Patient activation and its influencing factors among maintenance hemodialysis patients: a cross-sectional study in China

Abstract

Objective

To investigate the current status of patient activation and identify its influencing factors among patients undergoing maintenance hemodialysis (MHD).

Methods

A cross-sectional study was conducted among 200 MHD patients from four tertiary hospitals in Taiyuan, China, between May and July 2024. Convenience sampling was used. Data were collected using the Patient Activation Measure (PAM-13), Self-Efficacy for Managing Chronic Disease Scale (SECD6), Hemodialysis Self-Management Instrument (HD-SMI), Herth Hope Index (HHI), and Social Support Rating Scale (SSRS). Sociodemographic and clinical variables were also recorded. Univariate analyses, Pearson correlation, and multiple linear regression were employed to explore the relationships between patient activation and potential associated factors.

Results

The mean PAM score was 50.29 ± 12.01, with 71.5% of patients classified as Level 1 or 2, indicating low to moderate activation. Higher patient activation was significantly associated with higher education level, potassium-lowering drug use, full self-care ability, and target-range hemoglobin levels. Correlation analysis showed that PAM scores were positively correlated with SECD6, HD-SMI, HHI, and SSRS scores (p < 0.001). Multiple regression identified education level, potassium-lowering drug use, self-care ability, hemoglobin level, self-efficacy, self-management, hope, and social support as significant independent predictors, collectively explaining 41.7% of the variance in activation scores.

Conclusion

Patient activation among the surveyed MHD patients was generally low and influenced by a combination of clinical, functional, and psychosocial factors. These findings highlight the need for targeted, multidimensional interventions to enhance patient engagement and self-management in the MHD population.

Introduction

Maintenance hemodialysis (MHD) is a critical life-sustaining therapy for patients with kidney failure [1, 2]. As the global burden of chronic kidney disease (CKD) continues to rise, the number of patients dependent on long-term hemodialysis is steadily increasing [3, 4]. In China, the annual growth of the MHD population has placed significant pressure on both the healthcare system and affected families, with patients facing not only high treatment costs but also a range of physical and psychological challenges [5, 6]. Despite advancements in dialysis technology and standardization of clinical protocols, outcomes such as quality of life, complication rates, and hospitalization remain suboptimal for many patients.

Patient activation, defined as the individual’s knowledge, skills, and confidence in managing their health, has gained attention as a critical determinant of chronic disease outcomes [7, 8]. Higher levels of patient activation have been associated with improved treatment adherence, better self-management behaviors, reduced healthcare utilization, and enhanced quality of life [9]. In the context of MHD, where patients must adhere to strict fluid and dietary restrictions, manage vascular access, and cope with fatigue, emotional distress, and complex medication regimens, activation plays a particularly important role [10]. However, existing evidence suggests that patients undergoing MHD often exhibit lower levels of activation compared to those managing other chronic conditions, indicating a pressing need for further investigation in this population [11].

Moreover, little is known about the multifactorial influences on activation, including psychological resilience, self-management capabilities, health literacy, and social support. This study aims to assess the current status of patient activation in MHD patients and to identify key demographic, clinical, and psychosocial factors associated with activation levels. Understanding these relationships may inform the development of targeted interventions to enhance patient engagement and ultimately improve dialysis outcomes.

Methods

Study design and participants

This study employed a cross-sectional design to assess the status of patient activation and its influencing factors among MHD patients. The sample size was determined using the rule of thumb that requires a minimum of 5–10 participants per independent variable in regression analysis. With 26 independent variables considered, the estimated sample size ranged between 143 and 286, factoring in a 10% potential for invalid responses. Ultimately, 200 patients were enrolled through convenience sampling from four tertiary general hospitals in Taiyuan, Shanxi Province, between May and July 2024.

Inclusion criteria were: (1) diagnosed with kidney failure and receiving MHD for more than three months; (2) aged 18 years or older; (3) able to understand and complete the questionnaires independently; and (4) provided written informed consent. Exclusion criteria included: severe cardiovascular or cerebrovascular disease, malignancies, psychiatric disorders, severe hepatic disease, concurrent peritoneal dialysis, or communication impairments affecting questionnaire completion.

