Bladder cancer is one of the most prevalent tumors and a critical cause of tumor-induced death worldwide [1]. Over 60% of all bladder cancer cases and half of all 165,000 bladder cancer deaths occur in low-income regions worldwide [2]. Three-fourths of all bladder cancer cases occur in men [3]. According to Globocan 2012, bladder cancer ranks 12th among the top 20 most common cancers in China [4]. This cancer is 5 to 6 times more prevalent in men than in women in China [5]. Owing to its high morbidity and mortality rate, bladder cancer has attracted close attention from scientists [6]. A study has reported the causes of developing bladder cancer, including genetic mutations and external risk factors such as tobacco smoking, carcinogen exposure, chlorination of drinking water, and possibly cyclophosphamide [7]. On the basis of the examination of these external factors, we hypothesized that the prevention and treatment of bladder cancer could be affected by another essential factor, that is, patients’ attitudes, which should be investigated to deliver relevant interventions. On the basis of this hypothesis, we raised the following research questions: How many clusters can the study participants be grouped into based on their attitudes toward the prevention and treatment of bladder cancer, specifically their beliefs about the source of reinforcements for health-related behaviors if they develop this cancer? and Can such beliefs be associated with the participants’ demographic features and health literacy status?

Patients’ attitudes toward their conditions can be measured by the locus of control (LOC) construct, as extensively demonstrated by a large variety of pathologies [8-13]. It has been clinically observed that the LOC construct can mediate health status and outcomes [14-16], and it has proven helpful in predicting and explaining specific health-related behaviors [17]. On the basis of the LOC construct, individuals can be categorized into 2 main classes: those believing that their health status (or sickness) results from their own behaviors (health-internals) and those considering that their health status is generally determined by factors over which they have poor control, such as chance or powerful others (health-externals) [18]. Wallston et al [19] later proved the importance of separately assessing beliefs in the influence of chance and powerful others. More recently, Wallston et al [20] have shown that it is also helpful to distinguish between expectancies related to doctors and those related to significant others (eg, relatives and friends) within the “powerful others” construct. Health LOC has been conceptualized as a construct comprising at least 3 dimensions [19]. The most extensively used and validated instrument of LOC in health is the Multidimensional Health Locus of Control (MHLC) Scale [19]. This measure consists of 18 items rated on a 6-point Likert scale ranging from strongly disagree (n=1) to strongly agree (n=6). These 18 items are divided into three 6-item subscales that measure “Internal,” “Chance,” and “Powerful Others” LOC. Higher scores for each subscale indicate greater belief in that subscale domain in relation to health. There are 3 refined condition-specific versions of the MHLC Scale, namely, the MHLC Form A, the MHLC Form B, and the MHLC Form C [20]. The MHLC Form C consists of 4 subscales, “internal” (6 items), “chance” (6 items), “powerful other people” (3 items), and “doctors” (3 items) [19].

The MHLC Form C has been applied to enhance the knowledge about the point of view of patients with HIV on their complex health condition [21]; to determine how LOC relates to health care use, medication adherence, missed school, and readiness for transition to adult medical care for youth with chronic conditions [22]; and to explore the relationship between LOC and pregnancy [23-27]. Thus informed, we believe that this scale is most likely to solicit patients’ attitudes toward the prevention and treatment of bladder cancer based on which tailor-made education, intervention, and treatment could be delivered for the benefit of the prevention and treatment of this disease. However, it has never been used in this respect. Considering the magnitude of bladder cancer among men in China, it is imperative to examine patients’ attitudes toward and beliefs about the prevention and treatment of this disease to reduce its prevalence and the associated male mortality rate.

We aimed to classify male patients into different latent groups according to their attitudes toward and beliefs about bladder cancer prevention and treatment and to identify associated factors to provide implications for the delivery of tailored education and interventions and the administration of targeted prevention and treatment.

First, we designed the questionnaire. We then recruited male patients from the Qilu Hospital of Shandong University, China, who were invited to complete the web-based questionnaire. Finally, we used latent class analysis (LCA) to identify subgroups of men based on their categorical responses to items on the MHLC Form C.

