A socioecological model of multilevel determinants for proactive health behaviors among Chinese adults with chronic diseases Authors: Chaoyi Chen, Yanhua Hao, Weilan Xu, Ze Liu, Yuwei Zhang, Zhanchun Feng Published: 2025 ABSTRACT AbstractBackgroundChronic diseases have become a major public health challenge facing the world. Identifying key factors and developing effective management strategies to promote proactive health behaviors in patients is crucial for improving health outcomes.ObjectiveThis study aims to construct a comprehensive model of proactive health behaviors in chronic disease patients, elucidate multilevel determinants, and guide targeted policy interventions in China.MethodsA cross-sectional survey was conducted among 805 patients with chronic diseases in China. Latent profile analysis (LPA) was conducted to identify distinct profiles of proactive health behaviors among patients. Binary logistic regression analysis was used to verify and analyze the determinants affecting the proactive health behaviors of patients.ResultsAmong the 805 participants, 471 were classified as highly proactive, and 334 were classified as less proactive. The average score for proactive health behaviors was 70.37 ± 10.93. Several factors positively predicted proactive health behaviors: patients aged > 74 years (AOR = 8.85, 95% CI 2.06–39.45), married patients (AOR = 1.78, 95% CI 1.02–3.11), urban residents (AOR= 1.33, 95% CI 1.04–1.70), those with stronger health intentions (AOR = 1.42, 95% CI 1.28–1.60), higher self-efficacy (AOR = 1.12, 95% CI 1.04–1.20), positive health beliefs (AOR = 1.21, 95% CI 1.09–1.34)), and greater community support (AOR = 1.18, 95% CI 1.07–1.32). Regarding policy support, perceiving an adequate upper payment limit for drugs was associated with twice the odds of proactive health behaviors (AOR = 2.61, 95% CI 1.44–4.78). Additionally, age and the medication reimbursement policy for drug expenses exerted negative effects on proactive health behaviors (β = -0.507, P < 0.01).ConclusionGovernments should transform medical insurance from a passive payer into an active health investor. By incorporating behavioral economics principles, such a reform reallocates policy design, resources, and decision-making power toward disadvantaged populations. This shift breaks the “well-intentioned policy trap”, achieving lower medical costs alongside improved population health.Keywords: Chronic diseases, Proactive health behaviors, Patients, Policy, Health service Introduction Globally, chronic diseases remain the primary driver of mortality, with an estimated 41 million deaths each year—roughly 74% of all deaths [1]. The economic burden on health systems is equally alarming. In the United States, chronic diseases will drive nearly $2 trillion in healthcare costs by 2030, with an additional $794 billion lost annually due to reduced labor force participation [2]. Similarly, Australia’s National Health Survey (NHS) reports that over $38 billion are spent annually on chronic disease care [3]. These staggering figures highlight the profound and pervasive impact of chronic diseases on a global scale. With the increase in aging and changes in lifestyle and disease patterns, noncommunicable diseases have also shown a high prevalence in China. According to the Report on Nutrition and Chronic Disease Status of Chinese Residents (2020), deaths caused by chronic diseases in China account for 88.5% of the total deaths, of which cardiovascular diseases, cancers, and chronic respiratory diseases account for 80.7% of the deaths [4]. The data indicate a rising prevalence of chronic diseases in China, characterized by a heavy disease burden, a large affected population, high pressure on healthcare systems, frequent comorbidities, and management difficulties. To effectively reverse the current trajectory and fully realize the “Healthy China 2030” vision and the “Healthy China Action Plan”, a fundamental reorientation of individual health attitudes and behaviors is essential. However, the implementation of such a comprehensive target is hampered by the absence of suitable theoretical models of systematically integrating determinants spanning from the individual to the policy level. This study responds to that critical gap by constructing a socioecological model, which provides a cohesive, multi-level framework aligned with the strategy’s core principle of “integrating health into all policies.” Within this research framework, patients are encouraged to transition from passive recipients to proactive custodians of their own health—a shift essential to translating policy into sustained individual action. Evidence shows that for patients living with chronic illness, sustained proactive, health-promoting behaviors (such as regular exercise, a healthy diet, and stress coping) are decisive for achieving the best possible outcomes and lightening the burden of disease. By acting early, patients navigate the healthcare system more efficiently, secure fairer access to scarce resources and make better-informed choices—together propelling meaningful health gains. For example, timely screening and rigorous self-management slow or even avert disease progression. Therefore, proactive health behaviors can directly alleviate health disparities that may be exacerbated by resource scarcity by shifting limited medical spending from high-cost emergency treatment to low-cost prevention. In