Association Between Stroke‐Related Knowledge and Self‐Management Behaviours in Patients With Ischemic Stroke: Serial Mediation by Health Beliefs and Depression Authors: Yue Li, Lulu Li, Xuan Wang, Beilei Lin, Ziye Guo, Qilan Tang, Aixia Wang Published: 2026 ABSTRACT ABSTRACTObjectiveTo examine the association between stroke‐related knowledge and self‐management behaviours among patients with ischemic stroke and to assess the potential mediating roles of health beliefs and depression.MethodsA total of 382 patients with ischemic stroke were enrolled from a tertiary hospital in Henan Province between June 2024 and March 2025. Data were collected using a general information questionnaire, the Stroke Knowledge Questionnaire, the Health Belief Scale, the Self‐Rating Depression Scale and the Self‐Management Behaviour Scale. Spearman correlation analysis was performed to examine the associations among the variables. Serial mediation was examined using structural equation modelling with robust maximum likelihood estimation in the lavaan package (version 0.7–2).ResultsStroke‐related knowledge was positively associated with self‐management behaviours (total effect: B = 2.448, 95% CI [1.925, 3.015]). After accounting for health beliefs and depressive symptoms, the direct effect of stroke‐related knowledge on self‐management behaviours remained statistically significant (B = 1.124, 95% CI [0.552, 1.696]) and the total indirect effect through health beliefs and depressive symptoms was also statistically significant (B = 1.324, 95% CI [0.966, 1.752]), accounting for 54.08% of the total effect. Specifically, the indirect effect through health beliefs (B = 0.594, 95% CI [0.238, 0.970]), the indirect effect through depressive symptoms (B = 0.419, 95% CI [0.236, 0.683]) and the serial indirect effect involving health beliefs and depressive symptoms (B = 0.311, 95% CI [0.176, 0.510]) were all statistically significant.ConclusionsAmong patients with ischemic stroke, stroke‐related knowledge was significantly associated with self‐management behaviours, with health beliefs and depressive symptoms serving not only as separate mediators but also as a significant serial mediation pathway.Implications for Nursing PracticeNurses should provide individualized stroke education, assess patients' health beliefs and depressive symptoms and offer psychological support when needed. Integrating educational and psychological assessment into rehabilitation nursing may provide more comprehensive support for self‐management.Reporting MethodThis study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.Patient/Public ContributionPatients with ischemic stroke were surveyed to collect relevant data. Introduction Stroke has become the leading cause of death and disability in China. According to national surveillance data from 1599 hospitals, ischemic stroke accounted for 81.9% of hospitalized stroke cases in 2020 (Tu and Wang 2023). During recovery, many patients experience physical impairments and psychological problems (Rahman et al. 2023), making self‐management particularly important. Effective self‐management can promote functional recovery, prevent complications and improve quality of life (Correia et al. 2022; Klockar et al. 2023). However, self‐management behaviours among patients with ischemic stroke remain suboptimal (Li et al. 2025) and may be associated with factors such as stroke‐related knowledge, health beliefs and emotional status. Stroke‐related knowledge may facilitate self‐management by helping patients recognize health risks, understand recommended treatments and make informed decisions regarding their health and rehabilitation (Mun and Shim 2024). Previous studies have also reported positive associations between stroke‐related knowledge and health‐promoting behaviours (Chen et al. 2024). According to the Health Belief Model (HBM), health behaviours are associated with perceived susceptibility, severity, benefits, barriers and self‐efficacy (Janz and Becker 1984). Stroke‐related knowledge may shape health beliefs by influencing how patients interpret stroke‐related risks and evaluate recommended actions (Park et al. 2025). Cognitive appraisal theory proposes that individuals' evaluations of illness‐related threats, recommended actions and their coping resources may shape subsequent emotional responses (Lazarus 1991). Longitudinal evidence in stroke populations also suggests that declining self‐efficacy is associated with depressive symptoms at follow‐up (Volz et al. 2016). Depressive symptoms are common during stroke recovery, with a reported prevalence of up to 44.6% (Kwon et al. 2021). These symptoms may interfere with the motivation and concentration required to initiate and maintain self‐management behaviours and have been associated with poorer medication adherence among stroke survivors (Srithumsuk et al. 2021). Previous studies have separately examined the associations of stroke‐related knowledge, health beliefs and depressive symptoms with health‐related behaviours after stroke (Fitzsimons et al. 2022; Mahmood et al. 2023; Kim et al. 2023). Recent studies have also examined potential mediators between stroke‐related knowledge and health‐related behaviours. Health beliefs were reported as a mediator between stroke‐related knowledge and health behaviours among individuals at high risk of stroke, while self‐esteem and health information literacy were identified as mediators between stroke‐related knowledge and health‐promoting or self‐management behaviours among patients with stroke (Mun and Shim 2024; Chen et al. 2024; Zeng et al. 2025). Another study examined the serial mediating roles of psychological flexibility and anxiety in the association between stroke‐related knowledge and health