A national cross-sectional convergent mixed methods study was conducted using a self-administered electronic survey from November 2023 to January 2024 (Figure 1). Quantitative and qualitative data were collected concurrently and analyzed in parallel. This study followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines to ensure transparency and rigor in reporting [30] (Checklist 1).

The research objectives (RO) were explicitly aligned with the selected instruments and data sources to ensure conceptual and methodological coherence. Perceived digital health competence (RO1) was operationalized using the DigiHealthCom, which captures multiple dimensions of health care professionals’ digital health competence. Organizational and educational factors that facilitate or hinder competence development (RO2) were assessed using the DigiComInf, focusing on managerial support, organizational practices, and collegial influence. Perceptions of AI in professional practice (RO3) were explored qualitatively through an open-ended question, allowing participants to articulate expectations, concerns, and perceived implications of AI without predefined response constraints. Finally, to examine the association between digital health competence and AI attitude (RO4), AI question (AIQ) responses were quantified into 4 mutually exclusive attitude categories and, using Participant Identification Code, linked to DigiHealthCom for inferential testing using multinomial logistic regression, adjusted for age and sex.

Integration was conducted at the interpretation stage. Quantitative findings on competence levels were compared with qualitative themes regarding competence development, training needs, and governance concerns to identify convergence, divergence, and complementarity. The quantitated AI attitude variable enabled direct modeling of competence-attitude associations.

A voluntary convenience sampling strategy through professional bodies was used [31]. The target population was invited through national professional councils via institutional mailing lists and public dissemination channels. Participation was voluntary, and all eligible professionals who accessed the survey during the data collection period were included. The exact response rate could not be calculated because the survey was disseminated through multiple professional mailing lists.

In this study, the data were collected using an online survey structured into 3 parts. The first part included background questions that described this study’s sample (sociodemographic data, professional characteristics, and digital use).

The second part comprised the DigiHealthCom and DigiComInf [24]. The final part consisted of an open-ended question designed to explore participants’ perceptions of the use of AI in the workplace and its potential applications in professional contexts (the AIQ).

The DigiHealthCom comprises 5 domains (42 items) and is used to assess health care professionals’ digital health competence. It covers the following domains: human-centered remote counseling competence (16 items), digital solutions as part of work (9 items), information and communication technology (ICT) competence (5 items), competence in using and evaluating digital solutions (8 items), and ethical competence related to digital solutions (4 items). The DigiComInf comprises 3 domains (15 items) that measure educational and organizational factors associated with digital health competence, including support from management (6 items), organizational practices for digital competence development (4 items), and colleagues’ adoption and influence (5 items). Both instruments used a 4-point Likert scale for scoring (1=completely disagree, 2=partially disagree, 3=partially agree, 4=completely agree). Both instruments were originally developed and psychometrically validated in the Finnish population, demonstrating satisfactory face and content validity and reliability [24]. DigiHealthCom and DigiComInf were translated from English into Spanish for the international study by Mikkonen et al [27]. An independent back-translation into English was subsequently performed by 2 researchers, and the translated versions were compared with the original instruments to ensure conceptual and semantic equivalence.

The content validity of the final survey, including DigiHealthCom, DigiComInf, the AIQ, and the background questions, was evaluated by a panel of 10 national experts to ensure the instruments’ relevance and applicability in the Spanish context. Content validity was assessed using the content validity index, applying both the item-level method (I-CVI) and the scale-level averaging method (S-CVI/Ave) [32].

Data were collected between November 2023 and January 2024 using an online survey hosted on Google Forms, a secure web-based survey platform. The survey link was distributed via personalized email invitations disseminated through the official communication channels of the General Council of Nursing Colleges, the General Council of Medical Colleges, and national social work professional bodies in Spain. In addition, publicly accessible links were shared through institutional newsletters and professional mailing lists.

Participation was voluntary and anonymous. Professionals were invited once, and 2 to 3 reminder emails were sent at weekly intervals, depending on the response rate, to maximize participation during the data collection period. No financial incentives were provided to participants. Responses were reviewed for potential duplicate entries based on timestamp patterns. All submitted questionnaires were retained in the dataset. Descriptive analyses were based on all available data, whereas multivariable analyses were conducted using complete-case analysis, excluding participants with missing values for variables included in each model.

