This study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki and complied with all applicable national regulatory standards. The research protocol was reviewed and approved by the institutional review board of Hospital Santa Izabel, under the Certificate of Presentation for Ethical Consideration number 70726523.3.0000.5520, with approval opinion number 068562/2023 (Multimedia Appendix 1). All participants provided digital informed consent prior to inclusion. Consent was documented through electronic confirmation via REDCap (Research Electronic Data Capture). All procedures adhered to the Brazilian General Data Protection Law (Lei Geral de Proteção de Dados [LGPD]), guaranteeing anonymity, secure data handling, and restricted access to collected information.

This exploratory and descriptive study adopted a quantitative cross-sectional self-report survey design to evaluate patients with cancer’ perceptions regarding health data sharing, self-management, and trust in digital health solutions, considering the potential feasibility of blockchain as an enabling technology [36,37]. This study followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) reporting guidelines for cross-sectional studies. A completed STROBE checklist is provided as Checklist 1.

The study was conducted at Hospital Santa Izabel, located in Salvador, Bahia, Brazil, a national oncology reference center integrated within the ecosystem of scientific, technological, and innovation institutions. Data collection occurred between September and November 2023.

Recruitment followed a systematic convenience approach. Clinical staff—nurses and attending oncologists—identified eligible patients during outpatient appointments in both SUS and private insurance oncology clinics. Eligible patients were personally invited by a trained research team member who provided standardized verbal information about the study objectives, voluntary participation, and confidentiality. Those who agreed to participate were provided access to the questionnaire via a QR code displayed in waiting areas and consulting rooms. The questionnaire was self-administered on participants’ own personal devices or on an institution-provided tablet when needed, without direct researcher intervention during completion, ensuring participant autonomy and spontaneous responses. Participation had no influence on clinical care or the patient–health care provider relationship.

A convenience sample of 110 patients with cancer undergoing treatment or follow-up at Hospital Santa Izabel was included. Given the exploratory nature of the study and the absence of a primary hypothesis requiring formal power calculation, sample size was determined based on institutional feasibility and methodological guidance. A sample of n≥100 is sufficient to detect medium-sized effects (Cohen f=0.25) with approximately 80% power in ANOVA comparisons between 2 and 4 groups and to ensure sufficient diversity of sociodemographic profiles for descriptive and comparative analyses [39].

The inclusion criteria were as follows: adults aged 18 years or older; patients undergoing active oncology treatment or follow-up; and patients with an Eastern Cooperative Oncology Group (ECOG) performance status of 0, 1, or 2.

The exclusion criteria were as follows: patients with advanced or terminal disease (ECOG performance status 3 or 4) and patients who did not provide responses to the items comprising the 3 composite scoring domains (self-management, adherence, and governance).

The development and validation of the data collection instrument followed a structured multistep process.

Data were collected using the validated questionnaire hosted on the REDCap platform, a secure, Health Insurance Portability and Accountability Act (HIPAA)–compliant web-based application widely used for clinical and academic research. Access was provided via a QR code displayed in strategic institutional locations, including waiting areas of SUS, private oncology outpatient clinics, and physicians’ offices. The questionnaire was self-administered without direct researcher intervention during completion. The expression “in-person” used in previous versions of this manuscript referred exclusively to the recruitment context—face-to-face invitation by research staff—not to the data collection modality, which was entirely digital and self-administered. Participants who experienced difficulty with personal devices were offered an institution-provided tablet. All data were stored in encrypted REDCap servers with access restricted to the principal investigator.

All quantitative analyses were conducted using Stata (version 17.0; StataCorp). Descriptive statistics (absolute and relative frequencies, means, and SDs) characterized sample distributions. Independent-samples t tests compared composite domain scores between 2-group sociodemographic variables (gender, age group, educational level, income, and self-rated health); 1-way ANOVA was applied for race comparisons across 3 groups, with normality assumed based on the central limit theorem (n>30 per comparison). Pearson correlation coefficients assessed associations among the 3 composite domain scores. A Pearson chi-square goodness-of-fit test examined whether proportions expressing trust differed significantly across the 4 recipient-type categories (health care professionals, hospitals, the pharmaceutical industry, and the government). All tests were 2-tailed; α=.05 was the significance threshold. SEs were estimated under the assumption of a simple random sample, acknowledging the convenience sampling design as a limitation.

Three composite domain scores were calculated by averaging the constituent items (self-management: items 8‐11; adherence: items 12‐16; and governance: items 17‐20), rescaled to 0 to 100. Internal consistency was assessed using Cronbach α, with values ≥0.70 considered acceptable [39].

