Formation Mechanisms, Interrelationships, and Effects of Cognitive Factors on Diet and Physical Activity During the Post-Bariatric Surgery Period: A Cross-Sectional Study Based on Compensatory Carry-Over Action Model

Introduction

Obesity, defined as excessive fat accumulation in the body, is regarded as an epidemic in the 21st century. According to the latest epidemiological report published by the World Health Organization, more than 39% of adults globally are overweight (body mass index [BMI] ≥25 kg/m²), and 13% of them are obese (BMI ≥30 kg/m2).¹ Obesity is a significant public health issue, contributing to various metabolic diseases and imposing substantial social and economic burdens.² Recently, bariatric surgery has been considered the most effective treatment approach for obesity.³ The excess weight loss rate (EWL%) at 10 years after surgery can reach 60%~70%, the remission rate of type 2 diabetes at 1 year after surgery is 80%,⁴ and the average incident rate of hepatic steatosis has reduced from 37.4% to 16.0%.⁵ After bariatric surgery, patients are advised to maintain good lifestyle habits to prevent weight gain. Guidelines recommend a diet high in protein, low in carbohydrates and saturated fats, with daily supplementation of vitamins, iron, and trace elements.⁶ Physical activity (PA), defined as any bodily movement produced by the skeletal muscles that requires energy expenditure, is recommended for postoperative patients to achieve 150 minutes or more of moderate-intensity aerobic activity per week. However, during the post-bariatric period, only 11% of patients carry out adequate PA 1 year after surgery,⁷ 26% of patients fail to follow dietary recommendations, 73% of patients often eat high-fat foods,⁸ and 29% of patients reduce their PA by 50% compared to pre-surgery levels. These lifestyle habits will lead to unsatisfactory weight outcomes after surgery in the long term.⁹

Diet and PA directly influence energy intake and expenditure, and are subject to cognitive factors, environmental factors, interpersonal relationships, and socio-cultural contexts.¹⁰ Among these, cognitive factors are central in many behavior change theories, impacting behavior selection, execution, and maintenance, and are easier to intervene and adjust.¹¹ Bergh et al¹² suggested that cognitive factors such as self-efficacy, intention, and behavioral planning in patients undergoing bariatric surgery are related to the patients’ postoperative dietary compliance. Sarwer et al¹³ reported that self-efficacy, perceived benefits and barriers are the main cognitions that affect post-bariatric patients’ ability to perform PA. Dietary and PA management during the post-bariatric surgery period is often carried out simultaneously. Moreover, interrelationships exist between the two behaviors (diet and PA), which is mediated by cognitive factors. Geers et al¹⁴ demonstrated that PA and consumption of fruits and vegetables by obese individuals were significantly mediated by cognitive beliefs and subjective norms. As for compensatory belief (CB), obese people may use CB to relieve the guilt caused by executing an unhealthy behavior through the execution of another compensatory healthy behavior, thereby leading to poor overall health.¹⁵

Accumulating evidence in published literature suggests that post-bariatric patients may need dual behavioral management for diet and PA, and these behaviors are jointly affected by various cognitions. To explore the formation mechanism of such behaviors from perspective of cognitive factors, we chose to use the compensatory carry-over action model (CCAM) in the present study to provide a theoretical basis for this study.

Theoretical Framework and Research Hypotheses

CCAM was proposed by Professor Lippke¹⁶ in 2014 to provide better guidance for prevention and management of diabetes and obesity (Figure 1). Based on the health action process approach (HAPA) theory, CCAM combines specific cognitive factors to link multiple single behaviors together. CCAM’s primary constructs encompass intention, self-efficacy, CB, and transfer cognition (TC). CCAM assumes that CBs and TCs serve as principal antecedents for the association among diverse health behaviors, thus categorizing them as “antecedent cognitive factors”. CB is defined as “a belief that the negative effects of participating in unhealthy behaviors can be mitigated by the positive effects of another healthy behavior”.¹⁷ TC denotes factors embodying the carry-over mechanism of individual volition, wherein one healthy behavior facilitates the adoption of another wellness behavior.¹⁸ Experience, skills, knowledge, and self-efficacy typically constitute elements of this mechanism, intertwining diverse behaviors at the belief level. Each behavior’s development is directly driven by intention and self-efficacy. CB modulates the level of behavior intention, while TC modulates the level of self-efficacy. The adoption of healthy behaviors can enhance various subjective health outcomes.

