The Joiner Triangle
Quality. Scientific Approach. All One Team. Brian Joiner’s Triangle identifies three interdependent elements of effective management — and makes one observation that most organisations violate: taken separately they are not as effective. They are best used together. The Triangle is not three programmes. It is one system. When any element is separated from the others, the contradictions that prevent improvement remain intact.
Quality, Scientific Approach, and All One Team are not three separate improvement programmes. They are one interdependent system. Separating them — running Quality as a clinical governance programme, Scientific Approach as a data team function, and All One Team as an HR initiative — is what produces the contradictions that make improvement fail.
Bootstrap CUSUM is the Scientific Approach element applied to the honest test of whether Quality has actually improved — and whether the All One Team conditions are in place to sustain it. A Bootstrap CUSUM flat line after an improvement programme is evidence that one or more elements of the Triangle were absent.
☰ Contents
The three elements
Quality is defined by the customer — not just getting rid of what hassles them, but understanding their current and future needs. Eliminating complexity, waste, rework, and unnecessary work. Finding out and acting on what is important to customers and spending resources wisely.
In NHS terms: quality is not a CQC rating or a four-hour target. It is what matters to the patient in front of you — safe, timely, effective care without the harms that corridors and delays introduce.
Managing the organisation as a system — not silos. Developing process thinking. Basing decisions on data and facts. Understanding variation. Rapid learning and rapid improvement — no substitute for knowledge, finding the facts, using PDCA. Knowledge will create a stream of improvement ideas.
In NHS terms: Bootstrap CUSUM is the Scientific Approach test. Has the system structurally changed? The answer is in the data — not in the narrative, not in the action plan, not in the improvement team’s confidence.
Believing in people, treating everyone with dignity, trust, and respect. All working toward a win for all stakeholders — customers, staff, shareholders, and the community. No blame for staff. Staff involved in improvements. Trusted to make intelligent decisions on behalf of the company and its customers. Good policies.
In NHS terms: no blame culture, staff empowered to speak out when they see a problem, fear removed. Without All One Team, the information that Scientific Approach needs to function honestly does not flow upward.
Joiner’s original Triangle uses “Quality” as the apex — defined broadly as what matters to the customer, eliminating waste, rework, and unnecessary complexity. In applying the Triangle to NHS corridor care, I have adapted this to “Quality and Safety” together at the apex, because in a healthcare context the two are inseparable: a patient harmed by a delay has not received quality care, and a patient discharged unsafely has not received quality care. They are the same thing defined from the patient’s perspective.
In other contexts — particularly manufacturing and process industries — Quality (product conformance, efficiency, productivity) and Safety (worker and process safety, zero harm) are sometimes treated as separate dimensions with their own potential tensions. Paul O’Neill at Alcoa demonstrated that this separation is also a false assumption: fixing the processes that cause harm also fixes the processes that cause waste and inefficiency. Quality and Safety, properly understood, are not in conflict in any industry. They are both symptoms of how well the system is designed.
Why interdependence is the whole point
Most organisations treat the three elements as separate programmes. Quality becomes a clinical governance function. Scientific Approach becomes a data analytics team. All One Team becomes a staff engagement or HR initiative. Each has its own budget, its own leadership, its own reporting line, its own annual report.
