Gloucestershire Hospitals — Corridor Care Eliminated with Data
Gloucestershire Hospitals NHS Foundation Trust is the most quantitatively documented corridor care Bright Spot in the public domain. CEO Kevin McNamara presented the evidence to NHS England in June 2026: a 68% reduction in ambulance handover delay hours, average handover time falling from 80 to 26 minutes, discharge-ready patients reduced by 30% — all achieved while ED attendances were rising. SHMI fell for 14 consecutive months. More flow. Better safety. The contradiction dissolved.
Gloucestershire is the strongest quantitative Bright Spot in the public domain. The CEO’s presentation includes time-series charts showing sustained step changes in ambulance handover hours and discharge-ready patient counts, against rising ED demand. The visual evidence is consistent with genuine Bootstrap CUSUM change points. No formal pre-committed Bootstrap CUSUM analysis has been published. What that would require is explained at the bottom of this page.
☰ Contents
The results — what the data shows
Kevin McNamara, CEO of Gloucestershire Hospitals NHS Foundation Trust, presented the following results to NHS England in June 2026 as part of the “Best of the NHS” series on ending corridor care.
| Measure | Baseline | 2025/26 | Change | Significance |
|---|---|---|---|---|
| Ambulance handover delay hours | 3,343 hours/month (2023) | 1,106 hours/month (2025) | 68% reduction | Equivalent to returning 12 ambulance crews to the road every day |
| Average ambulance handover time | 80 minutes (2022/23) | 26 minutes (2025/26) | 67% reduction | Sustained across multiple seasons — not regression to the mean |
| Discharge-ready patients (DRD) | 169/day average (2023) | 118/day average (2026) | 30% reduction | Step change visible from mid-2023; external constraint still partially binding |
| SHMI (mortality indicator) | Above 1.1 (peaked at 1.17) | Approaching 1.0 (Nov 2025) | Falling 14 consecutive months | Balancing measure moving in right direction — more flow did not harm safety |
| ED attendances | 153,960/year (2019/20) | 165,110/year (2025/26) | +7% increase | Improvement achieved against rising demand — rules out demand reduction |
| Staff morale | Second lowest in the NHS (2023) | Significantly improved | Recovery | Staff recommending GHFT as a place to receive care declined 33% in 2023 — reversed |
Why rising demand matters
The single most important number in the table above is the ED attendances figure: up 7% while every other metric improved significantly. This rules out the two most common alternative explanations for corridor care improvement:
- Demand reduction — if fewer patients were arriving, improvement would be expected regardless of any intervention. ED attendances rose. Demand reduction is not the explanation.
- Regression to the mean — if 2023 was an unusually bad year, 2024/25 improvement might simply be the system returning to its historical average. But the improvement sustained and deepened through 2025 and into 2026, against rising demand. Statistical regression does not sustain and deepen over three years.
What remains is the most parsimonious explanation: the system structurally changed. The intervention reached the binding constraint. The improvement is real.
In 2023/24 Gloucestershire was in crisis. NCTR/DRD patients peaked at over 250 (against a total bed base of approximately 800 — nearly a third of all beds occupied by patients ready to leave). Staff morale was second lowest in the NHS. The trust had clear evidence of delay-related harm. Average ambulance handover time peaked at 2 hours 50 minutes.
The crisis created the pre-condition for structural change: the assumed contradiction — that flow and safety are in conflict — was no longer defensible. The cost of the assumption was visible to everyone. This is the Evaporating Cloud mechanism in practice: the assumption sustaining the conflict becomes visible only when the cost of the conflict becomes undeniable.
