Perspectives on Health Data Sharing Among Patients With Somatic and Mental Health Diseases: Focus Group Study

Introduction

Background

The German Health Data Utilization Act (Gesundheitsdatennutzungsgsetz [GDNG]) and the Act to Accelerate the Digitization of the Healthcare System (Digital-Gesetz; Digital Act), both enacted in March 2024, have created new legal regulations for using and sharing personal health data in the German health care system. The former is primarily intended to enable the use of health data for research and to facilitate its usability for purposes oriented toward the common good [1]. The latter includes the introduction of the electronic patient file. Thus, it explicitly promotes improved digital data exchange in the German health care system [2] and promises to integrate clinical data of patients across different health care institutions and throughout their lives [3]. The GDNG and the introduction of the electronic patient file as an opt-out solution are exemplary innovations that are not only relevant for health care providers and researchers but also bring opportunities and challenges for patients. In particular, the new measures aim to improve the overall quality of health care by ensuring data access for health care and research in accordance with the respective legal regulations for primary use (PU) and secondary use (SU) [4-7].

Personal health data refers to an individual’s health status and information that are uniquely identifiable to them, such as diagnoses or medication plans. It is collected during care delivery in the health care system and can be obtained from existing sources, such as electronic health records [8]. The PU of health data refers to its application to directly provide medical care to patients, such as by sharing personal health data with health care providers [9,10]. In contrast, the SU of health data extends beyond the primary goal of individual treatment. It involves the use of pseudonymized or anonymized health data for purposes other than direct patient care, including activities such as commercial and noncommercial research, public health monitoring, and developing health care policies [9,10].

Germany is pursuing a deliberate, data protection–focused approach compared with other countries’ more centralized or incentive-based health care systems. Unlike the first national centralized health data access bodies in France in 2019 (Health Data Hub) and in Finland in 2020 (Findata), Germany has launched decentralized initiatives, such as the German Medical Informatics Initiative (Medizininformatik-Initiative) in 2016 and the German Health Data Lab (Forschungsdatenzentrum Gesundheit) in 2020. The latter started to process the first requests for data use in 2025. These initiatives aim to link clinical data and medical research data to promote medical research and thus improve health care for all [11-13].

As in other European countries, patients in Germany value information, transparency, consent, and control regarding data sharing. However, Germany’s decentralized system and national federal laws such as the GDNG can make it challenging to meet these expectations consistently in all regions [14,15]. Additionally, these different framework conditions and strict German requirements for data exchange can be a challenge for the research community because there is no single, straightforward opt-out regime for all secondary research use. As a result, this can hamper collaborative research efforts, slowing progress in areas such as disease detection and treatment optimization, and reducing data usability. Overall, these challenges can lead to suboptimal individualized medical care [16,17].

Patients’ willingness to provide their health data for PU and SU is a highly topical issue for science, research, and public health care [14,18-21]. Data sharing can improve personalized care through closer cooperation between treating health care professionals and thus contribute to optimizing diagnostic and therapeutic outcomes [18]. Patients benefit from reduced treatment delays and medical errors, as well as more effective, efficient, and personalized care through access to comprehensive patient data [18,19].

In recent years, qualitative research has explored willingness to share data among vulnerable patient groups, such as patients with mental health diseases, to determine their needs, preferences, and concerns, which may differ from those of patients without mental health diseases [22-28]. These differences may require tailored approaches to promoting the responsible sharing of personal health data. While patients with mental health diseases may face stigmatization in society and in health care, which can lead to mistrust of health care providers and make comprehensive data sharing difficult, this stigmatization is less evident for patients with somatic diseases [22,25,28-30]. Compared with other patient groups, patients with mental health diseases may feel that their data are more sensitive and could be misused to their detriment or for discrimination. In contrast, patients with somatic diseases may feel that their data are less personal and more directly linked to physical health outcomes [22,23,28,31,32].

Based on the outlined new regulatory structures in Germany for digital progress, it is important to include the perspectives of patients from different target groups. Recent surveys have shown that both patients with somatic and mental health diseases in Germany are cautious and reserved about data protection [33-35]. However, the regulation of opt-out, which requires active objection, is particularly relevant for patients with mental health diseases due to the above-described potentially different concerns regarding health data sharing. Deeper insights into their attitudes are therefore crucial.

To our knowledge, there are currently no qualitative studies examining and comparing the perspectives of German patients with somatic and mental health diseases regarding sharing their personal health data for PU and SU, following the enactment of the new legal framework mentioned above.

Objective and Research Questions

This study reports a qualitative study comparing German patients with somatic and mental health diseases who were asked about their attitudes, preferences, and needs regarding their willingness to share their personal health data for PU and SU in 2 focus groups (FGs). It shows the similarities and differences in the perspectives of both groups of patients receiving outpatient treatment in Germany, considering the context of the new regulations.

