Through the iterative sequence of different phases of the research process, the sampling procedure (purposeful case selection), and the continuous revision of the survey instruments (e.g., interview guidelines), qualitative researchers can flexibly adapt their studies to changing research conditions [6, 7]. Thus, it is possible to respond more comprehensively than is the case in quantitative research, where once a random sample has been drawn, it can not be changed, and where researchers usually can not modify a research question during the quantitative research process . In contrast, for qualitative research projects, it is more the rule rather than the exception to continuously adapt the theoretical sampling and quota plans , survey instruments, and research questions to new findings or changing conditions in the research field [10, 11]. Within ADAPTIVE, we found that qualitative research designs are crisis-proof due to their flexibility. In contrast, the classic quality criteria of quantitative research – objectivity, reliability, and validity – are significantly related to adherence to a linear research process. However, the potential for adherence to the specific quality criteria of qualitative research – subject adequacy, empirical saturation, textual performance, and originality  – showed to be much more robust in the necessary adjustments to research designs since March 2020.
How does this translate to qualitative research under pandemic conditions? As in many other areas of society, one of the most widespread coping strategies in the health care sector is the use of digital technologies both by health care professionals and health care researchers, i.e., health services research [13, 14]. Accordingly, there has also been a massive increase in "digital" data collection in health services research since spring 2020 [2, 4, 15, 16]. Field access strategies  had to be reconsidered and adapted, interviews  and focus groups  were conducted by video call or at least by telephone [20–22]. By switching to purely digital or at least hybrid communication, research projects could be continued and completed. However, due to the virtual circumstances and common technical problems, additional context information is often lost, i.e., such as facial expressions and gestures of participants . Also, often other contextual factors are lost, which are crucial for qualitative research – their absence must be reflected upon in any case to deal with the new conditions methodically .
Further, it is essential to recognize that digital communication is neither comprehensive nor evenly distributed across society. For example, in Germany, small and medium-sized enterprises are significantly less digitized than large companies and younger, well-educated people still use digital technologies much more extensively and competently than older adults with lower education levels . In this regard, the researcher should consider the following questions: Who is structurally excluded or included from the sample by (not) having access to digital technology? Which groups of people are more likely to shy away from a digital interview, and which groups have an affinity for a video interview?
In response to these questions, within ADAPTIVE, we offered telephone interviews because we were not sure if all participants were familiar with video call technologies and if data protection policies in participating clinics forbid them. With this approach, we tried to avoid other significant problems in digital data collection such as weak internet connections, unfamiliarity with the technology on parts of the participants, dropouts due to possible insecurities regarding their being on camera, and the exclusion of specific clinics through their data protection concepts which would have again significantly reduced the basis for recruitment.
However, recruitment and scheduling of appointments proved to be consistently challenging due to the limited time capacities of potential interviewees. In order to make it easier for medical professionals to participate, we streamlined our study design down to only one shorter interview per participant and no focus groups. We also adapted the interview guidelines to save time and to include topics related to the pandemic. However, it might not always be feasible for all projects to dismiss work packages such as focus groups because, e.g., the funding partner may not agree to do so. Further, we agreed on telephone interviews. For ADAPTIVE, disadvantages of telephone data collection were offset by advantages in case selection. While there is a loss of nonverbal communication, respondents could be interviewed nationwide rather than in only one region, thus facilitating the recruitment of a sufficient number of participants. Further, participants always have their phones with them, so it was very time-saving and convenient to be interviewed this way, instead of, e.g., having to sit down at a computer for a video call. While video calls are also possible on smartphones, they often were not allowed in clinics due to concerns regarding data protection.
Nevertheless, Vindrola-Padros and colleagues  pose a crucial question about research in pandemic times: is it necessary and ethically justifiable to conduct research in pandemic times when caregivers are already under enormous pressure? The additional time and cognitive burden on healthcare staff must be ethically weighed against the benefits of the research results. While participation in the project was voluntary, 26 participants agreed to participate in a one-hour interview despite working extra hours, an increased workload in their daily work, and great professional and personal pressure. Their participation shows how valuable and necessary participants thought data collection during this time was and how useful they thought the data gathered would prove. By expanding the data collection to cover participants' insights on their situation during the pandemic without necessarily prolonging the interviews, participants also had the opportunity to discuss their worries and illustrate their hardships of the last weeks and months. Participants very much appreciated this option, and it generated more valuable data on handling the pandemic. In this context, it was also possible to investigate the importance of digitization in medical settings. The trend towards digitization in the medical field seemed to accelerate as a result of Covid-19. Our results convey the importance of the resulting networking with additional providers to the participants. In the context of these considerations, we decided to continue the study, to expand the guideline to include the experiences during the Covid-19 lockdown and the accompanying digitization measures, and to incorporate the resulting findings into the initial research question.
Like many analyses of the effects of new technologies in health care, exploratory research projects rely on collecting data "in the field" to gain a first impression of the field. Especially when analyzing the use of new technologies, such field visits (in the sense of ethnographic go-along) can help sharpen the researchers' focus of analysis. Also, in qualitative interview studies, there is usually the possibility of being shown the technologies under investigation by the users on-site or observing them in actual use. This possibility represents a critical (data) triangulation  in interview studies, which is not available in purely linguistic interview transcripts. This possibility decreases due to the pandemic-related restrictions and poses particular challenges for qualitative research focussing on health services and digitalization, such as ADAPTIVE.
Another challenge we had to overcome was how to collect data while (a) ensuring safety for participants and interviewers but also (b) ensuring high quality of the data. Within ADAPTIVE, we tried to combine masked face-to-face interviews with telephone interviews. However, as illustrated above, qualitative research is based on trust between interviewer and participant. An initial fear was that telephone interviews or wearing a face mask would result in a lack of legibility of nonverbal communication, resulting in difficulties in establishing a trusting relationship with the interviewer. Without a trust-based relationship, interviewees might not be as open and vulnerable with interviewers as they would be with a person they trust. The original research questions within ADAPTIVE only included participants' professional experience with digital technology and did not result in any particularly sensitive content. However, this may be a challenge to consider in studies with particularly sensitive interview content and research questions, and vulnerable target groups. We also feared unclear articulation due to wearing a face mask would result in unclear transcriptions of the interviews. This possibility should be considered for more detailed transcriptions, i.e., data preparation for hermeneutic evaluation (e.g., sequence analysis). For interviews conducted in ADAPTIVE, a verbatim transcription was sufficient, and all interviews, including those where interviewer and interviewee were wearing a mask, could be transcribed well.