Prompt Style Effects on LLM Differential Diagnosis Output: Case Vignette and Raw Model Outputs [Dataset]

De-identified clinical vignette and four raw large language model (ChatGPT-5.5) outputs generated under zero-shot, chain-of-thought, role/persona, and structured/constrained prompt conditions, for a pediatric psychiatric case adapted from Ho et al. (2026) under a Creative Commons Attribution license. Supporting data for the manuscript "Prompt Style Effects on Large Language Model Differential Diagnosis Output in a Pediatric Psychiatric Case." Citation: Ho AL, Liwag TM, Gonzalez J, Phan A, Wong D, Fu M. Mania with psychotic symptoms in an adolescent patient associated with ashwagandha use: Case report. Open J Clin Med Case Rep. 2026; 2402