Measures and instruments

1. General information questionnaire A self-designed demographic and clinical data questionnaire was used, informed by literature and study objectives. Items included age, sex, education level, marital status, employment status, dialysis duration, frequency, comorbidities, medication usage (including antihypertensive and potassium-lowering drugs), self-care ability, hemoglobin levels, and insurance type. Self-care ability was assessed using the six basic Activities of Daily Living (ADL): eating, dressing, transferring in/out of bed, toileting, indoor walking, and bathing. Each activity was rated as independently performed or not. Patients were classified as fully independent (all six activities performed independently) or partially dependent (one to three activities requiring assistance). The assessment was completed by trained nursing staff according to patients’ usual functional status during the study period.

2. Patient Activation Measure (PAM-13) Patient activation was assessed using the PAM-13, measuring knowledge, skills, and confidence in health management. Items are scored on a 5-point Likert scale, yielding a converted total score (0–100). Scores classify patients into four progressively active levels: Level 1 (≤ 47.0), Level 2 (47.1–55.1), Level 3 (55.2–67.0), and Level 4 (≥ 67.1). The Chinese version demonstrated high reliability (Cronbach’s α = 0.882) [12].

3. Self-Efficacy for Managing Chronic Disease 6-Item Scale (SECD6) The SECD6 evaluates confidence in managing symptoms [3, 13]. Items are rated from 1 to 10, with total scores (6–60) positively reflecting self-efficacy.

4. Hemodialysis Self-Management Instrument (HD-SMI) The 20-item HD-SMI assesses self-management across four dimensions (problem-solving, partnership, emotional processing, and self-care implementation) [14, 4]. Items are scored 1–4, with total scores (20–80) positively correlating with better self-management. The reliability in this study was Cronbach’s α = 0.862.

5. Herth Hope Index (HHI) The 12-item HHI measures hope across three domains using a 1–4 Likert scale [15]. Total scores range from 12 to 48, with higher scores indicating greater hope (Cronbach’s α = 0.85).

6. Social Support Rating Scale (SSRS) The 10-item SSRS assesses objective support, subjective support, and support utilization [5, 16]. Scores (0–66) are categorized into low (≤ 22), moderate (23–44), and high (≥ 45) social support (Cronbach’s α = 0.89–0.94).

Data collection procedure

Data collection was conducted by two trained nephrology nurses and two nursing graduate students. All investigators received standardized training on the use of measurement tools and administration procedures. After obtaining written informed consent, participants were provided with the questionnaire and completed it independently in a quiet clinical setting. A total of 211 questionnaires were distributed, and 200 valid responses were retained (valid response rate: 94.8%).

Ethical considerations

This study was conducted in accordance with the principles of the Declaration of Helsinki. The research protocol was reviewed and approved by the Ethics Committee of our hospital. Written informed consent was obtained from all participants prior to data collection. Participants were assured of the confidentiality of their responses and their right to withdraw from the study at any time without any impact on their medical care.

Statistical analysis

Data were entered and double-checked by two independent researchers and analyzed using SPSS 26.0. Missing data were handled via listwise deletion, meaning only fully completed and logically consistent questionnaires were retained for the final analysis. Continuous variables were expressed as mean ± standard deviation (SD), while categorical variables were reported as frequencies and percentages. Univariate analysis was performed using independent-sample t-tests and one-way ANOVA. Pearson correlation analysis was used to examine the associations between activation and psychosocial measures. Variables showing statistical significance in the univariate analyses (p < 0.05) or those considered clinically important based on prior literature and theoretical relevance (e.g., age, sex, dialysis duration, and comorbidities) were considered as candidate predictors. To balance model parsimony with the risk of omitting important confounders in this moderate-sized sample (n = 200), we primarily used a data-driven approach: variables with p < 0.05 in univariate analyses were entered into the multiple linear regression model using the enter method. Linear regression was chosen because the outcome variable was approximately normally distributed and the primary aim was to estimate adjusted associations between predictors and the continuous outcome, without assuming non-linear effects or complex hierarchical structures. Alternative approaches such as ordinal logistic regression would sacrifice granularity of the continuous scores.

