Regulation (EU) 2017/745, also known as the Medical Device Regulation (MDR), is highly debated. The focus of the literature concerning the MDR mainly centers on its challenges and complexity [1,2]. The MDR is associated with strict regulatory requirements [3], the involvement of multiple stakeholders [4], delayed market access [5], ambiguity [6], and a financial burden [7]. The expenses related to the MDR include costs for personnel undertaking activities to meet the requirements of the MDR. In addition, the setup, maintenance, and certification of the quality management system (QMS) are significant cost drivers. For products exceeding risk class I, fees for the conformity assessment conducted by a notified body must be considered [8]. A significant cost driver is the clinical investigation [9]. It requires a substantial amount of time, expertise, and potential fees for the study site. The MDR considerably impacts the business model of software as a medical device (SaMD) startups, which are challenged to incorporate the regulation into their business model. SaMD describes software intended to be used for at least 1 medical purpose, performing these purposes without being part of a hardware medical device [10]. SaMD ranges from solutions primarily used by patients for treating medical conditions, such as depression, to solutions that support physicians in diagnosing and preparing treatment plans [11]. SaMD also includes platforms used by physicians to monitor patients’ medical conditions [11]. Simpler apps used by patients are often associated with lower risk compared to complex diagnostic tools used by physicians. For lower-risk devices, the regulatory burden under the MDR is reduced, such as in the postmarket surveillance process [12] or the involvement of a notified body [12,13].

The requirements of the MDR apply to all medical device manufacturers irrespective of their company size [14]. Consequently, they seem to disproportionately impact small-to-medium enterprises and startups [9]. Often inexperienced and possessing limited resources, they face considerable barriers [9]. Startups struggle with the complexity of the requirements and their lack of regulatory understanding [12,15]. In addition, startups at an early stage find it difficult to grasp the implementation of a QMS that requires structured processes and extra effort [14]. Finally, the clinical investigation, if required, is especially challenging for many startups due to its financial costs and time investments [14,15].

Ironically, startups foster more radical innovation [16] and are, therefore, a driving force in the SaMD industry. The flexibility and adaptability of startups are a door opener for digital innovations [17]. The strict and complex MDR, on the one hand, and the innovative and flexible culture of startups, on the other, create the impression of a clash between 2 distinct cultures [14]. Understandably, this confrontation puts startups with obstacles many cannot tackle, thereby jeopardizing innovations in the SaMD industry. According to Stern [18], industry regulatory requirements are frequently linked to firms postponing or diminishing market entry due to increased time and costs. For smaller firms, it is more challenging to act as pioneers in medical device markets due to the relatively higher costs and the greater financial constraints compared to established companies [18]. Nevertheless, the MDR plays a crucial role in shaping industry standards for medical devices in the EU and globally. Bradford [19] describes the power of the EU to project its regulatory standards to the rest of the world as the “Brussels effect.” For example, Canada based its medical device regulations on the standards of the EU Medical Device Directives [20], the predecessor of the relatively new MDR. The MDR aligns with international industry standards, facilitating market access beyond the EU.

Although the literature on the challenges of the MDR, specifically for SaMD startups, is profound [1-3], possible solutions to mitigate them are missing. Business model frameworks are tools that facilitate business modeling. By focusing on the key aspects only, they reduce complexity while ensuring that no crucial business components are missing. Consequently, especially in complex settings, business model frameworks are perceived as powerful tools. The situation in which startups aiming to develop SaMD are embedded is intricate, with the MDR being a significant challenge. A business model framework that directly incorporates the MDR might assist them in deriving a sustainable business model and mitigating the inherent complexity of the MDR.

