We conducted a qualitative study on the perspectives of stakeholders about the adoption of VCTs. The Consolidated Criteria for Reporting Qualitative Research (COREQ) checklist for reporting qualitative research was followed (Multimedia Appendix 1) [25]. Approval for the qualitative study was granted by the Internal Review Board of TNO in April 2020 (reference number: 2020-034).

To base our investigation on a solid theoretical framework, we utilized Rogers’ diffusion of innovation theory. Rogers’ framework aims to explain how and why a new innovation or technology is, or is not, adopted, where adoption can be defined as “the full use of an innovation [when the innovation] is the best course of action available” [26]. A VCT can be considered a process innovation, which is defined as the implementation of a new or improved delivery method, including substantial changes in techniques, equipment, and/or software [27]. Zhang and colleagues [28] reported their investigation of the factors leading to a successful eHealth process innovation. They used Rogers’ theory to analyze patient acceptance and identify reasons for the utilization of eHealth innovations [28]. Several other recent studies have also used Rogers’ framework to better understand health technology adoption [28,29]. Therefore, Rogers’ theory was considered to be an appropriate framework for this study. Multimedia Appendix 2 [30-33] provides more details about Rogers’ theory and concepts, including the five stages of the innovation-decision process.

A purposive sampling strategy was used to select invitees [25]. Three stakeholder groups were approached via email: invitees were recruited from TNO (within which people working in the health division were approached), pharmaceutical organizations, and food and health organizations. Within TNO, the Microbiology and Systems Biology (MSB) department supports the research and development activities of companies in the agriculture and food, health, personal care, chemistry, biotechnology, and pharmaceutical sectors. The department collaborates with pharmaceutical, food, and health organizations as well as other companies to set up clinical trials [24]. Within the pharmaceutical industry, a distinction between sponsors and contract research organizations (CROs) can be made. Food and health companies can be defined as organizations that aim to improve health by healthy foods. These organizations study nutrition by conducting nutrition trials.

Twenty potential participants received an interview invitation. Due to the COVID-19 pandemic, all interviews were held by telephone or Skype. Invitees were enrolled in the study if they met the following inclusion criteria: speak Dutch or English fluently; work for one of the three stakeholder groups (research organization, pharmaceutical industry, or a food company) in a key position of clinical trial execution; have experience and/or interest in using VCTs; and are willing and able to participate and sign an online informed consent form.

The semistructured interviews were carried out by author RMHC. The interview guide consisted of themes and questions based on the five innovation stages as suggested by Rogers, namely: knowledge, persuasion, decision, implementation, and confirmation (Multimedia Appendix 3). The interview guide was discussed with and approved by all authors. Furthermore, the interview was piloted with 2 independent researchers at TNO [25]. Once the invitees had given consent, the interviews were conducted and audio recorded. Field notes were made during and directly after the interviews.

The audio recordings were transcribed immediately after the interview, new concepts were identified, and the degree of data saturation was measured [34-37]. In addition, to confirm the accuracy of content and key messages, participants were asked to review a brief summary of the interview and a couple of quotes, and comment if necessary [35,37]. The recordings were deleted directly after transcription, and the transcripts were stored as Microsoft Word files (Microsoft Corp) on a secured laptop per proper data and privacy protection measures. The data were analyzed by using the thematic analysis method suggested by Clarke and Braun [38], which consists of the following 6 steps: familiarization with the data, initial code generation, theme identification, theme review, theme definition and naming, and, finally, report production [38]. The data obtained from the interviews were coded by RMHC and merged into a code book. A 5% check (1 interview transcript and coding) was done by another researcher to ensure that coding was performed consistently. The data from the interviews were descriptively analyzed with Microsoft Excel 2016 (Microsoft Corp).

