... language model (LLM)-enhanced grounded theory procedure involving open, axial and selective coding. The study is framed as a mechanism-building reconstruction rather than a substitute for direct participant testimony. Findings Analysis of the reconstructed accounts identified two broad types...