summary, active self-care not only improves individual well-being but also serves as a potent lever for reducing systemic health inequities and shaping a more just healthcare landscape [5–7]. However, patients often neglect health issues because of multiple factors, such as economic constraints, life pressures, and limited disease awareness. For low- and middle-income households, the immediate need to secure wages and steady income consistently outweighs any willingness to invest in longer-term personal health [8]. Furthermore, the concept of passive health management, dominated by a “health deficit” mentality, remains deeply ingrained among patients. Individuals’ misconceptions, such as “health equates to the absence of disease” and “seeking treatment only after falling sick,” hinder their proactive health awareness and self-management capabilities. Learning how to reorient the factors that influence patients’ health behaviors toward proactive health is essential. Prior research has identified several potential determinants of proactive health behaviors, including gender [9, 10], age [11], living area [12], education [13, 14], marital status [15], and self-efficacy [16]. Researchers have mapped the determinants of health behaviors onto a two-tier schema: a “downstream” tier that encompasses individual biology and psychology, and an “upstream” tier formed by the social structure and macro-level governance in which people are embedded [17]. Upstream forces include the configuration of health-care systems, the contours of labour markets and the shocks that reverberate through both—such as the unemployment surges and social instability triggered by economic downturns. These macro-stressors can reshape health trajectories for years [18]. However, mapping the complex, multilevel pathways that shape such behaviors among adults living with chronic diseases remains an unfinished research task. While most studies concentrate on individual-level behavioral change [19, 20] or upstream influences such as social support [21] and the immediate social environment [22], macrolevel health policies and environmental conditions within the broader health system context have received comparatively little attention. Although disentangling these dispersed elements of the external macroenvironment and their interdependencies is inherently challenging, a systematic and comprehensive approach—grounded from a social-ecological perspective—is crucial for analyzing and intervening in the health behaviors of people with chronic diseases. Therefore, this study aims to develop an integrated socioecological model to elucidate the mechanisms underlying proactive health behaviors among chronic disease patients. The model will examine the impact of individual-level factors (e.g., health intentions, self-efficacy, subjective norms, health beliefs), institutional-level factors (e.g., community support, family doctor follow-up, access to medical services), and policy-level factors (e.g., drug pricing caps, drug reimbursement policies) on proactive health behaviors. By identifying multilevel determinants of proactive health behaviors among chronic disease patients in China, this research seeks to contribute to reduced morbidity and mortality, alleviate the economic burden of disease, and promote health equity. The subject and method of the study Study design This investigation was conducted from August to September 2023 and employed a multistage stratified sampling approach. Based on Yichang city’s geographical location and population distribution, the first stage involved randomly selecting 3 urban community health service centers (Heyi village, Dagongqiao, and Xiling Community Health Service Center) and 3 rural township health centers (Qilixincun in Xiaoting District, Lianpeng in Dianjun District, and Xiaoxita Street in Yiling District. In the second stage, patient recruitment was conducted at Primary Healthcare Institutions using a systematic random sampling approach. Eligible patients from each village were listed and every second patient was selected to achieve the target sample sizes of at least 120 per urban center and over 150 per rural center. The inclusion criteria for participants were as follows: (1) aged ≥ 18 years; (2) were diagnosed with at least one chronic disease (e.g., hypertension, heart disease, diabetes); and (3) were receiving routine daily medication for three months or more. Additionally, the participants needed to be conscious, free of mental illness, possess basic communication skills, and be willing to participate. Patients with unstable conditions, other serious comorbidities or complications, or those unable to cooperate with the study were excluded. The questionnaires were distributed to 851 participants. After 46 questionnaires were excluded because of incompleteness or logical inconsistencies, 805 valid responses were retained, yielding a valid response rate of 94.6%. Data collection tools The self-designed questionnaire included structured items about basic demographic variables, such as age, sex, education, marital status, region, height and weight. The independent variables assessed were as follows: community health support, which comprised four aspects: (1) the availability of a health information bulletin board in the community; (2) the existence of a community WeChat public account or other health education platforms for disseminating disease