Objective
To describe the methodological procedures of a multi-centre focus group research for obtaining content categories also suitable for categorical statistical analyses.

Methods
Inductive content analyses were performed on a subsample of 27 focus groups conducted in three different countries, the Netherlands (Utrecht), the UK (Liverpool) and Italy (Verona). The analyses of the subsample of focus group discussions were performed in five steps: (1) independent development of content categories in each of the participating centres, (2) obtaining consensus categories, (3) creation of a manual with coding rules and defining criteria for categories and subcategories, (4) assessment of inter-rater reliability to identify unreliable categories to be revised, and (5) repetition of inter-rater reliability assessment.

Results
The resulting coding system considers five areas: non verbal communication, process oriented expression, task oriented or problem focused expressions, affective or emotional expressions, and physician's personal characteristics. It contains 12 categories of acceptable inter-rater reliability and 41 subcategories.

Conclusion
The coding procedures show how focus group data, obtained in an international multi-centre study can be analysed in a systematic way combining scientific rigour with the richness of data obtainable from qualitative methodologies.

Practice implication
The applied procedures may be helpful for multi-centre focus group research on other topics.