The research questions of this study were (1) What factors influence German patients’ willingness to share their digital health data for PU and SU? (2) What similarities and differences exist in data sharing perspectives between German patients with somatic and mental health diseases?

Methods

This study was conducted as part of the PATH (Personal Mastery of Health and Wellness Data) project in January 2024.

Study Design

Following our overview of reviews on factors influencing patients’ willingness to share their digital health data for PU and SU [9], we conducted 2 FGs with German patients with somatic and mental health diseases. FGs are useful when individual motivations and perspectives need to be explored in depth [36]. Unlike standardized interviews or questionnaires, FGs are based on a topic guide with open-ended guiding questions. Participants are free to answer according to their subjective relevance, which is also a major advantage over standardized survey methods. The moderator also has the opportunity to follow up on individual aspects to explore them in more depth. In addition, the use of a topic guide allows the discussion to be steered thematically to maintain the focus of the discussion. Therefore, FGs are particularly suitable for gaining practical knowledge for implementing data sharing concepts.

The methods and results are reported according to the COREQ (Consolidated Criteria for Reporting Qualitative Research) checklist (Multimedia Appendix 1) [37].

Recruitment of Study Participants

Patients were recruited and enrolled at the University Hospital Dresden, Germany. To identify and recruit potential patients, physicians working in the outpatient departments of diabetology, neurology, dermatology, and psychiatry at the University Hospital Dresden were contacted in November 2023 by members of the research team in person and by email, explaining the project and planned details of the study.

Participants were selected from different clinical disciplines to obtain the most heterogeneous sample possible. Two FGs were formed, 1 with patients with somatic diseases (FG1) and another with patients with mental health diseases (FG2). Patients in both FGs were recruited by face-to-face and email with the support of project partners and staff in the respective outpatient clinics during their regular medical consultations. For this purpose, information flyers, a brief study information document, and a consent form were prepared for potential participants, containing the key details of the study (location with map, date, time, planned schedule, amount of compensation, and contact details for further questions). Participants for both FGs were recruited based on a predefined, purposive sampling plan and predefined criteria (Textbox 1).

Patients interested in participating were screened for eligibility, provided detailed information to establish a relationship prior to study commencement, and invited to attend the relevant FG by email. Consent forms were explained, and consent was obtained before the FG. Participants who attended the FG were offered financial compensation of €50 (a currency exchange rate of €1=US $1.16 is applicable).

Topic Guide

The topic guide for the FGs was developed based on our overview of reviews [9] on patients’ attitudes toward health data sharing for PU and SU. Guiding open-ended questions and topics were used in both FGs to allow comparisons (Multimedia Appendix 2). The list of topics (Figure 1) was presented as a word cloud on Microsoft PowerPoint slides to stimulate discussion 3 times in each FG: for health data sharing with health care professionals (PU), with research (SU), and with private companies (SU). The participants could choose 1 or more topics according to their own priorities, focus, and interests.

‎

Cognitive Pretests

Cognitive pretests were used to test and adapt the interview guide to ensure that all questions and items are understood and interpreted as intended and that no relevant aspects were missing. They also assessed whether the time required and the number of questions were acceptable [38]. The cognitive pretests were performed in October 2023 as individual online interviews with 7 participants with different diseases according to the predefined inclusion criteria (Textbox 1). The duration of the interviews ranged from 36 to 58 minutes. The pretests provided relevant insights and comments concerning the formulation of questions, comprehension of terms (eg, types of data), and conduct of the FGs.

Procedure

The FGs were held on 2 evenings in January 2024 in an intimate setting of a reserved room at the University Hospital Dresden. Both were audio recorded. Three research associates with experience in qualitative research took minutes (SF, MSc; female, and MS, Dr; female, for FG1; and SF and CL, Dr; female, for FG2). Another research associate and experienced moderator (Dr; female), who was not involved in the project and had been briefed in advance, used the topic guide to lead the participants through both discussions. A resident psychiatrist (FGV, Dr; male) also took part in FG2 (patients with mental health diseases) to ensure that participants felt confident. One patient’s relative was present in each FG due to private circumstances that made this necessary at short notice. All participants agreed that the relatives could remain present; however, they did not participate in the discussion. A short presentation was provided at the beginning of each session to introduce the project goal and the researchers who were present. However, no aspects relevant to the discussion were raised to avoid influencing the participants. At the end of the discussion, the participants were given a short sociodemographic questionnaire and a validated questionnaire [39] to determine their technology commitment score.

Data Analysis

The content of the FG discussions was transcribed verbatim, checked for accuracy, adjusted according to the rules of transcription [40], and all personal or personally identifiable information was removed. The data were analyzed using MAXQDA software (version 2018; Verbi Software GmbH) and qualitative content analysis according to Kuckartz and Rädiker’s [41] approach, in which the data are coded and evaluated according to content and thematic aspects in more detail than with Mayring’s [42] strongly theory-driven approach. This approach also enables the combination of deductive and inductive category development.

