Each of the domains as originally published was accompanied by a set of detailed underlying constructs, and suggested interview questions, which we developed into an interview guide. We revised this after early interviews revealed it was difficult to ask the questions in this way. For example, we merged questions on evidence into one.
TDF questions	Revised for interview guide
Knowledge: “what do they think the evidence is?” Social/professional role: “what do they think about the credibility of the source [of a guideline]?” Beliefs about consequences: “does the evidence suggest that doing x is a good thing?”	Tell me a bit about your views of the evidence base for PPI
Social Influences: “To what extent do social influences facilitate or hinder x?” “Will they observe others doing x?”	How do your colleagues feel about PPI? Do they approve of it or disagree?