Could NZ's hospital EDs see 95,177 more people on time?

At The Cause Collective NZ we were founded in 2016 by a health & wellbeing transformation dream articulated in a 2012 white paper that used cause & effect logic from the Theory of Constraints (TOC; Dr Eli Goldratt). TOC analysis and scrutiny tools use the scientific method. Consequently TOC is a principle and ethic in our thoughts on health sector matters.

Earlier this week on 17 February 2025 The New Zealand Herald: https://bit.ly/4b2z0cl reported that "All hospital emergency departments (EDs) are failing to assess patients with “imminently” or “potentially” life-threatening conditions on time," and "More than 300,000 patients with time-critical conditions were not seen within recommended time frames during the first six months of last year, according to official information obtained exclusively by the NZ Herald."

Can the principles of Critical Chain Project Management (CCPM) and the Theory of Constraints (TOC) be applied to hospital emergency departments (EDs) to alleviate bottlenecks and reduce patient queues? If so then both CCPM and TOC would focus on identifying and managing constraints to optimise flow and throughput.

1.????? Identifying the Bottleneck: The first step is to identify the constraint or primary bottleneck in the ED.

Is it triage?

Or physician assessment?

Or lab testing?

Or bed availability?

In the NZ Herald article, Sarah Dalton, the executive director of the Association of Salaried Medical Specialists (ASMS said that In New Zealand's EDs, "access block" due to a lack of free beds in wards is a contributing factor to long waits.

But .. an "access block" is a bottleneck.

So, the TOC approach is to resolve or eliminate the bottleneck by:

1. Exploiting the Bottleneck: eliminating the bottleneck requires us to maximise the efficiency of the constraint. For example, if physician assessment is the constraint, intervention measures would include streamlining assessment protocols or ensuring physicians are not delayed by non-essential tasks.
2. Subordinating Other Processes: Non-bottleneck areas must be aligned to support the bottleneck, as "a system of local optimums is not an optimum system at all; it is a very inefficient system". This would mean reducing unnecessary tasks performed by nurses or other staff to allow them to support the bottleneck area.

Dr Eric Topol writes:

.. that in USA, more than 20 percent of healthcare spending is related to administration. This is far from automated !!!

Consequently manual, human scheduling for operating rooms or staffing of inpatient and outpatient units in a hospital leads to gross inefficiencies. Much of the work related to patients calling in to schedule appointments could be accomplished with AI natural-language processing, using human interface as a backup.

Algorithms are already being used in some US health systems to predict no-shows for both hospital staffing and patient’s clinic appointments. This is a significant source of inefficiency because staff absences and patients missing appointments create resource contention. This is a damaging bottleneck.

It causes a catastrophic, chain-reactions network of suspended diagnoses & treatments which result in delays and deaths because hospital workflow systems feed the bottleneck instead of eliminating it.

No wonder ED personnel and associated staff are burnt out.

SO WHAT DO WE DO TO FIX ED WAITING TIME?

1. Elevate the Bottleneck: If bottlenecks still exist, consider adding resources or changing processes to increase capacity at the constraint. For example, creating a new assessment area to expedite triage. That is the only solution, to the observation by the New Zealand Chair for the Australasian College of Emergency Medicine, Dr Kate Allan. She says, “When it’s day after day and you’re unable to deliver care or you see harm happen that is when it gets really challenging. When you’re buffering a system that feels broken, that can be really challenging.”
2. Prevent Inertia: After addressing one bottleneck, keep checking for new constraints that may emerge.

BUFFERING

Buffers can be strategically placed. This could translate to creating a "patient buffer".

For example:

A capacity buffer ensures adequate staffing during peak hours. This would handle the situation described by HNZ chief clinical officer Dr Richard Sullivan. He said, “he was aware of the failures, acknowledged staffing, workload and patient surges in EDs contributed to the problem.”

A time buffer is built into the schedule to absorb unexpected delays.

A resource buffer ensures that critical equipment and personnel are available when needed.

AND SO BUT THEN ..

Applying TOC and critical chain thinking can help EDs move away from "pushing in more material (patients) than the system can convert into throughput," (resulting in "excess inventory") to "release the material for the red parts according to the rate at which the bottlenecks need material—and strictly at that rate".

HYPOTHETICAL IMPROVEMENT CALCULATION

Dr Eric Topol writes about efficiency gains with AI and predictive tools. For example Qventus, through using multimodal data to predict proceedings in a hospital’s emergency department, claims to have achieved a marked reduction in patients leaving the emergency room without being seen and the time it takes for a doctor to see a patient.

Topol also observes reduced readmission risk in that for every extra minute a home health visit lasts, there was a reduction in the risk of readmission of 8 percent. This does not directly relate to ED wait times. But it does indicated the impact of timely intervention on patient outcomes.

And improved palliative care efficiency through AI tools which predict time to death with unprecedented accuracy will influencing palliative care teams and improve efficiency without compromising care. The effect is to release partial FTE nursing staff and physicians to be included in the Resource Buffer.

Therefore if we estimate conservatively that Theory of Constraints interventions would cause an average improvement of 15% then 95,177 more people might have been seen on time in the first six months of last year (2024)!

Hmm .. worth a thought.

REFERENCES

Topol, E. (2019). Deep medicine: How artificial intelligence can make healthcare human again. Basic Books

Morrah, M., & Knox, C. (2025, February 17). How does your hospital emergency department ED rate? Herald investigation into the best and worst-performing EDs. NZ Herald.

For more information 18 contact me: [email protected]