Introduction
Malignant tumor is a significant public health problem threatening the physical and mental health of humans. According to the World Health Organization, in 2020, 19.3 million new cancer cases and almost 10 million cancer deaths were reported worldwide. By 2040, 28.4 million new cases of malignant tumors are expected, an increase of 47% over 2020 (Xia et al. Reference Xia, Dong and Li2022). Facing malignant diseases is one of the greatest challenges and pressures that patients face, and it can be a starting point of a life crisis that leads to a series of psychological problems (Wu et al. Reference Wu, Quan and Gao2021). Demoralization is one of the psychosocial problems of cancer, from diagnosis to treatment and gradual recovery (Bobevski et al. Reference Bobevski, Kissane and Vehling2022; Bovero et al. Reference Bovero, Botto and Adriano2019; Mishra Reference Mishra2023).
Demoralization is a psychological issue caused by a series of internal or external pressures (Wang et al. Reference Wang, Sun and Ji2023). It can be manifested as a sustained inability to cope with a stressful predicament, followed by a sense of helplessness, hopelessness, meaninglessness, subjective incompetence, and damaged self-respect, ranging from mild depression and depression to deep despair (Bockholt et al. Reference Bockholt, Mehnert-Theuerkauf and Vehling2021; Clarke and Kissane Reference Clarke and Kissane2002). Demoralization has become a common psychological problem among cancer patients (Airoldi et al. Reference Airoldi, Bovero and Botto2019). Studies have shown that the incidence of demoralization among cancer patients can be as high as 86.75%, and the average score of demoralization among cancer patients is higher than that in non-tumor ones (Gan et al. Reference Gan, Gong and Kissane2022; Nanni et al. Reference Nanni, Caruso and Travado2018). In addition, other systematic reviews have demonstrated that the prevalence of demoralization range from 20.6% to 94% in various diseases, including schizophrenic, progressive diseases (Costanza et al. Reference Costanza, Vasileios and Ambrosetti2022; Robinson et al. Reference Robinson, Kissane and Brooker2015; Tecuta et al. Reference Tecuta, Tomba and Grandi2015). Furthermore, demoralization is also strongly associated with several adverse outcomes, such as lower quality of life (QoL) and suicidal thinking and behavior (Chang et al. Reference Chang, Hung and Huang2022b; Costanza et al. Reference Costanza, Vasileios and Ambrosetti2022; Crespo et al. Reference Crespo, Rodriguez-Prat and Monforte-Royo2020; Robinson et al. Reference Robinson, Kissane and Brooker2017). Therefore, earlier identification of the demoralization of cancer patients by medical staff can more effectively improve the prognosis of patients and reduce the occurrence of adverse outcomes, such as depression and suicide.
A growing number of qualitative and quantitative studies have explored the factors associated with demoralization in cancer patients, but no consensus has been reached regarding the identification of relevant factors. For example, Belar et al. (Reference Belar, Arantzamendi and Rodríguez-Núñez2019) and Shao et al. (Reference Shao, Li and Lin2024) have reported disagreements about the relationship between religion and demoralization. The former researchers show that demoralization negatively correlates with religion, whereas the latter concludes the opposite. Moreover, no meta-analysis of studies related to demoralization in cancer patients can help systematically synthesize these data to provide strong evidence of the identifying factors associated with demoralization among cancer patients.
This meta-analysis aimed to explore the correlates of demoralization among cancer patients and provide recommendations for further research and policymaking.
Methods
This meta-analysis was conducted based on the Preferred Reporting Items of Systematic Reviews and Meta-Analyses statement. It was registered with PROSPERO on September 25, 2023, number CRD42023463663.
Search strategy
A systematic search was conducted from database inception to October 2023 across 8 databases: PubMed, Web of Science, CINAHL, Embase, the Cochrane Library, PsycINFO, CNKI, and Wan Fang. Reference lists of included articles and previous relevant systematic reviews were screened for secondary literature. The search terms were “cancer OR tumor OR neoplasms OR carcinoma” AND “demoralization OR demoralization syndrome.” In addition, we checked references in this review and previous relevant systematic reviews to identify additional studies. We excluded gray literature from the search because their diverse formats presented a challenge for the systematic search. Search strategies for all databases are described in Table S1.