First, a category system was deductively developed based on the discussion guide and our overview of reviews [9], which was applied as a frame to the transcribed material. The category system was modified and refined as the material was reviewed in more detail, as well as through the subsequent inductive formation of recurring major categories into additional categories. Two research team members (SF and CL) independently coded the material, continuously compared emerging themes and codes, discussed data saturation through an iterative process of data collection and analysis, and checked intercoder reliability. With the help of added quotes as anchor examples, the coders could optimally match participants’ responses to the defined categories.

Data saturation was reached when additional data no longer produced new categories or insights, indicating that the breadth and depth of relevant information were adequately captured. To confirm saturation, MAXQDA visualization tools were used to track the frequency and recurrence of codings across datasets. Furthermore, peer discussions were conducted to validate that no significant new information appeared in the later stages of coding. This approach ensured that the coding scheme was comprehensive and that the study findings were robust and grounded in the data. This transparent and systematic approach to assessing data saturation strengthened the credibility and trustworthiness of the qualitative findings. Moreover, the research team collaboratively developed and continually refined the category system through iterative discussions to ensure that participants’ diverse perspectives were accurately represented. Figure 2 illustrates the data analysis process and the sequence of activities.

A codebook with definitions, respective anchor examples, and the number and context of statements used for qualitative coding and analysis in MAXQDA is provided in Multimedia Appendix 3.

‎

Ethical Considerations

The study was approved on November 6, 2023, by the Ethics Committee of Technische Universität Dresden (reference BO-EK-310072023). All procedures were in accordance with the ethical standards of the institutional ethics committee and with the Declaration of Helsinki (2024 revision). Persons with mental health conditions, considered a vulnerable population, were included in accordance with the 2024 Declaration of Helsinki (paragraphs 19-20). Additional safeguards were implemented, including assessment of decision-making capacity prior to informed consent and a resident psychiatrist (FGV) acting as a supporting person where needed, as approved by the ethics committee. Participants were informed about the purpose of the study, the voluntary nature of participation, and their right to withdraw at any time without penalty. Participants received compensation of €50 for their time and travel expenses related to attending the FG discussions. This incentive was preapproved by the ethics committee. Written informed consent was obtained from all participants who attended the FG discussions. Confidentiality was assured in compliance with the EU-GDPR (EU General Data Protection Regulation). Only researchers involved in the data analysis had access to the pseudonymized data, which were stored securely on institution-protected servers and will be deleted 10 years after study completion. Pseudonyms were used in the transcripts to protect participants’ identities.

Results

Sample Characteristics

Sample Sizes and Diseases

FG1 comprised 7 participants, of whom 3 had type 2 diabetes, 2 had a neurological disease, and 2 had a dermatological disease. The discussion lasted approximately 82 minutes. FG2 comprised 6 participants, of whom 3 had bipolar disorder, 2 had paranoid schizophrenia, and 1 had schizoaffective disorder (all currently in remission or with residual symptoms). The discussion lasted approximately 78 minutes. Because recruitment took place as part of regular daily care, it was not possible to provide a figure on refusals. A total of 2 patients dropped out on the day of the event due to acute illness.

Sociodemographic Characteristics

All participants’ sociodemographic characteristics (Multimedia Appendix 4) were extracted from the questionnaire provided. Regarding education, most had completed an apprenticeship (7/13, 54%). Regarding residence, most lived in large cities (9/13, 69%), followed by a medium-sized city (3/13, 23%). Regarding income, most participants in FG2 (4/13, 67%) had a monthly household income of <€1000-€1500, whereas most participants in FG1 (5/13, 71%) had a monthly household income of €3001-€3500 or €4001-€5000.

Technology Commitment

The total scores for the self-rated technology commitment scale and its subdomains of technology acceptance, technology competence beliefs, and technology control beliefs were high. Out of a maximum possible score of 60, the mean total score was 50 (range 41-57) for patients with somatic diseases and 45 (range 38-56) for patients with mental health diseases. The scores for each participant and FG are shown in Multimedia Appendix 5.

Categories

Overview

We identified 10 overarching main categories (I-X) with 32 corresponding subcategories (Table 1). These main categories were defined to capture overarching themes relevant to the research question, whereas the 32 subcategories were developed to reflect more specific aspects and nuances within these broad themes. This approach allowed for a detailed and systematic structuring of the data, enabling the identification of patterns, differences, and relationships more precisely. The hierarchical organization enhanced the clarity and transparency of the analysis, facilitating an in-depth understanding of the participants’ perspectives. The categories in the Results section and in Table 1 were sorted according to the topic guide and participants’ response behavior. The complete codebook with definitions, respective anchor examples, and the number and context of statements used for qualitative coding and analysis in MAXQDA is provided in Multimedia Appendix 3. Each main category included in the final category system was discussed in both FGs, but with different relevance in each FG and emphasis by individual participants. Table 1 presents an overview on the number of statements per main category and subcategory in both FGs. While most statements were about consent management preferences, few were related to social involvement and influence, with only 1 comment in FG1.