We acknowledge that strict univariate screening at p < 0.05 can lead to the exclusion of potentially important confounders due to limited statistical power or suppressor effects in the univariate stage. Therefore, sensitivity analyses were performed by forcing theoretically relevant variables (regardless of univariate p-values) into the model, and the results remained largely consistent (data not shown). Model assumptions were thoroughly evaluated, and no evidence of multicollinearity was detected (all VIF < 5). To evaluate the validity of the model, key assumptions of multiple linear regression were examined. Multicollinearity was assessed using the Variance Inflation Factor (VIF) and tolerance statistics in SPSS. Linearity was assessed through scatterplots of residuals versus predicted values, normality of residuals was evaluated using histograms and normal probability plots (P-P plots), and homoscedasticity was examined via residual plots. Cook’s distance and leverage values were used to identify influential cases. All assumptions were adequately met. Standardized regression coefficients (β) were also calculated and reported to allow comparison of the relative importance of predictors. Statistical significance was set at p < 0.05.

Results

Patient characteristics

A total of 211 questionnaires were distributed and collected. Among these, 11 questionnaires were deemed invalid due to incomplete responses (n = 6), missing key variables (n = 3), or logical inconsistencies (n = 2), resulting in 200 valid questionnaires for analysis. The valid response rate was 94.8%. The sociodemographic and clinical characteristics of the 200 patients included in the analysis are presented in Table 1. In particular, the mean age was 54.79 ± 14.33 years, with 62.0% male and 38.0% female. The majority were married (85.5%), and 70.5% lived with their spouse. Regarding education, 37.5% had completed high school or vocational school, while 16.0% held a bachelor’s degree. Employment status showed that 16.5% were currently employed.

Table 1.

Demographic and clinical characteristics of MHD patients

Characteristic	n (%) or Mean ± SD
Age (years)	54.79 ± 14.33
Gender
- Male	124 (62.0%)
- Female	76 (38.0%)
Education level
- Elementary or below	24 (12.0%)
- Junior high	65 (32.5%)
- High school/Vocational	75 (37.5%)
- Bachelor’s degree	32 (16.0%)
- Graduate and above	4 (2.0%)
Marital status
- Married	171 (85.5%)
- Unmarried	18 (9.0%)
- Divorced or Widowed	11 (5.5%)
Employment status
- Employed	33 (16.5%)
- Unemployed/Retired	167 (83.5%)
Living arrangement
- With spouse	141 (70.5%)
- With children	25 (12.5%)
- Alone	19 (9.5%)
- Others	15 (7.5%)
Self-care ability
- Fully independent	157 (78.5%)
- Partially dependent	43 (21.5%)
Use of antihypertensive drugs
- Yes	148 (74.0%)
- No	52 (26.0%)
Use of potassium-lowering drugs
- Yes	105 (52.5%)
- No	95 (47.5%)
Hemoglobin within target
- Yes	94 (47.0%)
- No	106 (53.0%)
Dialysis frequency (per week)
- Twice	18 (9.0%)
- Thrice	174 (87.0%)
- Four times or more	8 (4.0%)

Clinically, 78.5% of patients reported full self-care ability. 74.0% were on antihypertensive medications, and 52.5% were taking potassium-lowering drugs. Regarding hemoglobin level status, the recommended target for maintenance hemodialysis patients was defined as 110–130 g/L, in accordance with clinical guidelines. Among the 200 patients, 94 (47.0%) had hemoglobin levels within the target range, while 106 (53.0%) were outside the target range. Of those outside the target, 92 patients (46.0% of the total sample) had hemoglobin levels below 110 g/L, and 14 patients (7.0% of the total sample) had levels above 130 g/L. The average dialysis duration was 3.4 ± 2.1 years, with 87.0% receiving dialysis three times per week.

Patient activation status

The overall patient activation level among the 200 MHD patients was relatively low. The total score on the Patient Activation Measure (PAM-13) ranged from 15.40 to 75.60, with a mean score of 50.29 ± 12.01, indicating that the average patient activation was at Level 2. According to the standardized PAM classification, the majority of patients fell into the lower activation tiers: 114 patients (57.0%) were classified as Level 1, 29 patients (14.5%) as Level 2, 26 patients (13.0%) as Level 3, and 31 patients (15.5%) as Level 4 (Table 2).