Eligibility criteria
Studies were eligible if participants (age ≥18 years) were diagnosed with cancer and reported their state of demoralization and factors associated with demoralization (e.g., demographic variables). Eligible study types were longitudinal studies, case–control studies, and cross-sectional studies. Studies that combined noncancer patients or did not report demoralization were excluded. In addition, editorials and reviews were also excluded.
Data extraction and quality assessment
Two investigators independently performed the literature review and data extraction. The review process included title reviews, abstract reviews, and full-text reviews. If there was any disagreement between the 2 investigators after discussion, a third researcher was invited to decide whether to include the study. The items extracted and coded included the first author, publication time, country, study design, sample size, age, type of cancer, demoralization assessment tool, prevalence/score of demoralization, and associated factors.
Regarding the quality of the included articles, 2 investigators independently evaluated by using the cross-sectional Study Quality Assessment Scale recommended by the Agency for Healthcare Research and Quality (AHRQ) (Rostom et al. Reference Rostom, Dubé and Cranney2004) and the Newcastle–Ottawa Scale (NOS) (Stang Reference Stang2010). The AHRQ scale was adopted to assess the quality of the cross-sectional study, including 11 items. The score of each question ranges from 0 to 1, and the scale’s total score ranges from 0 to 11. Studies were classified into low (0–3), medium (4–7), and high (8–11) quality according to their scores. We used the NOS to evaluate the quality of the longitudinal study. The NOS includes 8 items, and the assessment is conducted in 3 dimensions; the total score ranges from 0 to 9. Studies were classified into low (0–3), medium (4–6), and high (7–9) evidence. The low-quality studies were finally excluded.
Data synthesis and statistical analysis
R4.1.1 software was used to perform a meta-analysis on the relevant factors of the included studies; p < 0.05 was considered statistically significant. We adopted the Pearson correlation coefficients (r) as an effect analysis statistic. For studies that do not provide an r value, other convertible statistics, such as Spearman’s correlation coefficient (rs), odds ratio (OR), or standardized regression coefficient (β), can be converted into an r value (Peterson and Brown Reference Peterson and Brown2005; Sarraf et al. Reference Sarraf, Lepage and Sauvé2022).
We used I2 test to quantify the heterogeneity and used subgroup analysis to find the possible sources of heterogeneity. When significant heterogeneity (I 2 > 50% and p < 0.05) was detected, a random-effect model was used; otherwise, using a fixed-effect model. Sensitivity analysis was adopted to assess the impact of each study on the pooled results. We used Egger’s test to measure the potential publication bias.
Results
Study selection
We identified 1303 eligible articles through a database search, with 734 remaining after removing 569 duplicates. Among them, 646 were excluded by their title and abstract evaluation. The full text of 88 articles were searched. After further reading the full text, 38 publications were finally included (Fig. 1).
Flow chart of study selection procedure.
Study characteristic
A total of 38 articles (n = 9639) were included in the meta-analysis. The 38 included articles were published between 2014 and 2023. Most studies were conducted in China (n = 23) (Chang, Huang, et al., Reference Chang, Huang and Hsu2022a; Chang, Hung, et al., Reference Chang, Hung and Huang2022b; Cheng et al. Reference Cheng, Chen and Zhang2019; Chien et al. Reference Chien, Pang and Chuang2022; Kang et al. Reference Kang, Liu and Jiang2023; Ko et al. Reference Ko, Lin and Pi2018; Lai et al. Reference Lai, Huang and Zhu2022; Li et al. Reference Li, Liu and Xu2020; Li et al. Reference Li, Feng and Ma2023; Li et al. Reference Li, Ho and Wang2016; Lin et al. Reference Lin, Hong and Lin2022; Liu et al. Reference Liu, Wu and Wang2020; Peng et al. Reference Peng, Hsueh and Chang2021; Shao et al. Reference Shao, Li and Lin2024; Tang et al. Reference Tang, Li and Pang2020; Tang et al. Reference Tang, Lin and Wang2022; Wu et al. Reference Wu, Quan and Gao2021; Xu et al., Reference Xu, Hu and Liu2019). The studies included a sample size of 78–874 participants aged 18–97 years. The most commonly used tool in the studies was the Demoralization Scale (DS). The most frequently reported correlate was depression. Overall, the study quality scores were between 6 and 10 points, indicating that the quality of studies was at a medium to high levels (Tables S2 and S3). The characteristics of all included studies are summarized in Table 1.