behaviours (Ma et al. 2026). Despite these findings, evidence remains limited regarding the proposed serial indirect association involving overall health beliefs and depressive symptoms among patients recovering from ischemic stroke. Guided by the HBM and cognitive appraisal theory, this study aimed to examine the association between stroke‐related knowledge and self‐management behaviours and the independent and serial mediating roles of health beliefs and depressive symptoms among patients recovering from ischemic stroke. Accordingly, the following hypotheses were proposed: Stroke‐related knowledge is positively associated with self‐management behaviours. Health beliefs mediate the association between stroke‐related knowledge and self‐management behaviours, such that greater stroke‐related knowledge is associated with stronger health beliefs, and stronger health beliefs are associated with better self‐management behaviours. Depression mediates the association between stroke‐related knowledge and self‐management behaviours, such that greater stroke‐related knowledge is associated with fewer depressive symptoms, and fewer depressive symptoms are associated with better self‐management behaviours. Health beliefs and depression serially mediate the association between stroke‐related knowledge and self‐management behaviours, such that greater stroke‐related knowledge is associated with stronger health beliefs, stronger health beliefs are associated with fewer depressive symptoms, and fewer depressive symptoms are associated with better self‐management behaviours. Stroke‐related knowledge is positively associated with self‐management behaviours. Health beliefs mediate the association between stroke‐related knowledge and self‐management behaviours, such that greater stroke‐related knowledge is associated with stronger health beliefs, and stronger health beliefs are associated with better self‐management behaviours. Depression mediates the association between stroke‐related knowledge and self‐management behaviours, such that greater stroke‐related knowledge is associated with fewer depressive symptoms, and fewer depressive symptoms are associated with better self‐management behaviours. Health beliefs and depression serially mediate the association between stroke‐related knowledge and self‐management behaviours, such that greater stroke‐related knowledge is associated with stronger health beliefs, stronger health beliefs are associated with fewer depressive symptoms, and fewer depressive symptoms are associated with better self‐management behaviours. Subjects and Methods Participants and Data Collection This was a cross‐sectional study. We used convenience sampling to recruit patients with ischemic stroke from a tertiary hospital in Zhengzhou, China, between June 2024 and March 2025. Patients attending the neurology ward for reassessment or the neurology outpatient clinic for follow‐up were recruited by trained members of the research team. The inclusion criteria were as follows: (1) Patients aged ≥ 18 years; (2) Diagnosis of ischemic stroke confirmed by CT or MRI according to the 2019 Chinese guidelines (Chinese Society of Neurology and Chinese Stroke Society 2019); (3) Between 15 days and 6 months after ischemic stroke onset (Bernhardt et al. 2017); (4) Stable condition and basic reading and writing ability; (5) Awareness of their condition and voluntary agreement to participate. The exclusion criteria were as follows: (1) Patients with impaired consciousness; (2) Cognitive or language impairments that prevented communication; (3) Heart, liver and kidney failure or malignant tumours; (4) Participation in other studies. This study was approved by the Ethics Committee of REDACTED (Approval No. REDACTED). Written informed consent was obtained from all participants prior to data collection. A total of 33 variables were collected, including 13 participant characteristics and 20 scale‐derived variables. The structural equation model comprised 20 observed variables and three latent variables: stroke‐related knowledge, health beliefs and self‐management behaviours; depressive symptoms were entered as an observed variable based on the total SDS score. Using the conservative criterion of 10 participants per observed variable, the minimum sample size was 200. After allowing for a 20% invalid questionnaire rate, the required sample size was 250 [20 × 10/(1–0.20)], exceeding the commonly recommended minimum of 200 for structural equation modelling (Wolf et al. 2013). Therefore, during the study period, 430 potentially eligible patients were screened. Of these, 16 did not meet the eligibility criteria and 14 declined to participate. A total of 400 patients were enrolled, including 182 inpatients undergoing reassessment and 218 patients attending outpatient follow‐up. After 18 invalid questionnaires were excluded, 382 participants were included in the final analysis. Instruments General Information Questionnaire A self‐designed questionnaire was used to collect demographic and clinical data, including age, sex, residence, employment status, marital status, education level, type of medical insurance, income, smoking, drinking, number of chronic diseases, time since stroke onset and activity of daily living scale. Stroke Knowledge Questionnaire This questionnaire was developed by Yao (2016) and was used to assess patients' stroke‐related knowledge. It contains 40 items across six domains: warning signs, emergency responses, risk factors, safe medication, health behaviours and rehabilitation knowledge. Each item was scored as 1 for a correct answer and 0 for an incorrect or unknown answer, with total scores ranging from 0 to 40. Higher scores indicated better knowledge. The content validity index of the