Health behaviour abilities profiles in patients with haematological malignancies and its influencing factors: a cross-sectional survey.

STRENGTHS AND LIMITATIONS OF THIS STUDY

The use of latent profile analysis combined with multinomial logistic regression provides a robust, person-centred methodological approach.

Sampling across multiple distinct primary diagnoses ensures a methodologically comprehensive representation of haematological malignancies.

Data collection relied entirely on online self-reported questionnaires, introducing inherent response and recall biases.

Due to the cross-sectional design, causal relationships between these factors and health behaviour abilities cannot be inferred.

Recruitment was restricted to hospitals within a single city, potentially limiting the generalisability of the findings.

Introduction

Haematological malignancies (HMs) comprise a heterogeneous group of aggressive neoplasms originating from the haematopoietic system, 1 primarily affecting the blood, bone marrow and lymph nodes. These malignancies encompass primary or secondary disorders such as leukaemia, lymphoma and multiple myeloma, and are clinically characterised by high malignancy, rapid progression and substantial therapeutic challenges. Globally, HMs rank as the fourth most prevalent class of cancer. 1 According to global cancer statistics in 2022, approximately 20 million new cancer cases and 9.7 million cancer-related deaths occurred worldwide; among these, HMs accounted for 6.5% of new diagnoses and 7.1% of mortality, indicating an escalating global burden that severely threatens patient health. 2 Although breakthroughs in targeted therapy and immunotherapy have significantly prolonged survival rates, patients with HMs still face multiple challenges, including financial toxicity driven by exorbitant drug costs, 3 prolonged treatment cycles and psychological distress associated with disease relapse. 4 In this context, proactive self-care through health behaviours such as medication adherence, symptom monitoring and nutritional management can effectively optimise health outcomes and prevent adverse events. 5

According to the Health Action Process Approach theory, health behaviours are heavily governed by an individual’s psychological drive and coping planning. 6 In clinical oncology settings, these underlying motivational dynamics are directly manifested through a patient’s health behaviour abilities, which reflect an individual’s self-rated capability to implement specific actions (eg, treatment adherence or symptom management) when encountering distinct obstacles. Crucially, robust health behaviour abilities are well-documented as a critical determinant of superior quality of life and reduced symptom burden in cancer survivors. 7 Patients with higher abilities beliefs demonstrate greater adherence to physical exercise and self-care regimens, directly mitigating cancer-related fatigue and enhancing physical functioning. 8 For HM patients undergoing intensive treatments like stem cell transplantation, strong health behaviour abilities are essential to overcome severe treatment-induced physical deconditioning, accelerate haematological recovery and preserve muscle mass, 9 serving as a fundamental psychological cornerstone for long-term survivorship. 10

The health behaviour abilities are associated with multiple factors, involving four dimensions: patient characteristics, disease treatment, family environment and social context. Within the dimension of patient characteristics, attributes such as educational level 11 and sense of coherence (SOC) 12 are relevant. In the disease treatment dimension, key variables include disease duration, 13 14 comorbidities, 14 etc. The family environment dimension encompasses monthly household income, 15 place of residence and other factors. The social dimension involves social support, 16 insurance type and similar elements. Exploring the influencing factors of health behaviour abilities in haematological tumour patients can provide evidence for constructing targeted intervention strategies, thereby enhancing patients’ long-term health management efficacy.

However, current research on health behaviour abilities across diverse cancer populations, including HMs, still faces significant limitations: on one hand, most existing studies rely heavily on aggregate scale scores to assess overall abilities levels, thereby failing to fully capture the heterogeneous characteristics within specific populations. 6 For example, Hsia et al 6 evaluated health-promoting lifestyle in female breast cancer survivors, while Pitt et al 17 studied health behaviour abilities in broad paediatric oncology groups; both lines of research, however, overlook the unique disease-specific trajectories inherent to HMs. On the other hand, systematic research examining the comprehensive influencing factors of health behaviour abilities remains fragmented and largely focused on non-haematological subgroups. 6 For instance, Brouwer-Goossensen et al 18 investigated the determinants of efficacy changes specifically in brain tumour patients. Consequently, the extent to which these findings apply to HM populations remains unclear, highlighting a pronounced scarcity of disease-specific empirical evidence for this distinct patient group.