Meta-analysis results
Table 2 summarizes the meta-analysis results of all the reported correlates. Figures 2 and 3 respectively display the forest plots and Egger test plot of frequently reported correlates. The forest plots of variables reported in only 2 articles or with no statistical significance (Figures S1–S3), and the Egger test figures (Figure S4) are shown in the supplementary material.
Sociodemographic correlates
Three sociodemographic factors (income, education, and age) were found to be associated with demoralization. Results of the meta-analysis suggested that demoralization has a statistically negative correlation with income (z = −0.29, 95% CI: −0.51, −0.02; Fig. 2A), education (z = −0.11, 95% CI: −0.16, −0.05; Fig. 2B), and age (z = −0.45, −0.75, −0.01; Fig. 2C). These findings showed that the characteristics of the studies reported by income (I 2 = 97%, p < 0.01) and age (I 2 = 99%, p < 0.01) are highly heterogeneous, which may be attributed to the different tumor types and evaluation tools. Egger’s asymmetry test for income (t = −0.67, df = 4, p = 0.54; Fig. 3A), education (t = 1.51, df = 4, p = 0.207; Fig. 3B), and age (t = −0.21, df = 2, p = 0.85; Fig. 3C) suggested no evidence of funnel-plot asymmetry, indicating that there is no publication bias.
Clinical correlates
QoL was the most frequently reported correlate associated with demoralization, followed by symptom burden. Results of the meta-analysis suggested that demoralization has a statistically mild, negative correlation with QoL (z = −0.40, 95% CI: −0.54, −0.24; Fig. 2D). These findings demonstrated that the characteristics of the included 9 studies are strongly heterogeneous (I 2 = 94%, p < 0.01). Symptom burden (z = 0.37, 95% CI: 0.22, 0.50; Fig. 2E) had a small, positive correlation with demoralization. These findings demonstrated that the characteristics of the included studies are strongly heterogeneous (I 2 = 91%, p < 0.01).
Basic information of included studies
Note: DS = Demoralization Scale; DS‐II = Demoralization Scale II; DS‐6 = Demoralization Scale-short form; DCPR/D = Diagnostic Criteria for Psychosomatic Research‐Demoralization interview.
Correlates: 1. anxiety; 2. depression; 3. quality of life; 4. death anxiety; 5. symptom burden and bother; 6. coping style; 7. religious; 8.functional status; 9. spiritual interests; 10. suicidal ideation; 11. hopelessness; 12. age; 13. Income; 14. education level; 15. cancer stage (higher); 16. self-esteem; 17. residence (urban); 18. social support; 19. psychological resilience; 20. sleep quality; 21.self-efficacy; 22. relationship to health-care provider; 23.self-perceived burden; 24. spiritual well-being; 25. medical payment; 26. time of chemotherapy; 27. perceived cognitive impair; 28. illness intrusiveness; 29. at-work productivity loss; 30. psychological distress; 31. self-compassion; 32. having at least 1 child; 33. stress; 34. marital status; 35. treatment type; 36. psychological well-being; 37. source of income; 38. financial toxicity; 39. positive life orientation; 40. empirical avoidance; 41. gender; 42. the researchers’ perceptions about the demoralization; 43. perceived relatedness; 44. anxious attachment; 45. avoidant attachment; 46. life completion; 47.personality traits; 48. cancer type; 49. surgery type; 50. time of radiotherapy; 51. Hope.
Factors correlated with demoralization among cancer patients
Note: n = number of studies; n. s. = statistically nonsignificant; n. a. = not applicable due to the number of studies included.
In addition, Egger’s asymmetry test for QoL (t = −1.48, df = 7, p = 0.18; Fig. 3D) and symptom burden (t = 0.64, df = 5, p = 0.548; Fig. 3E) indicated no publication bias.
Psychological correlates
Sixteen studies were included in the meta-analysis examining the correlation between demoralization and depression (Fig. 2F). Results of the meta-analysis showed that demoralization has a statistically significant, strong, positive correlation with depression (z = 0.61, 95% CI: 0.54, 0.67; Fig. 2F). These findings demonstrated that the characteristics of the included 16 studies are strongly heterogeneous (I 2 = 92%, p < 0.01). Egger’s asymmetry test suggested no evidence of funnel-plot asymmetry, indicating that there is no publication bias (t = −0.90, df = 14, p = 0.381; Fig. 3F).
