Northern Health NHS Foundation Trust | 3 sites | January–December 2024

Conclusions

1. The Trust's four-hour breach problem is a process issue, not a capacity or single-site issue. Journey stage durations are nearly identical across all three sites (within 1-2 minutes on every stage). The same bottleneck exists everywhere, which means a single process fix has Trust-wide impact rather than needing three separate interventions.
2. Two stages account for the overwhelming majority of breach risk. Breaching visits spend +29 minutes waiting for a clinician and +38 minutes in treatment compared to non-breaching visits — together, ~85% of all excess time in a breach. Triage adds only 5 extra minutes and is not a meaningful driver.
3. Breach volume and breach risk point to two different priorities. Triage codes 3 ("Urgent") and 4 ("Standard") drive ~72% of all breaches by volume. But triage code 2 ("Very Urgent") has the highest per-patient breach rate (9.2%), despite lower volume. These require different responses: process fixes for codes 3/4, clinical review for code 2.
4. Site performance is inconsistent depending on how you measure it. Leeds General Infirmary has the highest average breach rate for the year (9.45%) but is consistently the most stable performer month-to-month. St James's has the best yearly average (7.22%) but the most volatile monthly swings, spiking above 40% in March.
5. Breach rates spike seasonally (March, June, August) across all three sites simultaneously — suggesting an external or Trust-wide driver rather than a site-specific cause.
6. Shift timing and staffing levels don't show a strong relationship with breach outcomes — but this may be a data limitation, not a true absence of effect. Time-to-clinician is broadly consistent across Day/Evening/Night shifts (76-82 min), with Leeds General's Evening shift a modest outlier (4 min slower than Day, 3 min slower than Night). The staffing-vs-breach-rate correlation is essentially flat (-0.04), but based on only 9 site/shift data points.
7. Arrivals peak in two distinct windows — one expected, one not. Early-to-mid afternoon (13:00-15:00) is predictable. An overnight cluster (03:00-05:00) is actually the single busiest hour in the dataset (107 arrivals at 04:00) — a window that typically overlaps with the thinnest staffing.

Recommendations

Priority 1 — Process improvement (highest confidence, highest impact)

• Investigate and redesign clinician allocation immediately after triage — the single largest driver of breach time (+29 min), consistent across all three sites.
• Review treatment-stage throughput and handoffs — the largest stage in absolute terms (+38 min).
• Target these fixes at triage codes 3 and 4 first, since they represent ~72% of breach volume.

Priority 2 — Clinical risk review

• Separately review triage code 2 ("Very Urgent") patients despite lower breach volume, since they carry the highest per-patient breach rate.

Priority 3 — Data & measurement gaps to close

• Link shift-level staffing records directly to patient visit records — the staffing-vs-breach analysis is currently inconclusive due to data granularity, not because staffing doesn't matter.
• Add a boarding/bed-availability timestamp to EPR data capture, if not already tracked — the original scope intended to measure this and the current dataset can't answer it as framed.

Priority 4 — Rota and demand alignment