In the literature, a couple of business model frameworks can be found. A widely recognized [21] and used tool [22] is the Business Model Canvas by Osterwalder and Pigneur [23]. Having no industry focus makes it a widely applicable but unspecific tool for business modeling. According to Albert and Van der Auwermeulen [24], general business model frameworks, such as the Business Model Canvas, do not sufficiently account for the specifics of the health care sector. The nuances of the health care sector require a specialized business model framework [24]. Therefore, there are increasing attempts to develop health-specific frameworks. One example is the framework for the telehealth sector by Velayati et al [25], which is based on interviews suggesting key components of a telehealth business model. Mueller [26] developed a mobile health business model framework that considers regulations. However, regulation is only included as a component and is not specified further. Outside of the digital health industry, regulations have been discussed. For example, Lindgren [27] researched the impact of General Data Protection Regulations on business models, but did not derive a business model framework to work with. Ademi et al [28] focused on the development of sustainable business models due to increased regulations and changing customer priorities. Still, literature regarding business modeling in regulated markets is in its infancy. More work is needed to understand the interplay between business models and regulations, particularly in highly regulated industries. To address this research gap, the primary objective of our work is to develop a comprehensive business model framework tailored for startups venturing into the development of SaMD in the EU. To achieve this pivotal goal, we formulated 3 research questions (RQs):

The conceptual business model framework developed in this work provides a condensed solution for startups developing SaMD. The framework incorporates the key components these startups need to consider while directly integrating the specifics of the MDR. Especially startups in the early stage tend to overlook key aspects [14]. A business model framework can be instrumental in this context. Our work can assist SaMD startups in successfully entering the SaMD industry in the EU with a sustainable business model. The hurdle to entering the industry and the risk of failing during the Conformité Européenne (CE) marking process can be reduced. Theoretically, this work adds to the current discussions on the MDR by investigating its impact on business modeling and systematically incorporating it into a business model framework. In the literature on the MDR, the focus seems to be primarily on its challenges. Our work aims to develop solutions to tackle some of these challenges, thereby contributing to a rather unexplored research avenue.

The remainder of this work is structured as follows: the subsequent section provides background information on the MDR and clarifies related terms. The Methods section presents the research methodology, followed by the study’s results, including the derived conceptual framework. Finally, the key findings are discussed before we conclude the work.

The World Health Organization defines digital health as the practice of using information and communication technology to meet health care needs [29]. Digital health includes concepts such as eHealth and emerging areas such as artificial intelligence [29]. Digital health technologies include SaMD [30]. In the EU, SaMD, which is not an in vitro medical device, is regulated under the MDR.

Aiming to increase the performance and safety of medical devices placed on the EU market [31], the MDR entered into force in 2021 [32], replacing the less stringent previous Medical Device Directive [33]. Multimedia Appendix 1 provides the definition of a medical device according to article 2 of the MDR. SaMD is software intended to be used for at least 1 medical purpose, performing these purposes without being part of a hardware medical device [10]. The software is intended to be used for medical purposes, such as diagnosing or treating diseases. The initial baseline for further steps in developing and approving medical devices is the determination of the intended purpose [34]. The intended purpose describes the designated use of the device as intended by the manufacturer [35]. For a device to qualify as a medical device, its intended purpose must be a medical purpose, in accordance with the definition of a medical device under the MDR (Multimedia Appendix 1).

A startup with a product that falls within the scope of SaMD in the EU needs to fulfill the safety and performance requirements specified by the MDR. These requirements include a QMS and technical documentation, among others [35]. The QMS captures processes and procedures to ensure the safety and performance of medical devices [31]. The technical documentation demonstrates their compliance with the defined safety and performance requirements [34]. Manufacturers are advised to apply harmonized standards and common specifications [35,36]. This helps prove compliance with the requirements and provides guidance on the necessary actions for the CE marking. Table 1 depicts the MDR’s key regulatory requirements and related standards relevant to SaMD startups. Notably, the standards mentioned in the table are not compulsory. However, they reflect the state of the art and are commonly used.