Key Findings

In both FGs, all participants expressed widely varying personal experiences with health data sharing, mainly regarding PU. In contrast, only those in FG2 discussed their personal experiences with employers and private companies. Regarding PU and individual usefulness for medical care, the benefit of comprehensive documentation was considered most relevant in both FGs. Regarding SU and public benefit for medical care, the advancement of drug and product development and the commercial interests of health data sharing were discussed extensively in FG2. Fears of discrimination and stigmatization were more prevalent in FG2. These fears were particularly related to disadvantages or dismissal in the workplace. Mental health diagnoses were also identified as highly sensitive in FG1. In addition, data security concerns, especially misuse, were raised more frequently in FG2 in response to the question about automatic data sharing.

It was not possible to clearly delineate participants’ views on the consent model. However, it became evident that participants in FG1 expressed only a few and rather general conditions regarding their acceptance of an opt-out model, whereas participants in FG2 attached very strict and more differentiated conditions to their acceptance of an opt-out model. University research institutions were sometimes more trusted than private and pharmaceutical research institutions, and the intended use or research project was crucial for consent preference. Participants in both FGs explicitly opposed the use of health data by “big players” such as Amazon, Google, and Apple.

Overall, technical safety measures of anonymization, or at least pseudonymization, in terms of health data sharing for SU were debated more intensively in FG1. Regarding legal and ethical framework conditions, the aspect of data protection was discussed more broadly in FG2. Aspects of standardization, regulation, monitoring, and authorization were addressed only in FG1. As facilitating factors, autonomy and control appeared more relevant in FG2, both in the context of PU and SU. Social influence, involving family, peers, and health care professionals, became much more important for participants in FG2.

Below, the results for each main category are described, with additional details and representative quotes for the corresponding subcategories from both FGs. At the end of the Results section, an overview of the key points regarding patterns and relationships between influencing factors and sociodemographic characteristics is provided.

Main Category I: Previous Experience With Data Sharing

Within this main category, participants distinguished between PU and SU experiences. While most reported positive experiences in terms of PU, 2 participants in each group indicated that there were bureaucratic barriers to creating an electronic health record with their health insurance companies and that the exchange of health data with their health care professionals was still paper based. Concerning SU, experiences of participating in clinical trials were more common among participants in FG2 (Textbox 2).

Main Category II: Individual Usefulness for Medical Care

A total of 5 subcategories emerged on individual usefulness for medical care when sharing health data. If a patient changes location or physician, the medical history is fully available to the next practitioner, which could prevent undue disease progression. Benefits were also perceived for obtaining a second medical opinion and to avoiding bringing paper copies of medical reports to the appointment. In addition, rapid assistance in an emergency and saving time for administrative formalities, such as pension procedures or medical history interviews, were discussed (Textbox 3).

Main Category III: Public Benefit

Regarding possible public benefits, drug and product development was discussed by 5 participants in FG2, which was linked to their gratitude that new drugs are being developed, meaning that treatment for patients with mental health diseases is improving. Regarding statistics and achieving gender-neutral, high-quality health care, there was agreement that health data should be collected automatically, especially if they are anonymized. The aspect of cost savings in health care by treating individuals well enough to enable them to remain active and able to work in the long term was raised by 1 participant in FG1 (Textbox 4).

Main Category IV: Personal and Privacy Concerns

Concerns about commercial interests were present in both FGs. Skepticism about possible conclusions from anonymized health data or data sharing without prior information and the associated potential for discrimination or stigmatization was particularly expressed in FG2. Contrary to the benefits of comprehensive and transparent documentation, concerns about being overstrained by the large amount of data available were also expressed (Textbox 5).

Main Category V: Data Security Concerns

Regarding data security, participants differentiated between concerns about data access and data processing. It was particularly emphasized in FG2 that health data are generally very sensitive. The data must be protected against unauthorized access and changes and should not be used, for example, for research purposes without notification. Contrary to existing concerns about data security, there was a statement from FG1 that the high level of data protection in Germany may sometimes be riskier for health than sharing one’s own health data (Textbox 6).

Main Category VI: Consent Management Preferences

Consent Scope

Regarding the scope of consent, a distinction was made between sensitive and nonsensitive health data. Communicable diseases, sexually transmitted diseases, mental health diseases, and health care professionals’ subjective assessments of patients in open notes, such as on their behavior during treatment, were categorized as highly sensitive and consequently associated with a restriction or rejection of sharing. As one possible approach, the creation of different categories of health data was mentioned. Health data in anonymized or pseudonymized form were also considered particularly notable, and emergency data were discussed as a special case in FG2 only (Textbox 7).