Figure 1 Framework of compensatory carry-over action model. The solid lines refer to the relation hypothesis in the original compensatory carry-over action model (CCAM); The dotted lines refer to supplementary hypothesis in this study. Abbreviations: PA: physical activity; QOL: quality of life. In this study, we set behavior A as diet, and behavior B as PA.

CCAM has been demonstrated by various behavioral science studies to effectively elucidate the intricate relationship between diverse health behaviors and their determinants. A retrospective cohort study revealed that changes in self-efficacy for PA predicted changes in self-efficacy for fruit/vegetable intake, and vice versa, this phenomenon was defined as “coaction” originating from CCAM.¹⁹ Tan et al explored the relationship among restful sleep, PA, and intake of vegetables by applying CCAM, and identified interactive relationships among these behaviors through path analysis.²⁰ In CCAM, CB and TC influence behavior development through intention and self-efficacy as mediating factors. However, direct relationships may also exist between both CB, TC and behaviors. Geers et al demonstrated that obese individuals shared subjective norms in fruit/vegetable intake and PA, which directly reflects the presence of the TC mechanism.¹⁴ In studies on alcohol consumption and smoking, CBs in these domains significantly detrimentally influenced healthy behaviors.^21,22 Moreover, CCAM primarily emphasizes subjective outcomes like perceived stress and happiness as health outcomes, given that cognitive factors are central to the model. Therefore, greater emphasis should be placed on individuals’ subjective experiences.¹⁶ In fact, several studies have widely demonstrated the effects of diet and PA on physiological indicators such as BMI and EWL%.^23,24 Therefore, this study placed greater emphasis on subjective outcomes. We also focused on patients’ quality of life (QOL) as QOL is a crucial assessment parameter in Bariatric Analysis and Reporting Outcome System (BAROS).²⁵ In summary, based on CCAM and comprehensive literature review, the present study proposed several hypotheses. Table 1 lists the original hypothetical paths.

Table 1 Original Hypothetical Paths Based on CCAM

Considering the aforementioned information, the present study applied CCAM to investigate the effects of cognitive factors, namely “intention, self-efficacy, CB, and TC”, on patients’ diet and PA during the post-bariatric surgery period, and to determine the effects of behaviors on subjective health outcomes. We developed a structural equation model (SEM) to elucidate the formation mechanisms and interrelationships between diet and PA.

Methods

Study Participants

This cross-sectional study was conducted in the Bariatric Surgery Center of the First Affiliated Hospital of Nanjing Medical University. According to the formula for calculating the sample size of cross-sectional studies , we set α = 0.05, β = 0.8, and considering that diet behavior (the Dietary Adherence Scale after Bariatric Surgery scale) to be the core variable in this study, in bariatric population the σ = 10.08, μ = 54.90,²⁶ and the difference of μ₀ in this study was expected to be 2.00, also considering a 10% dropout rate, the calculated sample size was at least 219 cases. A convenient sampling method was used to select the study subjects. Patient recruitment was conducted through the attending physician of our department during the postoperative follow-up. We opted to gather patient data at 3-month post-surgery, aligning with guidelines recommending the resumption of normal diet and initiation of structured resistance training during this period. This time point is critical for physiological recovery and the cultivation of healthy behavior patterns.