To address such intra-population heterogeneity, an increasing number of published studies have recently used latent profile analysis (LPA) to capture the distinct patterns of health-promoting behaviours and behavioural abilities across various oncological settings. For instance, in gastrointestinal cancer survivorship, recent multicentre evidence successfully identified potential health-promoting behavioural typologies in patients following oesophageal cancer surgery, uncovering boundary conditions driven by symptom distress. 19 Similarly, investigator-led studies on patients with colorectal cancer applied this person-centred approach to uncover heterogeneous patterns of health-promoting lifestyles, demonstrating that sociodemographic factors heavily dictated sub-optimal profile memberships. 20 Within the specific domain of HMs, emerging evidence has also confirmed that patients undergoing intensive chemotherapy exhibit substantial inter-individual variations in their self-care and behavioural capabilities. 21

Building on these methodological trends, LPA 22 can identify latent profiles based on multidimensional indicators and reveal the categorical characteristics within groups and their specific influencing factors. Therefore, this study intends to use the LPA method to deeply explore different categories of health behaviour abilities among HM patients and their population characteristics, and systematically analyse the influencing factors. The research results will provide a theoretical basis for clinical nurses to develop precise and individualised intervention strategies, which is of great practical significance for improving the health behaviour management and long-term prognosis of HM patients.

Materials and methods

Participants

This study employed a cross-sectional design. Between October and December 2024, patients hospitalised in the haematology-oncology departments of three provincial hospitals in Hunan Province, China, were recruited using convenience sampling. The inclusion criteria were (1) pathologically diagnosed with HMs; (2) currently hospitalised and undergoing treatment; (3) aged ≥18 years; (4) possessed adequate communication and cognitive skills to complete the survey. The exclusion criteria were (1) cognitive impairment or diagnosed with mental disorders; and (2) complicated with diseases of vital organs such as the heart, brain and kidneys.

Measurements

The selection and structural framework of the measurement variables in this study were systematically guided by the Social Ecological Model. 23 This theoretical approach conceptualises health behaviour abilities as a multifaceted outcome driven by multi-layered systemic influences. Consequently, the evaluated variables were structured into four hierarchical ecosystems: the intrapersonal level (such as age, gender, SOC), the microsystem level (such as residence, monthly family income), the exosystem level (such as clinical diagnosis, treatment stage) and the macrosystem level (such as social support, insurance).

These measured dimensions, along with their specific variables and instruments, are elaborated below and summarised in table 1 .

General information questionnaire

It includes 14 variables: age, gender, marital status, educational level, occupation, residence, monthly family income, clinical diagnosis, chronic diseases, disease course, treatment stage, relapse, insurance and complications.

Sense of coherence

The Sense of Coherence Scale-13 (SOC-13), designed and developed by Antonovsky 12 and translated by Bao and Liu. 24 The scale consists of 13 items across three dimensions: the comprehensibility dimension (five items), the manageability dimension (four items) and the meaningfulness dimension (four items). A 7-point Likert scale is employed for scoring, ranging from ‘never’ (coded as 1 point) to ‘frequently’ (coded as 7 points). Total scores on the scale range from 13 to 91, with higher scores indicating a higher level of SOC. Scores between 13 and 63 indicate a low level, 64 and 79 a moderate level and 80 and 91 a high level. The scale demonstrates a Cronbach’s α of 0.780.

Social support

We measured perceived social support using the 12-item Perceived Social Support Scale, originally developed by Zimet et al 25 and translated by Jiang et al. 26 The tool assesses the degree of support an individual perceives from family, friends and significant others. Each item is rated on a 7-point scale (1=strongly disagree to 7=strongly agree), yielding a total score between 12 and 84, the higher the score, the more social support the person perceives. According to the total score, support levels are categorised into three tiers: low (≤36), moderate (37–60) and high (61–84). In our study, the scale achieved a Cronbach’s α of 0.90.