original questionnaire was 0.95 and the Cronbach's α coefficient of 0.858 (Bernhardt et al. 2017). In the present study, the KR‐20 coefficient was 0.956 for the overall questionnaire and ranged from 0.812 to 0.940 across the six domains. Stroke Health Belief Scale This scale was revised by Peng et al. (2012) and was used to measure health beliefs related to stroke prevention. It contains 35 items across six dimensions. Each item is rated on a five‐point Likert scale. Items in the perceived‐barriers dimension were reverse‐scored, with higher scores indicating fewer perceived barriers. Higher total scores indicate stronger health beliefs regarding the adoption of relevant health behaviours. The Cronbach's α coefficient of the scale in this study was 0.949. Self‐Rating Depression Scale (SDS) The SDS was developed by Zung (1965) in 1965 and contains 20 items, including 10 positively scored items and 10 negatively scored items. Positively worded items are reverse‐scored, and the item scores are summed to obtain a raw score ranging from 20 to 80. The raw score was multiplied by 1.25 and rounded to the nearest whole number to obtain a standard index score ranging from 25 to 100. According to the Chinese normative study, a standard index score of 53 is used as the cut‐off for depressive symptoms: scores of < 53, 53–62, 63–72 and ≥ 73 indicate no, mild, moderate and severe depressive symptoms, respectively (Wang et al. 1986). Higher scores indicate more severe depressive symptoms. The Cronbach's α coefficient in this study was 0.905. Stroke Self‐Management Behaviour Scale This scale was adapted from the self‐management behaviour assessment scale developed by Wang (2012). It includes 50 items covering seven dimensions. Forty‐nine items are rated from 1 to 5. The second item in the diet‐management dimension consists of four subitems and yields a combined score ranging from 1 to 10. Therefore, the total score ranges from 50 to 255, with higher scores indicating better self‐management behaviours. The original total‐scale Cronbach's α was 0.835 and the Cronbach's α in this study was 0.959. Data Collection Method Data were collected through face‐to‐face structured questionnaire interviews during inpatient reassessment or outpatient follow‐up visits. A trained researcher administered all questionnaires using a standardized procedure, reading each item and response option verbatim without interpretation or prompting. Participants provided all responses themselves. To ensure confidentiality, questionnaires were identified by study numbers without names or other direct identifiers. Completed questionnaires were checked on site for missing or unclear responses. Of the 400 completed questionnaires, 18 invalid questionnaires were excluded, leaving 382 questionnaires for the final analysis, with a valid questionnaire rate of 95.5%. The final analytic dataset contained no missing values, and no data imputation was performed. Statistical Analysis Statistical analyses were performed using SPSS version 27.0 and R version 4.4.3. Normally distributed continuous variables were expressed as mean ± standard deviation, non‐normally distributed variables as median (P25, P75) and categorical variables as frequencies and percentages. Associations among variables were examined using Spearman's rank correlation. Serial mediation was examined using structural equation modelling with the lavaan package (version 0.7–2). The primary model was estimated using robust maximum likelihood estimation (MLR). Model fit was evaluated using the χ 2 statistic, degrees of freedom, p value, χ 2/df, RMSEA with its 90% confidence interval, CFI, TLI and SRMR. Direct, indirect and total effects were estimated within the structural equation model. An adjusted model incorporating relevant demographic and clinical covariates was examined as a sensitivity analysis. The adjusted model was estimated in two separate runs: MLR was used to obtain robust standard errors and model‐fit indices, whereas maximum likelihood estimation with 5000 nonparametric bootstrap resamples was used to obtain bias‐corrected 95% confidence intervals for the indirect effects. Path diagrams were generated using the semPlot package. A two‐sided p < 0.05 was considered statistically significant. Results Common Method Bias Test Harman's single‐factor test was conducted to assess potential common method bias. All self‐reported items were entered into an unrotated exploratory factor analysis. The results identified 24 factors with eigenvalues greater than 1. The Kaiser–Meyer–Olkin value was 0.926, and Bartlett's test of sphericity was significant (χ 2 = 40,554.676, df = 9453, p < 0.001). The first unrotated factor accounted for 24.954% of the total variance, indicating that no single factor explained the majority of the covariance among the self‐reported items. Assessment of Normality Univariate normality was assessed using the Kolmogorov–Smirnov and Shapiro–Wilk tests, together with skewness and kurtosis. Although the formal normality tests were significant for several variables, the absolute skewness and kurtosis values of the 20 variables included in the structural equation model ranged from 0.013 to 1.044 and from 0.232 to 1.619, respectively. Mardia's test indicated significant multivariate skewness (p < 0.001), whereas multivariate kurtosis was not significant (p = 0.110). No multivariate outliers were identified based on Mahalanobis distance using a significance threshold of p < 0.001. General Information of the Study Participants A total of 382 patients were included in the study, including 216 men and 166 women. The demographic and clinical characteristics of the participants are presented in Table 1. A total of 382 patients were included in the final