Table 2.

Distribution of patient activation levels in MHD patients

Activation Level	Score Range	Frequency (n, %)	Mean ± SD Score
Level 1	≤ 47.0	114 (57.0%)	41.87 ± 5.08
Level 2	47.1–55.1	29 (14.5%)	50.84 ± 1.89
Level 3	55.2–67.0	26 (13.0%)	61.44 ± 3.30
Level 4	≥ 67.1	31 (15.5%)	71.41 ± 2.99

Level 1 patients showed limited recognition of their role in health management and largely relied on passive care, while those at Level 4 demonstrated confidence and consistency in self-care, particularly under stress. The distribution suggests that a significant proportion of patients lack sufficient confidence, knowledge, and skills for active health self-management.

Univariate analysis of influencing factors

To identify potential variables associated with patient activation, univariate analyses were performed across demographic and clinical characteristics. The results indicated that education level, use of potassium-lowering drugs, self-care ability, and hemoglobin level status were significantly associated with activation scores (p < 0.05).

Specifically, activation scores were higher among patients with a bachelor’s degree or higher, those taking potassium-lowering medications, individuals capable of full self-care, and those with hemoglobin levels within the target range (110–130 g/L). These associations are visualized in Fig. 1A and D, which demonstrate progressively higher median activation scores in patients with favorable profiles across each factor.

Fig. 1.

Boxplots showing differences in patient activation scores by key influencing factors. (A) Education level: activation scores increased with higher educational attainment. (B) Potassium-lowering drug use: patients on potassium-lowering medications had higher activation scores than those not taking them. (C) Self-care ability: fully independent patients demonstrated significantly higher activation scores. (D) Hemoglobin level status: patients with hemoglobin within the target range (110–130 g/L) had higher activation scores than those with suboptimal levels

Correlation between activation and psychosocial factors

To explore the relationship between patient activation and key psychosocial variables, Pearson correlation analyses were conducted between PAM scores and the scores of four instruments: the SECD6, the HD-SMI, the HHI, and the SSRS.

The analysis revealed statistically significant positive correlations between patient activation and all four psychosocial measures (p < 0.001). Specifically, the correlation coefficients were as follows: SECD6 (r = 0.369), HD-SMI (r = 0.502), HHI (r = 0.450), and SSRS (r = 0.366), indicating that higher levels of self-efficacy, self-management ability, hope, and social support were each associated with greater patient activation (Table 3).

Table 3.

Pearson correlation coefficients between patient activation and psychosocial factors

Variable	Correlation Coefficient (r)	p-value
SECD6 (Self-efficacy)	0.369	< 0.001
HD-SMI (Self-management)	0.502	< 0.001
HHI (Hope level)	0.450	< 0.001
SSRS (Social support)	0.366	< 0.001

Multivariate regression analysis

To further identify the independent factors associated with patient activation, a multiple linear regression analysis was performed using the PAM score as the dependent variable. Predictor variables were selected based on their significance in the univariate analysis and included education level, potassium-lowering drug use, self-care ability, hemoglobin level status, and the scores from four psychosocial scales: SECD6 (self-efficacy), HD-SMI (self-management), HHI (hope level), and SSRS (social support).

The regression model demonstrated good explanatory power, with an F-value of 16.819 and a p-value of less than 0.001, indicating statistical significance. The adjusted R-squared was 0.417, suggesting that 41.7% of the variance in patient activation scores could be accounted for by the variables included in the model.

Higher education level was positively associated with activation (B = 1.482, β = 0.142, p = 0.039). Patients not taking potassium-lowering drugs had significantly lower activation scores (B = − 4.701, β = − 0.196, p < 0.001). Similarly, partial self-care ability (compared to full independence) was negatively associated with activation (B = − 4.126, β = − 0.148, p = 0.021), as was having hemoglobin levels outside the recommended target range (B = − 3.545, β = − 0.147, p = 0.008).

