Structured coding of dying patients linked to far higher advance care planning and community death rates, NHS audit finds

Harpham-Lockyer et al. (BMJ Supportive & Palliative Care; PMID 41130666) conducted a retrospective review of all adult inpatients at a 670-bed NHS acute hospital trust (The Dudley Group NHS Foundation Trust) who died within 12 months of a fixed index date. Of 216 patients who died, 104 (48%) had been coded onto the Gold Standards Framework (GSF) — a structured process for identifying patients approaching the end of life. Among patients coded GSF Green or Amber (n=71), 54% had a documented advance care plan, compared with 13% of non-coded patients, and 65% died outside hospital, compared with 45% of non-coded patients. The pattern of more out-of-hospital deaths among coded patients held for both cancer (76% versus 54%) and non-cancer (60% versus 40%) diagnoses. [Audit/Data]

GSF is an established, nationally recognised UK framework for identifying patients nearing the end of life. The audit does not report an international comparator, and none is invented here. In the absence of a wider benchmark, the strongest available standard against which to judge end-of-life identification practice is the presence, or absence, of a structured coding process itself.

The scale of the gap between coded and uncoded patients in this single-trust audit suggests that reviewing how consistently patients approaching the end of life are identified and coded is a lower-cost, more immediate lever on preferred place of death than a new clinical intervention would be. The retrospective, single-trust design, and the fact that the strongest comparison applies specifically to patients coded Green or Amber rather than to all coded patients, mean the finding is best read as evidence that identification processes can move outcomes substantially wherever they are audited, rather than as a generalised NHS-wide coding rate.