analysis, comprising 216 men and 166 women. The activities of daily living score was 96.00 (70.00, 100.00). The demographic and clinical characteristics of the participants are presented in Table 1. TABLE 1General characteristics of participants with ischemic stroke (N = 382).VariantOption n %VariantOption n %GenderMale21656.5Time since stroke onset15 days to 3 months19851.8Female16643.5> 3 months to ≤ 6 months18448.2Age18–447720.2Medical insurance typeProvincial employee insurance5815.245–5916242.4Municipal employee insurance6918.1≥ 6014337.4Urban–rural resident insurance21155.2Place of residenceUrban22959.9Others4411.5Rural15340.1Income≤ 30004311.3OccupationWorker9123.83000–500022859.7Farmer13334.8≥ 500011129.0Retired8522.3SmokingNo20453.4Others7319.1Yes17846.6Marital statusNot married5614.7DrinkingNo18347.9Married32685.3Yes19952.1Education levelJunior high school or below9925.9Number of chronic diseases110126.4218448.2Senior high school/Technical secondary school14738.5≥ 39725.4College/Junior college7519.6Activity of daily livingIndependent11429.9Mild dependence14638.2Postgraduate or above6116.0Moderate dependence8321.7Severe dependence3910.2 General characteristics of participants with ischemic stroke (N = 382). Descriptive and Bivariate Statistics The stroke‐related knowledge score among patients with ischemic stroke was 17.00 (IQR: 8.00–28.25). The median scores for warning signs, emergency response, risk factors, safe medication use, health behaviours and rehabilitation knowledge were 1.00 (0.00–5.00), 1.00 (0.00–3.00), 6.00 (1.00–11.00), 1.00 (0.00–3.00), 7.00 (2.00–10.00) and 0.00 (0.00–2.00), respectively. The health belief score was 126.00 (IQR: 110.00–145.00), with median dimension scores of 14.00 (11.00–17.00) for perceived susceptibility, 25.00 (20.00–30.00) for perceived severity, 21.00 (16.00–25.00) for perceived benefits, 24.00 (19.00–27.00) for perceived barriers, 27.00 (21.00–31.00) for health motivation and 18.00 (15.00–21.00) for self‐efficacy. The median item‐mean scores for perceived susceptibility, perceived severity, perceived benefits, perceived barriers, health motivation and self‐efficacy were 3.50 (2.75–4.25), 3.57 (2.86–4.29), 3.50 (2.67–4.17), 4.00 (3.17–4.50), 3.86 (3.00–4.43) and 3.60 (3.00–4.20), respectively. The median depression score was 57.50 (IQR: 50.00–67.50). The median self‐management behaviour score was 181.00 (IQR: 158.75–203.00), with median dimension scores of 36.00 (29.00–44.00) for disease management, 19.00 (15.00–22.00) for medication management, 30.00 (25.00–34.00) for diet management, 31.00 (25.00–36.00) for daily living management, 19.00 (15.00–22.00) for emotion management, 23.00 (18.00–27.00) for social functioning and interpersonal relationships and 26.50 (22.00–31.00) for rehabilitation management. The median standardized domain scores for disease management, medication management, diet management, daily living management, emotion management, social functioning and interpersonal relationships, and rehabilitation management were 56.82 (40.91–75.00), 70.00 (50.00–85.00), 59.46 (45.95–70.27), 71.88 (53.13–87.50), 70.00 (50.00–85.00), 70.83 (50.00–87.50) and 69.64 (53.57–85.71), respectively. Correlation Between Studied Variables Spearman's rank correlation analysis showed that stroke‐related knowledge was positively correlated with health beliefs (ρ = 0.553, p < 0.001) and self‐management behaviours (ρ = 0.551, p < 0.001) and negatively correlated with depressive symptoms (ρ = −0.562, p < 0.001). Health beliefs were positively correlated with self‐management behaviours (ρ = 0.539, p < 0.001) and negatively correlated with depressive symptoms (ρ = −0.533, p < 0.001). Depressive symptoms were negatively correlated with self‐management behaviours (ρ = −0.586, p < 0.001). The detailed correlation coefficients are presented in Table 2. TABLE 2Correlation among stroke‐related knowledge, health beliefs, depression and self‐management behaviours in patients with Ischemic Stroke.Stroke‐related knowledgeHealth beliefsDepressionSelf‐management behavioursStroke‐related knowledge10.553** −0.562** 0.551** Health beliefs0.553** 1−0.533** 0.539** Depression−0.562** −0.533** 1−0.586** Self‐management behaviours0.551** 0.539** −0.586** 1 Note: Spearman's ρ; N = 382.** p < 0.01 (two‐tailed). Correlation among stroke‐related knowledge, health beliefs, depression and self‐management behaviours in patients with Ischemic Stroke. Note: Spearman's ρ; N = 382. p < 0.01 (two‐tailed). Serial Mediation Effects of Health Beliefs and Depressive Symptoms on the Association Between Stroke‐Related Knowledge and Self‐Management Behaviours Among Patients With Ischemic Stroke Model Fit Structural equation modelling was performed using the lavaan package (version 0.7–2) in R. Mardia's test indicated that the assumption of multivariate normality was not met. Therefore, model parameters, their 95% confidence intervals and robust fit indices were estimated using the MLR. Stroke‐related knowledge was specified as the independent variable, self‐management behaviours as the dependent variable and health beliefs and depressive symptoms as the mediators. The hypothesized model demonstrated a good fit to the data: robust χ 2 = 192.996, df = 165, p = 0.067, χ 2/df = 1.170, robust RMSEA = 0.021 (90% CI: 0.000–0.033), robust CFI = 0.990, robust TLI = 0.988 and SRMR = 0.037. All indices met the prespecified criteria of χ2/df < 3.00, RMSEA < 0.05, CFI and TLI > 0.90 and SRMR < 0.08. Direct and Indirect Effects of the Structural Equation Model The structural equation model showed that stroke‐related knowledge was positively associated with health beliefs (B = 0.900, β = 0.640, p < 0.001) and self‐management behaviours (B = 1.124, β = 0.293, p < 0.001) and negatively associated with