What changed — the six interventions
Kevin McNamara’s presentation listed the interventions under two headings — “What changed?” and “What’s made the difference?” — and they are identical, which is itself informative: the same six things both caused the change and sustained it.
| Intervention | Joiner level | Framework | What it changed |
|---|---|---|---|
| Focus on harm not performance — willingness to quantify and publicise it | Level 3 | Making the invisible visible | Reframed the organisation’s accountability from performance metrics to patient harm. Made harm data visible to everyone simultaneously. Changed what the organisation attended to. |
| Working together with partners on resetting relationships in flow | Level 3 | Systems thinking — silo dissolution | Dissolved the organisational boundary between departments and partner organisations at the point of flow. Made the whole system visible to all parties simultaneously. |
| Remove the comfort blanket that beds create flow | Level 3 Deep | Evaporating Cloud — challenging the assumption | Tested the assumption that more beds equals better flow and found it false in some instances. The assumption was keeping the constraint invisible. Removing it opened the real constraint to view. |
| Clarity around Exec Tri role — no separation between quality, safety and flow | Level 3 Deep | Joiner Triangle — all three elements held simultaneously | The assumed contradiction between flow and safety dissolved at executive level. Quality, Scientific Approach, and All One Team operated as one system rather than three competing workstreams. |
| Better alignment from speciality to Board on expectations, roles and responsibilities | Level 3 | Going to the Gemba — information flow | Accountability made visible from front line to Board. The information pathway that fear had previously blocked was opened. Staff trusted to speak out. Decisions made by people who could see the constraint. |
| Dedicated clinical leadership with a focus on flow | Level 3 | Going to the Gemba — authority at the constraint | Senior clinical presence at the point of constraint with the authority to act. The visibility mechanism made continuous rather than periodic. The constraint visible to someone who could dissolve it immediately. |
The Joiner Triangle in practice
Brian Joiner’s Triangle identifies three interdependent elements of effective management: Quality (defined by the customer), Scientific Approach (managing with data and understanding variation), and All One Team (no blame, trust, staff empowered to make decisions). The observation that makes the Triangle powerful: taken separately they are not as effective. They are best used together.
Gloucestershire’s six interventions map directly and precisely onto all three elements held simultaneously:
“Focus on harm not performance — willingness to quantify and publicise it.”
Quality is not a CQC rating or a four-hour target. It is the patient in the corridor experiencing delay-related harm. Quantifying and publicising that harm made the quality definition visible to everyone — and made the gap between current reality and the desired state permanently uncomfortable.
“Remove the comfort blanket that beds create flow.”
Classic Scientific Approach: test the assumption empirically rather than accepting it as given. The assumption that beds create flow was challenged with data and found false in some instances. This is Deming’s PDSA cycle applied to a management belief rather than a process.
“Better alignment from speciality to Board — no separation between quality, safety and flow.”
Deming’s Point 8 applied structurally: fear removed, staff trusted to make decisions, information flowing honestly from front line to board. The Exec Tri holding all three elements as one system rather than three competing workstreams.
The dissolved contradiction
The most important result in the Gloucestershire data is not the 68% reduction in ambulance handover hours. It is the SHMI trend: mortality falling for 14 consecutive months while throughput increased.
This is the empirical dissolution of the assumed contradiction between flow and safety.
The Evaporating Cloud — flow vs safety at Gloucestershire
A — Objective: Deliver high quality patient care
B — Requirement 1: Hospital needs faster patient flow
C — Requirement 2: Patients need safe, thorough care
D: Speed up processes, reduce time per patient
D′: Maintain thoroughness, take the time needed
Conflict: Speed vs thoroughness
The assumption: Speed and thoroughness are in conflict — improving flow compromises safety
The injection: “No separation between quality, safety and flow.” When the Exec Tri held all three simultaneously as one system, the trade-off could not be maintained — because there was no separate team to accept the downside of either.
The empirical result: SHMI fell for 14 consecutive months while throughput rose. More flow. Better safety. The contradiction was not a natural law. It was an assumption about how the system had to be organised. Challenge the assumption, and the contradiction evaporates. Not through compromise — Paul O’Neill at Alcoa demonstrated the same principle in manufacturing thirty-five years earlier. Both safety and productivity improved simultaneously when processes were fixed at source.
The DRD data — what it tells us about the constraint
The discharge-ready patient (DRD) figure fell from 169 to 118 per day — a 30% reduction. This is significant. But the figure is still at 118, well above zero and above the trust’s own target.