Permission to use the DigiHealthCom and DigiComInf was obtained from the original authors. Ethical approval for this study was granted by the Research Ethics Committee of the University of Alicante, Spain (UA-2023-04-03). Participation was voluntary, and informed consent was obtained electronically before completing the survey. Participants were informed of their right to withdraw from this study at any time without consequences.

This study was conducted in accordance with the principles of the Declaration of Helsinki [40], ensuring respect for participants’ autonomy, privacy, and well-being. Data were collected through an anonymous online questionnaire, and researchers had no direct contact with participants at any stage of this study. All data were anonymized and handled confidentially in accordance with secure data management procedures. This study also complied with the EU’s General Data Protection Regulation [41], ensuring the responsible processing and protection of personal data.

A total of 229 health care professionals completed the survey (Table 1). The sample was predominantly female, with women representing nearly three-quarters of respondents, while approximately one-quarter were men. Participants were, on average, middle-aged, with a broad age range from early adulthood to late working age. Most respondents reported high levels of educational attainment. The majority held at least a bachelor’s degree, and a substantial proportion had completed master’s or doctoral studies, indicating a highly educated professional sample.

In terms of professional background, nurses were the largest group, accounting for more than half of the sample, followed by public health nurses. Other professions were represented in smaller proportions, reflecting a multidisciplinary workforce.

In the clinical context, more than one-third of participants worked in inpatient care settings, while outpatient and emergency services were also well represented. Smaller proportions were employed in administration, research, home-based care, or social care services, highlighting the diversity of work environments included in this study. Most respondents reported frequent contact with patients or clients, with nearly two-thirds indicating daily contact. However, a notable minority reported limited or no contact with current patients or clients at the time of data collection.

The DigiHealthCom scale showed excellent internal consistency at both total and domain levels (Table 2). Digital health competence was moderate overall, with higher scores observed in ICT-related competences and lower scores in domains related to the evaluation and implementation of digital solutions. Item-level results are in Multimedia Appendix 2.

No statistically significant differences were observed between women and men in the overall competence or across domains, with negligible effect sizes (Table 2). Age was not significantly associated with DigiHealthCom scores, indicating no linear relationship between age and digital health competence in this sample.

In contrast, the highest educational achievement was associated with slightly higher competence levels. This association was statistically significant but small in magnitude, suggesting limited practical impact.

The DigiComInf scale showed excellent internal consistency at both total and domain levels (Table 3). Organizational and educational factors supporting digital competence were moderate, with relatively stronger perceptions of collegial influence and weaker evaluations of structured organizational practices. Item-level results are presented in Multimedia Appendix 3.

No statistically significant differences were observed between women and men in overall scores or across domains, with negligible effect sizes (Table 3). Age was not associated with DigiComInf scores, indicating no linear relationship between age and perceived organizational or educational support.

Similarly, no significant differences in perceptions of organizational support for digital competence development were found, with trivial effect sizes, indicating consistency across subgroups.

Qualitative content analysis identified several interrelated themes and subthemes reflecting health care professionals’ expectations, concerns, and perceived implications of AI in the workplace and across health systems (Table 4). Their perceptions combined anticipated benefits with important ethical, professional, and organizational concerns. The most prominent theme, “adaptation and challenges,” highlighted the need to adjust to AI-enabled workflows, including training requirements, role changes, uncertainty, and concerns about depersonalization and unequal adoption.

The theme “patient care impact” reflected expectations of improved care quality, early detection, and personalized care, alongside awareness of potential effects on the patient-provider relationship. Similarly, “efficiency and workflow” highlighted anticipated gains in process optimization, reduced administrative burden, and support for evidence-based practice.

At the same time, participants emphasized “ethics and regulation” and “human oversight,” underscoring the need for clear governance, data protection, accountability, and the preservation of professional responsibility, with AI viewed primarily as a decision-support tool rather than a replacement for clinicians. Additional themes related to “resource constraints,” “sustainability,” and “accessibility” pointed to AI as a potential response to workforce shortages and to enhance system efficiency and access to care.

These themes together indicate a pattern of conditional acceptance of AI, shaped by perceived benefits but moderated by ethical, organizational, and professional considerations. Visual analyses supported these findings: the word cloud (Multimedia Appendix 4) emphasized central concepts such as patient care, workload, and professional roles, while the Sankey diagram (Figure 2) illustrated the prominence of adaptation-related challenges and their connections with training, uncertainty, and workflow changes.

