Your IVF Cancellation Rate Is a Lagging Indicator. Here's What to Track Instead.

Let me start with something that might be slightly uncomfortable to hear.

Your cycle cancellation rate, the one that shows up in your monthly report, your annual review, your DHA submission, is information that has arrived too late to be useful.

I don't mean it's not important. It absolutely is. But by the time that number appears on the page, everything has already happened. The medication was given. The monitoring appointments were used. The patient sat through stimulation that wasn't working. And then the cycle was stopped, and the number went up by one.

The cancellation rate is a measure of what you couldn't prevent. And if you're only tracking it after the fact, you're spending a lot of clinical energy looking in the rear-view mirror.

The question most clinics aren't asking

Here's the question that actually matters: which patients in your current cohort are most likely to cancel before you start treating them?

Not which patients cancelled last quarter. Which ones are going to cancel next month. and what can you change before the cycle begins?

That shift, from tracking to predicting, is where the real clinical and commercial value of IVF analytics lives. And it starts with something most clinics skip entirely: understanding where in the cycle their cancellations are actually happening.

Your cancellation rate is not one number. It's four.

This is the thing that surprises most medical directors when we first look at their data together.

A cycle can be cancelled at four completely different points in the treatment journey. Each one has a different cause. A different patient profile. A different fix. And when you lump them all into one headline percentage, all of that information disappears.

During stimulation. poor ovarian response. Not enough follicles develop to proceed to egg retrieval. This is where AMH does its most important work. because low AMH is identifiable at first consultation, before a single dose of medication has been given. If your clinic is using AMH as a baseline measurement rather than a predictive tool, you're collecting the right data and not acting on it.

Before retrieval. hyperstimulation risk. The cycle is stopped to protect the patient from OHSS. This risk is also predictable. Younger patients with high AMH and high antral follicle counts are over-represented in this category. Modified protocols and trigger decisions made at the stimulation planning stage can reduce this. but only if you know who these patients are before you start.

After retrieval. embryo failure. No viable embryos available for transfer. This one points directly at your embryology lab. Fertilisation rates, blastulation rates, maturation rates. tracked at the individual embryologist level. In clinics where these metrics aren't benchmarked individually, lab-level performance gaps stay hidden for months or years. The cancellation gets attributed to the patient rather than to the lab.

Administrative cancellation. The cycle stops for non-clinical reasons. finances, scheduling, the patient changed their mind. This is a completely different problem from the other three. And if it's being counted in the same number as clinical cancellations, your clinical cancellation rate is higher than it actually is.

Four categories. Four causes. Four solutions. One number tells you none of this.

The three metrics that should be sitting alongside your cancellation rate

If cancellation rate is the lagging indicator, here's what the leading indicators look like.

Cancellation rate by protocol, age group, and AMH category. An 8% overall cancellation rate might be hiding a 3% rate for patients under 35 with normal ovarian reserve and a 24% rate for patients over 38 with low AMH on a long protocol. The moment you break it down this way, you know exactly who needs a different clinical conversation at first consultation.

AMH-to-cancellation correlation tracking. AMH is the strongest single predictor of stimulation response. and by extension, cancellation risk. In analytical work on IVF cycle outcomes, removing AMH from a predictive model produces a larger performance drop than removing any other variable. Including age. Including AFC. Including FSH. Yet most clinics never analyse the relationship between their AMH distributions and their subsequent cancellation patterns. That relationship is sitting in your data right now.

First-consultation-to-cancellation rate by doctor. How many patients who proceed to stimulation after seeing a particular doctor ultimately cancel? Controlled for patient mix, this metric tells you whether first consultation conversations are appropriately calibrating treatment plans to individual risk profiles. A significant gap between doctors on this metric is a clinical signal worth investigating.

What happens when you add prediction

I want to be clear about what predictive analytics actually does here. because there's sometimes an assumption that it requires data you don't have.

It doesn't. The variables that go into a well-built cancellation risk model. AMH, AFC, FSH, LH, baseline estradiol, age, BMI, previous cycle history, diagnosis, protocol. are all collected at a standard first fertility consultation. No additional testing. No additional appointments. No additional cost.

What changes is what you do with that data.

Instead of using it to describe what happened after the cycle, you use it to predict what's likely to happen before the cycle starts. Patients are stratified into risk tiers. low, medium, high. based on their pre-treatment profile. High-risk patients get a different clinical conversation at first consultation. A different protocol. More intensive monitoring from day one. Realistic expectations that are set before treatment begins rather than revised after it fails.

In a well-built model, the majority of cycles that go on to be cancelled can be identified as elevated risk at the point of first consultation. That's not a guarantee of prevention. But it's the difference between finding out at day 8 that the stimulation isn't working and knowing at the first consultation that this patient needs a modified approach.

What a useful IVF analytics dashboard actually looks like

A useful dashboard doesn't show you a single cancellation rate. It shows you:

• Cancellation rate broken down by the four stages. separately, not aggregated
• Cancellation rate by protocol, age group, AMH category, and center
• Predictive risk stratification at first consultation. low, medium, high. based on pre-treatment markers
• Embryologist-level lab performance metrics. fertilisation rate, blastulation rate, maturation rate. individually benchmarked against each other and against ESHRE standards
• Monthly trend analysis so you can see whether your cancellation rate is moving in response to protocol changes
• Center-level comparison for multi-site clinics. because a 5-percentage-point performance gap between two centers is invisible in an aggregate number and obvious the moment you break it apart

This is what reporting becomes when it's designed to change clinical decisions rather than just document them.

Where to start

If your clinic currently has a single headline cancellation rate, the most immediate step is segmentation. Break it into the four stages. Break it by patient profile. Break it by protocol. Apply it to 12 months of cycle data.

What typically happens is that two or three patterns emerge immediately that weren't visible before. Which patient profiles are driving your cancellation rate. Which protocols are associated with higher risk. Whether the problem sits in stimulation response, lab performance, or patient management.

The number stays the same. What changes is your understanding of what it's telling you. and what you can actually do about it.