Financial toxicity

Developed by De Souza et al 3 and translated by Yu et al , 27 this scale is primarily used to assess the perceived financial toxicity of research participants over the past 7 days. The scale comprises three dimensions: active economic expenditure, passive economic resources and psychosocial responses, with a total of 11 items. A 5-point Likert scale is employed, ranging from ‘not at all’ to ‘very much’, scored 0–4. Items 1, 6, 7 and 11 are positively scored, while the remaining items are reverse-scored. The total score ranges from 0 to 44. A total score >26 indicates no financial toxicity, below 26 indicates low financial toxicity, below 14 indicates medium financial toxicity and 0 indicates high financial toxicity. The scale demonstrates a Cronbach’s α of 0.893

Psychological distress

Psychological distress was measured using the Patient Health Questionnaire-4 (PHQ-4) developed by Kroenke et al. 28 The Chinese version of the PHQ-4 was adapted from the validated Chinese versions of the PHQ-9 and Generalised Anxiety Disorder-7 and has demonstrated good psychometric properties in Chinese populations. 29 It evaluates patients’ emotional states over the past 2 weeks. Each item is rated on a 4-point scale: ‘not at all’ (0), ‘several days’ (1), ‘more than half the days’ (2) and ‘nearly every day’ (3), with total scores ranging from 0 to 12. Scores are interpreted as follows: 0–2 indicates no psychological distress, 3–5 indicates low psychological distress, 6–8 indicates medium psychological distress and 9–12 indicates high psychological distress. The scale demonstrated a Cronbach’s α of 0.896.

Health behaviour abilities

The Self-rated Abilities for Health Practices Scale 7 developed by Becker et al and translated by Hu and Zhou, 30 was used to measure participants’ abilities for health behaviours. The scale encompasses 28 items across four dimensions: nutrition, exercise, psychological adjustment and health responsibility. Each item is scored on a 5-point Likert scale ranging from ‘almost no confidence’ (0) to ‘absolute confidence’ (4), yielding a total score range of 0–112. Higher scores indicate greater abilities in performing health behaviours. The overall Cronbach’s α for the scale was 0.975, with subscale coefficients ranging from 0.902 to 0.953.

Data collection and quality control

After obtaining the consent of the hospital’s nursing department, data were collected by researchers and two trained research assistants. Before the investigation, a unified guideline was used to explain the purpose, significance, content and precautions of this study to the research subjects. All research subjects were required to independently complete the questionnaire anonymously within the prescribed time, and the patients filled it out by themselves. For the research subjects who had difficulty filling in the form, the investigators filled it out orally and based on the patients’ responses on their behalf. After filling out the questionnaire, check it on the spot. If any errors or omissions are found, return it promptly for filling and correction. A total of 500 questionnaires were distributed in this study, and 492 questionnaires were retrieved. Among them, 23 questionnaires were filled out by the researcher on behalf of others. After double-person verification, invalid questionnaires were excluded (four were answered regularly and two had an answering time of less than 5 min), and 486 valid questionnaires were retrieved, with an effective recovery rate of 98.78%.

Statistical method

Potential profile analysis was conducted using Mplus V.8.0 software. Taking the scores of the four dimensions of the Health Behaviour Abilities Rating Scale (nutrition, exercise, psychological adjustment and health responsibility) as explicit variables, the exploration started from one category and went up to four categories for analysis. The fitting indices of each item were compared and the best category model was selected in combination with practical significance, and the potential profile was drawn. The specific fitting indicators of the evaluation model are as follows: (1) information evaluation indicators: Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC) and the sample corrected BIC (adjusted Bayesian Information Criterion, aBIC); the smaller the statistical value, the better the model fit; (2) classification evaluation index: Entropy reflects the classification accuracy, with a value range of 0–1. The closer it is to 1, the more accurate the model classification is. (3) Likelihood ratio test: The Lo-Mendel-Rubin (LMR) and Bootstrap-based strapped likelihood ratio test (BLRT) were used as the comparison statistics of the model. If p<0.05, it indicates that the model of this category is superior to that of the previous category. The above evaluation indicators are for reference only. When determining the best model, the interpretability of each category should also be taken into consideration. Data were statistically analysed using SPSS V.28.0 software. Measurement data were described by x±s and M (Q1, Q3), and count data were described by frequency and rate. χ 2 test or rank-sum test was used to compare the general information and scores of each scale of patients with different health behaviour abilities categories, and indicators with statistically significant differences between categories were screened out. And logistic regression was used to analyse the predictive indicators of different categories. A two-tailed p value <0.05 was considered statistically significant.

Sample size

For sample size estimation, the events-per-variable criterion for multivariable logistic regression analysis was applied, which dictates a minimum ratio of 10–15 cases per independent variable. Given that 18 independent variables were examined in the current study, and accommodating an anticipated 20% invalid questionnaire rate, the minimum required sample size was determined to be 324 cases. Ultimately, a larger sample of 486 participants was achieved, satisfying the statistical power requirements.