prevention knowledge; (3) the implementation of community activities, such as free clinics and lectures focusing on chronic disease prevention and healthy lifestyles; and (4) the organization of community-based national fitness activities (e.g., aerobics, basketball games, hiking). Additionally, the frequency of family doctor follow-up was also evaluated. The policy support-related variables were (1) whether the upper payment limit for drugs (the maximum amount of the basic medical insurance fund) could cover an insured individual within a single policy year and (2) the degree to which the medication reimbursement policy for chronic diseases covers most of the medication expenses. Finally, the other relevant independent variables were measured by these scales as follows: Assessment of health beliefs The health behaviors of individuals are influenced by their perceptions and beliefs concerning health concerns, as well as their degree of acceptance regarding the adoption of specific actions. In this study, we employed the Health Beliefs Scale for Patients with Coronary Heart Disease in China [23], which comprises five dimensions: perceived susceptibility, perceived severity of the disease, perceived benefits, perceived behavioral barriers and cues to action. The scale consists of 5 items, and each item is rated on a five-point rating scale ranging from 1 (not at all) to 5 (totally agree). The sum score of all the items, ranging from 5 to 25, was calculated for analysis, with higher scores indicating stronger health beliefs. The Cronbach’s alpha coefficient for the scale was 0.875, and the Guttman split reliability was 0.776. Self-efficacy assessment Self-efficacy, a concept advanced by psychologist Albert Bandura [24], signifies an individual’s conviction in their capacity to effectively carry out particular goal-directed tasks or behaviors and thereby achieve the desired results. In measuring self-efficacy dimensions, the Chinese version of the Chronic Disease Self-Efficacy Scale was referenced [25], which encompasses five items: (1) I am confident in overcoming the physical exhaustion caused by the illness I have; (2) I am confident in overcoming the emotional stress caused by the illness, so as not to let it interfere with my daily life; (3) I will not be prevented from doing what I would like to do because of the physical symptoms or health problems caused by the illness; (4) I am confident that I can follow the advice of my healthcare provider, such as changing my diet or exercising; and (5) I am confident that I can do things other than taking medication to reduce the impact of my illness. Each item was measured via a five-point Likert scale. The total score ranges from 5 to 25. The scale demonstrated good reliability, with a Cronbach’s alpha coefficient of 0.934. Health intentions assessment The health behavioral intention of chronic disease patients was measured to assess their willingness to participate in health-related behaviors, motivation, and acceptance of new technologies, which included 4 items: (1) I am willing to spend time to improve my health; (2) I am willing to spend a portion of my income on chronic disease treatment; (3) I am willing to change unhealthy behaviors and lifestyles; and (4) I am willing to use new tools related to health management (health management APP/device, etc.). All the questions were measured via a 5-point Likert scale ranging from 1 = “Strongly Unwilling” to 5 = “Strongly Willing”; the Cronbach’s alpha coefficient for the Health Intention Scale was determined to be 0.679. The validity of this scale was good (KMO = 0.807, Bartlett’s Test P < 0.001, all factor loadings > 0.6, Average Variance Extracted is 0.514). Subjective norms assessment When the majority of people identify with a certain behavior adopted by an actor, it has a driving and facilitating effect on that individual. We designed entries to measure the subjective norms of patients with chronic diseases. These entry points are as follows: (1) I will feel guilty if I do not manage my health well; (2) Relatives and friends have a greater influence on me to adopt healthy behaviors (taking medication, diet, exercise, etc.); (3) Sick people around me have a greater influence on me to adopt healthy behaviors (taking medication, diet, exercise, etc.). Each topic in the scale was measured on a five-point Likert scale (1 = completely disagree, 5 = completely agree). The Cronbach’s alpha coefficient for the scale was 0.727. The validity has been verified (KMO = 0.692, Bartlett’s Test P < 0.001, all factor loadings > 0.6, Average Variance Extracted is 0.523). Proactive health behaviors scale for patients with chronic diseases The proactive health behaviors of patients with chronic diseases were measured primarily via the established Proactive Health Behaviors Scale for Chronic Disease Patients developed by Chen CY et al. [26]. This scale comprises 20 items, each rated on a 5-point scale, resulting in a total possible score ranging from 20 to 100 points. Higher scores indicate more proactive health behaviors. The scale demonstrated good internal consistency, with a Cronbach’s alpha coefficient of 0.829. Statistical analyses The sociodemographic characteristics of the participants are presented as frequencies and percentages for categorical variables. Exploratory factor analyses were used to explore the latent structure. Latent profile analysis (LPA) was conducted