depressive symptoms (B = −2.324, β = −0.332, p < 0.001). Health beliefs were negatively associated with depressive symptoms (B = −1.919, β = −0.386, p < 0.001) and positively associated with self‐management behaviours (B = 0.660, β = 0.242, p = 0.001). Depressive symptoms were negatively associated with self‐management behaviours (B = −0.180, β = −0.329, p < 0.001). The model explained 40.9%, 42.3% and 55.2% of the variance in health beliefs, depressive symptoms and self‐management behaviours, respectively. Detailed structural path estimates are presented in Table 3. TABLE 3Estimates of the direct structural paths.OutcomePredictorBSE z β 95% CI p Health beliefsStroke‐related knowledge0.9000.09010.0340.6400.725 to 1.076< 0.001Depressive symptomsStroke‐related knowledge−2.3240.483−4.809−0.332−3.272 to −1.377< 0.001Depressive symptomsHealth beliefs−1.9190.322−5.962−0.386−2.550 to −1.288< 0.001Self‐management behavioursStroke‐related knowledge1.1240.2923.8530.2930.552 to 1.696< 0.001Self‐management behavioursHealth beliefs0.6600.2083.1760.2420.253 to 1.0670.001Self‐management behavioursDepressive symptoms−0.1800.033−5.428−0.329−0.245 to −0.115< 0.001 Note: The R 2 values for health beliefs, depressive symptoms and self‐management behaviours were 0.409, 0.423 and 0.552, respectively.Abbreviations: β, standardized path coefficient; B, unstandardized path coefficient; CI, confidence interval; SE, standard error. Estimates of the direct structural paths. Note: The R 2 values for health beliefs, depressive symptoms and self‐management behaviours were 0.409, 0.423 and 0.552, respectively. Abbreviations: β, standardized path coefficient; B, unstandardized path coefficient; CI, confidence interval; SE, standard error. The total effect of stroke‐related knowledge on self‐management behaviours was statistically significant (B = 2.448, 95% CI [1.925, 3.015]). After health beliefs and depressive symptoms were included in the model, the direct effect of stroke‐related knowledge on self‐management behaviours remained statistically significant (B = 1.124, 95% CI [0.552, 1.696]), accounting for 45.92% of the total effect. The total indirect effect of health beliefs and depressive symptoms was also statistically significant (B = 1.324, 95% CI [0.966, 1.752]), accounting for 54.08% of the total effect. All three specific indirect pathways were statistically significant. The indirect effect through health beliefs was 0.594 (95% CI [0.238, 0.970]), accounting for 24.26% of the total effect. The indirect effect through depressive symptoms was 0.419 (95% CI [0.236, 0.683]), accounting for 17.12% of the total effect. The serial indirect effect through health beliefs followed by depressive symptoms was 0.311 (95% CI [0.176, 0.510]), accounting for 12.70% of the total effect. The 95% CIs for all three specific indirect effects did not include zero. Standardized path coefficients are shown in Figure 1, and the direct, indirect and total effects are presented in Table 4. Serial mediation model of health beliefs and depressive symptoms in the association between stroke‐related knowledge and self‐management behaviours. Values shown are standardized path coefficients. TABLE 4Direct and indirect effects of health beliefs and depressive symptoms on the association between stroke‐related knowledge and self‐management behaviours.Effect pathwayBSE95% CI p β Proportion of total effect (%)Stroke‐related knowledge → Health beliefs → Self‐management behaviours0.5940.1870.238–0.9700.0020.15524.26Stroke‐related knowledge → Depressive symptoms → Self‐management behaviours0.4190.1130.236–0.683< 0.0010.10917.12Stroke‐related knowledge → Health beliefs → Depressive symptoms → Self‐management behaviours0.3110.0850.176–0.510< 0.0010.08112.70Total indirect effect1.3240.1960.966–1.752< 0.0010.34654.08Direct effect1.1240.2920.552–1.696< 0.0010.29345.92Total effect2.4480.2771.925–3.015< 0.0010.639100.00 Note: The direct, indirect and total effects and their 95% confidence intervals were estimated using the MLR estimator.Abbreviations: β, standardized effect estimate; B, unstandardized effect estimate; CI, confidence interval; SE, standard error. Direct and indirect effects of health beliefs and depressive symptoms on the association between stroke‐related knowledge and self‐management behaviours. Note: The direct, indirect and total effects and their 95% confidence intervals were estimated using the MLR estimator. Abbreviations: β, standardized effect estimate; B, unstandardized effect estimate; CI, confidence interval; SE, standard error. Sensitivity Analysis To examine the robustness of the primary findings, a covariate‐adjusted path model was additionally fitted using the observed total scores of the study variables. Gender, age, place of residence, educational level, income, time since stroke onset and activities of daily living were simultaneously included as covariates. Among the potential covariates, recruitment setting was not included because inpatients were admitted for reassessment during recovery rather than for acute stroke treatment. After adjustment for these demographic and clinical characteristics, the directions and statistical significance of the principal pathways remained unchanged. Stroke‐related knowledge was positively associated with health beliefs (B = 0.860, β = 0.425, p < 0.001) and self‐management behaviours (B = 0.215, β = 0.079, p = 0.006) and negatively associated with depressive symptoms (B = −0.302, β = −0.281, p < 0.001). Health beliefs were negatively associated with depressive symptoms (B = −0.149, β = −0.280, p < 0.001) and positively associated with self‐management behaviours (B = 0.104, β = 0.078, p = 0.005). Depressive symptoms remained negatively associated with self. The adjusted model