From the 20 potential participants who were approached, 4 individuals were excluded from this study; 2 did not respond to the invitation, and 2 were not experienced enough in VCTs and thus did not met the inclusion criteria. In total, 16 individuals were interviewed between April and June 2020. Six employees from a research organization, 6 pharmaceutical employees, and 4 individuals from food companies were enrolled in the study. Among the pharmaceutical participants, one was an employee of a CRO. This person was grouped into the pharmaceutical cohort since the results of the interviews between the CRO and pharmaceutical companies were much aligned. All interviewees from the research organization lived in The Netherlands, while 2 pharmaceutical interviewees and 2 interviewees from food companies were based in Ireland, Denmark, France, and Germany (Table 1). In Table 2, the type of organizations where interviewees were employed is shown.

Of the 6 individuals affiliated with the research organization, 5 were business developers. Business developers were included since they are the initial point of contact with business partners and discuss different options for conducting studies. They have ideas about how VCTs could be interesting for their clients and whether the research organization should be involved in VCTs. The remaining individual was a senior scientist conducting clinical research in collaboration with pharmaceutical companies and universities.

The 6 participants from the pharmaceutical companies had various professional backgrounds. Two were country study managers leading a team that conducts clinical trials. Moreover, they are responsible for the quality of the research. Two participants worked within an innovation department. One was the medical lead of innovation, and the other worked as the head of clinical technology and innovation. The medical lead of innovation was responsible for several (clinical trial) innovation projects within the company. The head of clinical technology and innovation was closely involved with conducting VCTs. Another interviewee was a clinical scientist, who was involved in conducting clinical trials. The last participant was a medical advisor and data scientist, and was responsible for coordinating clinical research and analyzing clinical trial data.

Three of the 4 interviewees from the food companies were (clinical) scientists and/or study leaders. They were responsible for conducting or leading clinical trials. The remaining participant was head of scientific marketing within a food company, who conceptualized, planned, and implemented clinical research. Figure 1 displays the professional backgrounds of the participants.

A total of 12 interviews were conducted by telephone and 4 by Skype (with video). Interviews lasted approximately 45 minutes. After the 15th interview, data saturation was reached. Data collection was continued for one more interview to confirm that no new concepts were mentioned.

Table 3 illustrates current barriers and facilitators, according to stakeholder perspectives.

A point mentioned by participants across all organizations was that adoption depended on the type of study to be performed. In principle, if a research question was suitable, for instance, if a researcher was interested in studying lifestyle changes measured through a wearable or collected via an online questionnaire at home, a VCT could provide a significantly richer data set (including real-life data) when compared to a conventional clinical trial. On a more practical note, for many respondents, the decision to adopt VCTs also depended on the costs of the trials. If VCTs could in effect reduce costs, some respondents would be inclined to conduct a VCT rather than a conventional clinical trial.

The three stakeholder groups made several suggestions on what types of tools could be used and how VCTs could be implemented. Contemporary smart devices or wearables could be used to measure health parameters (eg, heart rate, temperature, oxygen saturation, respiration rate, blood pressure, blood glucose level) and monitor patients while the investigational intervention or drug is studied. Next to noninvasive measurements, the collection of small samples, such as drawing blood by a finger prick, is also possible. By extension, multiple blood droplet analysis devices exist that enable consumers to collect data (eg, blood cholesterol, glucose, or ferritin) themselves and at home. Furthermore, participants mentioned the possibility of conducting online tests consisting of questionnaires or cognition assessments that could be used in, for example, nutrition studies, where participants consume a daily dietary supplement. The effect of the supplement on satiety or gastrointestinal function could be monitored easily online and does not require clinical assistance.

Communication between patients and researchers/doctors could be executed virtually (ie, online). Visits to the clinic, which usually would be planned in conventional clinical trials, could be done remotely, especially for follow-up visits once the patient is considered disease-free. Respondents from the pharmaceutical companies and the research organization explained that a VCT could be implemented when a product or service was close to market introduction. Such trials might investigate how the product or service works in the real-life environment of patients. The hybrid variant was mentioned as the most promising form of a VCT. This means that such trials would have virtual and conventional clinical trial components. For instance, conventional clinical trials could have a hybrid locality approach by collecting some data outside of the central research facilities (eg, plasma and tissue samples are collected by affiliated labs and sent to central locations for testing). Figure 2 shows agreement among the interviewees about potential VCT options.