This is exactly what Joiner warned against. Separated, the three elements undermine each other:
| What gets separated | What goes wrong |
|---|---|
| Quality without Scientific Approach | Improvement claims are made without honest data. Seasonal variation is attributed to programmes. Regression to the mean is mistaken for genuine improvement. The question “is this improvement real?” has no rigorous answer. Bootstrap CUSUM is not applied before attribution is made. |
| Quality without All One Team | Staff know what is wrong but do not say so. The information that would identify the real constraint stays at the front line. Improvement programmes are designed by people who cannot see what the people doing the work can see. Fear makes the constraint invisible. |
| Scientific Approach without All One Team | Data is collected but gamed. Metrics are chosen to show improvement rather than to test it. Staff report what the system rewards rather than what is true. The data does not reflect reality because fear has distorted the signal. |
| All One Team without Scientific Approach | Good culture, willing staff — but no method for testing whether what the team is doing is actually working. Effort is not distinguished from impact. The team works harder on interventions that are not reaching the constraint, without the data to tell them so. |
| Scientific Approach without Quality focus | Data without purpose. Metrics are measured because they are available, not because they reflect what matters to the patient. A trust can have excellent data infrastructure and still be measuring the wrong things — process measures rather than outcome measures, activity rather than improvement. |
The contradiction the Triangle dissolves
The most persistent contradiction in NHS management is the assumed conflict between flow and safety and quality:
The assumed contradiction
“We need more flow — faster throughput, shorter stays, quicker discharge. AND we need quality and safety — which requires time, thoroughness, and careful handover. These are in conflict. Increasing flow compromises safety. Maintaining safety slows flow.”
This contradiction is accepted as an unavoidable fact of life in most NHS trusts. Flow and safety are managed separately, by different teams, with different metrics, and with an implicit assumption that progress on one comes at the expense of the other.
The Joiner Triangle says this is a false contradiction — produced not by the nature of hospital care but by the way the system is organised. Separate the three elements of the Triangle, and flow and safety become competing priorities. Hold them together, and they become the same thing.
Quality (defined by the patient) requires both safety and timeliness. A patient harmed by a delay has not received quality care any more than a patient harmed by a rushed discharge. Quality is not a trade-off between flow and safety — it requires both simultaneously.
Scientific Approach applied to both flow and safety data shows that the assumed trade-off often does not exist in the data. When corridor care decreases, safety events typically decrease too — because the care environment improves. The trade-off was an assumption about how the system had to work, not an empirical observation about how it does work.
All One Team removes the organisational structure that created the separation in the first place. When quality, safety, and flow are held by the same people, in the same room, with shared accountability, the trade-off cannot be maintained — because there is no separate team to accept the downside.
The Triangle in NHS practice
The Triangle maps directly onto the most common failure modes in NHS improvement:
All One Team Fear prevents honest information flow. Staff can see the constraint — the boarded patient on the wrong ward, the discharge that isn’t happening, the protocol that nobody follows because it doesn’t work — but do not report it because the management system punishes bad news. The Scientific Approach cannot function without honest data. Quality cannot improve when the information about what is wrong is suppressed by fear. Deming’s Point 8 (drive out fear) is the prerequisite for the other two elements of the Triangle to function.
Scientific Approach Improvement claims without honest testing. An intervention is applied. The metric moves. The improvement is claimed. Nobody asks whether the movement is structural and sustained, whether it would survive a Bootstrap CUSUM test, whether it is seasonal, or whether it is regression to the mean after a bad period. The Scientific Approach element requires the pre-committed prediction — the question asked before the data arrives, not after. Without it, every intervention appears to work and none can be distinguished from noise.
Quality definition Measuring what is measurable rather than what matters. A trust can have excellent data on four-hour performance, ambulance handover times, and bed occupancy — and still have no idea whether the patient in the corridor is being harmed. The Quality element requires the organisation to define quality from the patient’s perspective, to measure harm directly, and to hold itself accountable for what matters rather than what is reported. Gloucestershire’s “focus on harm not performance” is this element being applied.
Gloucestershire — the Triangle in action
Gloucestershire Hospitals NHS Foundation Trust provides the most complete public evidence of the Joiner Triangle operating as a system in an NHS context. Chief Executive Kevin McNamara’s June 2026 presentation to NHS England maps almost exactly onto the three elements.