was additionally estimated using maximum likelihood with 5000 nonparametric bootstrap resamples to obtain bias‐corrected 95% confidence intervals for the indirect effects. The adjusted indirect effect through health beliefs was 0.089 (95% CI [0.025, 0.158]), the adjusted indirect effect through depressive symptoms was 0.086 (95% CI [0.045, 0.146]) and the adjusted serial indirect effect through health beliefs and depressive symptoms was 0.036 (95% CI [0.018, 0.065]). The total adjusted indirect effect was 0.212 (95% CI [0.136, 0.289]), while the adjusted direct and total effects were 0.215 and 0.426, respectively. None of the confidence intervals for the indirect effects included zero, indicating that the primary mediation findings remained robust after adjustment for the measured demographic and clinical covariates. Discussion Current Status of Stroke‐Related Knowledge, Health Beliefs, Depression and Self‐Management Levels in Patients With Ischemic Stroke The stroke‐related knowledge score was 17.00 (8.00–28.25), with considerable variation among patients, consistent with previous findings (Wang et al. 2018). Some patients may lack the knowledge needed to recognize stroke‐related risks and promote recovery. This may be related to sociodemographic characteristics, health‐related behaviours, access to health information and previous exposure to stroke education (Wang et al. 2025). Therefore, stroke education should be tailored to individual patient characteristics, with particular emphasis on stroke recognition, medication safety, rehabilitation and prevention of recurrent stroke. The health belief score was 126.00 (110.00–145.00), with differences observed across the six dimensions. Based on the median item‐mean scores, the perceived‐barriers dimension and health motivation had relatively higher scores, whereas perceived susceptibility and perceived benefits had relatively lower scores; self‐efficacy and perceived severity were intermediate. These findings suggest that although patients may recognize the seriousness of stroke, their perceptions of the risk of recurrent stroke and the benefits of health behaviours, as well as their confidence in engaging in these behaviours, may remain insufficient. Previous research has suggested that limited opportunities for health education and inadequate access to health information may be associated with insufficient health beliefs (Cabezas et al. 2025). Therefore, stroke education should not focus solely on the provision of disease‐related knowledge but should also help patients recognize the risk of stroke recurrence and the benefits of health behaviours and strengthen their confidence in initiating and maintaining recommended behaviours. Appropriate involvement of family members may also provide support for maintaining health behaviours during recovery. In this study, the prevalence of depressive symptoms was 70.9%, including mild symptoms in 123 patients (32.2%), moderate symptoms in 90 (23.6%) and severe symptoms in 58 (15.2%), which was higher than that reported in previous studies (Xiao et al. 2024). Differences across studies may be related to variations in the stage of stroke recovery, severity of functional impairment and level of social support among study populations. Post‐stroke functional limitations and changes in social roles may increase patients' psychological burden during recovery, particularly when difficulties with daily activities or reduced social participation persist (Wijeratne and Sales 2021). Therefore, psychological status should be considered in stroke rehabilitation care, with greater attention to the early identification of depressive symptoms. Patients with moderate or severe depressive symptoms should receive further assessment, with appropriate psychological support or referral for additional intervention based on the assessment findings. In addition, the total self‐management behaviour score was 181.00 (158.75–203.00), with substantial differences across dimensions. Among the seven dimensions, daily living management and social functioning and interpersonal relationships had the two highest standardized scores, whereas disease management and diet management had the two lowest standardized scores, consistent with previous findings (Ruksakulpiwat and Zhou 2021). This suggests that patients may experience difficulties in translating disease‐related health recommendations into sustained self‐management behaviours. Disease management involves multiple behaviours, including symptom monitoring, vascular risk‐factor control, rehabilitation exercises and communication with healthcare professionals, whereas diet management requires sustained changes in established dietary habits and long‐term adherence to dietary recommendations. These demands may make these behaviours more difficult to maintain. Therefore, stroke rehabilitation care should pay greater attention to disease and diet management by providing individualized guidance on symptom monitoring, risk‐factor control and dietary management, together with continued follow‐up after discharge to facilitate the long‐term maintenance of recommended behaviours. Relationships Among Illness Perception, Health Beliefs, Depression and Self‐Management Levels in Patients Recovering From Ischemic Stroke In this study, stroke‐related knowledge was positively correlated with health beliefs and self‐management behaviours and negatively correlated with depression. These findings highlight the close interrelationships among stroke‐related knowledge, health beliefs, depressive symptoms and self‐management behaviours. Therefore, nurses can provide individualized education according to patients' knowledge levels, use visual educational materials and repeat key rehabilitation