Although the requirements illustrated in Table 1 apply to all medical devices, the degree of detail is proportionate to the risk class. Medical devices can be classified into the following risk classes: I, IIa, IIb, and III [34]. The higher the risk class, the more stringent regulatory scrutiny is required. Compared to higher-risk classes, risk class I devices do not require a notified body for conformity assessment [13]. This saves the SaMD startup time and money. For risk classes III and implants, the MDR requires a clinical investigation [35] to demonstrate the safety and performance of the medical device. In addition, risk class III devices require a scrutiny procedure, increasing the regulatory burden. Thus, the risk class determines the regulatory burden for SaMD startups. It also affects the business model, for example, in financial terms, due to the required clinical investigation. Furthermore, it is significant to note that the requirements of the MDR are highly interdependent. For example, risk management and software development impact each other and must, therefore, be closely integrated to fulfill the safety and performance requirements of SaMD.

Our research design comprises 3 phases, as depicted in Figure 1. The basis for our study is the analysis of MDR requirements for SaMD. To understand MDR-related aspects, we mainly included the MDR itself [35], the standards mentioned in Table 1 [37-41], and guidance documents provided by the Medical Device Coordination Group. The guidance documents covered the documents on clinical investigation [42] and risk classification [43]. Additional sources included the works of Baumgartner et al [44] and Hastenteufel and Renaud [34].

Armed with the regulatory background information, we conducted a systematic literature review. We supplemented the results with qualitative data gathered through interviews with startups and experts in the EU SaMD industry. The input from the MDR, together with the results derived from the literature research and the expert interviews, served as the foundation for the development of the SaMD business model framework. An initial prototype was discussed within the research team, and feedback from potential users was gathered, which led to the framework presented in this work. Figure 1 depicts the overall research design of our study.

A systematic literature review is a structured and transparent approach to synthesizing research [45]. Kraus et al [46] define 4 steps of a systematic literature review, which were followed in this work: planning the review, identifying and evaluating studies, extracting and synthesizing data, and disseminating the review findings.

We conducted semistructured interviews to identify aspects of the business model that startups and experts in the EU SaMD industry perceived as important. The literature on business modeling for SaMD is a relatively new research avenue. Therefore, we had to expand the literature review to encompass literature on digital health in general. Interviews were crucial for our study to account for SaMD specifics. They were valuable in driving the development of the framework and enriched it with practical insights, including lessons learned and hints.

We searched for interviewees through web searches. Our focus was on SaMD startups in the EU and experts, such as startup coaches in incubators or accelerators in the SaMD industry. We aimed for a well-balanced mix of founders and industry experts. Potential interviewees were contacted via email or professional networking platforms. For the interviews, we followed a semistructured approach [92]. Semistructured interviews allow for flexibility in structure [92,93] and follow-up questions by the interviewer [94]. Therefore, complex topics can be explored, and an in-depth investigation of a topic is possible. Most interviews were conducted with chief executive officers and founder or cofounders of SaMD startups, which was ideally suited for our study and significantly augmented its value. The product range of the respective startups included various solutions, such as software to monitor and treat eye diseases or software for tracking skin conditions. In addition, opinions from experts with experience across a wide range of SaMD were included. An overview of the interviewees is provided in Multimedia Appendix 3. The interviews were recorded and transcribed. Subsequently, they were inductively coded following the approach by Gioia et al [95]. To ensure the objectivity of the coding, 2 researchers analyzed the transcripts and carried out the coding. Inconsistencies were discussed to derive common, aggregated dimensions.

The interviews were conducted in accordance with the relevant ethics guidelines and were exempt from undergoing an ethical review. All participants provided informed consent and agreed to the interview. Data collected from the interviews were deidentified to ensure confidentiality.

