The problem with maternity queues

Every week, NHS maternity units face the same difficult problem. They have a fixed number of theatre slots for planned caesarean sections and more women who need them than capacity allows. In many units, a simple rule determines who gets scheduled first: the woman with the nearest due date.

It sounds clinically sensible. In practice, it quietly fails both groups of women it is meant to serve.

For the past few years, I have been studying this problem with LBS PhD student Yuhang Du and my clinical collaborators Professor Catherine Aiken and Dr Pedra Rabiee at the University of Cambridge. The solution, it turns out, requires no additional clinical capacity.

Two patients, one broken queue

At Cambridge University Hospitals — the unit at the centre of our research — demand for elective caesarean theatre slots exceeds supply by a ratio of 1.32 to 1. For every 100 planned slots available, around 132 women need one. That gap creates a hidden unfairness.

The hospital classifies patients into two groups. Early-engaging patients are women who know months in advance they will need a caesarean section, perhaps because of a previous caesarean, maternal preference, or another clinical reason. They enter the system early, their situation is predictable, and they do what the hospital asks.

Late-engaging patients are those whose need emerges close to term: a breech presentation detected late, a condition that deteriorates unexpectedly. They arrive unpredictably, with less time and often greater clinical complexity.

Both groups wait in the same queue. And the earliest-deadline-first rule treats them identically: whoever's due date is nearest gets the next slot.

For early-engaging women, this means months of uncertainty. They are known to the hospital, but because their deadline is still distant, they are repeatedly passed over. By the time their due date approaches, they are competing for slots alongside everyone else.

Engaging early, it turns out, buys no certainty at all.

The cost of waiting

In an internal survey at Cambridge, early-engaging patients expressed near-zero satisfaction with the pathway — not low satisfaction, zero. Most did not receive a confirmed surgery date until roughly a week before delivery. The uncertainty affected childcare arrangements, work plans, and anxiety levels across an entire pregnancy.

A single queue can look fair because everyone is in the same line. But when patients have different deadlines and different risks, one line hides inequity rather than solves it.

For late-engaging patients, the problem runs in the opposite direction. Despite often carrying higher clinical risk, they have no protected access to capacity when they enter an already crowded queue. In our data, around 20 percent of these cases became what we call case losses — planned caesareans that could not be delivered through the elective pathway and had to be escalated to emergency.

Why changing the priority rule is not the answer

The obvious response is to change the scheduling rule. But no alternative works within a single shared queue. First-come-first-served would favour early engagers while potentially crowding out urgent late-arriving patients. Strict clinical priority simply recreates the current problem.

The deeper issue is structural. One shared queue cannot satisfy two different fairness objectives when capacity is insufficient. The problem is not that hospitals have chosen the wrong rule. It’s that the system makes it impossible to serve both groups well at the same time.

The fix: split the queue

Our solution is deliberately simple: divide planned theatre capacity into two protected streams – one for early-engaging patients, one for late-engaging patients. Each stream has its own reserved slots and neither group can displace the other.

This is not a proposal to add operating theatres or ask already-stretched teams to work harder. It is instead about allocating existing capacity more deliberately.

The practical question becomes: how much capacity should each stream receive? Our model allocates slots to the point where the next available slot creates equal marginal benefit in each stream. In plain terms, put the next slot where it prevents the greatest harm. For Cambridge, the model recommends roughly two-thirds of planned capacity for early engagers and one-third for late engagers.

Once early engagers have their own protected stream, something else becomes possible: the hospital can confirm a surgery date at the point of clinical request, potentially months before the due date, without worsening overall system performance.

This was one of our most counterintuitive findings. Many organisations delay committing to appointments in the belief that flexibility has operational value. Our model shows that in this case, the delay creates anxiety and administrative burden without providing any benefit. The hospital is withholding certainty it can already provide, at no cost.

What the data shows

We back-tested the model against approximately 2,500 elective caesarean cases from Cambridge University Hospitals across two years (2024-2025) and the results were striking. Under the current system, satisfaction among early-engaging patients was effectively zero. With the reservation policy, at least 70 percent would receive a confirmed date at the point of request, while case losses among late-engaging patients fall from around 22 percent to below 10 percent. Notably, the total case-loss cost can be lowered by approximately 9 percent, scaling to savings of over £1.9 million per year if applied across NHS.

There is a trade-off. The model estimates that one additional early-engaging patient per week may need to be escalated to emergency under the new policy. Our clinical partners commented this is acceptable: early-engaging patients are generally less clinically complex, and the risk profile of an unplanned caesarean for this group is substantially lower than for a late-engaging patient.

Cambridge University Hospitals recently began preparations for a one-year pilot. We built a simple interactive planning tool: they input the expected number of early and late-engaging cases for each week, along with available theatre slots, and the system calculates the optimal slot reservation split.

No additional staff or additional theatre capacity are needed. Just a number, and two buckets.

The broader lesson

This research goes beyond elective caesarean scheduling. Its insights apply to any overloaded system that uses a single rule to patients with fundamentally different needs and deadlines.

Queues are not neutral. A shared waiting list can look fair, everyone in one line, same rule applied to all, while quietly distributing the costs of scarcity in ways nobody intended and nobody is monitoring. Identical treatment in an overloaded system is not the same as fair treatment.

In maternity care and beyond, the most important question is not always: what is the right order within this queue? Sometimes it is rather, should this be one queue at all?