via Mplus 8.0 to identify distinct profiles of proactive health behaviors among patients. This approach allowed for the exploration of potential subgroups exhibiting unique patterns of proactive health behaviors, providing a nuanced perspective on this phenomenon. Model fit was assessed via goodness-of-fit indices, including the Akaike information criterion (AIC), Bayesian information criterion (BIC), adjusted BIC (aBIC), entropy, bootstrap likelihood ratio test (BLRT), and number of cases per profile. Multiple and regression imputation were applied to handle missing values. Binary logistic regression analysis was employed to examine associations between independent variables and proactive health behaviors, reporting results as odds ratios (ORs) with 95% confidence intervals (CIs). Model fit was assessed via goodness-of-fit indices (AIC, BIC, aBIC, entropy, BLRT) and profile sizes. All statistical analyses were performed using R (version 3.1.0), with statistical significance set at p < 0.05. Ethical considerations This research was approved by the Ethics Committee of Tongji Medical College of Huazhong University of Science and Technology (Approval No. 2023S065). Trained investigators invited potential participants, and informed consent was obtained from participants prior to their participation. Participants were assured that they had the right to withdraw at any point during the study without providing any reason and without suffering any adverse consequences. All collected personal identification information (such as name, contact information, etc.) was immediately replaced with a unique anonymous code during the data entry stage. All research data were stored on an encrypted hard drive, with access restricted to our research team. Results Participant characteristics A total of 805 chronic disease patients participated in the study. The descriptive statistics of the participants are presented in Table 1. The average age of the enrolled patients was 61.93 ± 13.21 years, with 59.75% aged ≥ 60 years. Most participants (86.09%) had obtained less than an undergraduate education. With respect to residence, 539 patients (66.96%) lived in villages/towns, whereas 33.04% resided in cities. In terms of perceived financial status, 25.47% of patients reported their daily financial situation as challenging. The proportion of participants who purchased medical insurance was 95.78%. Table 1Demographic variables of the study sampleVariableClassificationNo.Proportion(%)GenderMale27734.41Female52865.59Age (years)< 6032440.2460–7434943.35> 7413216.40EducationBelow undergraduate69386.09Bachelor degree and above11213.91MarriageUnmarried13016.15Married67583.85ResidenceRural53966.96City26633.04Economic statusDifficulty20525.47Average50662.86Rich9411.67Medical InsuranceNone344.22Have77195.78 Demographic variables of the study sample The average score for proactive health behaviors among chronic disease patients was 70.37 ± 10.93. After standardizing the variable directions and calculating the correlations among key variables, the results of the Pearson correlation analysis (Fig. 1) revealed that the proactive health behaviors of patients with chronic diseases were positively correlated with health belief (r = 0.38, P < 0.001), health intention (r = 0.45, P < 0.001), and self-efficacy (r = 0.27, P < 0.001). Furthermore, proactive health behaviors were positively correlated with subjective norms (r = 0.18, P < 0.001), community health support (r = 0.26, P < 0.001), family doctor follow-up (r = 0.15, P < 0.001), the upper payment limit for drugs (r = 0.13, P < 0.001), and the medication reimbursement policy for drug expenses (r = 0.06, P < 0.001). Correlation analysis of variables influencing proactive health behaviors among patients This study employed latent profile analysis (LPA) to identify varying characteristics of proactive health behaviors and explore the sociodemographic differences among patients with different proactive health behavior patterns. The analysis used scores from 20 items in the Proactive Health Behaviors Scale for Chronic Disease Patients as observed indicators. Starting with a 1-latent-class baseline model, LPA was conducted on patients’ proactive health behaviors. Latent class models with 1 to 4 categories were fitted, and their fit indices are presented in Table 2. The Class 2 model had the highest entropy value of 0.892, with the LMR and BLRT values reaching significance (P < 0.001), indicating precise classification. Consequently, according to the above indicators, the dependent variable (patients’ proactive health behaviors) was grouped into two categories. Table 2Analysis of fit indices of the latent profile model of proactive health behaviors (N = 805)ModelAICBICaBICEntropyPClass ProbabilityLMRBLRTClass 166500.2966706.6966566.96Class 263866.2864180.5663967.800.892< 0.001< 0.0010.42/0.58Class 363214.8563637.0363351.230.8770.1834< 0.0010.12/0.43/0.45Class 462705.8963235.9662877.120.8850.7488< 0.0010.10/0.42/0.35/0.13 Analysis of fit indices of the latent profile model of proactive health behaviors (N = 805) Figure 2 shows the distributions of the proactive health behaviors characteristics of the two categories of patients. The two groups generally have similar trends in terms of the distribution of changes in each indicator. Nevertheless, the scores of all indicators in latent category group 1 (breakpoint) are lower than those in latent