The north-east question.

The ICER is a rate.

ICER stands for incremental cost-effectiveness ratio, and every word earns its place. It's incremental — built from differences versus the comparator, never from totals. It's a ratio — one thing divided by another. And what it measures is cost per unit of effect:

Because we measure health in QALYs, the ICER comes out in £ per QALY — pounds paid for each extra quality-adjusted life-year the technology buys. If a drug's ICER is £18,000 per QALY, that's the price tag: every additional QALY it delivers costs the system £18,000 more than standard care.

That's the whole idea. The ICER turns "it costs £12,000 more and adds 0.8 QALYs" into a single, comparable rate — a price for health you can weigh against a budget.

The ICER is a slope.

Here's where the number meets the picture. Put the technology's point on the plane, draw a straight line from the origin — the comparator — up to it, and the ICER is the steepness of that line. Rise over run: ΔCost over ΔEffect. Exactly the ratio.

That makes the ICER easy to read at a glance:

• A shallow line means a lot of health for not much money — cheap QALYs, a low ICER.
• A steep line means you're paying a lot for each extra QALY — expensive QALYs, a high ICER.

So "is this cost-effective?" becomes a question about steepness — and steepness needs something to be compared against. That something is the threshold, and it's a slope too.

The slope and the ruler.

Drag the point. The solid line from the origin is the ICER — its steepness is the price per QALY. The dashed line is the cost-effectiveness threshold, fixed here at £20,000 per QALY. When your line is shallower than the dashed ruler, you're buying QALYs more cheaply than we're willing to pay — cost-effective.

Two things to notice. Cost-effectiveness is just your slope versus the ruler's slope. And the moment you drag off the north-east quadrant, the ICER goes negative — and a negative ICER shows up in the best corner and the worst. Hold that thought.

Computing it, step by step.

Let's do one by hand, because in a dossier you'll never be handed the ICER — you'll be handed four numbers and have to build it.

A new oral therapy is compared with standard care over a patient's lifetime:

• Standard care: £8,000, 5.0 QALYs
• New therapy: £20,000, 5.8 QALYs

Notice the discipline: increments first, ratio second, comparison last. Skip the increments and divide the totals (£20,000 ÷ 5.8) and you get a meaningless number — the ICER only ever lives in the difference between two options.

Now you.

The trap: one number, two lost signs.

The ICER is powerful because it squeezes two numbers into one. It's dangerous for exactly the same reason: squeezing them together throws away which direction each one pointed.

Watch what the sign of the ICER does across the four quadrants:

Read the ICER-sign column. A positive ICER means you're in the north-east or the south-west — quadrants that demand opposite decisions. A negative ICER means you're dominant or dominated — the best and worst corners on the plane, sharing one meaningless sign.

So the sign of an ICER tells you nothing about the decision. −£40,000 per QALY could be the finest technology you'll see all year or the one you reject on sight. The ratio remembers the size of the trade; it forgets the direction.

On this number alone, what can you conclude?

An ICER is meaningless without three things.

The ICER is a genuinely useful number — but only when it travels with its context. Strip any of three things away and it stops meaning anything:

• Its comparator. An ICER is always "versus something." The same drug measured against standard care and against an active competitor has two different ICERs. There's no such thing as "the drug's ICER" in the abstract — change the comparator, change the number.
• Its quadrant. As you just saw, the sign alone is ambiguous. You need the individual signs of ΔCost and ΔEffect — i.e. where the point sits — before the ratio means anything.
• A cost-effectiveness threshold. £15,000 per QALY is neither good nor bad until you compare it to what the system is willing to pay. Where that threshold comes from is the next lesson.

And one last subtlety, for the south-west. There, the ICER is positive but the decision rule inverts: you're giving up health to save money, so you'd only accept it if you save more than the threshold per QALY lost. Same positive number, opposite inequality. It's the sharpest reminder that a raw ICER is never a decision — it's an input to one.

There's a way to carry both numbers without collapsing them, and it dodges every trap on this screen. That's net benefit, two lessons from now.

Why this matters for HTA

The single most quoted figure in any economic dossier is the ICER — usually in bold, usually in the executive summary, usually presented as the verdict. Your job is to treat it as the start of an argument, not the end.

• A headline ICER without its comparator is a half-argument. The first question is always "versus what?" A flattering ICER against a weak comparator can hide a poor deal against the one that matters.
• A negative or "dominant" ICER earns more scrutiny, not less. Because the sign is ambiguous, a dominance claim has to be verified on both axes — cheaper and better — not accepted because the number looks striking.
• The ICER tells you the price, never whether to pay it. That decision needs the threshold and the budget, which is why the ICER is one input into an appraisal, not the appraisal itself.

When someone hands you a single number and calls it the answer, your first move is to ask what it forgot. With the ICER, it forgot the signs.

The ICER, in one breath.

• The ICER is a rate: ΔCost ÷ ΔEffect, measured in £ per QALY — the price of one extra unit of health versus the comparator.
• On the plane it's a slope: the line from the origin to the point. Cost-effective means your slope is shallower than the threshold's.
• Compute it in order: increments first (new minus comparator), then divide, then compare to the threshold.
• Its fatal flaw: collapsing two numbers into one loses the signs. A negative ICER is dominant or dominated; a positive one is north-east or south-west. The number alone is never a verdict.

The ICER prices the trade. It can't tell you which way the trade runs — for that, you always go back to the plane.

Next, we answer the question this lesson kept deferring: that £20,000 ruler — where does the threshold actually come from? It isn't what we're willing to pay. It's what we're forced to give up.