Further subcategories were developed, differentiating consent scope by recipient and purpose and dividing statements related to PU (health care professionals) and SU (research, pharmaceutical, and private companies). Regarding PU, while some participants expressed the desire for limited data sharing only and the relevance of a complete dataset for an optimal, holistic treatment, others expressed agreement with full sharing. Opinions included limited selection with prior consent to the treating physician, sharing only with relevant treating physicians, refusal of automatic sharing, and emergency situations as an exception. In FG1, health insurance companies and company physicians were discussed as special recipients for whom exceptions should be possible to avoid disadvantages at work or regarding health insurance fees. Regarding SU, the private companies mentioned included suppliers of medical devices and medical technology (Textbox 8).

Consent Model

The consent model subcategory covers whether participants preferred an opt-in or opt-out solution, reflecting our guiding question on automatic health data sharing for PU and SU. By definition, active consent and permission to use and share health data, or explicit and flexible consent, were coded as opt-in. Standard consent with an active and explicit right to refuse the use and sharing of health data was coded as opt-out. The answers in both FGs were very heterogeneous, and no clear picture could be derived. However, it became apparent that patients in FG1 tended to mention only a few and rather general conditions for their agreement to an opt-out model, whereas patients with mental health diseases set strict and more differentiated conditions for their acceptance (Textbox 9).

Main Category VII: Technical Safety Measures

Anonymization, or at least pseudonymization, serves as a facilitator of health data sharing for SU. A central data repository, such as the storage of anonymized health data on secure server platforms, preferably in Germany, or in a registry, would support controlled access for SU. Data encryption, possibly managed by entering a personal identification number, was considered relevant to the design of related software (Textbox 10).

Main Category VIII: Legal and Ethical Framework Conditions and Requirements

The most frequently mentioned aspect of this main category, especially among participants in FG2, was privacy. The importance of employers not having access to health data was especially highlighted. There was a call for effective data protection under the EU-GDPR. The paradox was also described, as some individuals reveal everything on social media but insist on high data protection when sharing health data. It would be important to create a harmonized legal framework, binding regulations, and a standardized digital data recording system valid throughout Germany, but this was viewed as hardly feasible. Participants in FG1 would be more willing to share their health data for SU if the appropriate and ethical use of health data were ensured through authorization procedures and monitoring by an independent committee (Textbox 11).

Main Category IX: Informational Self-Determination

This main category included key aspects that enable individuals to make and act on their own decisions: transparency, trust, autonomy, and control. The latter was considered relevant in the context of PU and the self-determined decision of which physicians will see individual health data. In particular, participants in FG2 considered the implementation of robust control mechanisms important because of anticipated limited public trust in sharing health data. Regarding SU, autonomy should be respected so that, in principle, everyone can decide whether to share their health data for research. Participants also expressed interest in protocols for transparently reporting information to each individual about their contribution and support of research activities, possibly combined with a mandatory disclosure of research results (Textbox 12).

Main Category X: Social Involvement and Influence

The topic of social influence was not addressed at all in FG1, apart from 1 statement on data sharing with the employer. In FG2, it was argued that family members or peers could also be involved in decision-making and that family members should also be allowed to release health information in an emergency. Health care professionals could play a supportive and consultative role, such as by providing information about data sovereignty and personal responsibility for the electronic medical record (Textbox 13).

Patterns, Differences, and Relationships Regarding Sociodemographic Characteristics

Gender

The category of personal and privacy concerns was predominantly articulated by female participants. The concerns about discrimination or stigmatization expressed were concrete and related to disadvantages in professional life and social contexts. By contrast, male participants emphasized the category of informational self-determination the most, especially the points of autonomy and control. Statements such as “But that everyone can then decide whether to approve data sharing or not” (FG2.5) illustrate a desire for fundamental sovereignty and the ability to make independent decisions. Furthermore, male participants called for clearer rules and more rigorous monitoring within the legal and ethical framework conditions and requirements. Their priority was to establish a reliable, externally controlled system that could be trusted.

Age

For older participants (aged 49-79 years), the focus was on data security concerns and informational self-determination principles. As with male participants, autonomy and control were central to this group. The core issue was the fundamental right to control how one´s own data are used. There were general but strong concerns voiced about the overall security of the system, which often failed to address technical details. Concerns regarding data access emerged leading in this group with statements like “But the abuse that results when someone who has no actual responsibility for it gets hold of the date is the fundamental fear” (FG2.1). Conversely, younger participants (aged 25-43 years) focused on concrete use cases and the practical individual usefulness for medical care. Emergency assistance was a frequent topic of discussion, highlighting a willingness to share health data in a clearly defined context. Younger participants expressed very granular preferences regarding the consent model. Within the subcategory of recipient and purpose, they specified which health data should be visible to which health care professionals (eg, general practitioner vs specialist), indicating a desire for flexibility and adaptability within the scope of consent.

