A clinician prompt pack turns a general LLM into a documentation drafting aid that respects clinical structure. Rather than free-typing a request, you choose your specialty and the document you need — a SOAP note, a patient letter, a literature summary, or a care-plan draft — and copy a prompt engineered to stay accurate, structured, and safe. Crucially, it is built to work with de-identified inputs only.
How it works
Each template encodes clinical conventions. The SOAP note prompt organises your inputs into Subjective, Objective, Assessment, and Plan, and flags missing data instead of inventing it. The patient letter prompt rewrites clinical content into plain language at a readable level while preserving accuracy. The literature summary prompt structures evidence into findings, strength of evidence, and clinical relevance, marking anything uncertain. The care-plan draft prompt produces problem-oriented goals and interventions for your review. You set the specialty and how much context you will provide; the tool builds the instruction block so the model stays anchored to your field and your data.
Tips and examples
De-identify everything before you paste it — strip names, dates, NHS/MRN numbers, and any rare detail that could re-identify a patient. Treat every output as a first draft that requires full clinical review; the prompts are written to surface gaps, but they cannot replace your judgement. For literature summaries, ask the model to mark claims as “verify” and check them against the primary source, since LLMs can misremember study results. For patient letters, read the plain-language version aloud — if it would confuse a non-clinician, refine the prompt’s reading-level instruction.