In terms of psychosocial predictors, higher scores on the SECD6 self-efficacy scale were associated with higher activation (B = 0.146, β = 0.132, p = 0.048). The self-management score (HD-SMI) was the strongest predictor among psychosocial factors (B = 0.271, β = 0.289, p < 0.001), followed by hope level (HHI; B = 0.384, β = 0.167, p = 0.018), and social support (SSRS; B = 0.221, β = 0.154, p = 0.029). All of these variables showed significant positive associations with activation. No evidence of multicollinearity was detected (all VIF < 5, tolerance > 0.25). All of these variables showed significant positive associations with activation, suggesting that internal and external psychosocial resources play a crucial role in promoting proactive health behaviors in MHD patients (Fig. 2). Sensitivity analyses were performed by forcing theoretically relevant variables into the model regardless of univariate significance. The direction and statistical significance of the primary predictors remained materially unchanged, supporting the robustness of the main findings.

Fig. 2.

Multivariate regression coefficients and 95% confidence intervals for factors associated with patient activation

The plot illustrates the unstandardized regression coefficients (B values) and their 95% confidence intervals derived from the multiple linear regression model. Positive values indicate a direct association with higher patient activation scores, while negative values indicate an inverse relationship. The most influential factors included self-management (HD-SMI), hope level (HHI), education level, and use of potassium-lowering drugs. The red dashed line represents the null effect (B = 0).

Discussion

This study revealed that patient activation among MHD patients in China remains suboptimal, with over 70% of participants scoring at Levels 1 or 2 on the PAM-13. These findings are consistent with previous international studies indicating low activation among dialysis patients, possibly due to the complex, high-burden nature of treatment regimens and the psychological toll of chronic kidney disease [17, 18, 6, 7]. The mean activation score in our cohort (50.29 ± 12.01) is comparable to findings from Belgium and the Middle East, suggesting a global need to enhance self-management support for this vulnerable population [11].

Among the demographic and clinical factors examined, higher educational attainment emerged as a factor significantly and positively associated with activation. This aligns with findings from Gopal et al., who reported that more educated patients tend to demonstrate better understanding and engagement in disease management, possibly due to greater health literacy and information-seeking behaviors [19]. The association between use of potassium-lowering drugs and higher activation scores is noteworthy but should be interpreted cautiously. Potassium-lowering medications are typically prescribed for patients with persistent hyperkalemia despite dietary and dialytic management, which may reflect greater disease severity or higher treatment intensity. Thus, the higher activation observed in this subgroup could reflect more intensive clinical monitoring, increased patient-provider interactions, or greater proactive engagement in care, rather than a direct causal effect of the medication itself on patient activation. In the Chinese healthcare context, these drugs are generally reimbursed under national medical insurance schemes, minimizing financial barriers. Future studies should further explore whether patient activation influences adherence to complex regimens or moderates the need for such therapies.

In the Chinese healthcare context, potassium-lowering drugs (primarily traditional resins such as sodium or calcium polystyrene sulfonate) are generally reimbursed under national and provincial medical insurance schemes for kidney failure patients, minimizing out-of-pocket costs and reducing the likelihood of income-related disparities in access. Although we did not collect data on household income, any potential confounding by socioeconomic status is likely limited for these standard therapies. The observed association may instead reflect greater clinical monitoring and patient-provider interactions in those requiring potassium-lowering medications.

The association between full self-care ability and higher activation emphasizes the importance of functional independence in chronic illness management. This supports prior observations by Mirmazhari and Ghafourifard, who found that self-efficacy and physical capability play pivotal roles in fostering patient activation among MHD patients [20, 10]. Hemoglobin levels within the target range were also positively associated with activation, suggesting that physiological stability may enhance patients’ confidence and motivation to manage their health proactively. In the Chinese context, erythropoietin-stimulating agents and iron supplements for renal anemia are generally reimbursed under national medical insurance schemes for end-stage kidney disease patients, with high coverage rates minimizing major out-of-pocket burdens. The high proportion of patients outside the hemoglobin target range (53.0% in our study) is thus more likely related to clinical factors such as inflammation, iron deficiency, erythropoietin resistance, and variations in dialysis adequacy rather than primary access limitations.

