Study design
The study was a qualitative study using Nominal Group Technique (NGT) modified to adapt to a virtual scenario (mNGT). The NGT is a structured method for the development of consensus, using a focus group modality in which the contribution of the participants is obtained through the presentation of a predetermined guiding question during the discussion. Answers are obtained from the participants through the presentation of a question on the topic addressed. NGT allows participants to identify, evaluate and rank important aspects of a specific topic or problem (24). The decision to use the mNGT considered the study objectives, the need for consensus, practicalities and associated limitations, such as participant’s time and availability, and Brazil's geographical dimensions. Compared to the Delphi Technique, NGT has a lower drop-out rate throughout the process (25,26). Given the challenges of running in-person event during the COVID-19 pandemic (27) and the desire to include participants from different Brazilian regions, a careful adaptation of the traditional technique was used to allow for its application online Supplementary file (Additional file 1) (28).
Participants
Nurses working professionally in Brazil in different contexts were invited, including members of hospital infection prevention and control (IPC) committees, public health organizations, professional councils, such as the Regional Nursing Council. Academics and researchers as well as frontline nurses from primary health care (PHC), outpatient and general hospitals were also invited.
In order to ensure a proportional distribution of the guests according to their work context among the five regions of Brazil, a proportion matrix was constructed according to the percentage distribution of health facilities and university hospitals in the country. The proportion matrix included four categories: PHC and outpatient clinics, general hospitals, universities, and the "other" category included health surveillance agencies, professional councils, and professional associations in IPC. To assist in the construction of the matrix, a search was carried out in the database of the National Register of Health Establishments (CNES) and the Ministry of Education, looking for the percentage of health facilities and universities distributed in the five regions of Brazil. Details of the distribution of healthcare facilities, and the proposed matrix of participant’s is available in the Supplementary file (Additional file 2).
In order to ensure equitable participation and a more efficient discussion, it is recommended that the number of NGT participants range between 7 and 10 guests (24,29). The number of guests was carefully planned with the expectation that approximately 30 individuals would participate in the consensus group. By utilizing the recommended group size, it was envisioned to establish three groups, each consisting of a maximum of ten professionals.
Ethical approval
The study received ethical approval from the Research Ethics Committee of the School of Nursing of University of São Paulo, with approval nº 5.381.334. All participants have signed an informed consent form before taking part in the virtual workshop.
Recruitment of participants
Initially, the recruitment of participants involved inviting nurses nominated by the research executive team. Subsequently, snowball sampling was employed to identify potential participants. Participants were then invited to identify others who could fulfil the research inclusion criteria. Those individuals who consented to participate in the research were notified via email of the invitation letter, consent form, and additional details regarding the virtual workshop, including the program, explanations about mNGT, and a list of publications on the subject.
Data Collection
The data was collected during a virtual workshop called “The Role of Nursing in Antimicrobial Stewardship Programs” using the Google Meet® platform with an upgrade to Google Workspace®. The workshop duration was for three four-hour sessions (12 hours total) over three days, with an interval of two weeks between each day. The aim of the workshop was to identify and prioritize knowledge gaps about nurses' engagement in ASPs in the Brazilian context. All sessions were recorded (audio and video), and field notes were taken. The authors of this paper actively participated in the workshop from its initial planning stage, serving as lecturers, hosts, facilitators, and supporters during the NGT, and as manager of information and communication technology (ICT).
On the first day of the workshop, the guests were provided with information about AMR, the actions to tackle the problem and the potential nurses’ role in ASPs. The agenda included short lectures and moments dedicated to questions and clarifications on the subject. After day 1, the participants received, via email, a link to a digital form containing the guiding question, “What do you consider to be an important research gap to be explored regarding the role of nursing in antimicrobial stewardship?”. The electronic form contributed to the idea generation stage of mNGT. Considering their own views and work experience, the participants were instructed to answer the form with up to three ideas considered research gaps to be explored in future investigations (Figure 1).
Figure 1. Planning the online workshop
After a two-week interval, the second day of the four-hour workshop was held. Day 2 aimed to explore the preliminary ideas proposed by the participants on the first day and to reach a consensus on the priorities for future research. All participants were received in the main virtual room, called “plenary”. The host explained the importance of the research question and used the FINER criteria—feasible, interesting, novel, ethical, and relevant—to ensure that appropriate research questions could be formulated (30,31). Participants were advised that ideas entered into the electronic form during the idea generation stage should be converted into research questions using the FINER criteria.
To clarify and discuss the ideas previously identified, participants were randomly allocated into three small groups (A, B, and C), with up to 10 participants in each group. The division of the groups considered the excessive time and dispersion caused by tiredness that could be generated by discussing all the ideas with all the participants. Therefore, the discussions took place independently in each thematic room, limiting themselves to the ideas generated by their participants. Each group was assigned a facilitator and a supporter.
The purpose of the presentation and clarification phase was to clarify the ideas entered into the electronic form. The ideas were then converted into research questions. If the ideas were similar and the group agreed, the facilitator made adjustments, such as combining them or editing or deleting repetitions.
The voting round took place in two stages using the SurveyMonkey® virtual platform: the first in the breakout room, where participants voted anonymously choose up to three relevant research questions for future investigations according to their point of view. The second round of voting took place in the plenary room, where each participant and the facilitators could vote individually on a single most relevant research question. The choice of two rounds of voting was made so that all participants could discuss and vote on the most relevant research questions in each breakout room. At the end of the process, the participants prioritised the research questions democratically.
After a three-week break, day 3 was held. The aim was to develop the three main research gaps prioritised in the consensus groups during day 2 into preliminary research projects. The methodology used was based on a group discussion following a previously prepared roadmap. The participants were divided into three groups where each group worked on one of the three prioritised research questions. At the end, the participants presented their projects in the plenary room.
Data analysis
The mNGT results and the demographic data of participants were presented using descriptive statistics (frequencies, mean, standard deviation, and range). Content analysis was performed to the research questions after the idea’s clarification stage to characterise and identify common topics and to create categories in the research priorities (32,33). Two researchers carried out the analysis independently in the categorisation phase (AMSF and VCLG). Conflicting categorisations were discussed and agreed upon. A third researcher (MCP) then assessed the relevance of the categorisation.