| What Gloucestershire did | Triangle element | Result |
|---|---|---|
| Focus on harm not performance — quantify and publicise it | Quality (patient-defined) | Reframed the organisation’s accountability from performance metrics to patient harm. Made harm visible to everyone simultaneously. |
| No separation between quality, safety and flow (Exec Tri role) | All three elements held together | The Triangle operated as a system at executive level. The contradiction between flow and safety could not be maintained when the same people were accountable for both. |
| Remove comfort blanket that beds create flow | Scientific Approach | Tested an assumption empirically and found it false. Classic Scientific Approach: challenge the belief with data rather than accepting it as a given. |
| Dedicated clinical leadership with focus on flow | Scientific Approach + Quality | Clinical expertise applied to the flow problem — not administrative management of a clinical process but clinicians owning the system design. |
| Better alignment from speciality to Board | All One Team | Information flowed honestly from front line to board. The silo between speciality and executive was dissolved. Accountability became shared rather than separated. |
| Challenging historic practices, culture, behaviours, expectations | All One Team + Quality | Fear removed. Historic practices that nobody had questioned were made visible and challenged. Staff trusted to make intelligent decisions. |
Gloucestershire’s Summary Hospital-level Mortality Indicator (SHMI) fell for 14 consecutive months while throughput increased — ED attendances rose from 153,960 to 165,110 per year and ambulance handover time fell from 80 to 26 minutes. More flow. Better safety. The contradiction was not a natural law. It was an assumption about how the system had to be organised. When the Joiner Triangle operated as a system — Quality, Scientific Approach, and All One Team held together without separation — the contradiction evaporated.
This is Goldratt’s Evaporating Cloud in empirical form. The cloud was drawn. The assumption was identified (“flow and safety are in conflict because they are managed separately”). The assumption was challenged (“no separation between quality, safety and flow”). The contradiction dissolved — and the data confirmed it. See the full analysis: Corridor Care Bright Spots — Watford and Gloucestershire.
The Triangle and Joiner’s levels of fix
The Joiner Triangle and Joiner’s levels of fix are two parts of the same framework, used at different moments in the improvement process.
The levels of fix answer: at what level is the intervention applied? Level 1 fixes the output. Level 2 fixes the process. Level 3 fixes the system. Bootstrap CUSUM tells you whether you reached the right level.
The Triangle answers: what conditions must be in place for a Level 3 fix to succeed? A Level 3 intervention requires all three elements of the Triangle to be operating simultaneously:
- Quality must be defined precisely enough to know what a successful Level 3 outcome looks like — which outcome measure, moving in which direction, by how much, sustained over how long
- Scientific Approach must be applied to confirm the Level 3 change point — Bootstrap CUSUM, pre-committed prediction, honest reporting of a flat line if that is what the data shows
- All One Team must be in place for the Level 3 constraint to be visible — because Level 3 constraints are almost always structural features of the system that only become visible when fear is removed and information flows honestly
A Level 3 intervention applied in an organisation where All One Team is absent will fail — not because the intervention is wrong, but because the information needed to design and implement it correctly is being suppressed by fear. The Triangle is the prerequisite for Joiner’s levels to produce lasting structural change.
Bootstrap CUSUM as the Scientific Approach test
The Scientific Approach element of the Triangle requires that decisions are based on data and facts, and that the organisation understands variation. Bootstrap CUSUM is the operational instrument for both requirements.
It answers the specific question the Scientific Approach demands: has the system structurally changed, or is what we are seeing common cause variation? Without that answer, Quality improvement claims cannot be validated and All One Team cannot be given honest feedback about whether what they are doing is working.
📝 The pre-committed prediction as the Triangle test
A trust applying the Joiner Triangle correctly will, before implementing a Level 3 intervention, make a pre-committed prediction: which metric will change, in which direction, within what timeframe, at what confidence level. This is the Scientific Approach element applied prospectively.
Bootstrap CUSUM then tests the prediction. If a change point appears at the predicted time and in the predicted direction — the Triangle is working. If a flat line appears — one or more elements of the Triangle was absent or the intervention did not reach the constraint. A flat line is not a failure to be explained away. It is the most honest output the Scientific Approach can produce: the system has not structurally changed. The loop begins again from Step 1.
Test whether your improvement has produced structural change
Upload your outcome metric data to the StepChange Analyzer. Bootstrap CUSUM will tell you whether a structural change point appeared — the Scientific Approach element of the Joiner Triangle applied to your data.
▶ Open the StepChange Analyzer