Of 110 patients approached, 104 (94.5%) were included in the final analytic sample (n=6, 5.5% excluded for not providing responses to the composite scoring domain items). Within the 104 included participants, item-level missing data varied by variable (range 1.0%‐7.7%), as reported in Table 1. The sample was predominantly female (63/98, 64.3%) and self-identified as pardo (mixed race; 64/98, 65.3%), with a predominance of lower-income participants (55/96, 57.3% earning less than twice the minimum wage).

The results indicate high acceptance of digital technologies: 61.4% (68/110) of participants considered the use of technology to integrate health information essential and 86.4% (95/110) were willing to use applications to store health data. Willingness to use prevention-focused applications reached 89.1% (98/110), and 86.4% (95/110) of participants believed that sharing health data is crucial for advances in medical research.

Regarding trust in data sharing, 79.1% felt comfortable sharing information with health care professionals, while trust in hospitals was moderate (57/110, 51.8%) and lower regarding the pharmaceutical industry (17/110, 15.5%). The lowest trust was for sharing with the government ( 11/110, 10%). A Pearson chi-square goodness-of-fit test indicated a statistically significant difference between these proportions (χ²₃=210.4; P<.001). Guaranteed anonymity and encryption considerably increased comfort in sharing, with 83.6% (92/110) of patients reporting greater willingness under these conditions.

Table 2 presents mean scores for the 3 composite domains—self-management, adherence, and governance—cross-referenced with sociodemographic variables. No statistically significant differences were observed in domain scores based on income, gender, educational level, or self-rated health status (all P>.05). Patients aged 18 to 59 years showed significantly higher adherence scores compared with those aged 60 years or older (mean 76.49, SD 19.16 vs mean 65.83, SD 26.66; t₉₅=2.29; P=.02), suggesting a greater affinity for digital solutions among younger participants.

Race-based comparisons did not reach statistical significance for any domain (all P>.05). As shown in Figure 2, Black participants exhibited numerically higher mean scores across all the self-management, adherence, and governance domains; however, these results must be interpreted with considerable caution given the small subgroup sizes (Black: n=15; White: n=20), which substantially limit statistical power and preclude reliable between-group estimates. This is acknowledged as a key limitation (refer to the Limitations section).

Table 3 presents the mean scores, SDs, and internal consistency (Cronbach α) for each of the 3 composite domains. The adherence domain achieved the highest Cronbach α (α=0.8807), indicating excellent reliability, followed by self-management (α=0.8504) and governance (α=0.7576), the latter still meeting the accepted threshold for health research [39]. The higher mean score for adherence (71.53, SD 23.12) compared with governance (61.25, SD 18.19) suggests that patients felt more personally engaged with digital tools than they trusted institutional actors to manage their data responsibly.

This study provides empirical evidence that patients with cancer in Brazil demonstrate high acceptance of digital health technologies and significant willingness to share clinical data, conditional on robust privacy protections and transparent governance. These findings directly address our 3 research questions, confirming the feasibility of patient-centered digital health initiatives within the Brazilian SUS context. As detailed in the Methods, the instrument assessed attitudes toward the functional properties that blockchain enables rather than the technology by name; the findings should therefore be interpreted as reflecting acceptance of blockchain’s underlying value proposition—particularly security and transparency—rather than informed endorsement of the technology itself, a distinction acknowledged in the Limitations section.

The strong inclination toward digital tool adoption (95/110, 86.4%) is consistent with Brazil’s high smartphone penetration [40-42]. The marked disparity in trust across recipient types—79.1% (87/110) for health care professionals versus 10% (11/110) for the government—reflects a pattern documented in the literature [7,11,43-45]. This asymmetry is explained by 3 interacting factors: the personalized patient–health care provider relationship, institutional distance, and historical privacy scandals that have eroded public trust [6,43,46,47].

These findings have a critical design implication for blockchain systems: the technology’s inherent auditability and immutability may be uniquely positioned to address institutional distrust by allowing patients to monitor who accessed their data, when, and for what purpose [21,29,30]. Technical accountability must be complemented by participatory governance frameworks and regulatory alignment with the LGPD and RNDS standards to achieve social legitimacy [48-50].

Studies indicate that public opinion often associates large institutions with economic interests and decisions disconnected from individual needs [6]. Blockchain-based data-sharing agreements represent essential complementary instruments that translate technical transparency into legally enforceable governance frameworks [51]. Their integration within the RNDS infrastructure—which uses Health Level Seven Fast Healthcare Interoperability Resources (HL7 FHIR) standards and aligns with both LGPD and General Data Protection Regulation principles—provides regulatory scaffolding for ethical blockchain implementation in Brazilian oncology [48,52-54].

