A narrative discussion was used to synthesize the studies according to the four predefined themes.
The form included four predefined themes: policy, practice, education, and research.
The predefined themes were explored and an inductive process was used to derive subthemes from the main themes.
The moderator asked the participants initially to define what they meant by guidelines, encouraged participants to discuss freely, and only probed to clarify and to secure that all the predefined themes were discussed.
Predefined themes were addressed in the theory-based interview scheme, which was derived from multiple socio-cognitive theories i.e., the Extended Parallel Process Model, the Reasoned Action Approach and the socio-psychological view on stigma.
One non-predefined theme emerged bottom up from the data.
Narrative data were analysed using framework analysis [ 16], identifying predefined and emerging themes and subthemes from the raw data and coding data transcripts accordingly, using manual data management techniques, such as highlighting and cutting out of themed sections.
Issues related to study aims were identified and coded without predefined categories capturing the main themes and concepts.
Qualitative data were analysed using a 'mixed' thematic approach with the identified themes deriving from predefined and emergent issues relevant to the study objectives.
All of the transcripts of field reports, IDIs and FGDs, were imported into the qualitative data analysis software NVivo 2 (QSR International Pty Ltd. Version 2, 2002) and coded according to a matrix of themes that had been predefined for all community liaison data collection.
The data were categorised using themes identified in the raw data (predefined categories or 'taxonomy' were not used) [ 23] using a thematic analysis approach.
