L1 and L2. The core absent conditions in both configurations were perceived exercise benefits, exercise self-efficacy, and exercise rehabilitation literacy, while exercise prescription emerged as a core present condition. The peripheral present conditions differed: L1 included fear of exercise, whereas L2 included multiple social support. These findings illustrate a key principle of causal asymmetry, which posits that the absence of an outcome is not simply the inverse of the conditions that lead to its presence (Cangialosi, 2023). This stands in contrast to traditional linear models, where predictors are assumed to exert symmetrical and continuous effects on outcomes. In our results, high adherence configurations required the presence of motivation-related conditions as a necessary or core condition. However, in low adherence configurations, the absence of those motivational and capability factors, even when structural support like exercise prescription was present, still led to poor adherence. This asymmetry highlights the non-linear and complex nature of health behavior: interventions that promote adherence cannot be designed merely by reversing the factors linked to non-adherence. Rather, it underscores the importance of tailored, configurational approaches that account for how different combinations of factors interact under varying contexts.

In conclusion, there is no singular model for promoting exercise adherence among patients with CHF; rather, the four pathways are emblematic of “divergent routes leading to the same endpoint.” Based on the dissolution of the high-adherence configuration and the confirmation of the low-adherence configuration, this study concludes that motivation are pivotal factor in both the initiation and maintenance of exercise. Furthermore, enhancing exercise adherence among the majority of patients with CHF necessitates multidimensional systemic interventions, which should be grounded in the establishment and reinforcement of motivation. Kaushal et al. identified that the factors of perceived benefit and self-efficacy are modifiable and responsive to intervention in ethnically diverse patients with CHF (Kaushal et al., 2023). This finding indicates that healthcare professionals’ interventions targeting patients’ reflexive motivation may be more effective than those focusing on autonomous motivation, such as habits, emotions, and impulses. Reflective motivation interventions are designed to enhance the patient’s conscious effort in making accurate assessments, formulating detailed plans, and making rational decisions regarding exercise rehabilitation. Simultaneously, owing to inter-individual variability, not all patients require the presence of all three factors to attain the desired behavior. Identifying patients who necessitate particular triggers to achieve adherence to exercise is a crucial component of the clinical intervention process. This approach facilitates the implementation of tailored interventions to optimize outcomes. Furthermore, the findings of this study demonstrate the existence of an asymmetric and complex causal relationship concerning exercise adherence behavior in patients with CHF. Specifically, the antecedent conditions and outcomes do not consistently manifest or vanish simultaneously, and various antecedent conditions exhibit differential effects when combined in distinct ways. This insight offers a novel perspective for future research on the mechanisms underlying the exercise adherence behavior.

This research examined the mechanisms underlying exercise adherence in patients with CHF grounded in the COM-B model through the application of qualitative comparative analysis. By adopting a configuration perspective that highlights the interactions among influencing factors, the study identified six pathways characterized by multiple equivalences, with comprehensive explanations provided for each pathway. The study indicates that healthcare providers should account for the diverse pathways of exercise adherence and the varying roles each factor plays among different patients when designing an intervention program. This approach aims to identify the critical factors and strategically allocate intervention resources to enhance patient adherence and optimize the outcomes of exercise rehabilitation.