Patient and public involvement

Patients and/or the public were not involved in the design, conduct, reporting or dissemination of this research.

Results

General information

A total of 486 patients completed the survey, among whom approximately half (49.6%) were aged between 40 and 59. The other demographic characteristics are shown in table 2 . The average score of health behaviours abilities in HMs patients was 47.96±23.29 points. The average scores of the four dimensions of nutrition, exercise, psychological adjustment and health responsibility were 11.68±6.36, 12.35±7.16, 9.29±7.78 and 14.64±6.40, respectively, as shown in table 3 .

Latent profile analysis

Taking the four dimensions of health behaviour abilities of HMs patients as explicit indicators respectively, a total of four models were fitted. The fitting indicators of each model are shown in table 4 . With the increase of categories, the values of AIC, BIC and aBIC are constantly shrinking. The entropy value is the largest in the three-category model, followed by the four-category model. When the fitting category is 4, both the LMR (p) and BLRT (p) values are >0.05, indicating that the fitting degree of this type is poor. For the fitting index results of category 3, the LMR and BLRT (p) values are <0.05. Therefore, it can be considered that category 3 is the best-fitting model. Based on the classification of the potential profile of category 3, the distribution characteristics of the health behaviour abilities of patients with HMs in each dimension are drawn as shown in figure 1 . The scores of each section were 26.25, 55.15 and 887.62, respectively. According to the cross-sections, it can be found that the motion dimensions of the following two sections are at a relatively low level. Combined with the assignment of health behaviour abilities, each section was named ‘exercise-weak survival profile’ (203 cases, 41%), ‘exercise-deficient developmental profile’ (215 cases, 45%) and ‘balanced robust profile’ (68 cases, 14%). The total scores and dimension scores of different profiles are shown in table 3 .

Univariate analysis

Univariate analysis was used to compare the general data of patients with HMs in the three categories between groups. The results show that there are statistically significant differences among the three categories of health behaviour abilities in terms of age, educational level, occupation, residence, monthly family income, clinical diagnosis, SOC, social support and financial toxicity. The results are presented in table 2 .

Multivariate analysis

Taking the variables with statistically significant differences in the univariate analysis as the independent variables (there was no multicollinearity among the independent variables, and the variance inflation factor value <5), and taking the three categories of health behaviour abilities determined by the potential profile analysis as the dependent variables.

When choosing the ‘balanced-robust profile’ (profile 3) as the reference category, patients with lower educational attainment (junior high school and below) and lower levels of social support were significantly more likely to be classified into the ‘exercise-weak survival profile’ (profile 1). Conversely, having no financial toxicity or low financial toxicity, and experiencing no psychological distress or low psychological distress, were associated with a heavily reduced likelihood of belonging to profile 1.

In the comparison between ‘exercise-deficient developmental profile’ (profile 2) and profile 3, patients with a junior high school education were more likely to be members of profile 2 rather than profile 3. Interestingly, younger age, absence of financial toxicity or low financial toxicity significantly decreased the probability of belonging to profile 2.

When directly comparing profile 1 to profile 2 (with profile 2 as the reference), younger patients (0–19) and older patients (60–79) both demonstrated a lower likelihood of belonging to profile 1, meaning they were more frequently assigned to profile 2. Furthermore, lower educational level was associated with a reduced probability of belonging to profile 1 over profile 2, whereas individuals with no psychological distress or low psychological distress were substantially more likely to be classified into profile 1 than profile 2 ( table 5 ).

Discussion

The results of this study showed that the health behaviours abilities score among 486 haematological malignant patients was 47.96±23.29, lower than the mean score of breast cancer patients reported in the study by Hsia et al 6 (90.40±17.98). This may be related to the inherent complexity of HMs (such as long-term chemotherapy and repeated hospitalisations), treatment-related side effects (such as fatigue and infection risk) and prognostic uncertainty, which collectively undermine patients’ confidence in self-management. 31 Dimension-specific analysis revealed the lowest score in the psychological adjustment dimension (9.29±7.78), which aligns with the finding that 69.34% of patients experienced psychological distress in this study. This suggests that psychological stress induced by the disease may weaken patients’ abilities in emotional management and stress coping. 32 The health responsibility dimension achieved the highest score (14.64±6.40), indicating that patients generally possess the awareness to proactively seek disease-related information and medical support. However, the effectiveness of translating this awareness into actual health behaviours requires further improvement.