Socioeconomic Status

The characteristics of education and income revealed overlapping patterns in the FG discussions and are therefore analyzed together. The analysis showed that socioeconomic status strongly influences the nature of the perceived positive and negative consequences of sharing health data. Participants with higher education and/or higher income (>€3000) made arguments for or against sharing data on a more systemic level. Participants in this group discussed technical safety measures in detail, citing anonymization and pseudonymization as prerequisites for consent.

When personal and privacy concerns were raised, they primarily fell within the subcategory of concerns relating to commercial interests:

It will always result in advertising and similar things. So, I think there will be enough creative ways within the industry to use health data to reach target groups even more precisely, to market certain products even more accurately, and so on.
[FG1.4]

The 2 participants with the highest income, who have a university degree and are in the middle of their professional lives, explicitly supported the idea of an opt-out solution. However, they linked this to clear conditions that meet their needs for both individual usefulness and individual control.

By contrast, participants with lower and/or middle levels of education and/or lower incomes raised more specific concerns related to personal matters and privacy. As with female participants, the subcategory of concerns related to discrimination or stigmatization was extremely dominant. The fear of job loss or disadvantages with insurance was existential. Another focus was on trust: while the positive social influence of health care professionals and trust in the treating physician were cited as anchors, trust in anonymous, digital systems was low.

Discussion

Overview

For this study, we conducted 2 FGs and analyzed the data using qualitative thematic analysis to explore the perspectives of German patients with somatic and mental health diseases on health data sharing. We explored similarities and differences between the 2 patient groups regarding factors influencing their willingness to share health data for PU and SU. Analogous to the order of the Results presentation, the findings are now discussed in the context of the current state of research.

Discussion of Principal Findings

Previous Experience With Data Sharing

The majority of participants were already experienced in sharing health data, with a predominantly positive response. While 3 participants from each FG reported some negative experiences, they were still willing to share their health data with health care professionals. This suggests that the individual usefulness for medical care can outweigh these negative experiences. In the context of PU, an interview study of the general population in Poland in 2024 showed that positive experiences and frequent use of eHealth solutions are important facilitators for health data sharing [43]. Similar findings regarding SU have been reported in the literature, where 98% of these participants considered the altruistic and public benefits of data sharing to outweigh the potential risks [44].

Individual Usefulness for Medical Care

Patients in both FGs were highly supportive of sharing health data with their health care providers, especially to support personalized medical care. This finding is consistent with previous qualitative studies [27,45,46] postulating benefits for better treatment through comprehensive documentation, such as avoiding multiple diagnoses. However, a scoping review [47] concluded that relevant sensitive information about patients’ mental health was regularly missing from electronic health records because they follow a more narrative reporting style, negatively impacting the flow of information in the treatment process. Nonetheless, in this study 1 patient per FG argued that an electronic patient file should include only objectifiable data, such as laboratory parameters, and not subjective assessments.

These findings are also consistent with 2 qualitative studies [22,48] regarding the potential to increase personal informedness and simplify processes. Personal access to medical records can help patients with mental health diseases improve their understanding of their diagnosis, their mental health status over time, and the reasons for their medical treatment, as well as help them remember their previous visits to the physician. In addition, access to personal data increases the sense of control over one’s own health care [49]. Emergency assistance and time-saving aspects were frequent arguments for health data sharing, especially for younger participants, and also echo previous qualitative studies [43,50].

Public Benefit

In both FGs, the potential for improvements in medical care through the sharing of health data was seen as a key public benefit. This was considered in 2 ways: in terms of drug and product development and in terms of gaining knowledge, for example, for evidence-based public policy decision-making and deriving preventive measures from health statistics. A current study using German survey data and a survey experimental vignette design confirmed that individuals understand how valuable their data are for researchers in terms of SU and possible public benefits [51].

Participants in FG2 particularly supported the work of the pharmaceutical industry and acknowledged the improvements that have been achieved over the years through the development of drugs for the health care of patients with mental health diseases. An interview study involving patients with mental health diseases in the United Kingdom revealed a desire to alleviate their suffering, as well as the expectation of personalized medicine, tailored treatments, and improved disease management in the population [28].

Personal and Privacy Concerns

In this study, especially participants with higher levels of education and/or higher incomes had major concerns about commercial interests, particularly regarding the use of health data by technology companies, which were discussed extensively by FG2. These findings are consistent with the existing literature of other FGs and an online survey in the United Kingdom [46,52,53]. Similarly, concerns about discrimination and stigmatization were more prevalent among the FG2 sample, particularly among participants with lower and middle levels of education and/or lower incomes, and among women. These fears, particularly discrimination by insurance companies and employers in the context of SU, have been previously reported [28]. In a qualitative study, mental health service users expressed concerns about institutional discrimination [52]. The stigma associated with mental health diseases has hindered patients’ acceptance of electronic communication and data sharing, and limited the broader use of this technology among underserved populations [29], making the reduction of stigmatization particularly important. Moreover, participants in both FGs also expressed concerns about being overstrained by the amount of accessible documentation and the wide availability of data, findings that are consistent with those of a qualitative study in Belgium involving vulnerable patients [22].

