2.3 | Data analysis
All interviews were recorded and transcribed verbatim into mandarin Chinese. A phenomenological approach was adopted and inductive thematic analysis utilized to analyze the discourses produced by self-referral patients with advanced kidney disease and health professionals22. Nvivo 11 software was used to facilitate and manage data23. Data were analyzed following these steps: (1) Familiarizing with data, reading repeatedly transcripts and notes to obtain rough understanding and preliminary comment. (2) Generating initial codes, coding interesting features of the data across the entire data set. (3) Searching for themes, collating codes into potential themes. (4) Reviewing themes, checking if the themes work in relation to transcripts. (5) Defining and naming themes, ongoing analysis for refining each theme and overall story. When there were differences in perspectives, all author negotiated and discussed until agreement were reached. Two researchers (Xu S, Jiang XX) conducted the initial data coding. The interviewer (Gao YJ) and the research assistant (Li x) co-coded for themes. At least, the team discussed each theme and came to a consensus.