In this study, the health behaviour abilities of haematological tumour patients were classified into three categories: ‘exercise-weak survival profile’ (profile 1), ‘exercise-deficient developmental profile’ (profile 2) and ‘balanced-robust profile’ (profile 3), indicating significant group heterogeneity. Among them, profile 1 accounted for 41%. Patients in this category had significantly lower scores in all dimensions (especially the exercise dimension), which may be related to immune suppression associated with haematological tumour treatment, physical activity limitations due to repeated hospitalisations, and psychological avoidance of infection risks. Their health behaviour management mostly focuses on disease survival needs. 33 Profile 2 accounted for 45%. Although patients in this category had moderate scores in the nutrition, psychological comfort and health responsibility dimensions, similar to profile 1, their exercise dimension was still limited. This may be because although such patients have gradually established health management awareness, they have failed to achieve behavioural transformation due to cancer-related fatigue or cognitive biases about exercise risks. 34 Profile 3 accounted for 14%. Patients in this category had better overall health behaviour abilities, but their nutrition dimension was relatively weak, possibly influenced by factors such as gastrointestinal reactions to chemotherapy, dietary restrictions and lack of nutritional knowledge. Based on this, it is recommended to implement stratified interventions in clinical practice, carry out targeted interventions for different categories of patients and establish interdisciplinary teams to monitor dynamic behavioural changes during the treatment-rehabilitation period.

Regarding age, our findings revealed nuanced behavioural patterns across different developmental stages. When directly comparing profile 1 to profile 2, both younger patients (0–19 years) and older individuals (60–79 years) were more frequently assigned to profile 2. This suggests that age-extreme groups, while not yet achieving optimal abilities, still possess a degree of adaptive potential and developmental space rather than completely abandoning physical activity. 6 Interestingly, in the comparison between profile 2 and profile 3, patients aged 0–19 years showed a powerful propensity towards profile 3. This remarkable resilience among youth might be driven by the subgroup of older adolescents (13–19 years) who possess growing autonomy and better digital/health literacy to actively acquire self-management knowledge, 35 facilitating a direct leap to the robust profile when properly supported.

Regarding educational attainment, our findings demonstrated that lower educational levels were powerful predictors of suboptimal behavioural profiles. Specifically, when using profile 3 as the reference, patients with junior high school or below education exhibited a higher likelihood of being classified into profile 1. This heavily aligns with Yeom’s study, 36 which links lower education to decreased health abilities. Individuals with lower educational levels often face profound challenges in actively retrieving and comprehending complex exercise rehabilitation guidelines, causing their health cognitions to remain superficially restricted to ‘prioritising treatment’ while neglecting exercise benefits. 11 Furthermore, when facing professional health guidance, a lack of foundational health literacy impedes their ability to translate medical information into concrete physical actions, ultimately resulting in low exercise compliance and compromised abilities. 37 Interestingly, when directly comparing the two non-optimal profiles (with profile 2 as the reference), lower-educated patients were less likely to belong to profile 1 than profile 2, suggesting that while lower education hinders patients from achieving the optimal robust status (profile 3), they still lean towards a developmental trajectory (profile 2) rather than complete behavioural abandonment (profile 1).

Regarding psychosocial determinants, social support emerged as a powerful protective factor for cultivating optimal behaviour. When using profile 3 as the reference category, patients with low social support and moderate social support were both significantly more likely to be classified into profile 1. This underscores that insufficient or unstable social support is a critical risk factor that undermines patients’ health behaviour abilities. Although patients with moderate support receive certain resources, such assistance may lack continuity (eg, occasional care from relatives) or fail to meet their actual needs, which is inadequate to fully alleviate psychological distress or resolve practical barriers in disease management. 16 Consequently, only a high level of continuous social support can provide patients with sufficient emotional and instrumental backing, thereby fostering their confidence and sustaining a robust exercise abilities profile during recovery.