Anxiety (z = 0.65, 95% CI: 0.56, 0.73; Fig. 2G) and suicidal ideation (z = 0.48, 95% CI: 0.34, 0.60; Fig. 2H) were also found to have strong positive associations with demoralization. These findings demonstrated that the characteristics of the included studies are strongly heterogeneous (I 2 = 71%, p = 0.02; I 2 = 89%, p < 0.01, respectively). Egger’s asymmetry test for anxiety (t = −2.23, df = 2, p = 0.156; Fig. 3G) and suicidal ideation (t = 0.43, df = 4, p = 0.687; Fig. 3H) suggested no evidence of funnel-plot asymmetry, indicated no publication bias.
Results of the meta-analysis showed that demoralization has a statistically significant small, negative correlation with social support (z = −0.39, 95% CI: −0.51, −0.26; Fig. 2I). These findings demonstrated that the characteristics of the included studies are strongly heterogeneous (I 2 = 88%, p < 0.01). Egger’s asymmetry test suggested evidence of funnel-plot asymmetry, indicating that a publication bias is likely (t = 0.31, df = 1, p = 0.37; Fig. 3I).
Forest plot of correlates of demoralization.
Egger test plot of correlates of demoralization.
Descriptive results
Given that the following variables were reported in only 1 literature, only descriptive analysis is reported here: demoralization was correlated with sociodemographic variables (medical payment (r = 0.749) (Kang et al. Reference Kang, Liu and Jiang2023), having at least 1 child (r = 0.043) (Liu et al. Reference Liu, Wu and Wang2020), marital status (r = 0.11) (Y. C. Li et al. Reference Li, Ho and Wang2016), source of income (r = 0.588) (Shao et al. Reference Shao, Li and Lin2024), gender (r = 0.041) (Ou et al. Reference Ou, Qi and Hu2021)), clinical variables (time of chemotherapy (r = 0.61) (Kang et al. Reference Kang, Liu and Jiang2023), treatment type (r = 0.12) (Y. C. Li et al. Reference Li, Ho and Wang2016), surgery type [breast-conserving surgery (r = −0.028), breast plastic surgery (r = 0.029) ](J. Li et al. Reference Li, Liu and Xu2020), time of radiotherapy (r = −0.802) (P. L. Tang et al. Reference Tang, Lin and Wang2022), cancer type [digestive and liver (r = 0.044), head and neck (r = 0.011), breast (r = −0.011), gynecological (r = 0.061)] (Ko et al. Reference Ko, Lin and Pi2018)), positive psychological variables (psychological well-being (r = −0.67) (Peng et al. Reference Peng, Hsueh and Chang2021), self-compassion (r = −0.299) (Liu et al. Reference Liu, Wu and Wang2020), spiritual well-being (r = −0.718) (Andrea Bovero et al. Reference Bovero, Opezzo and Tesio2023), hope (r = −0.695) (A. Bovero et al. Reference Bovero, Opezzo and Botto2021), positive life orientation (r = −0.471) (L. L. Tang et al. Reference Tang, Li and Pang2020), life completion (r = −0.45) (An et al. Reference An, Lo and Hales2018), perceived relatedness (r = 0.763) (Philipp et al. Reference Philipp, Mehnert and Müller2020), anxious attachment (r = 0.54) (An et al. Reference An, Lo and Hales2018), avoidant attachment (r = 0.37) (An et al. Reference An, Lo and Hales2018), empirical avoidance (r = 0.661) (T. Li et al. Reference Li, Wang and Wang2022)), negative psychological variables (psychological distress (r = 0.422) (Ko et al. Reference Ko, Lin and Pi2018), stress (r = 0.42) (Peng et al. Reference Peng, Hsueh and Chang2021), illness intrusiveness (r = 0.50) (Eggen et al. Reference Eggen, Reyners and Shen2020), perceived cognitive impaired (r = 0.38) (Eggen et al. Reference Eggen, Reyners and Shen2020)), social variables (at-work productivity loss (r = 0.074) (Kim et al. Reference Kim, Kissane and Richardson2022), financial toxicity (r = −0.616) (Shao et al. Reference Shao, Li and Lin2024), personality traits [extraversion (r = −0.395), agreeableness (r = −0.235), conscientiousness (r = −0.164), neuroticism (r = 0.475), openness to experience (r = −0.233)] (Ghiggia et al. Reference Ghiggia, Pierotti and Tesio2021), and researchers’ perceptions about demoralization (r = 0.64) (Belar et al. Reference Belar, Arantzamendi and Rodríguez-Núñez2019).
