Friday 21 Feb 2025

[anchit-chandran]

I asked chatgpt to summarise and succinct-ize:

Transfer Behaviour

Current Implementation

A "transfer" currently works as follows:

1. patient_1 is associated with PDU_old.
2. When moving out of the area, PDU_old marks date_leaving_service and reason_leaving_service in the Transfer model (via Edit Child Details form).
3. Once patient_1 appears in a Submission under PDU_new, they are included in KPI calculations.

Key Concern

This implementation doesn't explicitly "transfer" a child but relies on:

• PDU_old marking them as left (to be excluded from KPIs).
• PDU_new including them in a Submission (to be counted in KPIs).
• This process is user-dependent, making it non-guaranteed but previously accepted.

KPI Calculation Issue

Currently, KPI calculations ignore Transfer.date_leaving_service and Transfer.reason_leaving_service, meaning patient_1 may still be included in PDU_old's KPIs even after being marked as left.

Proposed Fix: Exclude Patients Marked as Left

Modify the base queryset for KPI calculations to exclude patients with a date_leaving_service and reason_leaving_service:

self.patients = Patient.objects.filter(
    paediatric_diabetes_units__paediatric_diabetes_unit__pz_code__in=pz_codes,
    paediatric_diabetes_units__date_leaving_service__isnull=True,
    paediatric_diabetes_units__reason_leaving_service__isnull=True,
)

Example Impact of This Change

PDU	KPI1 (Current)	KPI1 (Updated)	KPI5
PDU_old	Included	Excluded ✅	Excluded
PDU_new	Included	Included	Included

• Currently, PDU_old includes patient_1 in KPI1 even after leaving.
• The proposed change ensures patient_1 is excluded from all calculations in PDU_old once marked as left.
• KPIs like KPI5 already exclude these patients correctly.
• A change to this base patient queryset will have knock on effects on almost all subsequent KPIs: essentially, if filtered out, they are ignored for almost all KPI calculations

Discussion Points

• Does this fix align with expected transfer behaviour?