Regarding financial toxicity, our study revealed it as a powerful determinant of latent profile membership. When choosing profile 3 as the reference category, patients with high financial toxicity were significantly more likely to be classified into the non-optimal profiles, as evidenced by the finding that the absence of financial toxicity heavily reduced the likelihood of belonging to profile 1 and profile 2. This confirms that high financial toxicity serves as a major barrier to robust health behaviour abilities. Chronic financial stress from cancer treatment directly restricts patients’ ability to access essential supportive care and medical resources. 3

Psychological distress manifested a profound and nuanced impact on patients’ behavioural profiles. Compared with the optimal profile 3, the absence of psychological distress substantially decreased the risk of membership in profile 1, indicating that severe psychological distress heavily paralyses patients’ abilities. 38 Interestingly, when directly comparing profile 1 to profile 2, patients with no distress or low distress were more frequently assigned to profile 1 rather than profile 2. This novel insight suggests that a complete lack of psychological distress may conversely lead to behavioural complacency. Without a moderate level of health-related anxiety to serve as a catalyst, patients might lack the proactive motivation required to progress from profile 1 towards profile 2. 38

The findings of this study offer meaningful implications for both oncology nursing practice and future research in HMs. In clinical practice, healthcare providers should transition from a ‘one-size-fits-all’ education model to risk-stratified, targeted interventions based on patients’ latent profiles. Given that lower educational attainment and poor social support are potent predictors for dropping into profile 1, clinical nurses should prioritise these vulnerable sub-populations by providing easily understandable, low-literacy health materials and actively mobilising family or community support systems. For future research, our results underscore the necessity of developing and testing personalised digital health interventions, such as interactive decision aids or tailored smartphone applications, that dynamically adapt to patients’ psychological states and financial barriers. Furthermore, subsequent longitudinal trials are warranted to track the fluctuations of these behavioural profiles across different treatment phases, and future studies should incorporate more granular age stratifications (young, middle-aged and older adults) to uncover stage-specific behavioural determinants during cancer survivorship.

Limitations

This study has several limitations. First, this study only investigated patients from provincial hospitals in Changsha City, which might introduce regional selection bias and limit the generalisability of our findings to patients in less-developed municipal or rural county-level medical settings. Second, the age categorisation in our current dataset lacked sufficient granular distinction; subsequent research should consider dividing age into young, middle-aged and older adult groups to explore age-specific determinants more comprehensively and provide precise evidence for tailored interventions. Third, the evaluation of health behaviour abilities and psychosocial variables relied exclusively on online self-reported questionnaires, which are inherently susceptible to response bias or recall bias. Finally, owing to the inherently cross-sectional design of this study, causal relationships between the sociodemographic or psychosocial factors and health behaviour abilities cannot be definitively inferred.

Conclusion

This study found that the health behaviour abilities in HMs patients are not high, which can be divided into three potential profiles. Patients of different categories have differences in age, educational level, SOC, educational level, financial toxicity and psychological distress status. It is suggested that when formulating intervention measures, nursing staff should implement precise health management based on the potential categories to which the patients belong: for elderly and low-educated patients, the popularisation and education of health knowledge need to be strengthened; psychological counselling and economic assistance should be integrated for patients with high economic toxicity or psychological distress. For those with a lower SOC, attention should be paid to enhancing their confidence in coping with the disease.

Acknowledgements

We sincerely thank all the patients who participated in the questionnaire survey.

Funding: This work was supported by the Natural Science Foundation of Hunan Province (grant numbers: 2024JJ9236 and 2025JJ60535), Xiangya Joint Fund for Early-Career Nursing Faculty (grant number 2025XYHLQNLHJJ07).

Prepublication history for this paper is available online. To view these files, please visit the journal online ( https://doi.org/10.1136/bmjopen-2025-110342 ).

Data availability free text: The data that support the findings of this study are available from the corresponding author on reasonable request.

Patient consent for publication: Consent obtained directly from patient(s).

Ethics approval: This study was approved by the Ethics Committee of the Third Xiangya Hospital, Central South University (Ethics Review and Approval No: 2024-S425) (xy3irb@163.com). Informed consent was obtained from all participants before data collection, and the clinical data and basic information of all participants were strictly confidential. All procedures complied with the Declaration of Helsinki.

Provenance and peer review: Not commissioned; externally peer reviewed.

Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

References

1 Taylor J Xiao W Abdel-Wahab O Diagnosis and classification of hematologic malignancies on the basis of genetics Blood 2017 130 410 23 10.1182/blood-2017-02-734541 28600336 PMC5533199

2 Bray F Laversanne M Sung H et al Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries CA Cancer J Clin 2024 74 229 63 10.3322/caac.21834 38572751

3 De Souza JA Aschebrook-Kilfoy B Grogan R et al Grading financial toxicity based upon its impact on health-related quality of life (HRQol) JCO 2016 34 16 10.1200/jco.2016.34.3_suppl.16

4 Kuczmarski TM Roemer L Odejide OO Depression in patients with hematologic malignancies: The current landscape and future directions Blood Rev 2024 65 101182 10.1016/j.blre.2024.101182 38402023

5 Knips L Bergenthal N Streckmann F et al Aerobic physical exercise for adult patients with haematological malignancies Cochrane Database Syst Rev 2019 1 CD009075 10.1002/14651858.CD009075.pub3 30702150 PMC6354325

6 Hsia H-H Tien Y Lin Y-C et al Factors Influencing Health Promotion Lifestyle in Female Breast Cancer Survivors: The Role of Health Behavior Self-Efficacy and Associated Factors Semin Oncol Nurs 2024 40 151622 10.1016/j.soncn.2024.151622 38522967

7 Ma Q Luo J Cao H et al Social support, health behavior self-efficacy, and anxiety on physical activity levels among lung cancer survivors: a structural equation modeling J Cancer Surviv 2026 20 1 11 10.1007/s11764-024-01626-y 38865009

8 Jackson ES Tucker CM Herman KC Health value, perceived social support, and health self-efficacy as factors in a health-promoting lifestyle J Am Coll Health 2007 56 69 74 10.3200/JACH.56.1.69-74 17711829

9 Pinto BM Papandonatos GD Goldstein MG et al Home-based physical activity intervention for colorectal cancer survivors Psychooncology 2013 22 54 64 10.1002/pon.2047 21905158

10 Mishra SI Scherer RW Geigle PM et al Exercise interventions on health-related quality of life for cancer survivors Cochrane Database Syst Rev 2012 2012 CD007566 10.1002/14651858.CD007566.pub2 22895961 PMC7387117

11 Bennardi M Diviani N Gamondi C et al Palliative care utilization in oncology and hemato-oncology: a systematic review of cognitive barriers and facilitators from the perspective of healthcare professionals, adult patients, and their families BMC Palliat Care 2020 19 47 10.1186/s12904-020-00556-7 32284064 PMC7155286

12 Antonovsky A The structure and properties of the sense of coherence scale Soc Sci Med 1993 36 725 33 10.1016/0277-9536(93)90033-Z 8480217

13 Rodríguez-García E Barnes-Ortiz S Pérez-Mármol JM Self-Efficacy, Pain Intensity, Rheumatic Disease Duration, and Hand Functional Disability on Activities of Daily Living Nurs Res 2020 69 E208 16 10.1097/NNR.0000000000000466 32804710

14 Nuraeni A Sugiharto F Anna A et al Self-Efficacy in Self-Care and Its Related Factors Among Patients with Coronary Heart Disease in Indonesia: A Rasch Analysis Vasc Health Risk Manag 2023 19 583 93 10.2147/VHRM.S427488 37691747 PMC10492567

15 Ali R Draman N Mohd Yusoff SS et al Self-Efficacy for Coping with Breast Cancer in North-Eastern State of Peninsular Malaysia Asian Pac J Cancer Prev 2020 21 2971 8 10.31557/APJCP.2020.21.10.2971 33112556 PMC7798160

16 Kiajamali M Hosseini M Estebsari F et al Correlation between social support, self-efficacy and health-promoting behavior in hemodialysis patients hospitalized in Karaj in 2015 Electron Physician 2017 9 4820 7 10.19082/4820 28894541 PMC5586999

17 Pitt E Cashion C Rumble S et al Associations between Health Behaviors, Health Self-Efficacy, and Long-Term Outcomes in Survivors of Childhood Cancer: A Cross-Sectional Study Semin Oncol Nurs 2023 39 151434 10.1016/j.soncn.2023.151434 37147151

18 Brouwer-Goossensen D van Genugten L Lingsma HF et al Self-efficacy for health-related behaviour change in patients with TIA or minor ischemic stroke Psychol Health 2018 33 1490 501 10.1080/08870446.2018.1508686 30596258

19 Wang J Yu Y Liu L et al Latent profile analysis of potential health-promoting behaviors in patients after esophageal cancer surgery: a multicenter cross-sectional survey Support Care Cancer 2026 34 10.1007/s00520-025-10249-7 PMC12696073 41369798

