Study setting, design and participants
Our study is based on interviews with health care professionals (physicians, registered nurses, assistant nurses) employed in the Swedish health care system, which is mainly publicly funded although private health care also exists. All residents are insured by the government, with equal access to health care for the entire population. Out-of-pocket fees are low and regulated by law. Sweden has 21 regions, which are responsible for providing health care.
We conducted semi-structured individual interviews with 30 health care professionals: 11 physicians, 12 registered nurses and seven assistant nurses (Table 1). The participants were employed in six different health care units, all located in cities, with 67,000, 135,000 and 150,000 inhabitants, respectively, an all located in south-eastern Sweden.
We used a purposeful sampling strategy to achieve a heterogeneous sample of health care professionals working in different health care facilities in Sweden, ranging from primary care to hospital and tertiary care, with patients who varied in terms of health status and duration of stay. The aim was to achieve a sample of health care professionals that represented a broad spectrum of perceptions and experiences concerning changes in health care.
We recruited frontline health care professionals through an e-mail that briefly described the study. The e-mail request was sent to the manager of each work unit, with a request that they forward our request to physicians, registered nurses and assistant nurses. To those who responded, we then sent an informational letter describing the study. None declined involvement after receiving the information letter. We scheduled interviews at a time (between January and September 2018) and in a location of the participants’ choosing where they could feel comfortable about speaking honestly (e.g. office with a closed door).
Data collection
The data collection applied an inductive approach, using a semi-structured interview guide developed by the authors. We generated the questions based on the existing literature on organizational change and change responses [15, 18–20]. The questions concerned the participants’ experiences and perceptions of any changes that they considered to have affected their work, regardless of whether these changes were “objectively” large organizational changes, e.g. a re-structuring of the organization, or small, e.g. modification of an already existing workplace routine, thus covering both broader, more general changes and more specific examples of changes such as the merging of the informant’s work unit with another unit, introduction of new information technology systems and moving to new localities.
Individuals’ subjective experience may not correspond with more objective measures of organizational outcomes of changes. However, it is crucial to understand health care professionals’ views on changes in health care because their attitudes towards changes, can influence the extent to which changes are successful [21, 22]. Given the potential influence of health care professionals’ views on the success of change, rather than asking about specific changes or providing lists or examples of changes, we allowed the participants to discuss any changes they considered to have relevance for their work. This conceptualization of change stems from research that shows that individuals experience change in unique ways; the same change may be attractive and imply advantages for some and be a source of stress and disadvantages for others [23].
Each interview began with some questions about the participant, the content of his or her work and some information about the workplace. The participants were then asked to describe examples of organizational changes they considered had achieved the desired results, and they were asked to explain why they believed these changes were successful (other questions were also asked but were not analysed for this study). We asked a final question concerning whether the participant had anything to add to what had been discussed.
We pilot tested the questions in two interviews with regard to meaningfulness for participants and clarity of concepts. The pilot interviews indicated that the questions were generic enough to be used in different health care contexts, that the wording was clear and that the interview did not exceed 60 minutes (which was deemed to be the maximum considering the participants’ work schedule). The two pilot interviews were included in the study.
Individual interviews were conducted by all the authors except SB, who does not speak Swedish, and were digitally recorded. Before the start of an interview, the participant was asked to re-read the information letter and give written informed consent to participate. Each interview lasted between 28 and 104 minutes (mean, 50.5 minutes). The interviews were transcribed verbatim by a professional transcription agency and were then reviewed by the researcher who conducted the interview.
Data analysis
Using an inductive approach, participants’ responses were analysed using directed content analysis according to descriptions by Hsieh and Shannon [24]. All authors except SB read the transcripts of the interviews individually to create a holistic view of the material. In the next step, each researcher performed a first analysis condensing meaning-bearing units and creating codes and subcategories. PN, IS, CE and KS then met to discuss and compare their respective interpretations of the material. Tentative findings were reported to and discussed with SB. Following her input, PN, IS, CE and KS met again (as SB is located in the US) to discuss the preliminary findings. This discussion led to a proposal concerning the categories of analysis, which was then fed back to SB for her comments. Eventually, consensus was reached on the categories and PN suggested labels which were accepted by the whole group. Representative quotations for reporting were jointly identified by PN, IS, CE and KS. PN, who is fluent in English, then translated the quotations from Swedish to English, which were then examined by IS, CE and KS for accuracy. Finally, SB, whose first language is English, reviewed the English-language quotations for clarity.