category group 2 (solid line). Thus, category group 1 is named the “low proactive health behaviors group”, comprising 334 patients (42%); category group 2 is named the “high proactive health behaviors group”, including 471 patients (58%). Conditional probability distributions of the proactive health behaviors levels of the two groups The fully integrated social-ecological model 5, which included predictors from all five dimensions (individual, interpersonal, community, societal, and policy), provided the best account of the data. It was confirmed that all independent variables had variance inflation factor (VIF) values below 5, indicating no substantial multicollinearity. The model 5 demonstrated excellent goodness-of-fit, which achieved a 41.3% reduction in AIC (1059.2 to 621.52) compared to the intercept-only model, highlighting the synergistic interplay of multi-level determinants. The parameters of the final models were estimated via the full dataset and are presented in Table 3. Furthermore, age and the medication reimbursement policy for drug expenses had a negative interaction (β=−0.507, P < 0.01) on proactive health behaviors among patients. Table 3Logistic regression results for proactive health behaviorsVariableModel1Model2Model3Model4Model5Gender (Male)0.1120.16670.0480.1850.203Age (< 60 years)60–74 years0.749***0.564**0.574**0.4691.492***> 74 years0.751**0.3330.3410.1872.180**Education (Below undergraduate)Bachelor degree and above−0.301−0.321−0.525*−0.438−0.373Marital (Unmarried)Married0.4070.479*0.593*0.551*0.577*Residence (Rural)Urban0.502***0.416***0.429***0.293*0.282*Medical Insurance (Have)None1.181**1.047*0.7680.6930.855BMI (< 18.5)18.5 ≤ BMI < 250.065−0.050−0.294−0.534−0.46125 ≤ BMI < 30−0.081−0.233−0.343−0.525−0.425≥ 300.1930.2470.1470.0750.101Health Intention (Low)High0.389***0.341***0.349***0.354***Self-efficacy (Low)High0.114***0.104***0.116***0.110**Health Belief (Low)High0.198***0.206***0.092**0.189***Subjective Norm (Low)High0.0770.0670.092*0.092Community Support (Low)High0.147*0.164**0.110Family doctor follow-up (Seldom)Often0.165***0.0980.117Access to health services (≤ 15 min)16–30 min0.4640.545> 30 min0.2130.255Upper payment limit for drugs (Not Enough)Average−0.0160.066Enough0.822**0.958**Medication reimbursement for drug expenses (Not Enough)Average−0.659*0.278Enough−0.676*1.138Age* Medication reimbursement for drug expenses−0.507**AIC1059.21042.3722.81628.6621.52 Logistic regression results for proactive health behaviors Logistic regression analysis was conducted to identify the key factors influencing proactive health behaviors across patients with chronic diseases. The adjusted ORs (AORs) and 95% CIs from binary logistic regression analysis are displayed in Fig. 3. The results revealed that patients aged > 74 Forest plot of factors influencing proactive health behaviors among patients with chronic diseases years (AOR = 8.85, 95% CI 2.06–39.45) exhibited better proactive health behaviors. Married individuals (AOR = 1.78, 95% CI 1.02–3.11) and urban residents (AOR= 1.33, 95% CI 1.04–1.70) show better proactive health behaviors. Specifically, a high level of health intention (AOR = 1.42, 95% CI 1.28–1.60) was associated with a higher odds ratio of proactive health behaviors. Self-efficacy (AOR = 1.12, 95% CI 1.04–1.20), health belief (AOR = 1.21, 95% CI 1.09–1.34) and good community support (AOR = 1.18, 95% CI 1.07–1.32) were contributing factors for patients’ proactive health behaviors. At the policy support level, patients who perceive adequate upper payment limits for drugs (AOR = 2.61, 95% CI 1.44–4.78) demonstrate increased engagement in proactive health behaviors. To further illustrate and verify the age-medical insurance reimbursement policy for drug expenses interaction observed in the forest plot, Fig. 4 visually depicts this relationship. The three lines, corresponding to patients’ perceptions across different medication reimbursement levels, converge, indicating a significant negative interaction between patient age and medication reimbursement for drug expenses (β = −0.507, P < 0.01). Interaction diagram of patient age and medication reimbursement for drug expenses Discussion The proactive health behaviors score(70.37 ± 10.93 points) observed among chronic disease patients in this study was slightly lower than the health-promoting lifestyle score reported for patients in the old town of Guangzhou (72.02 points) [27]. This difference may refiect regional variations. Our sample was recruited mainly in Yichang, Hubei, where average education, income and health literacy are lower than in Guangzhou; these socioeconomic characteristics can depress lifestyle scores. Even so, the mean score in the present study exceeded the national benchmark of 64.81 points among 18–65 years old residents reported by Cheng Jingmin et al. [28]. This difference likely reflects the distinct composition of the study populations. By focusing exclusively on adults living with chronic illness, our sample was composed of individuals who—because of the daily, long-term demands of their conditions—have developed a heightened sense of personal health responsibility. This constant, disease-driven vigilance fosters stronger self-monitoring habits, more consistent medication and dietary adherence, and greater receptiveness to preventive counseling, all of which translate into appreciably higher proactive-health behavior scores than those