information during hospitalization and follow‐up visits to improve patients' understanding of stroke recovery. Health beliefs were also positively correlated with self‐management behaviours. From the perspective of the Health Belief Model, greater perceived benefits and self‐efficacy, together with fewer perceived barriers, may strengthen patients' motivation and persistence in adopting recommended health behaviours. This mechanism is particularly relevant during stroke recovery, when medication use, lifestyle modification and rehabilitation exercises require sustained engagement (Taft et al. 2021). Health beliefs were also negatively associated with depressive symptoms, consistent with previous findings that greater self‐efficacy and perceived control are associated with fewer depressive symptoms among stroke survivors (Alyafei and Easton‐Carr 2025). More favourable health beliefs may help patients view stroke‐related risks and rehabilitation demands as manageable, thereby reducing negative expectations and strengthening confidence in coping with recovery. Nurses may therefore promote adaptive health beliefs by setting achievable rehabilitation goals, providing feedback on recovery progress and encouraging participation in self‐care activities. Depressive symptoms were negatively associated with self‐management behaviours in this study, consistent with previous studies reporting poorer adherence and lower engagement in health behaviours among stroke survivors with depressive symptoms (Hisan et al. 2025). Depressive symptoms may impair concentration and information processing, reduce motivation and perceived capacity for action and undermine the persistence required to follow medication, lifestyle and rehabilitation plans (Blöchl et al. 2019). Nursing care should therefore combine health education with emotional assessment and support. Early identification of depressive symptoms and timely psychological support may help patients maintain self‐management throughout recovery. Serial Multiple Mediation Effects of Health Beliefs and Depressive Symptoms on the Association Between Stroke‐Related Knowledge and Self‐Management Behaviours Stroke‐related knowledge was directly associated with self‐management behaviours, accounting for 45.92% of the total effect. Patients with greater stroke‐related knowledge may have a better understanding of disease management and recovery needs, which may support their engagement in rehabilitation and daily self‐management (Kharbach et al. 2020; Sakakibara et al. 2017). During stroke recovery, patients are required to manage multiple aspects of daily life, including functional exercise, dietary control, medication use and blood pressure monitoring. Therefore, nurses should provide individualized education and support patients in translating stroke‐related knowledge into daily self‐management behaviours. The total indirect effect accounted for 54.08% of the total effect. Among the three specific indirect pathways, the pathway through health beliefs made the largest contribution, accounting for 24.26% of the total effect. This finding suggests that health beliefs may partly account for the association between stroke‐related knowledge and self‐management behaviours, consistent with previous research (Kim et al. 2024). Greater knowledge may help patients understand the necessity, expected benefits, potential barriers and feasibility of rehabilitation and other self‐management behaviours. More favourable health beliefs, particularly greater perceived benefits and self‐efficacy and fewer perceived barriers, may in turn be associated with greater motivation and persistence in self‐management. Different dimensions of health beliefs may, however, play different roles, with perceived benefits and self‐efficacy potentially facilitating self‐management and perceived barriers potentially hindering it (Klockar et al. 2023). Therefore, nurses should assess not only patients' knowledge but also their perceptions of the benefits, barriers and feasibility of recommended behaviours. Goal setting, barrier‐focused counselling and feedback on rehabilitation progress may help patients translate knowledge into sustained self‐management. Depressive symptoms represented a conceptually distinct indirect pathway, accounting for 17.12% of the total effect. Greater stroke‐related knowledge was associated with fewer depressive symptoms, which were in turn associated with better self‐management behaviours. This finding is consistent with previous evidence linking depressive symptoms to poorer adherence, reduced rehabilitation participation and less effective self‐management among stroke survivors (Gnanaprakasam et al. 2024). Depressive symptoms may interfere with self‐management through fatigue, impaired concentration and executive functioning, anhedonia, reduced expectations of treatment benefit and social withdrawal (Caetano et al. 2023). These difficulties may make it more challenging to initiate and maintain medication use, lifestyle modification, rehabilitation exercises and other self‐management behaviours. Therefore, emotional assessment and timely psychological support should complement health education throughout stroke recovery. The serial indirect pathway through health beliefs and depressive symptoms accounted for 12.70% of the total effect. This significant serial indirect effect suggests that health beliefs and depressive symptoms may be jointly involved in the association between stroke‐related knowledge and self‐management behaviours. Patients with greater stroke‐related knowledge may develop more favourable perceptions of the benefits and feasibility of recommended behaviours and greater confidence in managing their recovery. Such health beliefs may be associated with