Consent Management Preferences

Overall, the results of this study show that no clear consensus has emerged regarding preferences for consent management, but fundamental differences were found between the characteristics of diseases and sociodemographic data. The aspects described below may explain our inconsistent findings and highlight the importance of granular personal control over data access, limited datasets, restricting specific information, purposes and recipients, and choices regarding health data sharing. A qualitative study exploring the public’s hopes and fears about health data sharing has shown that these perspectives are strongly influenced by an individual’s personal circumstances [54].

Regarding PU, some patients in both FGs expressed a desire for limited data sharing, while others agreed to full sharing of data. Younger participants expressed very specific preferences regarding which health care professionals should have access to which health data (eg, general practitioner vs specialist). Among other things, mental health diseases were categorized as highly sensitive in both FGs, and consequently associated with restrictions on data sharing or even rejection. A systematic review including survey and interview data also demonstrated greater agreement to share nonsensitive digital health data, with mental health, drug or alcohol use, and sexually transmitted diseases being considered particularly sensitive [55]. Regarding SU, our results suggest that university research often received greater support than pharmaceutical research, and participants in both FGs were explicitly opposed to the use of health data by Amazon, Google, and Apple, although artificial intelligence could transform patient care through its applications and augmentation of medical knowledge [56]. Our results on differentiation by recipient and purpose reflect the findings from studies in Germany [35,57] and other European countries [53,58]. A recent German cross-sectional survey also showed that only about 29% of the population would share their health data with private research institutions, and 66% are unwilling to share their health data with commercial companies, especially not technology companies [45].

While 2 participants with the highest income, who have a university degree and are in the middle of their professional lives, expressed a preference for opting out with their own personal options, others felt that the ability to make fundamental, independent choices was vital. This aligns with studies indicating that greater openness to data sharing is associated with higher digital literacy and socioeconomic status when control mechanisms are in place [18,35,59]. However, concerns about the reliability of our findings must be raised, as the distinction between opt-in and opt-out may have been ambiguous. At times, it appears that opt-out was interpreted as “inconvenient,” as noted by 1 participant in FG1, and might indicate insufficient knowledge on this issue. In an Irish FG study involving 85 participants by Flaherty et al [58], opt-out was considered more convenient because it avoids the need for repeated consent for similar studies and is more cost-effective. Similarly, participants in a recent interview study in Germany emphasized the convenience and effectiveness of opt-out [57]. An interview study with health care professionals for patients with mental diseases also supports the hypothesis of a possible lack of knowledge, where most respondents (75%) felt that their patients did not fully understand the process and consent forms for data sharing [24]. The identified knowledge gap found in this study and elsewhere underscores the need for thorough public information about the different consent options and existing regulations, some of which have already been implemented. At the European level, the European Health Data Space (EHDS) regulation, in force since March 2025 and applicable from March 2027, encompasses different key milestones [60]. By 2029, each member state must have established a national health data access body [61]. Details of the EHDS are currently being elaborated in European projects (eg the Joint Action Towards the European Health Data Space 2) [62]. A workshop focused on the perspectives of different stakeholders on the opt-out procedure within the EHDS, as well as how Germany can structure the opt-out procedure and rights of objection within this framework [62,63].

Legal and Ethical Framework Conditions

The issues of data protection and ethical justification for data sharing were particularly noteworthy to both of our FGs. The participants in FG2 strongly emphasized that health data are generally very sensitive and need special access protections. Patients with somatic diseases, particularly male patients, requested special monitoring and authorization measures. Although data protection in Germany is already very robust and a deliberate, data protection–focused approach is followed, there appear to be evident concerns and a lack of knowledge [11]. In an FG study in the Netherlands, Wetzels et al [64] showed that patients had limited knowledge of data protection practices. These results demonstrate the need for broader and targeted public relations work to better educate the public on this issue. Germany provides information for the interesting public on websites, interactive learning platforms, and in journals, which also include information on data protection [65-70].

Informational Self-Determination

Especially the points of autonomy and control were fundamental for the male and older participants aged 49-79 years. These 2 aspects, as well as trust, were consequently identified as facilitators of health data sharing in both FGs and elsewhere [71,72]. A Finnish study using survey data to develop a theoretical model found that trust and control are key mediators between attitudes and willingness to share health information, with trust being more influential than control [73]. However, a questionnaire study from Denmark revealed a strong desire for control over personal health data, even when there is trust in the institutions using it [74]. In a computer-assisted telephone interview study conducted in Germany, trust emerged as the most significant factor to increase the willingness to share personal health data [59].