observed in community-wide surveys. Our study used latent profile analysis to classify patients’ proactive health behaviors into two categories: the low-proactivity group (category 1: 42%) and the high-proactivity group (category 2: 58%). Significant differences were observed between these two groups in five aspects: individual health responsibility, interpersonal support, coping with psychological stress, and the other two dimensions. However, both groups had relatively low scores for physical exercise. This may be because the surveyed population primarily consisted of middle-aged and elderly individuals. Their physical function decreases, their balance ability decreases, and their cardiopulmonary endurance and other motor skills decrease because illness constrains the exercise capacity of chronic disease patients [29]. Our study revealed that older patients tend to exhibit more proactive health behaviors. In contrast, younger people, often brimming with vitality and possessing remarkable physiological regenerative capacities, may frequently overlook health-related concerns. Conversely, older patients with chronic diseases inevitably experience a decline in physical function. This decline makes the connection between health status and quality of life much more pronounced. For elderly individuals with low health reserves, the need for health services is significantly greater [10]. Life’s experience prompt a profound shift in perspective, leading them to value health above all else and inspiring a proactive commitment of time and effort to its preservation. Marriage (AOR = 1.78, 95% CI 1.02–3.11) also promoted proactive health behaviors among patients. This is in line with Witte Jr’s and Roger Antabe’s studies. They reported that spouses are more likely to encourage their partners to adopt healthy behaviors such as quitting smoking, moderating alcohol consumption, and seeking medical assistance, thereby creating an external mechanism of constraint [30, 31]. Furthermore, the intimacy inherent in marriage can mitigate stress and reduce unhealthy behaviors that stem from anxiety. Health beliefs encompass an individual’s perceptions of health threats and their conviction in the efficacy of actions to mitigate these threats. Social psychologists developed the health belief model in the 1950s to explain preventive health behavior [32]. Alani Abdallah et al. reported that individuals’ beliefs regarding their susceptibility, perceived benefits, and self-efficacy are strongly linked to their adoption of preventive measures against common treatable diseases, such as cervical cancer [33]. Lihong Wan et al. demonstrated that patients with stronger health beliefs perceive health threats as more significant, which in turn increases their likelihood of adopting proactive health behaviors [34]. Bandura posits that self-efficacy is a pivotal predictor of behavior change, asserting that “self-efficacy determines whether or not people act on outcome expectations” [35], positioning them as proximal determinants that influence human motivation and action through specific intervening processes [36]. Building on this foundation, Zulkosky clarifies the distinction between self-efficacy and locus of control: the former concerns a person’s belief in their capability to perform behaviors, while the latter refers to the belief that outcomes can be controlled through those behaviors. While locus of control pertains to attributions of outcomes to internal or external forces, self-efficacy is the belief in one’s capability to perform a task. A sense of accomplishment from acting on self-efficacy reinforces behavioral change and is a manifestation of an internal locus of control [37]. Our study moves beyond the traditional investigation of self-efficacy in natural states by conceptualizing it through the lens of post-traumatic growth. Specifically, we define self-efficacy among Chinese patients with chronic diseases as their capacity to manage disease-related stress and foster resilience. Accordingly, this definition serves as the basis for designing the scale items in the self-efficacy Assessment section. Our study mainly revealed that health-related self-efficacy (AOR = 1.12, 95% CI 1.04–1.20) positively predicts proactive health behaviors among chronic disease patients. This aligns with findings by Jasmine et al. [38], whose research demonstrated that self-efficacy significantly enhances exercise engagement among elderly chronic disease patients, continuously improves health-related knowledge, and strengthens the capacity to manage health risks. Hwang et al.’s investigation into exercise patterns among older adults revealed that exercise self-efficacy is positively correlated with perceived exercise benefits. Older adults with higher self-efficacy tend to select appropriately challenging tasks commensurate with their abilities, persistently overcoming barriers to health actions [39]. Behavioral intention is a robust predictor of health behavior. The current research explored the factors predicting action in those with already strong intentions. The theoretical foundation is provided by the Theory of Reasoned Action (TRA), which posits that the likelihood of engaging in a health behavior is determined by one’s behavioral intention [40]. Behavioral intentions do not drive health behaviors in a single dimension, but rather exhibit a complex dual effect. The direction of their