An economist staggers into a hospital...

Thoughts on an emergency admission

This post’s title might sound like the lead-in to a bad joke, but my recent arrival at the emergency department (ED) of a provincial New Zealand hospital was not. But even when suffering the symptoms and side effects of appendicitis, the “economist” in my brain refused to switch off. Anyone with a different background would have noticed different things, and perhaps had a different subjective experience.¹ That said, I think my observations are worth sharing, and hope you agree.

Waiting outside the emergency department

Emergency hospital care is zero-priced in New Zealand. Don’t get me wrong, I think that’s a good thing. Hospitals should not be turning away people with appendicitis because they cannot, or are unwilling to, pay the cost of care. But zero-pricing almost always has consequences. When demand exceeds supply — as it inevitably does in hospital emergency departments — non-price rationing takes over.

Economics tells us that, other than price, there aren’t that many choices for rationing mechanisms. The ED appears to use two mechanisms in combination.²

• Queuing allocates resources to people in the order they arrive. It replaces the willingness-to-pay criterion of price allocation with a willingness-to-wait criterion.

• In rules-based allocation, a human (or computer) applies pre-specified rules (and sometimes professional judgement) to decide who goes next.

The gatekeeper to ED was a “triage” nurse. ED ration via a process called triage, which uses rules to allocate incomers into three queues. Those in higher priority queues always receive treatment before those in lower priority queues.

From what I observed, the three queues were:

1. Likely to die in the waiting room.

2. Won’t die in the waiting room, but is in dire need of treatment they will only get at this hospital.

3. Could get treatment elsewhere or cope without it.

Those in queue 1 went straight through. Queues 2 and 3 stayed in the waiting room until called. For me — presumably allocated to queue 2 — a 2.5 hour wait was unpleasant, but without clinical consequences. The even longer waiting times for queue 3 acted to discourage those who could afford alternative treatment.³

Inside the emergency department

Inside, it was busy. Frantic even. People rushed in all directions. It appeared to be controlled from a raised area brimming with staff. Perched above them were large computer screens, supporting coordination of the apparent chaos into a smooth, well-oiled machine … or not. Every one of those large, presumably expensive, screens displayed nothing but the current date and time.

I was perplexed. Computers excel at supporting complex coordination tasks. Yet this department seemed driven by oral instructions. And the computer screens offered no more than would a mechanical clock.⁴

For me, things did move more quickly. I was now occupying a scarce resource (a bed in ED), rather than queuing for one. And the faster they could move me elsewhere, the quicker they could clear that resource.

It was in ED that I encountered more quirks of the system:

• Demarcation of activities by profession. I needed to be fitted with a cannula (a thin tube inserted into a vein in my arm for blood sample collection and IV antibiotics). My first cannula was fitted by a 4th year medical student. He found the task difficult, so relied on advice and instruction from a (clearly much more experienced) nurse. But at this hospital, cannula insertion is the work of doctors, not nurses, irrespective of skill or expertise.

• A robust firewall between the primary (ie, general practice) and secondary (ie, hospitals) healthcare systems. An experienced GP had called the hospital ahead of my arrival, and sent me there clutching a written diagnosis. These communications went into a black hole. Further, the admission process did not record the name of my GP, and on release I was told not to seek follow-up from my GP.⁵

In the surgical ward

Around me, there was a lot of data collection going on. But, from an operations perspective, it didn’t seem to be acted on. By way of example, I was given a questionnaire to fill out on day 2. I answered no to the question “is the information on your wrist bracelet correct”, as it contained an error. I assumed this would be quickly followed up and the bracelet corrected, as I thought that the bracelets are part of a patient safety system, helping to ensure that the right people get the right medications and operations. My assumption was incorrect, and my bracelet details were still incorrect when I left the hospital on day 4.