Psychosocial factors accounted for a substantial portion of the variance in activation scores, consistent with theoretical frameworks that conceptualize activation as influenced by behavioral, cognitive, and social domains [21, 13]. Among these, self-management (HD-SMI) had the strongest association, highlighting the centrality of practical health management skills in driving activation. Hope, social support, and self-efficacy also significantly contributed to activation levels. These findings underscore the multifactorial nature of activation. Given these results, interventions aiming to boost activation should be multidimensional—incorporating educational support, psychological counseling, social networking opportunities, and tailored self-management training. Programs designed with sensitivity to patients’ literacy levels, functional capacities, and psychosocial contexts are likely to yield better outcomes. Although the psychosocial scales used in this study are interrelated, they assess distinct constructs. PAM-13 represents a broad continuum of knowledge, skills, beliefs, and confidence for proactive health management. SECD6 is narrower, focusing on task-specific confidence; HD-SMI evaluates enacted hemodialysis-specific behaviors (partnership, self-care, problem-solving, and emotional management); HHI measures future-oriented positivity and interconnectedness; and SSRS captures perceived external resources. The observed associations highlight how these complementary factors—internal confidence/motivation, behavioral enactment, emotional resilience, and interpersonal support—collectively foster greater patient activation in this population.

However, this study has several limitations that should be considered when interpreting the findings. First, the cross-sectional design precludes any causal inferences regarding the relationships between patient activation and the identified factors. Second, convenience sampling from only four tertiary hospitals in a single city (Taiyuan) limits the external validity and generalizability of the results. The sample may not be representative of MHD patients in other regions of China, particularly those in rural areas or smaller hospitals. Additionally, the exclusion criteria (e.g., severe comorbidities, communication impairments, and cognitive issues) likely resulted in the exclusion of more vulnerable and complex patients, further restricting the applicability of our findings. Third, although we used univariate screening (p < 0.05) to select variables for the multiple regression model to maintain parsimony given our sample size, this approach carries the inherent risk of omitting important confounders that may not reach statistical significance in univariate tests. To address this, we conducted sensitivity analyses by forcing clinically relevant variables (such as age, sex, and dialysis vintage) into the model; the primary findings remained robust. We recommend that future studies with larger samples consider a priori inclusion of all theoretically important confounders or alternative selection methods (e.g., change-in-estimate criteria or penalized regression). Fourth, although we performed standard regression diagnostics (including checks for multicollinearity, linearity, normality of residuals, homoscedasticity, and influential cases), some degree of model uncertainty remains. Fifth, some conceptual overlap may exist between PAM-13 and psychosocial constructs such as self-efficacy and self-management. Although these instruments are theoretically distinct and assess complementary domains, partial overlap in confidence- and behavior-related content may have inflated the observed associations. Therefore, the relationships identified should be interpreted with appropriate caution. Finally, we did not formally assess cognitive function, which is highly prevalent in MHD patients and could have influenced questionnaire responses. Future multicenter, longitudinal studies with random sampling, more comprehensive variable inclusion, and cognitive screening are warranted to confirm and expand upon these results.

In summary, our findings suggest that patient activation within this surveyed cohort of MHD patients is shaped by an interplay of educational, clinical, and psychosocial factors. Future longitudinal and interventional studies are needed to validate causal pathways and test the effectiveness of activation-enhancing strategies tailored to this population.

Author contributions

X.L.: Guarantor of integrity of the entire study, Study Concepts, Study Design, Definition of Intellectual Content, Clinical Studies, Experimental Studies, Statistical Analysis, Manuscript Preparation, Manuscript Editing, Manuscript Review.M.N.S.: Literature Research.Y.Z.: Data Acquisition.G.Z.: Data Acquisition.Z.M.Y.: Data Analysis.All authors reviewed and approved the final manuscript.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Data availability

The data that support the findings of this study are not publicly available due to privacy reasons but are available from the corresponding author upon request.

Declarations

Ethical approval

Approval for the study was obtained from ethics committee of Shanxi Provincial People’s Hospital [Date:2024.9.30; No.2024 − 810], and we adhered strictly to the principles of the Declaration of Helsinki and ensured that patient confidentiality was maintained throughout the research process.

Patient consent statement

All of the patients had consented to research authorization for record review, and the study was approved by the institutional review board.

Competing interests

The authors declare no competing interests.