Discussion
Demoralization is a common mental problem affecting the prognosis of cancer patients. This systematic review highlights sociodemographic, clinical, and psychological factors associated with demoralization among oncology patients. Although demoralization varies across types of cancer, the current findings suggest that the nature of sociodemographic, clinical, and psychological factors associated with demoralization is similar across types of cancer. Thus, demoralization associated with cancer may be universal.
This study assessed the methodological quality of the included evidence using the AHRQ and the NOS. The summarized evidence ranged from medium to high quality. Specifically, almost all studies identified the source of treatment, criteria for inclusion and exclusion of patients, time to identification, and included continuous subjects. In addition, all studies described assessments undertaken and how confounding was assessed, which could contribute to the reality of the results. However, 9 studies did not summarize patient response rates, which could introduce bias into the results. Based on the findings above, the systematic review of the studies obtained a medium to high methodological quality.
Specifically, for demographic correlates, results show that demoralization has a mildly negative association with income and education levels. This finding is similar to a study in Taiwan (Li et al. Reference Li, Ho and Wang2017), revealing that education and monthly income are protective correlates of demoralization among cancer patients. The reason may be that the patient’s economic status or income directly determines their treatment methods and ability to afford medical expenses (Yu-Chi et al. Reference Yu-Chi, Chung-Han and Hsiu-Hung2017). Low-income cancer patients have limited access to medical resources, causing them to suffer from demoralization. Bailey et al. also found that education is an effective resource for dealing with serious diseases, especially in improving disease-related health literacy, which plays a crucial role (Bailey et al., Reference Bailey, Doyle and Dearin2020). Thus, patients with low education levels have poor access to disease-related knowledge, resulting in a lower ability to cope with the disease, and cancer patients are more likely to suffer from demoralization. Age is also negatively correlated with demoralization; younger patients are more likely to experience higher levels of demoralization, consistent with previous studies (Bailey et al. Reference Bailey, Doyle and Dearin2020; Cimilli Reference Cimilli2020). This may be related to younger patients taking on more social and family roles, leading to role conflicts caused by the disease. However, the result is contradictory in other studies that find no significant correlation between age and demoralization (Grassi et al. Reference Grassi, Costantini and Kissane2017; Vehling and Mehnert Reference Vehling and Mehnert2014). A recent study has further found that elderly patients have higher scores for demoralization (Vehling et al. Reference Vehling, Lehmann and Oechsle2012). Therefore, the correlation between age and demoralization remains controversial, and more high-quality studies are needed to evaluate the relationship between the 2 in the future.
The most frequently reported factors for clinical correlates are QoL and symptom burden. QoL generally refers to the state assessment of individual physiological, psychological, and social functions, which is a multidimensional concept. A moderate negative correlation between demoralization and QoL is detected among cancer patients, which has also been confirmed in other diseases (Zhu et al. Reference Zhu, Kohn and Patel2021). This may be related to the intensity of demoralization associated with physical symptoms (e.g. fatigue, limited physical activity, pain) (Gan et al. Reference Gan, Gong and Kissane2022). As well as being related to negative emotions, the results highlight the need for psychological intervention, which can be needed for health-care professionals to adequately evaluate cancer patients with demoralization and provide targeted interventions to improve their QoL. A recent study also found that Parkinson’s patients with demoralization report lower levels of QoL (Zhu et al. Reference Zhu, Kohn and Patel2021). In addition, symptom burden, including pain, lymphedema, and so on, is proven to be positively correlated with demoralization. These findings are consistent with those reported by Bailey et al. (Reference Bailey, Doyle and Dearin2020). Treating cancer is a long process where patients are prone to experience pain, fatigue, hair loss, and other symptoms, thereby seriously affecting their mental health (Kwekkeboom Reference Kwekkeboom2016).