Social Involvement and Influence

Positive social influence of health care professionals and trust in the treating physician can encourage health data sharing according to 2 participants in FG2 with lower levels of education and/or lower incomes. The existing literature certainly indicates a substantial influence of physicians’ behavior on patients’ willingness to share their health data [75]. In patients with mental health diseases, physicians may underestimate the importance of physical complaints and attribute them to a mental disorder, which can lead to misdiagnosis and poorer treatment outcomes. This phenomenon of “diagnostic overshadowing” can lead to a loss of trust in the treating physician, particularly in those with mental health diseases, reducing their willingness to share their health data [52,76-78]. Positive encouragement and clear, open communication strategies from health care professionals can be important facilitators and were found to significantly influence outcomes [79]. Specifically for this purpose, a patient handout in various and simplified languages, as well as waiting room videos and posters for German medical institutions, were developed [70]. Moreover, German statutory health insurance companies provided their patients with comprehensive information materials [80], and an interactive learning platform for experts is gradually being established [68].

Strengths and Limitations

A major strength of this study is the diversity of its cohort in terms of gender, age, and somatic and mental health diseases. The diversity of these characteristics was considered during the recruitment process, as it allows access to different perspectives from patients with different chronic and mental health diseases. In particular, the opinions of patients with mental health diseases are still rarely prioritized in research on health data sharing, even though they represent more than one-third of the population [81]. The group dynamics and observable interactions in FGs provide valuable and deeper insights than face-to-face interviews. Another strength of this study is that the FGs were able to show how participants’ views are formed and how they change through social interaction. This allowed us to better contextualize the verbal data.

However, our findings may have been biased by problems recalling experiences in the distant past. Our participants also represent a convenience sample, as participants were recruited from the University Hospital in Dresden, Germany. In addition, our cohort may have been affected by selection bias due to the financial compensation of €50. Moreover, investigated relationships between sociodemographic characteristics like age, gender, education, and income levels as possible influencing factors may only be considered of limited significance due to the small sample size. Furthermore, our recruitment of only relatively tech-savvy patients who used a smartphone or tablet daily to increase FG efficiency may have introduced selection bias, potentially resulting in higher technology acceptance scores. Besides, our results may have been influenced by volunteer bias. In addition, our results may have been influenced by social desirability, as participants may have feared being judged by other FG members if they provided truthful answers. Finally, qualitative research findings are not designed to be generalizable. However, an exploratory study design enables patient participation, as emphasized in the GDNG [1], and allows for a deeper and more comprehensive understanding of patients’ perspectives, as well as facilitating the development of hypotheses to inform future research and practice.

Implications for Future Research and Practice

Further validation of the findings of our exploratory qualitative study on data sharing can be achieved through future research with a representative sample. This research could also reveal possible correlations with education, income, or place of residence, which have already been identified at the international level [9]. Future qualitative and quantitative research comparing the attitudes of patients with somatic and mental health diseases, as well as healthy individuals, could deepen our understanding of their unique needs and concerns. Investigating the attitudes of individuals with poor mental health could also be beneficial, despite the potential challenges of recruitment.

Social influence of family members and especially health care professionals had a much greater impact on patients with mental health diseases and for those with a lower socioeconomic status. Therefore, the latter should be involved in future research and, after the aforementioned training measures, be empowered in their role as multipliers to inform patients about the potential benefits and risks of sharing health data. Research should also assess the effectiveness of involving family members, peers, and health care professionals in decision-making processes. This research can refine governance frameworks and improve patient-centered approaches to health data sharing. It can also lead to powerful medical artificial intelligence systems, such as clinical decision support systems, that are trained on higher-quality and more representative data.

The issue of data protection was more important for patients with mental health diseases. Both they and the female participants expressed greater fears of discrimination and stigmatization. To reduce the identified knowledge deficits, which have even become apparent among tech-savvy patients, it is crucial to implement broad education and public relations initiatives that provide easy-to-understand information about the very high data protection standards and the procedure for storing and transferring health data. In addition, there needs to be greater awareness of the different levels of consent options, such as only allowing certain physicians to access individuals’ electronic health data.

Conclusions

The use of health data for PU and SU is a valuable capability of a modern health care system and the society it serves. Therefore, patients’ and health care professionals’ perspectives must be considered when designing the regulatory framework. This study provides insights into the different perspectives and preferences of patients with somatic and mental health diseases regarding data sharing. In particular, those with mental health diseases viewed their health data as highly sensitive due to previous experiences and fears of discrimination and stigmatization. Therefore, tailored consent management and involving patients in health data sharing were highly preferred. Participatory approaches involving family members, peers, and health care professionals can also help to reduce fears and increase acceptance. Health care professionals play an important role, as they can ensure transparency through targeted educational work and actively sensitize patients to the principles of data sharing, which can be supportive, even if commercial interests are involved. The knowledge deficits identified in this study, which also existed among tech-savvy patients, indicate that broad and easily understandable awareness-raising and public relations initiatives are needed. They can help inform the population comprehensively about the very high data protection standards, the procedure for storing and transferring health data, and the personalized consent options.