20 Zhong M Wu X Xie J et al Patterns of health-promoting lifestyles among patients with colorectal cancer: A latent profile analysis Asia Pac J Oncol Nurs 2026 13 100957 10.1016/j.apjon.2026.100957 42180173 PMC13193774

21 Ma G Long N Mao P et al A latent profile analysis of financial toxicity and its association with anxiety and depression in patients with hematologic malignancies Asia Pac J Oncol Nurs 2025 12 100761 10.1016/j.apjon.2025.100761 40896752 PMC12391807

22 Shi B Zhang J Wang H et al Latent Profile Analysis and Related Factors of Colorectal Cancer Knowledge and Health Beliefs and Their Associations With Screening Behavior and Intention Among Urban Populations in China Cancer Nurs 2026 49 129 38 10.1097/NCC.0000000000001479 39982942

23 Stokols D Translating social ecological theory into guidelines for community health promotion Am J Health Promot 1996 10 282 98 10.4278/0890-1171-10.4.282 10159709

24 Bao L Liu J The Reliability and Validity of Chinese Version of SOC-13 Chinese Journal of Clinical Psychology 2005 399 401 10.1176/appi.ajgp.13.7.581

25 Zimet GD Powell SS Farley GK et al Psychometric characteristics of the Multidimensional Scale of Perceived Social Support J Pers Assess 1990 55 610 7 10.1080/00223891.1990.9674095 2280326

26 Jiang QJ Perceived social support scale Chin J Behav Med Sci 2001 10 41 3

27 Yu HH Bi X Liu YY Reliability and validity of the Chinese version on Comprehensive Scores for Financial Toxicity based on the patient-reported outcome measures Zhonghua Liu Xing Bing Xue Za Zhi 2017 38 1118 20 10.3760/cma.j.issn.0254-6450.2017.08.024 28847066

28 Kroenke K Spitzer RL Williams JBW et al Patient health questionnaire-4 BMC Psychiatry 2002

29 Meng R Jiang C Dzierzewski JM et al Longitudinal measurement invariance and psychometric properties of the Patient Health Questionnaire-Four in China BMC Psychiatry 2024 24 517 10.1186/s12888-024-05873-2 39039478 PMC11265176

30 Hu Y Zhou L A Study on the Reliability and Validity of the Chinese Version of the Health Behavior Ability Self-Assessment Scale Chinese Journal of Nursing 2012 47 261 2

31 Yang H Wu B Hu R et al Symptom experiences and self-management strategies of patients with haematological malignancy undergoing chemotherapy: A qualitative study Asia Pac J Oncol Nurs 2024 11 100563 10.1016/j.apjon.2024.100563 39239475 PMC11374969

32 Solberg Nes L Ehlers SL Patten CA et al Self-regulatory fatigue in hematologic malignancies: impact on quality of life, coping, and adherence to medical recommendations Int J Behav Med 2013 20 13 21 10.1007/s12529-011-9194-1 21928059 PMC4454389

33 Kansara RR Speziali C Immunotherapy in hematologic malignancies Curr Oncol 2020 27 S124 31 10.3747/co.27.5117 32368182 PMC7193997

34 Abel GA Klepin HD Frailty and the management of hematologic malignancies Blood 2018 131 515 24 10.1182/blood-2017-09-746420 29141942

35 Olsen JA Chen G Lamu AN The relative importance of education and health behaviour for health and wellbeing BMC Public Health 2023 23 1981 10.1186/s12889-023-16943-7 37821861 PMC10568892

36 Yeom H-E Lee J The association of education level with autonomy support, self-efficacy and health behaviour in patients with cardiovascular risk factors J Clin Nurs 2022 31 1547 56 10.1111/jocn.16008 34453378

37 Vallerand JR Rhodes RE Walker GJ et al Understanding strength exercise intentions and behavior in hematologic cancer survivors: an analysis of the intention-behavior gap J Cancer Surviv 2016 10 945 55 10.1007/s11764-016-0540-9 27052642

38 White LL Cohen MZ Berger AM et al Self-Efficacy for Management of Symptoms and Symptom Distress in Adults With Cancer: An Integrative Review Oncol Nurs Forum 2019 46 113 28 10.1188/19.ONF.113-128 30547965

Data availability statement

Data are available upon reasonable request.