The analysis of the final 45 papers demonstrated a substantial number of papers that developed business models based on the Business Model Canvas by Osterwalder and Pigneur [23], such as the papers by Bartels et al [47], Chen et al [48], and Christie et al [49]. Apart from the Business Model Canvas, the service, technology, organization, and finance (STOF) model by Bouwman et al [96] was applied to specific use cases [50,51]. Only a few authors derived their own framework [52,53]. Hwang and Christensen [52] developed a business model framework for health care, incorporating 4 components: value proposition, resources, processes, and profit formula. Lin et al [54] established a telemedicine framework and identified value proposition, partnership, resources, capability, service process, and cost structure as components. The remaining papers referred to other existing business model tools [97-102]. Multimedia Appendix 4 [25,26,47-89] provides an overview of the 45 papers, including their applied or derived business model frameworks.

Following the approach of Webster and Watson [90], we derived 14 business model components summarizing the constructs mentioned in the 45 papers. Multimedia Appendix 5 [25,26,47-89] depicts the concept matrix. Table 2 describes the 14 derived business model components. Of the 45 papers identified, only the study by Mueller [26] included regulation as a component, underlining its underrepresentation in the current digital health business model literature. The MDR, in particular, was not mentioned in any of the papers. Logically, no publication specifically focused on the business model of SaMD, underpinning the research gap. As shown in the concept matrix, value propositions and customer segments were frequently mentioned in the papers. This highlights their importance as key business model components.

The general business model components identified through the systematic literature review, on their own, do not sufficiently account for the specifics of the health care system. As Albert and Van der Auwermeulen [24] stated, the health care system is complex. Generic business model approaches, such as the Business Model Canvas, are inadequate to fully address its intricacies [24]. Therefore, to derive a business model framework tailored to SaMD startups, we enriched the results from the systematic literature review with expert interviews from the SaMD industry in the EU.

Multimedia Appendix 6 depicts the inductive coding results of the interviews, according to Gioia et al [95]. The interviews made the MDR a focal aspect of SaMD business model (interviewee 10, interviewee 12, and interviewee 11). It was advised that the MDR should be considered from the beginning (interviewee 12). The interviews unequivocally demonstrated the impact of the MDR on other business model components, such as human resources, as part of core assets (interviewee 2 and interviewee 4). The MDR requires expertise that many startups lack. Therefore, ensuring the availability of human resources with the necessary knowledge to fulfill the requirements of MDR is crucial. In the interviews, the position of the quality manager was emphasized as essential (interviewee 6 and interviewee 4). It was recommended that a full-time position for a quality manager should be planned (interviewee 4). The MDR mandates the provision of “sufficient clinical evidence” [35]. However, the term sufficient is left open to interpretation, leading to ambiguity (interviewee 8). In such cases, startups often find it beneficial to seek advice from experienced consultants. Consultants can provide valuable insights specific to their case. In addition, the immense financial burden of the MDR was highlighted (interviewee 5 and interviewee 10). The CE marking process is lengthy (interviewee 8). During this period, financial sustainability must be maintained. Furthermore, clinical investigations have been mentioned as significant cost drivers (interviewee 5 and interviewee 8). With its limited resources, startups often turn to consulting companies seeking assistance with the MDR (interviewee 8). This adds to the already high overall costs associated with the CE marking. A financial analysis (interviewee 11), including follow-up funding (interviewee 4), seems essential. In addition, securing follow-up funding early (interviewee 4) was mentioned to avoid falling into a financial gap. Finding investors is challenging due to the long CE marking process and the resulting delayed revenue generation (interviewee 10). Investors need “some element of faith or guided imagination of where it can go” (interviewee 7). Therefore, it is essential to start the search process early on (interviewee 7).

Furthermore, the revenue model was emphasized as pivotal for the business model of SaMD startups (interviewee 8). Having the digital solution CE certified opens the door for additional revenue pathways, such as reimbursement by health insurance companies (interviewee 10). Because reimbursement systems are country specific, possible reimbursement pathways need to be considered for the target country (interviewee 10). Further revenue options are the self-payment market (interviewee 11), which is also country specific. For example, Germany was mentioned as a country with a low self-payment mentality (interviewee 10). On the contrary, Poland was named as one characterized by a relatively high self-payment mentality, where people are accustomed to spending money on their health care (interviewee 10). Thus, awareness of revenue sources and their potential in the target country is central to a viable revenue model.