Surgeons and theatres are the scare resources. To make sure they are not left idle, it appears that the hospital overbooks them. (Similar reasons lead airlines to overbook seats on planes.) A consequence is that each day more patients are prepped for surgery than are likely to actually receive an operation. My surgery, scheduled for day 2, was bumped to day 3.

Lots of machines that go ping (but don’t talk to one another)

The hospital seemed full of machines. A lot of them were noisy. The hi-tech machine operating my IV line beeped loudly (a) if the line got blocked (presumably a moderately urgent situation), (b) if the IV bag emptied (presumably much less urgent), and (c) if I disconnected the power cord so I could maneuver myself and machine to the bathroom. (c) was decidedly non-urgent, as the machine was rated to run 8 hours on its internal battery. The machines didn’t talk to the hospital’s IT systems, so staff tended to ignore beeping machines, instead relying on the patient to press the staff call button in situation (a).

From an IT perspective, having sophisticated machines that don’t talk to one another is decidedly primitive. A missed opportunity for both efficiency and better patient care.

Patched up, and out the door

Service businesses know that customer perceptions of good quality are essential to getting new and repeat business. So much so that many bombard their customers with feedback forms, ratings and questionnaires. While too much of this is annoying (yes, Airbnb I’m talking about you), too little can tell you a lot about which customers matter to the business. I received no follow-up, of any kind, after leaving the hospital.

Does economics offer a plausible explanation for what I observed?

I saw staff working hard, doing their best and, for the most part, staying cheerful. The hospital’s infrastructure appeared reasonably modern and sound, and clinical care seemed good. But its systems seemed very creaky and inefficient.

Theory tells us that monopolies overcharge, under-innovate, and are neither efficient producers nor responsive to their customers.

Provincial public hospitals are monopolies, but zero-charging complicates this story. These hospitals have two types of customers: patients (who receive services, but don’t pay), and government (which pays, but doesn’t receive the services). Hospitals face endless demand from patients, but their government customer is crucial to their existence. So, you’d expect the hospital to be more attentive to its funder than its patients.

However, both types of customers find it difficult to evaluate quality of clinical care. This creates a situation of asymmetric information, as medical professionals know more about the quality of care provided than do their customers. Mechanisms to resolve asymmetric information problems can be costly and, at best, are only partially effective. But they are important. Similar information asymmetries arise in education and social services, and these, together with health, form a substantial, and costly, part of the economy. Furthermore, they seem resistant to the efficiency and productivity improvements that information technology has enabled in the wider economy.

Could more funding address these problems? Certainly, well-targeted systems improvement could increase hospital capacity and throughput. But decreased waiting times may induce additional demand, leaving the hospital just as busy as it was the day I staggered in.

The New Zealand Government is undertaking a costly top-down restructure of the health system, focusing on funding and governance. Whatever its merits, I can’t help but think the restructure won’t address the issues I experienced. But I’m keen to hear other views — comment below or email me directly.

Your editor, on location

1. I do not have a professional background in healthcare, and so cannot offer an evaluation of the clinical care I received. That said, they patched me up, sent me on my way and I’m feeling much better without that pesky appendix. So, a big thankyou to everyone involved in my care.

2. Random allocation is another possibility, but not to my knowledge used in hospitals. The rules applied in emergency situations are typically based on an assessment of the clinical status of the patients, and their likely response to treatment. Other rules-based allocation systems can use a wider range of criteria — including status, patronage, who-you-know, side payments, and group membership.

3. Some people abandoned the waiting room in frustration and disgust while I was there. It left me wondering if the hospital could somehow credibly signal expected wait times in advance, and leave people to decide whether that exceeded their willingness to wait. Such a system would lead to shorter queues — and fewer frustrated people.

4. I can’t rule out a simple explanation, like “the software isn’t ready yet”. But it did look like a missed opportunity for significant efficiency gains.

5. I understand that better integration of primary and secondary healthcare has been a goal of public policy for decades. So I was surprised to see just how poorly integrated the two were.