Atlas ti 7

Field notes were prepared during the observations using predetermined scales for assessment of clinical and soft skills. Changes in the two skill domains were analysed by repeated measures ANOVA.
Audio recordings of interviews with participants from Arm 1 were transcribed. After familiarising themselves with the interview transcripts, the researchers conducted the coding process through Atlas.ti 7 software (ATLAS.ti Scientific Software Development GmbH, Berlin), using predefined codes that were consistent with the objectives of the SRCC and the semi-structured interview guide. In addition, the number of participants who reported applying new knowledge and skills in diabetes care and other clinical areas was tabulated. Relevant quotations from interviews were extracted to support the findings.

The qualitative data was analysed using the approach from Mayring (2010) for summarizing content analyses with inductive category assignment. This approach is described as very fruitful to allow a true description and understanding of the original material “without bias owing to the preconceptions of the researcher” Mayring, 2014). After the definition of the material a first phase of paraphrasing and generalization of the data was performed. To develop a summarizing coding system, the material was further reduced and abstracted (first and second reduction according to Mayring, 2010). Finally, the summarizing coding system was re-checked with the original material to ensure that the category system matched it well. Findings and discrepancies in the interpretation of the data were regularly discussed in the research group. The analytical software Atlas.ti 7 assisted the analysis process.

A thematic analysis of the data was performed. Atlas.ti. 7 software was used to manage the data. Two of the researchers independently analyzed the transcriptions, following the 6 steps presented by Braun and Clarke [26 (link)], and met after Step 2 to ensure that the coding was the same. A second meeting took place with a third researcher to define and name themes. The complete process of ensuring trustworthiness is described in Table 4, as stated by Nowell et al. [27 (link)].

Participants’ feedback on the EQ-5D-5L, the AQL-5D, and the ASD was collated and summarized using the interview field notes or transcript data as available and supplemented by the audio recordings, if needed. The concept-elicitation transcript data from interview sample 2 were analyzed thematically using Atlas.ti 7 coding software (Atlas.ti; Berlin, Germany). An initial coding frame was applied iteratively to the transcript data and was updated as themes were refined and new codes were developed. The analysis was conducted by two researchers: one researcher undertook the primary summarization or coding, and the second researcher read the field notes and/or transcripts and reviewed the summaries and applied codes. Any discrepancies were resolved by the two researchers.
The output from the qualitative analysis was a descriptive summary of participants’ feedback on the questionnaires and of the issues discussed in relation to the impact of asthma. Concept saturation in the concept-elicitation data was assessed by the emergence of new analysis codes across successive sets of three interview transcripts [36 (link)].
Participants’ ratings of the relevance of each instrument were summarized by the mean rating. Participants’ selections of the most relevant instrument and individual items to describe the effect that asthma has on their lives were summarized by frequency counts.

The ODK platform data were exported to a Microsoft Excel spreadsheet and then imported into STATA Version 15 for further statistical analysis. ArcGIS version 10.4 was used to investigate spatial patterns and identify hotspot areas of non-resilient food insecurity. To investigate the spatial pattern of non-resilient food security across the entire study area, a global spatial autocorrelation (GSA) analysis was performed using the Global Moran's I statistic. Moran's I near 1, 0, − 1 indicates that the spatial distribution of resilience to food insecurity is clustered, randomly distributed, and dispersed, respectively-resilience. The Getis-Ord Gi* statistic was used in GSA analysis to detect local clusters in the presence of clustering. The findings were presented in the form of text, tables, and graphs. For descriptive data, proportion and mean were employed. To identify the factors associated with resilience to food insecurity, a binary logistic regression model was fitted. At a P-value of 0.05, the adjusted odds ratio (AOR) was used to assess the strength of the association. ATLASTi7 for qualitative data coding was used to prepare the data for thematic analysis in the qualitative aspect.