We designed a four-section questionnaire (Multimedia Appendix 1), including section 1: age, gender, and education; section 2: the “Communicative Health Literacy” subscale of the All Aspects of Health Literacy Scale (AAHLS) [28]; section 3: the eHealth Literacy Scale (eHEALS) [29]; and section 4: the MHLC Form C [19]. The “Communicative Health Literacy” subscale is defined as more advanced skills to actively participate in everyday activities, extract information, derive meaning from different forms of communication, and apply new information to changing situations [28]. The 8-item eHEALS evaluates study participants’ knowledge and skills that are essential for using eHealth resources and interventions [29]. It does not have subscales. Literacy in health information is becoming a critical factor essential for health status [30]. The 18-item MHLC Form C comprises 4 subscales that measure “internal,” “chance,” “doctor,” and “powerful others” LOC, that is, beliefs that the source of reinforcements for health-related behaviors is primarily internal, a matter of chance, or under the control of doctors or powerful others, respectively [19]. Such beliefs can motivate health behavior, which refers to taking voluntary actions to promote health, reduce health risks [31], and mediate health status [15,32]. Individuals categorized as having an “internal” LOC are more likely to engage in health behaviors and are more knowledgeable about their health problems [33,34]. Considering that the MHLC Form C is a “general purposes, condition-specific locus of control scale that could easily be adapted for use with any medical or health-related condition” [20], we adapted it for use with bladder cancer in the questionnaire. Informed by relevant studies [15,30-34], we hypothesized that the participants’ status of health belief and self-confidence measured by the MHLC Form C in section 4 could be closely associated with information collected through sections 1-3.

The study participants were recruited from the Qilu Hospital of Shandong University, China, using randomized sampling. Participants who had met four predefined inclusion criteria were invited to participate in this survey: (1) being aged ≥18 years, (2) having at least primary education (Year 6 schooling) to understand the questions in the questionnaire, (3) being patients rather than relatives accompanying patients, and (4) participating in the survey voluntarily. We made face-to-face contact with Mandarin Chinese–speaking patients who were attending the outpatient clinic and those who were hospitalized to identify those who satisfied the inclusion criteria, explain to them the purpose of the survey, and ask them to participate in the web-based survey as scheduled. We identified 988 eligible patients.

The survey lasted 1 month, from July 20 to August 19, 2022. The questionnaire was administered via wenjuanxing [35], the most popular web-based questionnaire platform in China. Participants filled out the administered questionnaire on the web. Returned questionnaires were considered valid only when all question items included were answered according to our predefined validation criterion. On August 20, 2022, the returned questionnaires were downloaded in the format of an Excel file (Microsoft Corp) from wenjuanxing. Out of 988 patients, a total of 718 (72.7% response rate) answered questionnaires were returned. We double-checked the returned questionnaires and found all of them to be valid.

On August 20, 2022, the answered questionnaires were downloaded in the format of an Excel file from wenjuanxing (Multimedia Appendix 2). We double-checked the validity of the returned questionnaires before coding valid data using the predefined coding schemes (Multimedia Appendix 2) based on Likert scales with varying score ranges for different questionnaire items. Subsequently, we used LCA (Latent GOLD 5.0) to identify subgroups of men based on their categorical responses to the items on the MHLC Form C and to ascertained factors contributing to the low self-efficacy group identified.

LCA is increasingly being applied in social and health sciences. LCA has methodological advantages over traditional clustering techniques [36-39]. A notable benefit of LCA is the probabilistic attribution of latent class membership to study participants using maximum likelihood estimation [36]. Therefore, each observed participant attained a probability of belonging to a certain latent class. For example, within a 2-class LCA solution, a study participant can have 2 probabilities associated with either latent class. On the basis of the conditional independence assumption of LCA, the combined probabilities of class memberships sum to 1, based on the conditional independence assumption of LCA. The probabilistic nature of LCA adds to the complexity of interpreting the results. However, in practice, the more flexible, intuitive approach of LCA when compared with “hard, rigid” clustering techniques allows researchers more insights into the impact of predictor variables on latent class membership fluidity and dynamics, and the susceptibility of class memberships to the definition and selection of probability thresholds to suit different research purposes.

This study was approved by the Ethics Review Board of the Qilu Hospital of Shandong University, China (KYLL-202208-026). The study data were anonymized to protect the privacy and confidentiality of the study participants. As the participants voluntarily participated in the survey to support and promote academic research, no compensation was provided for them as per the common practice in China.