Understanding the health care market was described as challenging (interviewee 6). Comprehending how the market works and identifying the biggest players and their relations were mentioned as significant obstacles for startups entering the health care sector (interviewee 6). The intricate stakeholder constellation was noted by several interviewees (interviewee 10). Understanding the interests of stakeholders, together with operational dynamics, was recognized as an obstacle (interviewee 10). In this regard, health insurance companies were highlighted as stakeholders whose functioning, interests, and expectations toward SaMD startups were difficult to grasp (interviewee 10). As a result, startups struggle to derive a value proposition for health insurance companies (interviewee 10). Because value propositions are to be considered for each customer segment from all angles (interviewee 11), the health insurance company needs to be addressed in case of a reimbursement revenue model.

Interviewee 9 highlighted the importance of competitor analysis. Distinct competitive advantages should be identified early (interviewee 9) to align the product accordingly. The complexity of the health care sector becomes evident once more when looking at customer segments. For SaMD startups, the user is often the patient, while the payer is the health insurance company (interviewee 6). This triangle of SaMD startup, health insurance company, and patient requires consideration. Interviewee 3 and interviewee 5 mentioned customer relationships, indicating the importance of reaching out to customers as early as possible to obtain feedback (interviewee 5). Interviewee 3 recommended creating a lock-in effect to bind the customer to the product. Key MDR-related activities mentioned were the QMS (interviewee 4) and the clinical evaluation (interviewee 5). The clinical evaluation was seen as unclearly formulated in the MDR. Terms such as “equivalent device” [24] were insufficiently described and left room for interpretation (interviewee 10). Apart from MDR-related activities, further key activities mentioned included research (interviewee 9) and the setup of a sales team (interviewee 5).

The MDR is a complex and challenging topic, especially for startups with limited resources. As the expert interviews demonstrated, the impact of the MDR on business modeling for startups aiming to obtain the CE marking for their SaMD is vast; however, it is underresearched, as the systematic literature review has illustrated. Therefore, the objective of this study was to understand the relationship between MDR and business modeling and to derive a business model framework that supports SaMD startups in developing a business model. To achieve this goal, we followed a 3-step approach. First, we conducted a systematic literature review to identify business model components relevant to the digital health sector. In total, 45 papers were identified. The review highlighted the lack of analysis of business models specific to SaMD startups. To make our conceptual framework tailored to SaMD startups, we conducted, in a second step, 13 expert interviews with specialists in the EU SaMD industry. We enriched the results with feedback from potential users and the research team on an initial prototype. Furthermore, we supplemented the results with the requirements of the MDR, including related literature and state-of-the-art standards. On the basis of this input, we conceptualized the SaMD business model framework depicted in this work.

The interviews and the literature unveiled that existing business model frameworks, such as the Business Model Canvas by Osterwalder and Pigneur [23], reach their limits regarding industry-specific cases. SaMD startups are embedded in a complex digital health care system with multiple stakeholders and a tricky customer constellation. The situation of SaMD startups is challenging due to many factors, and implementing the MDR in addition to these factors makes the challenge even greater. For business modeling, the MDR can be seen as a driving force impacting the entire business model, while being a pivotal component itself. Therefore, the SaMD business model framework is designed from a startup’s perspective, with the aim of developing SaMD business model. The MDR is considered in each derived business model component, facilitating business modeling for SaMD startups. In addition, the visual layout of the framework delineates the relation between business model components and demonstrates how the MDR fits into this constellation.