Overall, 281 patients who attended a follow-up visit in our surgery center at the third month after bariatric surgery (have returned to a normal diet) between August 2021 and August 2022 were selected, and relevant patient data were collected. The inclusion criteria were as follows: (1) patient age between 18 and 60 years old; (2) bariatric surgery being performed for the first time; (3) ability to understand the communicating language; and (4) volunteering to participate in this study. The exclusion criteria were as follows: (1) presence of severe postoperative complications (gastrointestinal bleeding, anastomotic leakage, intestinal obstruction); (2) presence of serious primary diseases (diseases of the cardio-cerebrovascular system, lungs, kidneys, etc.); (3) limited mobility due to disability; and (4) inability to eat properly due to oral problems.

Measures

Subjects completed the questionnaire under the guidance of the researcher, and data of general patient characteristics, status of diet and PA, status of behavioral cognitive factors, and status of subjective health outcomes were collected.

General Characteristics

Information of sex, age, working status, education level, operation method, preoperative BMI, current BMI, comorbidities, and complications was collected.

Status of Diet and PA

The Dietary Adherence Scale after Bariatric Surgery,²⁶ designed by Zhu, contains 16 items, assessing 4 dimensions of diet control, nutrient intake, fluid intake, and eating habits. It uses a 5-point Likert scale, wherein the total score is 16–80 points. The higher the score, the better the dietary compliance. The Cronbach’s α value was 0.825.

International Physical Activity Questionnaire-Short Form,²⁷ developed by the International Physical Activity Measurement Working Group, consists of 7 items that evaluate the frequency and duration of high-intensity PA, moderate-intensity PA, walking, and sedentary behaviors of individuals. It uses metabolic equivalent (MET) assignment according to the intensity of the activity to calculate the amount of PA of individuals within 1 week. To facilitate data analysis using SEM, this study used the amount of PA (MET -min/w) to reflect the status of PA.

Status of Behavioral Cognitive Factors

Attitude-Social influence-Efficacy Questionnaire after Bariatric Surgery, developed by Zhu,²⁶ consists of 28 items, assessing 4 dimensions of intention, attitude, social influence, and self-efficacy. It uses the 5-point Likert scale. This study used the intention (4 items, total score 4–20 points) and self-efficacy (7 items, total score 7–35 points) subscale. The higher the score, the higher the dietary intention and the higher the self-efficacy. The Cronbach’s α value was 0.930. Considering potential dietary habit differences between Eastern and Western populations, we selected bariatric surgery patients’ dietary behavior and cognitive factors assessment tools developed by Chinese researchers.

Physical Activity Intention Scale, designed by Lippke,²⁸ consists of 5 items. It uses the 4-point Likert scale, and the total score ranges from 5 to 20 points. The higher the score, the stronger the willingness of the individual to engage in PA. The Cronbach’s αvalue was 0.785.

Self-Efficacy for Exercise scale, developed by Lee,²⁹ consists of 9 items, and each item is scored from 0 (no confidence) to 10 (very confident), with the total score ranging from 0 to 90. The higher the score, the stronger the confidence gained by the individual in adhering to PA despite facing difficulties. The Cronbach’s α value was 0.878.

Compensatory health Beliefs Scale, developed by Knauper,¹⁷ is a 17-item scale that assesses the extent to which individuals believe that the negative effects of unhealthy behaviors can be compensated by engaging in healthy behaviors. In the present study, we used a simplified version of this scale, designed by Fleig,¹⁸ consisting of 4 items that are specific to diet and PA. It uses a 5-point Likert scale, with the total score ranging from 4 to 20 points. The higher the scores, the stronger the CBs the individual had. The Cronbach’s α value was 0.880.

Transfer Cognition Scale, developed by Fleig,¹⁸ consists of 6 items that evaluate an individual’s tendency to apply knowledge or abilities acquired in one behavior in another behavior. It uses the 5-point Likert scale, with the total score ranging from 6 to 30 points. The higher the scores, the stronger the TC. The Cronbach’s α value was 0.650.

Status of Subjective Health Outcomes

Perceived Stress Scale, developed by Cohen,³⁰ consists of 14 items that assess 2 dimensions of loss of control and tension. It uses the 5-point Likert scale, with the total score ranging from 14 to 70 points. The higher the scores, the greater the stress. The Cronbach’s α value was 0.750.