In this study, participants acknowledged the innovative value of VCTs and agreed that the development of VCTs should be further explored and promoted. However, various aspects have to be improved before VCTs can be widely adopted, such as improving effective collaboration between all stakeholders, clinically validating smart devices and wearables for data collection, and ensuring patient compliance. The data obtained from our interviewees demonstrate that pharmaceutical companies and the research organization in particular are aware of the benefits and disadvantages of VCTs. However, the applicability of VCTs in research is not yet well known. Therefore, informing stakeholders of the advantages of VCTs is an important factor to speed up their consideration and adoption. This is relevant especially now during the COVID-19 pandemic, due to which many conventional randomized controlled trials have been halted; here, VCTs can make a substantial contribution to the continuation of certain types of crucial clinical research [22].

In order to improve the adoption of VCTs, it is important that all stakeholders collaborate with each other [51]. One option for enhancing collaboration among all stakeholders and exploring if a VCT is a preferred option is an multicriteria decision analysis (MCDA) [52]. Although we have not yet implemented this method, an MCDA might be a good option for researchers who are considering VCTs as a data collection tool. An MCDA is a set of techniques that can help decision-makers take into account and integrate multidimensional data (eg, attributes of benefit) and rank different decision alternatives [52]. For example, the preferences of different criteria or parameters relevant to clinical trials could be elicited a priori and subsequently “preloaded” as preference profiles into dedicated MCDA software. This would potentially result in a more transparent and rapid decision process, and more effective support for decision-makers in determining if a VCT, or some of its components, would be beneficial for their trial. Although an MCDA gives a transparent methodology to compare different decision alternatives, it is an extensive technique that might not be applicable to every context; one setting might be more appropriate for an MCDA than another. For instance, it could be more easily implemented in larger organizations with, in relative terms, larger budgets than in smaller organizations. Next to the MCDA, stakeholders should invest in the use of validated diagnostics to obtain the most reliable results in VCTs [45]. The adherence of patients can be improved by properly devising tools that can motivate them (eg, in a playful way, that is, gamification) or via online monitoring. Because a VCT is a new form of clinical trial with a digital approach, additional assistance might be necessary for some participants. For this reason, we suggest the installation of helpdesks to support trial participants by answering their questions and providing them with support 24/7.

Underpinned by a well-known theoretical framework of innovation, our qualitative research approach allowed us to elicit and characterize the broad experiences of interviewees with the VCT adoption process [35]. Furthermore, we made use of the COREQ checklist and an interview guide, which improved the validity and standardization of the interviews, while allowing improvised follow-up questions based on our interviewees’ responses [25,35]. Whereas most studies concerning VCTs are conducted in the United States, our research focused on Europe, which is unique.

The conduct of this research in the European Union could be framed as a strength but also as a limitation. Data sampling was restricted to Europe, mostly to The Netherlands. Since the United States is more advanced in conducting VCTs, follow-up studies should also include this country in their data sample. Furthermore, this pilot study only included pharmaceutical companies (including a CRO), food companies, and a single research organization. Health care professionals, patients or participants, regulatory authorities, CROs, and payers are also important stakeholders in VCTs. According to previous studies, these stakeholders should also be included in the process of clinical trial development [42,53]. Therefore, future research should focus on these stakeholders, and their preferences, value points, and perspectives on VCTs. Lastly, inclusion of only one research organization introduced selection bias, leading to a less representative study sample.

This study used a qualitative research approach to identify the barriers and facilitators behind the adoption of VCTs and explored how this process can be improved. Collaboration among all stakeholders in VCT development is essential to increase knowledge and awareness. Organizations should invest in accurate data collection technologies, and compliance of patients in VCTs needs to be ensured. Furthermore, we suggest conducting an MCDA to explore whether a VCT is a preferred option by stakeholders; this can considerably enhance the decision-making process. The findings of this study can be a good vantage point to accelerate the development and widespread implementation of VCTs.