The MDR is not the only regulation that SaMD startups need to consider. Country-specific laws on marketing, taxes, or labor, as well as EU-wide regulations, such as the General Data Protection Regulation, need to be considered. Undoubtedly, these regulations have a ripple effect on business modeling as well. However, among all these regulations, the MDR seems to hold a special position because of its complexity, setting a series of consequential tasks. Evidently, in the interviews with SaMD startups and experts in this industry, the MDR has been highlighted as a hurdle in business modeling (interviewee 10 and interviewee 11). Furthermore, the regulatory value arc depicted in Figure 3 illustrates the close interrelation between the intended purpose, value propositions, and customer segments, underlining the embeddedness of the MDR in the business model of SaMD startups. Consequently, the MDR’s pivotal role needs to be directly integrated into the business modeling of SaMD startups.

Furthermore, we want to mention the identified regulatory value arc as another principal finding of our study. The trio of intended purpose, value propositions, and customer segments is the centerpiece of the framework and warrants particular attention. Although the link between value propositions and customer segments is well known, we identified the notable impact of the MDR on this duo, highlighting the extensive role of the MDR in business modeling. The intended purpose describes the designated use of the device as intended by the manufacturer and needs to cover at least 1 medical purpose to qualify as a medical device [35]. The user, use environment, potential contraindications, and claims, along with possible disclaimers, are formulated apart from the medical purpose. The description of the users and the link to customer segments are evident. Besides, claims are statements that can be used to advertise SaMD. Claims should reflect the value offering desired by the customer to ensure customer needs are met with the SaMD. Therefore, the intended purpose, as the cornerstone of the MDR, is the pivotal element bridging value propositions and customer segments, forming the trio “regulatory value arc”. With the intended purpose, a connection between the company-internal side of the business model framework and the external side of the framework is established. Claims formed in the context of the intended purpose are carried out via channels and customer relationships from the company to the customers. Thus, for a startup aiming to develop SaMD, it is imperative to consider the regulatory value arc.

In the interviews and the literature, the clinical investigation was mentioned as a significant hurdle in the context of the MDR (interviewee 10). Studies are expensive and time consuming. They slow down the launch of the SaMD on the market. Data from clinical investigations are mandatory for risk class III and implantable medical devices [109]. For the remaining risk classes, a clinical investigation does not need to be undertaken if sufficient clinical data for performing the clinical evaluation can be obtained from literature or other sources from equivalent devices. Following this line of reasoning, innovative products are unlikely to refer to existing clinical data, requiring them to conduct their own clinical investigations. In that sense, startups with innovative products have ironically a disadvantage with the MDR compared to less innovative products. For less innovative products, clinical data are often already on the market.

Each clinical benefit claim included in the intended purpose needs clinical evidence. Reflecting on the challenges of the clinical investigation mentioned earlier, it is worth trying to circumvent the clinical investigation by adapting the intended purpose. This consideration must be thoroughly evaluated. Possible advantages include cost and time savings, while potential disadvantages may involve an adjusted product focus and challenges in collaborating with health insurance companies.

Our work has several theoretical and practical contributions. From a theoretical perspective, our work contributes to both digital health literature and business model literature. Although the literature on business models is vast, the impact of regulations on business modeling has not been sufficiently investigated. Business model frameworks, such as the Business Model Canvas by Osterwalder and Pigneur [23], do not explicitly include regulation in the framework. This can be fatal when regulations such as the MDR occupy a pivotal role that significantly impacts the business modeling process. In our work, we investigated the impact of the MDR on business modeling for SaMD startups. Our work clearly showed that the MDR plays a central role in the business model of SaMD startups and needs to be considered from the beginning. This study lays a foundation for a research stream examining business modeling for SaMD startups specifically and the relationship between regulations and business models. Furthermore, the literature on SaMD startups mainly focuses on their challenges, especially in the context of the MDR. Concrete solutions for how these challenges can be tackled are missing. Our work provides a way to address some of these hurdles. This work lays the groundwork for a solution-oriented approach to tackling SaMD startups’ challenges by providing a SaMD business model framework.