Table 1 presents the descriptive statistics of the data collected from the patient participants. All data in the 718 returned questionnaires were valid. The patients had a mean age of 46.41 (SD 10.143) years. In total, 718 patients were men. The mean score for education was 2.89 (SD 1.398), indicating that their average educational level was between Year 9 and Year 12. The mean scores of the 3 communicative items on the AAHLS were 1.84 (SD 0.763), 1.91 (SD 0.751), and 1.95 (SD 0.744) for item 1, item 2, and item 3, respectively. These mean scores indicate that they “sometimes” needed help to read and comprehend health information and complete official documents and were “sometimes” able to identify and secure others’ help. The scores for the 8 items on the eHEALS ranged from 2.68 (SD 1.200) to 2.86 (SD 1.239), indicating their uncertainty about their skills to use eHealth resources and interventions. As with their scoring performance on the MHLC Form C, they scored on average 18.05 (SD 4.390), 16.80 (SD 4.435), 9.75 (SD 3.333), and 8.70 (SD 2.757) on the “internal,” “chance,” “doctor,” and “powerful others” subscales, respectively. The determined response of “slightly disagree” for the “internal” subscale indicates that they somehow did not believe in their internal drives to maintain health. The determined response between “moderately disagree” and “slightly disagree” for the “chance” subscale implies that they were generally less likely to attribute their health to a matter of luck. The determined response between “slightly disagree” and “slightly agree” for the “doctor” subscale means that they were generally uncertain about the role of doctors in the maintenance of their own health. The determined response between “moderately disagree” and “slightly disagree” for the “powerful others” subscale means that they generally did not believe in the role of others in the maintenance of their own health.

Table 2 and Figure 1 show the model fit statistics of the LCA. The Akaike information criterion and Bayesian information criterion provide the measures of model performance. Smaller Akaike information criterion and Bayesian information criterion values indicate better model performance. Indexes such as the Lo-Mendell-Rubin likelihood ratio test and the bootstrap likelihood ratio test examined whether adding clusters would significantly improve model performance. We considered all the factors and decided to opt for a 2-cluster solution for better model performance and simplicity to guide subsequent qualitative analyses, as shown in Multimedia Appendix 3.

We identified 2 subgroups—low and moderate self-efficacy groups—which represented 75.8% (544/718) and 24.2% (174/718) of the total sample, respectively. People in the low self-efficacy cluster (cluster 1: 544/718, 75.8%) had the following characteristics: they had low posterior probabilities on the “internal” and “doctor” subscales, suggesting that they were less likely to believe in their own capability or doctors’ advice to achieve optimal outcomes in bladder cancer prevention and treatment; they had higher posterior probabilities on the low (4-9) to medium (10-14) sections of the “other people” subscale, suggesting that most people in cluster 1 were not likely to see the influence of others’ behaviors on their health as significant; and finally, male patients in the low self-efficacy cluster had mixed beliefs about the role of chance and good luck on their health outcomes, as indicated by the wide-ranging sum scores on the “chance” subscale. Male patients in the moderate self-efficacy cluster (cluster 2: 174/718, 24.2%) had distinct psychological traits. They had high scores on the “internal” subscale, suggesting that they had stronger beliefs in their own capability to manage their health. They also had higher scores on the “doctor” subscale, suggesting that they had moderate to high levels of trust in health and medical professionals and the importance of adherence to their advice to achieve better health outcomes. Surprisingly, they had very high (score=15-16) scores on the “other people” subscale, suggesting that some of them were more likely to associate their own health outcomes with other people in their lives. Finally, we found that a polarized view regarding the role of chance in health and well-being existed among this group of male patients, despite being internally driven and having stronger beliefs in medical professionals.

The characteristics of the 2 clusters mentioned above could not be compared with relevant findings reported in previous studies because of the lack of relevant literature. However, we could relate the factors associated with the low-efficacy cluster identified in our study to relevant factors correlated with cancer risk, view of cancer, and cancer prevention, ascertained in a limited number of studies. As revealed in our study, limited communicative and digital health literacy contributed to low self-efficacy, that is, being less likely to believe in one’s own capability or doctors’ advice to achieve optimal outcomes in bladder cancer prevention and treatment. This finding aligns well with that reported by Morris et al [40] that those with low health literacy tended to feel less control over risks to their health and take more fatalistic attitudes toward cancer and cancer prevention, thus avoiding visiting doctors and being less likely to be up-to-date on cancer screening. The high-efficacy cluster in our study, in contrast, preferred to seek cancer prevention information from others and web-based resources, suggesting that they were more likely to believe in their own abilities, doctors’ recommendations, or others’ behaviors to achieve optimal outcomes in bladder cancer prevention and treatment. This finding somewhat confirms that of Morris et al [40].