General Well-Being schedule, developed by Campbell,³¹ consists of 18 items that assess 6 dimensions of well-being, self-control, vitality, depression, anxiety, and general health. It uses a 6-point response scale for the first 14 questions and a 10-point rating scale for the remaining 4 questions. The higher the score, the stronger the sense of well-being. The Cronbach’s α value was 0.850.

The 12-item Short Form Health Survey, developed by the Boston Institute of Health Education (USA),³² consists of 12 items that evaluate the effects on health, which is a measure of health-related QOL of individuals. It assesses 8 domains of physical functioning, role-physical, bodily pain, general health, vitality, social functioning, role-emotional, and mental health. After calculating in different ways, the physical component summary (PCS) and mental component summary (MCS) can be measured. The higher the scores, the better the health-related QOL. The Cronbach’s α values for PCS and MCS were 0.856 and 0.864, respectively.

Statistical Analyses

We used IBM SPSS Statistics 26.0 and AMOS 24.0 (Armonk, NY, USA) to perform statistical analyses of the collected data. We first performed linear regression analysis to analyze whether or not the influencing relationships in the original hypothetical paths were valid and then established an SEM based on the preliminarily verified model for path analysis. We finally tested the mediation relationships between cognitive factors and behaviors. Root mean square error of approximation, goodness-of-fit indices, comparative fit index, and χ²/df were calculated to evaluate the fitness of the model. A 2-sided p value of <0.05 was considered to indicate a statistically significant difference.

Results

General Information of the Study Participants

In this study, a total of 281 questionnaires were distributed, of which 239 (85.05%) were valid, 42 questionnaires were excluded because the response of regularity or unanswered items exceeded 20%. Among the 239 participants, 69.5% were female (n=166), and the average age was 30.92±7.41 years. The average preoperative BMI was 38.17±7.26 kg/m², and the average BMI at the third postoperative month was 28.48±5.37 kg/m². Table 2 presents data of the general information of the study participants. Table 3 presents data of the cognitive, behavioral, and subjective outcome status of the study participants.

Table 2 Demographic and Clinical Information of the Study Participants

Table 3 Status of Participants’ Cognitive, Behavioral, and Subjective Health Outcome

Preliminary Inspection of the Path Relationships in the Original Hypothetical Model

According to the results of linear regression analysis, the influences of TC on diet (H1d, p=0.131), CB on PA (H1e, p=0.522), TC on PA (H1h, p=0.071), PA on well-being (H3f, p=0.104), and PA on physical aspects of QOL (H3g, p=0.354) did not show any statitically significance. Therefore, we eliminated these paths in the subsequent SEM fitting. Paths in the other hypotheses were initially verified to be significant (p<0.05). Table 4 presents the standardized coefficients and the corresponding hypotheses. Figure 2 shows the adjusted hypothetical paths in the preliminarily verified model.

Table 4 Results of Linear Regression Analysis Based on CCAM

Figure 2 Adjusted hypothetical paths in the preliminary verified model. H1a indicates hypothesis 1a; the other hypotheses were designated in the same way. Abbreviations: PA, physical activity; QOL, quality of life.

Interrelationships Among Cognitive Factors, Behaviors, and Subjective Outcomes

The maximum likelihood method was used for parameter estimation. Table 5 presents the results of model fitting indicator values and reference values. The results suggested that the revised model fitted well. Table 6 presents the results of path analysis based on SEM. In hypothesis 1, among the antecedent cognitive factors, only CB had a significant effect on diet (H1a, p<0.001), and the paths of the other antecedent cognitions that influenced behaviors were eliminated based on the results of the linear regression model (H1d,e,h). Both intention and self-efficacy formed direct associations with their corresponding behaviors (H1b,c,f,g, p<0.05), which was consistent with the assumptions of other classical behavioral theories.^11,33 In hypothesis 2, CB exerted significant effects on intention, and TC exerted significant effects on self-efficacy (H2a-d, p<0.001), which were consistent with the original hypotheses in CCAM. In hypothesis 3, PA demonstrated significant influence only on perceived stress (H3e, p=0.004), and the path of PA to mental aspects of QOL was non-significant (H3h, p=0.077), and the remaining paths (H3f, H3g) were eliminated based on the results of the linear regression model. By contrast, diet demonstrated significant effects on all subjective outcomes (H3a-d, p<0.05). Figure 3 shows the standardized path coefficients in CCAM.

