fix discharged no treatment tpma

[tomjemmett]

resolves 2 issues @swyatt-SU pointed out with the discharged no treatment or investigation TPMA

• fixes Treatment/Investigation codes flipped in A&E "Is Discharged with No Treatment or Investigations" #276 - codes were transposed between investigations/treatments in the AAE dataset extract
• fixes Add "patient given written advice" code to "Is Discharged with No Treatment or Investigations" #277 - adds additional code in for the ECDS dataset extract

[Copilot]

Pull request overview

Updates the “discharged with no treatment or investigations” logic in the raw-data extracts to align with the intended code semantics for both AAE (HES A&E) and ECDS, addressing the two reported issues.

Changes:

• Fixes the AAE extract filters so treatment codes 22/99 are excluded from the “treatments present” set, and investigation code 24 is excluded from the “investigations present” set (resolving the prior transposition).
• Extends the ECDS extract logic to treat SNOMED 413334001 (“patient given written advice”) as part of the “no treatment / investigations” classification.
• Refactors the ECDS predicate to use centralized lists for the relevant SNOMED codes (via a small in-function mapping).

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
src/nhp/data/raw_data/aae.py	Corrects the treatment vs investigation filter logic used to derive is_discharged_no_treatment in AAE.
src/nhp/data/raw_data/ecds.py	Adds the written-advice SNOMED code to the discharged-no-treatment logic and switches to .isin(...) over a maintained code list.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[yiwen-h]

Does #277 impact current model data? How significant will the change be?

[tomjemmett]

yes, counts below are rounded number of attendances which are flagged, with the percentage shown of all attendances

	before	after
count	7,069,600	8,888,900
pcnt	26.1%	32.9%

@swyatt-SU for your information, as this does materially change the TPMA counts. Probably should go to dev board

[swyatt-SU]

Happy to take it to Dev Board next week. Looks like we will need to make the change, but others will need to work out implications for schemes, how it is communicated, etc. Maybe hold off authorising the pull request until I've got the green light.