In our study, we found that being aged ≥44 years was one indicator of participants’ low self-efficacy in bladder cancer prevention and treatment, which supports the findings by Taber et al [41] and Katapodi et al [42] that older age was correlated with lower perceived cancer risk, inducing older adults to take relatively negative attitudes toward cancer prevention and treatment. Similarly, Schroyen et al [43] reported that self-perception of aging and views of cancer could be seen as markers of vulnerability among older people with cancer. Similarly, the report by Schroyen et al [43] implied that older age and attitude toward cancer were correlated with low self-efficacy. Deeks et al [44] found that older study participants regarded a stable home life and relationships as important factors influencing health, which is contrary to our finding that older age was closely associated with low self-efficacy with regard to bladder cancer prevention and treatment, as shown by their perception of others’ behaviors as insignificant in influencing their health. Deeks et al [44] also discovered that having a disease prevention strategy was perceived as one of the most influential factors impacting health compared with our finding that patients in the low self-efficacy cluster displayed negative beliefs in their internal motivations to engage in health improvement.

Leung et al [45] found that fatalistic beliefs were negatively associated with cancer-related information-seeking behaviors, that is, patients who viewed the role of chance and good luck in their health outcomes as significant were less willing to seek information from health professionals and media. This finding supports our finding that patients in the low self-efficacy cluster were unlikely to regard the impact of others on their health as significant.

We identified low educational attainment as a significant contributor to low self-efficacy, specifically to negative attitudes toward the “internal,” “doctor,” or “others” LOC. However, we did not find such an association in the literature. This warrants further studies to ascertain the role of education, as “there exists a strong educational gradient in cancer risk, which has been documented in a wide range of populations” [46].

This study has some implications for clinical practice, health education, medical research, and public health policy-making. The 2 self-efficacy classes and 4 factors contributing to low self-efficacy can serve as important indicators for screening male patients with low self-efficacy to deliver more targeted education and more effective interventions to enhance their self-efficacy. Knowledge, skills, beliefs, and practices associated with the low self-efficacy class and the contributing factors could be integrated into public health education about and interventions in health beliefs about bladder cancer prevention and treatment among male patients to enhance their self-efficacy. Medical researchers can gain some insights into the topic of low self-efficacy and the contributing factors. Informed by this study, they could identify patients with low self-efficacy among their ethnic and socioeconomic groups, verify the contributing factors ascertained in this study, and identify more contributors in future studies. Finally, our research results and findings can provide some implications for public health policy-making in the future.

This study has some limitations. First, the generalizability of our results and findings may be limited. The recruitment of patients from only 1 hospital was most likely to make the results and findings less generalizable to populations in other provinces in China and to patients in different linguistic and cultural communities worldwide. Further research is warranted to validate the results and findings among populations of diverse ethnic and sociocultural backgrounds. Second, the self-reported nature of the collected data may result in some bias. As claimed by van der Varrt et al [47], self-reported literacy skills are not necessarily consistent with the actual abilities to comprehend, use, and appraise web-based health information. This is true for the self-reported literacy skills on the functional subscale of the AAHLS and the self-reported health beliefs and self-confidence on the MHLC Form C used in this study. More objective measures need to be developed to increase the reliability and consistency of assessments of various health literacy and health beliefs and self-confidence among culturally and linguistically diverse people. Third, comparisons could not sufficiently be made with previous studies because of the scarcity of relevant literature. Hopefully, our study will attract close attention from researchers who can further examine this topic to add to the body of literature and expand knowledge, which could promote academic conversation around such a topic of social significance.

This was the first study to investigate the attitudes and beliefs regarding bladder cancer treatment among male patients in China. Using latent class modeling, we identified 2 classes of LOC groups among male patients, the low self-efficacy group and the moderate self-efficacy group. Four factors contributing to low self-efficacy were identified, including (1) limited education (Year 6 to Year 12), (2) aged ≥44 years, (3) limited communicative health literacy, and (4) limited digital health literacy. These contributing factors can provide some implications for clinical practice, health education, medical research, and health policy-making.