Table 5 Results of Model Fitting Indicator Values and Reference Values

Table 6 Results of Path Analysis Using CCAM

Figure 3 Standardized path coefficients in the compensatory carry-over action model. *P<0.05, **P<0.001. The dotted lines indicate non-significant results. Abbreviations: PA: physical activity; QOL: quality of life.

Table 7 presents the results of mediating effects of cognitive factors on behaviors (hypothesis 4). We used the bootstrap method involving 5000 repeated iterations in AMOS to analyze the mediation effects. All 95% confidence intervals of the indirect effect estimate did not include 0, which suggested that all mediating effects were significant (H4a-d, p<0.05). Considering the previous results of hypothesis analyses (H1a,d,e,h), it can be judged that intention exerted partial mediating effects on the relationship between CB and diet (H4a) and exerted complete mediating effects on the relationship between CB and PA (H4b); self-efficacy exerted complete mediating effects on the relationship between TC and diet (H4c) and between TC and PA (H4d).

Table 7 Results of Mediating Effects of Cognitive Factors on Behaviors

Discussion

This study evaluated the relationships and mediating effects between cognitive factors and diet or PA during the post–bariatric surgery period, explored the effects of diet and PA on subjective health outcomes, and attempted to identify the hypothetical paths using CCAM through an empirical study. This study represents the first complete validation of hypotheses using CCAM. Originally designed for managing obesity or diabetes, CCAM aims to understand the relationship between diet and PA for more effective prevention or management of these conditions.¹⁶ Given that bariatric surgery patients typically belong to the severely obese population, with a high prevalence of diabetes,⁴ decoding their behaviors post-surgery using CCAM holds significant clinical implications for lifestyle management. According to the results, the SEM showed an excellent fitting, confirming the effectiveness of most path relationships between cognitive factors and behaviors in CCAM. Core cognitive factors (CB, TC, intention, and self-efficacy) assessed using CCAM were associated with individual cognitions and attitudes towards diet and PA. This implies the possibility of explaining the interaction between diet and PA through cognitive factors.

Our study also found that the intentions and self-efficacy regarding diet and PA influence their respective behaviors, consistent with previous research in bariatric surgery based on behavioral change theories. Zhu et al³³ investigated factors affecting dietary compliance post-bariatric surgery using the Attitude-Social influence-Efficacy Model, and reported that dietary intention, attitude, and self-efficacy were primary predictors of dietary compliance. Ren et al³⁴ conducted a randomized controlled trial of exercise intervention in patients after bariatric surgery based on Cross-theoretical Model. They evaluated patients’ exercise self-efficacy and intention, identifying stages of behavioral changes to implement targeted interventions and improve exercise compliance.