This is analytically important. It tells us precisely where the remaining constraint sits:
| What the DRD data shows | Interpretation |
|---|---|
| DRD fell from 169 to 118 (30% reduction) | The internal constraint — departmental silos, coordination failures, discharge process design — has been substantially addressed. Level 3 interventions within the trust boundary worked. |
| DRD remains at 118, above target | The external constraint — social care capacity, community bed availability, local authority discharge coordination — is still partially binding. It cannot be addressed by trust-level action alone. |
| Ambulance handover improved 68% while DRD only improved 30% | The internal coordination constraint was more binding than the external discharge constraint for ambulance handover. Dissolving the internal constraint released most of the handover improvement without requiring social care reform. |
Gloucestershire’s DRD data is the clearest public evidence that the corridor care constraint has two components: internal coordination (addressable within the trust boundary) and external discharge capacity (requiring system-level action). The 30% DRD improvement shows the internal component being substantially addressed. The remaining 118 DRD patients show the external component still binding.
This is the disaggregation that every national programme is missing. Not “is the constraint social care or internal dysfunction?” — it is both, in proportions that vary by trust. Bootstrap CUSUM applied to trust-level data, before any national intervention, would tell each trust which component is most binding in their specific system.
Sustainability assessment
| Intervention | Sustainability risk | Structural indicator |
|---|---|---|
| No separation between quality, safety and flow (Exec Tri) | Medium. Depends on the executive team maintaining the shared accountability structure. Vulnerable to leadership change or financial crisis that pulls teams back toward separate priorities. | Is the shared accountability written into governance structures and performance frameworks, or does it depend on the current executive team’s commitment? |
| Focus on harm not performance — quantified and publicised | Low-Medium. Harm quantification is now public and embedded in reporting. Harder to reverse than an internal policy change. But “publicising” requires ongoing leadership confidence. | Is harm data reported publicly and independently, or through trust-controlled channels that could be quietly narrowed? |
| Remove the comfort blanket that beds create flow | Low. This is a belief change backed by data. Once the assumption has been tested and found false, it is difficult to reinstate without contradicting the evidence. The most durable of the six changes. | Is the evidence of the assumption’s failure documented and available to future leaders? |
| Dedicated clinical leadership on flow | Medium. Depends on the individual remaining and the role being maintained at sufficient seniority. Gloucestershire’s model is more robust than Watford’s if the clinical leadership role is embedded in the structure rather than held by one person. | Is the role defined structurally or personally? Can it survive the departure of the current post-holder? |
| Staff morale — trusted, connected, listened to | High risk. Culture is always one leadership change away from reverting. The 33% decline in staff recommending GHFT as a place to receive care took one bad period to produce. Recovery required sustained leadership. Reversal could be faster. | Is psychological safety embedded in recruitment criteria, appraisal, incident reporting culture? Or dependent on the current CEO modelling it personally? |
| Partner relationships reset on flow | High risk. Cross-boundary relationships are the most fragile of all. Budget pressures, personnel changes, or political shifts in local authority priorities can dissolve them without any action by the trust. | Are the relationships formalised in joint accountability structures, or informal and personal? |
Bootstrap CUSUM — the honest position
Gloucestershire is the closest of the five Bright Spots to a Bootstrap CUSUM-confirmable result. The CEO’s presentation includes time-series charts — ambulance handover hours per month from June 2023 to March 2026, and DRD daily counts from April 2022 to April 2026. Both show what visually appears to be a sustained downward step change beginning in late 2023 or early 2024. The improvement sustained and deepened over three years against rising demand.
What would complete the Bootstrap CUSUM evaluation:
Of all five corridor care Bright Spots on this site, Gloucestershire provides the strongest public evidence of genuine structural improvement. Rising demand rules out demand reduction. Fourteen months of SHMI improvement rules out the safety-flow trade-off. The DRD data separates internal from external constraint. The time-series charts are consistent with Bootstrap CUSUM change points. The only missing element is the pre-committed prediction — and the evidence is strong enough that the absence of a formal pre-commitment does not seriously undermine the conclusion.
The honest statement: we are highly confident the improvement is real. We cannot prove from the public evidence which of the six interventions was the binding causal mechanism. Both claims matter — the first for practice, the second for replication.
Apply Bootstrap CUSUM to your trust’s flow data
If you work at a trust with monthly ambulance handover hours or daily DRD counts, upload your data to the StepChange Analyzer. Bootstrap CUSUM will identify whether a structural change point has appeared — and when.
▶ Open the StepChange Analyzer