Our study also reveals a remarkably positive correlation between anxiety, depression, suicidal ideation, and demoralization. This result agrees with a previous one (Mehmood Reference Mehmood2015; Nikoy Kouhpas et al. Reference Nikoy Kouhpas, Karimi and Rahmani2020; Ou et al. Reference Ou, Qi and Hu2021), indicating that cancer patients are more likely to have anxiety, depression, demoralization, and other negative emotions owing to the disease diagnosis and a series of reasons such as surgery, radiotherapy, and chemotherapy (Xu et al. Reference Xu, Hu and Liu2019). Depression is extensively recognized as a psychological problem in cancer patients, and its correlation with demoralization is widely supported (Shao et al. Reference Shao, Li and Lin2024). A systematic review by Tang et al. has shown that demoralization and depression are significantly positively correlated, with a risk ratio as high as 9.65 (Tang et al. Reference Tang, Wang and Chou2015). Our study has also demonstrated a negative correlation between demoralization and social support. As an available or usable social resource, social support can buffer the social psychological stress response of individuals under stress and help cancer patients rid themselves of negative emotional experiences to promote physical and mental health (Falak et al. Reference Falak, Safdar and Nuzhat Ul2020). These results suggest that medical staff should actively guide patients in seeking social support to relieve their negative symptoms caused by stress.
Clinical implications
Understanding the factors associated with demoralization among cancer patients can guide the formulation of improved strategies for managing demoralization and addressing its sources. First, clinicians should pay more attention to younger, less educated, and lower-income cancer patients, as they are more likely to experience higher levels of demoralization. Second, medical staff should take measures to relieve patients’ symptom burden and improve their QoL. Furthermore, clinicians could provide psychological interventions to reduce anxiety, depression, and suicidal ideation, potentially alleviating negative psychological and improving their QoL. Some psychological measures to improve self-esteem and mental resilience can also reduce demoralization in cancer patients. In addition, emphasizing social support can also help cancer patients manage demoralization and improve QoL.
Additionally, clinicians need to understand and distinguish between demoralization, anxiety, and depression. Demoralization and depression show common clinical features such as low mood or low self-esteem, and studies have found that when demoralization is severe, it may coexist with anxiety and depression (Grassi et al. Reference Grassi, Costantini and Kissane2017). Another study also demonstrated that demoralization may increase the risk of depression and anxiety (Bobevski et al. Reference Bobevski, Kissane and Vehling2018). However, there are some clinical differences between demoralization and depression. Recent studies used network and exploratory graph analysis and latent class analysis to examine the relationship between demoralization and depression (Bobevski et al. Reference Bobevski, Kissane and Vehling2022). The findings indicated that demoralization and depression exhibited specific common symptoms, including a low mood and suicidal thoughts. In contrast, symptoms specific to depression, such as inattention and insomnia, were clustered in a separate and stable community. Therefore, clinicians can use the new technology of network analysis to identify the main dimensions between demoralization and depression in cancer patients and evaluate their correlation more comprehensively. This can provide precise targets for psychotherapy and ultimately enhance the overall well-being of cancer patients.
Strengths and limitations
To our knowledge, this study is the first quantitative analysis of the sociodemographic, clinical, and psychological factors associated with demoralization among cancer patients, including different types of tumor. This meta-analysis may provide evidence for preventing and managing demoralization among cancer patients.
However, it still has the following limitations. First, different demoralization assessment tools are integrated into this meta-analysis, and various studies have different assessment tools for psychological distress, resulting in high heterogeneity. Second, the literature search is limited to studies published in English and Chinese, with the possibility of additional research published in the language of the country under investigation.
Conclusion
This study identifies sociodemographic, clinical, and psychological correlates of demoralization among cancer survivors. The results highlight a significant and growing demoralization among cancer patients, and such negative emotion may be overlooked. Demoralization should elicit the attention of health-care professionals and be evaluated to improve the health of cancer patients.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1478951524001597.
Data availability statement
All data are available in the manuscript.
Acknowledgments
The authors are special grateful to all researchers who participated in this study during the study period.
Author contributions
Wen-zhen Tang conceived the study, analyzed the data and drafted the manuscript. Wen-zhen Tang and Shi-li Cheng participated in the literature search, data extraction, and methodological quality assessment. Azlina Yusuf, Yulita Hanum P Iskandar, and Ernest Mangantig commented on the final manuscript. Kui Jia contributed to the study design. All authors have checked manuscripts and approved the publication of the protocol. Wen-zhen Tang and Shi-li Cheng contributed equally to this study.
Funding
The authors declare that no funds were received during the preparation of this manuscript.
Competing interests
None of the authors declare a conflict of interest.
Ethical approval
Not applicable.
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.
Provenance and peer review
Not commissioned; externally peer reviewed.
Open access