AI attitudes were predominantly positive, followed by uncertain or unclear, ambivalent, and negative orientations. Intercoder agreement for attitude classification was excellent (κ=0.949). No significant differences in AI attitudes were observed by sex or age, indicating that attitudinal variation was not explained by these demographic factors (Multimedia Appendices 5 and 6).

In contrast, digital health competence differed significantly across attitude categories (Table 5). Participants with a positive attitude toward AI reported higher competence levels than those with negative or uncertain attitudes, reflecting a moderate effect size.

Multinomial logistic regression analysis confirmed this pattern: higher digital health competence was independently associated with greater odds of expressing a positive attitude toward AI than a negative one, after adjusting for age and sex. No significant associations were observed for ambivalent or uncertain attitudes (Multimedia Appendix 7).

These quantitative findings are consistent with the qualitative results, which indicated a pattern of conditional acceptance of AI, shaped by perceived benefits but moderated by concerns about training, governance, and professional responsibility. The findings suggest that participants reporting higher levels of digital health competence were more likely to express positive and confident attitudes toward AI, whereas lower competence was associated with uncertainty or more cautious attitudes.

This convergent mixed methods study examined digital health competence, associated factors, and attitudes toward AI among health care professionals in Spain. Three principal findings emerged.

First, digital health competence was moderate overall, with stronger performance in ICT-related domains and weaker performance in evaluative and implementation competencies. Second, perceived organizational support for competence development was limited. Third, AI attitudes were predominantly positive but conditional, reflecting both expected benefits and concerns related to training, governance, and professional responsibility. Importantly, quantitative and qualitative findings converged to show that higher digital health competence was associated with more positive attitudes toward AI, suggesting that digital health competence may represent one of several factors associated with professionals’ attitudes toward AI.

The overall competence profile observed in this Spanish sample aligns with prior international evidence. Participants reported relatively strong ICT competence but lower scores in domains related to evaluation, implementation, and the critical appraisal of digital solutions. This pattern is consistent with systematic reviews indicating that health care professionals often demonstrate confidence in operational digital skills while reporting gaps in higher-order competencies such as critical evaluation and strategic integration [7,8].

Cross-national analyses using DigiHealthCom have similarly identified heterogeneity across domains and persistent weaknesses in advanced competence areas [27]. The confirmatory factor analyses in our study supported the structural validity of both DigiHealthCom and DigiComInf, reinforcing the robustness of the competence construct across contexts [21,24,25].

Associated factors of digital competence were perceived as only moderately developed, with structured planning and training consistently scoring lowest. This pattern aligns with prior research emphasizing the central role of managerial support, institutional practices, and formal training pathways in sustaining competence development [24,27]. The lack of significant associations with age and sex further suggests that variation in digital competence is more likely driven by structural and organizational factors than by individual demographic characteristics.

Qualitative findings revealed a pattern of conditional optimism toward AI. Most participants anticipated improvements in efficiency, decision support, and quality of care. However, these expectations were consistently linked to potential contributors such as adequate training, human oversight, regulatory clarity, and preservation of professional responsibility and patient-centered care.

Exclusively negative attitudes were uncommon. Rather, ambivalence reflected the coexistence of perceived opportunities and ethical or organizational concerns. This configuration aligns with prior studies indicating that health care professionals’ acceptance of AI is influenced by trust, perceived reliability, governance, and clear professional roles [4,11,12,23]. It is also consistent with empirical evidence showing that health care professionals’ perceptions of AI are generally favorable but nuanced, combining expectations of improved efficiency and clinical support with concerns about safety, accountability, and implementation context [42,43].

Participants’ emphasis on accountability and oversight resonates with the evolving regulatory framework under the EU Artificial Intelligence Act [29], which explicitly requires human oversight and adequate AI literacy in high-risk health care applications [44,45]. Although regulatory knowledge was not directly assessed, these findings suggest that professionals’ concerns reflect issues addressed in the current European regulatory framework.