Regarding antecedent cognitions, both CB and TC primarily exerted their impact on intention and self-efficacy, but most of their influencing paths had no significant direct relationship with behaviors, and only CB had a negative effect on diet. Previous studies have noted behavioral compensation in diet³⁵ and other behaviors including alcohol consumption, sedentary behavior, and smoking.^21,22 CB is a cognitive coping strategy used to alleviate guilt arising from unhealthy behaviors by engaging in compensatory healthy behaviors.¹⁵ Many bariatric surgery patients exhibited unhealthy dietary behaviors, such as overeating and grazing. The effects of weight loss caused by the surgery are usually extremely significant in the early postoperative period, some patients engaging in PA (such as walking or swimming) but exhibiting suboptimal dietary control may perceive that “weight loss continues due to participation in PA, even with occasional dietary lapses”,³⁶ this perception leads to an elevation in CB, further exacerbating dietary behaviors. Concurrently, poor dietary habits may attenuate CB’s direct effect on PA. This might explains why CB negatively affects only the intention of PA, while influencing both dietary behavior and intention. TC, originating from pedagogy research, has gained increasing attention in health behavior management.³⁷ A systematic review revealed that individuals engaged in optimal exercise levels are more motivated to enhance their eating behavior.³⁸ Based on this finding, Fleig investigated TC’s influence on factors within the HAPA theory, finding that TC affects self-efficacy, with high TC levels associated with frequent use of self-regulation strategies.³⁹ However, post-bariatric patients typically exhibit suboptimal diet and PA behaviors, and the behavior management requirements during follow-up may deplete patients’ willpower. This may limit TC’s positive effect on behaviors to self-efficacy rather than directly impacting behaviors.

We analyzed the mediating effects of antecedent cognitions on behaviors. Except for the CB and diet relationship (H4a), intention and self-efficacy were found to be complete mediators in the relationship between antecedent cognitions and behaviors in the other three hypotheses (H4b, H4c, and H4d). This suggested that intention and self-efficacy play crucial roles in mediating the influence of CB or TC on behaviors, and interventions targeting individual CB or TC may affect behaviors primarily in the pre-contemplation or contemplation stage. Particularly, intention partially mediated the CB and diet relationship, possibly due to patients’ poor dietary status postoperatively, leading to the activation of CB when patients refused to correct their unhealthy dietary habits.³⁵ This implies that implementing intervention measures to inhibit CB could enhance dietary behavior during the post–bariatric surgery period. Moreover, interventions targeting CB and TC can benefit individuals requiring multiple health behavior changes, such as bariatric surgery patients, due to their indirect associations with multiple behaviors. During the early postoperative period, healthcare providers can educate patients about the prevalence and adverse effects of “compensatory eating” by providing negative examples (eg, patients frequently consume high-sugar snacks after exercise as compensation, leading to excessive calorie intake exceeding energy expenditure from PA and hindering weight loss), aiming to suppress patients’ CB.¹⁵ TC can be enhanced by adjusting the order of behavioral management. Engaging in exercise can foster individual motivation and confidence in dietary control, while increasing fruit and vegetable intake or reducing sweets consumption can promote patients’ interest in PA.⁴⁰ In practical interventions, TC’s carry-over effects can be maximized based on patient preferences. For instance, patients inclined towards walking can be encouraged to consume fruits and vegetables following their activities, thus reinforcing their exercise enthusiasm.

Regarding the impact of behaviors on subjective health outcomes, diet significantly influenced perceived stress, well-being, and QOL, whereas PA only affected perceived stress. This discrepancy may stem from the extensive changes in gastrointestinal tract structure after bariatric surgery.⁴¹ Adherence to dietary guidelines not only facilitates rapid weight loss but also reduces gastrointestinal symptoms like vomiting and acid reflux, leading to an improved weight loss experience.³³ The main purpose of postoperative PA is to enhance energy expenditure and prevent muscle loss, thereby improving body composition. However, PA-induced weight loss often requires a long time to be effective, and most of patients do not regularly engage in PA. Consequently, the subjective benefits derived from energy dissipation through PA are limited.⁴² Our study focused on psychological outcomes, and dietary behavior was also a primary determinant in obesity studies with physiological status as main outcomes. In Gils et al’s study,²⁴ regardless of whether the patients engaged in PA, those who adhered to the Mediterranean diet achieved more weight loss than those who adhered to PA alone during the short-term after bariatric surgery. Similarly, Brown et al’s factorial control trial⁴³ in obese breast cancer patients found that those adhering to a controlled diet alone or combined with exercise achieved significant fat loss compared to those with uncontrolled behaviors, while exercising alone had limited effects on body weight.