The significant age-related difference in adherence scores corroborates findings identifying aging as a factor associated with lower digital engagement [55,56]. Lower eHealth literacy among older adults, combined with functional and cognitive changes and reduced social support for technology use, increases hesitation toward digital platforms [56,57]. This reinforces the need for age-adapted digital literacy programs and simplified interface design for older patients with cancer.

The absence of significant differences by income, educational level, and self-rated health status may reflect the relative socioeconomic homogeneity of this single-institution sample (all P>.05). However, this does not negate the well-documented role of digital equity in broader populations [58-60]. Digital health programs must proactively incorporate accessibility features and humanized technical support [55].

Race-based differences in domain scores did not reach statistical significance and must be interpreted with particular caution given the small subgroup sizes (Black: n=15; White: n=20), which substantially limit statistical power. The numerical trend of higher scores among Black participants may reflect local context-specific factors—such as community engagement initiatives at Hospital Santa Izabel—rather than generalizable patterns. Future studies with larger, multisite samples should examine racial disparities in digital health acceptance within Brazilian oncology [61,62]. In the absence of adequate statistical power, any interpretation of these results is exploratory and should not be extrapolated. The use of “race” terminology throughout is consistent with Brazilian Institute of Geography and Statistics categorization adopted in the questionnaire.

The analysis of data-sharing conditions highlights that anonymization and encryption substantially increased willingness to share (92/110, 83.6%), providing direct support for blockchain architectures incorporating privacy-preserving cryptographic mechanisms [26,27,30]. When patients retain technical control over who accesses their data, the psychological barrier of institutional distrust is substantially mitigated [29].

The correlation among the 3 domains—self-management, adherence, and governance—suggests that perceived information security acts as a central element in the willingness of patients with cancer to share data in digital environments. Note that these correlations are descriptive associations and do not test formal hypotheses, as the study did not prespecify directional hypotheses; accordingly, we refrain from confirmatory language in their interpretation.

This study has several limitations. First, the convenience sample from a single oncology center in Salvador, Bahia, limits generalizability to the broader, diverse Brazilian cancer population. Probabilistic, multisite sampling is recommended for future studies. Second, the racial subgroup analyses are substantially limited by small cell sizes (Black: n=15; White: n=20), which preclude definitive conclusions. Third, self-selection bias may have inflated acceptance estimates [63]. Fourth, the cross-sectional design precludes assessment of how perceptions evolve over time [64]. Fifth, participants’ prior knowledge of blockchain technology was not formally assessed before questionnaire completion. Therefore, it was not possible to determine the extent to which baseline familiarity with the technology may have influenced perceptions and acceptance. It is plausible that limited blockchain literacy influenced response patterns, such that participants’ acceptance reflected attitudes toward the underlying principles of data governance (eg, security, transparency, and patient control) rather than toward blockchain technology itself. Future studies should assess blockchain literacy at baseline to disentangle knowledge effects from attitudinal responses. Finally, the absence of formal factor analysis to confirm the 3-domain structure is acknowledged; future validation studies should use confirmatory factor analysis with larger samples. In addition, the instrument did not undergo formal cognitive pretesting or debriefing with respondents; although expert review and pilot testing were conducted, cognitive interviewing would have further strengthened content validity and is recommended for future studies. Additionally, SEs were estimated under the assumption of a simple random sample, whereas a convenience sample was used; this assumption may underestimate variance and further limit the generalizability of the results.

Future research should incorporate probabilistic, multiregion sampling and longitudinal designs to track the evolution of patient trust and digital acceptance. Mixed methods studies integrating in-depth interviews and focus groups would provide deeper insight into contextual factors shaping data-sharing decisions. Intervention studies evaluating real-world blockchain-based platforms—with patient-controlled consent mechanisms and transparent audit logs—should assess their impact on trust and clinical outcomes. Studies should formally assess blockchain literacy at baseline. Targeted digital inclusion programs for older adults and socioeconomically disadvantaged patients are essential to ensure equitable access to health innovations.

This study demonstrates that patients with cancer in Brazil exhibit high acceptance of digital health technologies and significant openness to clinical data sharing, provided that strong privacy protections, transparent governance, and patient-centered control are guaranteed. Critically, patient support for the proposed blockchain-based framework reflects acceptance of its functional principles—security, anonymization, and auditability—rather than informed familiarity with blockchain as a technology. This distinction has direct implications for future implementation strategies: educational components addressing blockchain literacy should be integral to any deployment. The marked trust asymmetry across recipient types underscores that technological innovation alone is insufficient. When implemented with adherence to LGPD regulations and integration with existing health infrastructure (RNDS and Conecte SUS), blockchain-based health data management systems represent a feasible and socially acceptable approach to transforming oncology data governance in Brazil.