influence depends on the nature of the intention itself—whether it is a positive motivation toward health goals or a negative tendency to avoid healthy actions, which directly leads to the adoption or avoidance of healthy behaviors [41]. These findings underscore the mechanism proposed by theoretical model, elucidating how behavioral intention translates into health behavior. However, our research focuses on the positive effects of patients’ health behavior intentions. It draws a critical distinction between these spontaneous, health-oriented actions and the reactive behaviors typically studied under the broad umbrella of disease management. It extends previous theoretical research by examining how health behavior intention shapes proactive health behavior—a key form of positive health behavior characterized by self-initiated and preventive actions. Health intention (AOR = 1.42, 95% CI 1.28–1.60) was significantly associated with proactive health behaviors. When an individual is motivated for health behaviors and perceives the necessity of beneficial actions, actual behavioral changes are often prompted by internal or external contextual factors [42]. Our findings align with existing evidence: the influence of health intentions on proactive health behaviors is moderated by an individual's change motivation and their perceived cost-benefit ratio of action [43]. Those with strong health intentions are usually more likely to resist external temptations or surmount internal obstacles, as they recognize these behaviors as crucial for achieving their health goals. Patients with greater health intentions are more inclined to accept support from others and hold positive attitudes toward proactive health behaviors. Our research indicates that a health-supportive community environment can provide individuals with access to health resources and empowerment opportunities within communities, families, and workplaces. This environment helps establish underlying health norms that encourage residents to adopt positive lifestyles. It enables individuals to shift from a “reactive health - participate when required” mindset to a “proactive health-participate voluntarily” approach. Consequently, individuals become empowered to identify and equitably harness community health resources, actively choosing the services that best meet their needs. This supportive environment not only sustains their motivation to engage with local health offerings but also bolsters their confidence and self-driven ability to resolve illness [44, 45]. Our findings underscore that upper payment limits (UPLs) on medications (AOR = 2.61, 95% CI 1.44–4.78) exert a decisive influence on proactive health behaviors. When these ceilings are set generously, they act as a financial buffer, shielding households from escalating drug costs—especially during economic downturns [46]. Illustrative reforms in Sanming city in China have raised annual payment caps for patients with overlapping chronic conditions such as hypertension and diabetes while granting automatic access to catastrophic disease insurance once the quota is reached, thereby averting treatment interruptions. By markedly lowering out-of-pocket expenditures, UPLs relieve fiscal pressure, freeing patients to channel their attention toward preventive care and long-term health management rather than worrying about affordability. Individuals who perceive their coverage as adequate report heightened financial security and greater confidence in navigating costly medical episodes. In turn, this perceived economic stability fosters a sense of personal agency, enabling proactive health behaviors choices that safeguard health and well-being. The main implication of our research is that there is a significant negative interaction (β=−0.507, P < 0.001) between age and medication reimbursement for drug expenses. Specifically, we observe that as age increases, a higher reimbursement ratio correlates with lower proactive health behaviors. There are several potential explanations for this phenomenon. First, according to adverse selection, for chronic disease patients with a relatively high reimbursement ratio, economic pressure is no longer the primary driver of health management behaviors. Instead, they might overlook prevention because of an overreliance on medical services [47, 48]. Additionally, elderly patients with chronic diseases experience more hospitalizations; despite a high reimbursement rate, their absolute medical burden may still escalate due to the accumulation of out-of-pocket costs from frequent hospital visits [49]. Given the complexity of their conditions, elderly patients face heightened vulnerability to cycles of recurrent hospitalization following illness. This pattern consequently diminishes the efficacy of reimbursement policies in fostering proactive health behaviors [50]. Policy implications The social ecological model has been widely used in the medical and health fields. Using this model, we can explore the influencing factors of proactive health behaviors in chronic disease patients, thus providing new ideas for the prevention and control of chronic diseases. The key implications include the following: (1) At the individual level, patients should establish health responsibility, strengthen health beliefs, and enhance proactive health behaviors. (2) At the institutional/societal level, governments should establish collaborative