The central analytical contribution of this study lies in demonstrating an association between digital health competence and AI attitudes using validated instruments and multivariable modeling. Higher digital health competence was independently associated with a significantly greater likelihood of expressing a positive AI orientation, even after adjusting for age and sex. Although technology acceptance was not assessed directly, the observed association could be interpreted considering technology adoption frameworks such as TAM and UTAUT, which propose that perceived capability and facilitating conditions may influence engagement with new technologies [17,19]. In the present study, these frameworks were used to provide a theoretical context for interpreting the findings rather than to evaluate technology acceptance itself. Rather than representing a component of technology acceptance, digital health competence can be viewed as a complementary construct associated with, but conceptually distinct from, technology acceptance. This interpretation is complemented by the qualitative findings, in which participants consistently emphasized the importance of training, critical appraisal, and human oversight when discussing AI in clinical practice. Consistent with the exploratory nature of a single open-ended survey question and previous research, these qualitative findings primarily contextualize and support the quantitative results rather than providing novel conceptual insights.

Beyond technical proficiency, health care professionals reporting higher digital health competence may also perceive themselves as better able to interpret probabilistic outputs, recognize algorithmic limitations, evaluate data provenance, and integrate AI recommendations into clinical reasoning. These capacities may be related to greater perceived usefulness and trust, two factors consistently associated with AI adoption in health care. Conversely, lower perceived competence may also be associated with greater perceived risk, uncertainty, and concerns regarding liability or loss of autonomy, themes frequently highlighted in qualitative studies of AI implementation [13,23]. However, these potential mechanisms were not directly examined in the present study and may also reflect broader individual or organizational factors.

These findings indicate that digital health competence is associated with attitudes toward AI. However, the mechanisms underlying this relationship remain uncertain and are likely to involve broader individual, organizational, and contextual factors that were not examined in this study.

Participants reported a lack of structured institutional strategies for developing digital competence, with training largely dependent on individual initiative. This gap is consequential in a regulatory environment that requires not only formal compliance but also demonstrable capacity for oversight, accountability, and risk management.

The lowest competence levels were observed in domains related to evaluating and implementing digital solutions, precisely those required for meaningful human oversight under the EU Artificial Intelligence Act [29]. At the national level, Spain’s AI strategy for the National Health System (2025) similarly emphasizes training and capacity-building. The present findings provide empirical support for prioritizing advanced competence domains within such policy frameworks.

The association between digital health competence and AI attitudes indicates that competence development may support more constructive engagement with AI in the workplace. Training should therefore prioritize advanced competencies, including the critical evaluation of digital tools, the interpretation of algorithmic outputs, their integration into clinical decision-making, and ethical-legal aspects such as accountability and data protection, in line with emerging European regulatory requirements [29,45].

At the organizational level, individual training must be complemented by structured strategies, protected time, interdisciplinary collaboration, and clear governance mechanisms to ensure safe and accountable implementation. Consistent with previous research [24,27], strengthening organizational support for digital health appears essential to align workforce preparedness with regulatory expectations and enable sustainable AI integration in health care.

This study combined psychometrically supported instruments with confirmatory factor analysis (CFA) and a convergent mixed methods design, allowing both structural assessment and qualitative contextualization of health care professionals’ digital health competence and attitudes toward AI. The intercoder agreement achieved during AI attitude classification further supports the consistency of the qualitative analysis.

However, several limitations should be acknowledged. The use of a convenience sample may have introduced self-selection bias, and the predominance of nurses, together with the inability to calculate an overall response rate, limits the representativeness of the sample. As this study was conducted exclusively in one country, the findings should be generalized cautiously to other health care settings. Digital health competence was assessed through self-report and may reflect perceived rather than actual competence. Consequently, the observed association with AI attitudes may partly reflect generalized technological confidence rather than actual differences in digital health competence. In addition, the qualitative findings were derived from a single open-ended survey question, providing useful contextual information but less depth than interviews or focus groups. Finally, the multinomial regression model was adjusted only for age and sex, and other professional and organizational factors may also have influenced attitudes toward AI. Together with the cross-sectional design, this limits causal interpretation of the observed associations. Future longitudinal studies incorporating a broader range of explanatory variables are needed to clarify these relationships.

Digital health competence was positively associated with health care professionals’ attitudes toward AI in this national sample. Although the cross-sectional design precludes causal inference, the findings suggest that professionals reporting higher levels of digital health competence are more likely to express positive attitudes toward AI. Additionally, lower scores in evaluative, implementation, and ethical competence domains, together with limited organizational support, highlight areas that may benefit from targeted educational and organizational strategies.

Future longitudinal and intervention studies are needed to clarify the direction of this association and to determine whether changes in digital health competence are accompanied by changes in attitudes toward AI and its implementation in health care.