In summary, valid paths in CCAM suggested that interventions targeting cognitive factors, particularly through dietary modifications, are easier and more effective. Inhibiting compensatory eating behaviors may promote overall behavioral patterns, potentially leading to improved PA levels and a better weight loss experience for patients. The link between behaviors and subjective health outcomes highlights that diet played a more significant role in overall health than PA during the postoperative period. Therefore, from the perspective of joint promotion of behaviors after bariatric surgery, prioritizing dietary intervention over PA is advisable based on CCAM analysis. During intervention, besides focusing on intention and self-efficacy, managing antecedent cognitions including CB and TC may positively influence the cooperative promotion of diet and PA. In particular, inhibiting CB levels may hold greater importance.

Previous studies have also examined the diet and PA interactive relationship at a physiological level. Hazell et al⁴⁴ found that exercise-induced muscle metabolism can lead to changes in cytokine release, insulin concentration, and lactic acid production, which may mediate alterations in anorexia signal peptide tyrosine-tyrosine and glucagon-like peptide-1. Consequently, changes in appetite hormones further suppress ghrelin levels, prolonging eating latency and reducing energy intake. A meta-analysis revealed that low-intensity exercise accelerated gastric emptying, leading to higher subsequent energy intake. However, when exercise intensity exceeded 70% VO₂ max, gastric emptying was delayed, potentially contributing to post-exercise appetite suppression.⁴⁵ Regarding the impact of diet on PA, Castro et al⁴⁶ found in their randomized controlled trial that obese patients receiving a very low-calorie ketogenic diet exhibited increased PA levels during the intervention. Diet composition also affected subsequent PA time. Bray et al⁴⁷ found that when additional energy intake came from fats, there was no increase in 24-hour energy expenditure (24EE), whereas excess dietary protein intake notably stimulated 24EE and increased energy expenditure during sleep. Many studies have elucidated the physiological mechanisms and mediating factors through which PA affects diet, but the reciprocal effect of diet on PA remains unclear. Our study investigated this behavioral interaction at the psychological level, revealing that cognitive factors, particularly CB, mediated the impact of diet on PA, which may help explain the balance between energy intake and subsequent energy expenditure.

This study has several limitations. Firstly, this was a cross-sectional study, hence we were unable to obtain the baseline status of cognitive factors and behaviors, nor observe how these variables changed at 3 months postoperatively, so the applicability of our results is primarily optimal at this time point. While psychological research has indicated that personality traits such as cognitive factors exhibit stable consistency throughout an individual’s lifespan,⁴⁸ future longitudinal studies will still be needed to observe dynamic changes and make stronger causal inferences. Secondly, our study measured overall dietary compliance and various durations of PA post-bariatric surgery, but did not differentiate dietary composition or types of PA. Thirdly, while this study’s theoretical basis mainly drew from research on bariatric surgery populations in Europe and America, we did not adequately account for the dietary habit disparities between Eastern and Western populations. This oversight may introduce potential cultural bias and impact the generalizability of the results. Finally, CB and TC were core variables in the CCAM, and this study and previous research mainly focused on their impact on behaviors. However, CB and TC measurement relied on patients’ personal attitudes towards behaviors. Therefore, future studies should further investigate how behaviors influence CB and TC to better understand behavioral patterns post-bariatric surgery and enhance behavioral interventions accordingly.

Conclusion

Drawing from the framework of CCAM, this study utilized SEM to examine the interplay between behavioral cognitive factors, diet, PA, and subjective health outcomes among post-bariatric surgery patients. Given the suboptimal behavioral patterns observed after surgery, prioritizing dietary interventions over PA interventions is recommended for maximizing behavioral synergies and optimizing subjective health outcomes. Notably, the carry-over effect of TC did not directly impact patient behaviors, underscoring the importance of targeting CB inhibition in interventions. Patients should be educated about forming correct cognitive views about dietary recommendations after bariatric surgery, discovering behavioral compensation tendency, reducing the extent of damage of CB to behavioral patterns, consolidating the weight loss effects of bariatric surgery with positive healthy behaviors, and avoiding the occurrence of re-obesity.