platforms involving medical institutions, enterprises, communities, and resident representatives. Integrating diverse social forces through such platforms can establish standardized frameworks for proactive health management. Furthermore, partnerships with local Centers for Disease Control (CDC) to implement “Health Science Outreach” initiatives targeting villages, communities, schools, and households are essential. These measures foster a community culture centered on proactive health while simultaneously addressing residents’ fundamental health needs and personalized service demands. (3) At the policy level, the impact of the policy environment, as a distal factor within the social-ecological model, on chronic disease comorbidity among the elderly merits attention. Medical insurance policy design could establish a reimbursement model that actively incentivizes healthier behaviors. For example, individuals adopting healthier routines—starting with first-contact care at community clinics and adhering to evidence-based management plans—could qualify for progressively higher reimbursement ratios. Additionally, governments or insurers could implement top-up payments or copay subsidies to encourage the sustained use of grassroots healthcare services and promote proactive health behaviors. Limitations The present study has several limitations. Owing to the cross-sectional design and limited sample size, we could infer only associations between variables rather than establish causal relationships. Second, the analyzed data were self-reported, which may introduce the possibility of recall bias. In addition, the sample only from Yichang City, Hubei Province, may not be fully representative of the broader populations, potentially introducing selection bias. Furthermore, despite controlling for numerous covariates based on previous literature, unmeasured confounding remains a possibility. Variables such as genetic predispositions, unrecorded environmental exposures, or personality traits could still influence the outcomes and confound the associations. Last, quantitative methods alone fail to penetrate the deep mechanism and underlying causal chains of behavior or to illuminate the social–cultural milieu and subjective meanings that shape individual motivation. Therefore, further longitudinal studies are needed to establish causal pathways and trace the evolution of proactive health behaviors over time. This research should integrate qualitative methodologies in the future, thereby uncovering the underlying social mechanisms embedded within data and phenomena. Conclusion In conclusion, proactive health behaviors in chronic disease patients are influenced by health beliefs, health intentions, self-efficacy, community health support, and medical insurance policy factors. Critically, our findings demonstrate that a significant inverse interaction between age and medication reimbursement for drug expenses affects proactive health behaviors. Therefore, government departments should advance the transformation of the medical insurance system from a passive reimbursement model to a dynamic health investment mechanism. Infusing behavioral-economic insights into the reform dismantled the “well-intentioned policy trap” by systematically tilting policy design, resource allocation and decision-making power toward disadvantaged groups—thereby achieving the twin goals of containing healthcare spending and increasing the population-wide quality of life. Acknowledgements The authors are grateful to those who participated in the study and are grateful for the help of the investigated hospitals’ main liaisons and coordinators in Yichang city, Hubei Province. Abbreviations Latent profile analysis Akaike information criterion Bayesian information criterion Adjusted odds ratio Confidence interval Authors’ contributions Chaoyi Chen contributed to the study design, data analysis, manuscript writing and supervised the study. Yuwei Zhang and Weilan Xu were responsible for data collection and curation. Yanhua Hao contributed to project administration and supervision. Zhanchun Feng contributed to ethical statements and applications. Ze Liu contributed to the investigation, data analysis, and revised the manuscript. All the authors contributed to the development of the manuscript and approved the final manuscript. Funding This study was funded mainly by the National Natural Science Foundation of China (Grant Number: 72274046) and the China Postdoctoral Science Foundation (Grant Number: 2025M770709). Data availability The entirety of the data from this study are not publicly accessible due to ethical constraints. If there were any inquiries about data access, the corresponding author of this article was contacted. Declarations Ethics approval and consent to participate This study was in accordance with the Declaration of Helsinki and received ethical approval from Tongji Medical College of Huazhong University of Science and Technology (Approval Number: 2023S065). Informed consent was obtained from all study participants. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Yanhua Hao, Email: hyhyjw@126.com. Ze Liu, Email: liuzejhwsx@163.com. References Associated Data Data Availability Statement The entirety of the data from this study are not publicly accessible due to ethical constraints. If there were any inquiries about data access, the corresponding author of this article was contacted.