Embedded Files
training were invited to participate for the duration of the study, involving 3 interviews at 6 monthly intervals over the course of 18 months. Of the 26 nurses, 16 volunteered to participate. All potential participants were informed of the aim of the study and given time to consider volunteering, and enrolled if they wished to participate. All signed informed consent prior to the study. Data collection The first phase of data collection was conducted in August 2014 following the completion of the neonatal nursing educational intervention in spring 2014. Two further phases of interviews were conducted at 6 monthly interviews. The same group of researchers conducted the interviews at each time point using the same semi structured interview questions. Interview schedules for both nurses and parents are available as Additional files 1 and 2 with this paper. Interviews took around 20 min to complete and were scheduled over the course of a week at each time point. Both parent and nursing interviews were audio recorded and took place in a quiet room on the neonatal unit. All study documentation was produced in English by the guiding researchers, and translated into Vietnamese by the local research team. Interviews were conducted in Vietnamese. Analysis All interview transcripts were transcribed into English by the local research team prior to data analysis. Data was entered into the data management package NVivo v14 to facilitate analysis. Data were analysed using manifest content analysis, a method which focuses on the analysis of clearly evident components of the text in the interview data [14, 15]. Content analysis is a useful methodology when conducting exploratory work in a relatively unknown area, as it allows for the reporting and analysis of common issues in the data [16]. It also provides a useful method to quantify the data through counts of these issues [17, 18], allowing for comparison over a longitudinal study period. As this was the first research of its kind it Vietnam, from both the researchers and the participants perspectives, this approach was determined to be the most feasible in order to explore the research questions under investigation. Content analysis allows a systematic way of describing a phenomenon through the ability to distil words (text) into fewer content related categories, or themes [19, 20], with the aim of providing a new insight towards the research question through the development of these categories of description. At each time point, transcripts were read and re read to get a sense of immersion with the data (data was analysed between participants in each group separately (parents and nurses) at each phase of data collection). Initial coding of the data was then undertaken by the UK based researcher (KG) which involved recognising basic concepts within the text which reflected the research questions (meaning units of text). These codes were then grouped together, or condensed, into categories which reflected specific areas of discussion using descriptions as close to the text as possible. Frequency counts were assigned to each area to determine the commonality of the responses (in retrospection if discussed at T2 and not T1). The development of the codes and categories were discussed at each time point with the global research team to determine the credibility of the findings. Across the time points, the condensed units were compared and grouped into sub-themes representing the data. These sub-themes were finally placed into themes which represented the data as a whole, identifying separate areas of discussion which the participants Gallagher et al. BMC Pediatrics (2017) 17:161 Page 3 of 11 discussed throughout the interviews (Table 1). The findings were again shared between the global research team to discuss any differences in opinion in categorisation and provide rigour to the study. This also provided an interesting approach to reflexivity in the study, as the analysis and interpretation were based upon different cultures and contexts. This minimised the risk of any individual biases towards the results and provided a depth of analysis which may not have been achieved otherwise. Results A total of 83 individuals participated in
Report abuse