From a practical perspective, the derived SaMD business model framework provides a hands-on and all-in-one solution for business modeling. The framework enables the description and design of a business model for SaMD startups. The medical device industry is complex and is associated with hurdles for startups. Maintaining oversight in this complex setting is challenging. The framework can reduce complexity and minimize the risk of overlooking key aspects. It covers all factors relevant to developing a business model for SaMD. From the perspective of SaMD startups, each business model component directly incorporates MDR-specific information. Furthermore, with the visualization of the framework, an overall picture of business modeling for SaMD startups becomes evident. Business model frameworks provide a level of customizable abstraction that can be adjusted to fit the specific use case [110]. They provide a common language [110] supporting mutual understanding within the startup and its partner network. Moreover, business model frameworks are often used for pitching to investors. This framework can assist in pitching in front of them. In this way, the issue of attracting suitable investors in the SaMD industry is addressed in this work.

Ideologically, the concepts of startups and MDR seem to be colliding [14]. Startups, often young, dynamic, and packed with innovative ideas, see themselves confronted with a strict, documentary-rich MDR, which slows down innovation. Startups often lack regulatory experience and have limited resources, whereas the MDR requires expertise and substantial financial and human resources. The SaMD business model framework aims to bring these worlds together for the benefit of both parties. The framework assists startups in integrating the seemingly opposite MDR world into their daily business.

The framework has been derived based on the needs of SaMD startups and, therefore, provides a tailored, practical solution to them. Beyond stand-alone SaMDs, hybrid models involving software and hardware are possible. Combining hardware with software makes it necessary to clarify whether hardware, software, or both fall under the category of medical devices. In addition, with a hardware component, the MDR requires the consideration of specific hardware standards, such as the International Electrotechnical Commission (IEC) 60601-1 on medical electrical equipment. Therefore, core activities in a hybrid model differ from pure stand-alone software. For example, hardware production must be considered, resulting in higher investments compared to stand-alone software. Distribution to customers is different for hybrid solutions as well. The derived SaMD business model framework, in its basic structure, can be applied to hybrid models that combine software and hardware. However, the guiding questions do not entirely align with the needs of startups working with hybrid models and would require adjustments.

This work is not without limitations. First, our work focused only on the EU region. The framework might not be applicable for startups aiming to bring SaMD to the market in non-EU countries. This is due to differences in the setting, such as a distinct regulatory environment. Second, our research focused on stand-alone software. Therefore, we did not incorporate cases that include hardware. Future research could adapt the derived framework to be tailored to hardware medical devices or embedded software. Third, our work focused on Regulation (EU) 2017/745. Regulation (EU) 2017/746 on in vitro diagnostic medical devices was not targeted. Fourth, the digital health sector is fast paced and changes rapidly. Our work provides a snapshot of the current setting for SaMD startups in the EU. Finally, the framework presented in this work has not been evaluated yet. Apart from experts in the EU SaMD industry, seed and early-stage SaMD startups, as the target users of the framework, should apply it in practice to validate its usefulness. The framework could, for example, be shared with accelerators, where startups develop the business model for their SaMD. An observation protocol could gather insights into how startups work with the framework. In addition, these startups could be interviewed to evaluate the framework based on predefined criteria. In addition, interviews could be conducted with startups that are in a later stage. These SaMD startups have already developed a business model for their SaMD and can offer valuable insights and feedback based on their experience. The validation of the framework could be a potential avenue for future research.

The field of SaMD is rapidly evolving and offers various opportunities. In our work, we analyzed business modeling in the EU SaMD industry, focusing on startups as a driving force in this industry. Initially, we conducted a systematic literature review and derived a concept matrix. The work was then supplemented with expert interviews. In addition, the requirements of the MDR relevant to SaMD startups and related state-of-the-art standards were considered to develop the SaMD business model framework. Our study is a first step toward mitigating the challenging situation of SaMD startups. Our contributions offer valuable insights into the relationship between MDR and business modeling. The work provides a hands-on solution for startups aiming to develop SaMD under the MDR. Our study could assist SaMD startups in their business modeling process and reduce complexity.