fewer depressive symptoms, which may in turn be associated with greater engagement in self‐management behaviours. Previous studies in chronic disease populations have linked disease‐related knowledge and depressive symptoms with health beliefs and self‐management, but evidence for their serial association remains limited (Wang et al. 2023; Derese et al. 2024). Studies in stroke‐related populations have also reported indirect associations between stroke‐related knowledge and health behaviour or self‐management involving health beliefs, self‐esteem, health information literacy, psychological flexibility and anxiety (Zeng et al. 2025; Ma et al. 2026). However, previous studies have primarily focused on a single mediator or examined serial pathways involving other psychological constructs, without simultaneously including overall health beliefs and depressive symptoms. Therefore, the present study addresses this gap by examining a proposed serial indirect association involving overall health beliefs and depressive symptoms among patients recovering from ischemic stroke. In clinical practice, nurses should assess not only patients' stroke‐related knowledge but also their perceived benefits, barriers, self‐efficacy and depressive symptoms. Nurses should also screen for depressive symptoms early, provide timely emotional support and coordinate referrals to psychological or psychiatric services when necessary, thereby helping patients initiate and sustain self‐management throughout recovery. Conclusion Stroke‐related knowledge, health beliefs, depressive symptoms and self‐management behaviours were significantly associated among patients recovering from ischemic stroke. Stroke‐related knowledge was associated with self‐management behaviours directly and through three significant indirect pathways involving health beliefs, depressive symptoms and health beliefs followed by depressive symptoms. The pathway through health beliefs accounted for the largest specific indirect effect, while the remaining direct association indicated partial mediation. These associations remained robust after adjustment for selected demographic characteristics, time since stroke onset and activities of daily living. Future nurses should strengthen support for self‐management during stroke recovery by integrating stroke‐related education with routine assessment of patients' health beliefs and depressive symptoms. Limitations This study provides new insights into self‐management among patients recovering from ischemic stroke. However, it has several limitations. First, its cross‐sectional design precludes conclusions about the temporal or causal ordering of stroke‐related knowledge, health beliefs, depressive symptoms and self‐management behaviours. Although the hypothesized model positioned health beliefs before depressive symptoms, alternative or reciprocal relationships remain possible and require longitudinal investigation. Second, convenience sampling from a single tertiary hospital may have resulted in selection bias and limited the representativeness of the sample and the generalizability of the findings. Exact time since stroke onset was not recorded, although categorized time since onset and activities of daily living were included in the sensitivity analysis. Third, unmeasured clinical factors, including stroke severity, lesion location, recurrence status, rehabilitation treatment and antidepressant use, may have resulted in residual confounding. Fourth, the use of self‐report questionnaires at the same time point may have introduced response, social desirability and common method biases, which cannot be excluded by Harman's single‐factor test alone. Future multicenter longitudinal studies should use repeated assessments and collect detailed clinical information. Author Contributions Yue Li: conceptualization, writing – original draft, data curation, investigation. Lulu Li: data curation, writing – review and editing, supervision, funding acquisition. Xuan Wang: data curation, investigation. Beilei Lin: writing – review and editing, supervision. Ziye Guo: writing – review and editing, investigation, data curation. Qilan Tang: writing – review and editing, supervision, funding acquisition. Aixia Wang: writing – review and editing, supervision, funding acquisition, resources. Funding This work was supported by the Health Commission of Henan Province, and the funding programme is the Henan Provincial Medical Science and Technology Research Program (SBGJ202102088), the Henan Provincial Federation of Social Sciences, and the funding programme is the Research Project of the Henan Provincial Federation of Social Sciences (SKL‐2025‐1364) and the First Affiliated Hospital of Zhengzhou University, and the funding programme is the Nursing Research Project of the First Affiliated Hospital of Zhengzhou University (HLKY2023007). Disclosure Statistical Statement: One of the authors (Yue Li) has received formal training in statistics and was responsible for the statistical analysis in this study. The authors confirm that the statistical methods used were appropriate for the study design and data, and that the results were correctly interpreted. Ethics Statement This study was approved by the Scientific Research and Clinical Trials Ethics Committee of the First Affiliated Hospital of Zhengzhou University (Approval No. 2024‐KY‐0556‐002; approval date: June 3, 2024). All participants were fully informed of the study's purpose and procedures. Written informed consent was obtained from all participants prior to data collection. All methods were carried out in accordance with relevant guidelines and regulations. Consent The authors have nothing to report. Conflicts of Interest The authors declare no conflicts of interest.