The pain scale everyone shares: what the McGill chart actually measures

Short answer

The chart ranking the worst pains out of 50 is not the McGill Pain Questionnaire. The real instrument has a maximum of 78, and it was never built to rank conditions against each other. The number 50 is the Y-axis of one 1984 graph, misread as a ceiling.

If a doctor has ever met your pain with a number that felt far too small, you were not imagining the gap. Pain scores are anchored to your own experience, which is exactly why they do not transfer between people.

Note: this article explains what pain scales measure and is not medical advice. Severe, sudden, or changing pain needs assessment, not a score. Sources are cited so you can read them yourself. Retrieved 23 September 2026.

Do you need a device to measure pain?

No, and nothing measures pain directly. There is no sensor for it — not a wearable, not a blood test, not a scan. Every pain scale in clinical use, including the McGill, works by asking you.

What a device can do is measure the things around pain: sleep, resting heart rate, heart-rate variability, activity. Those are context, not pain. Used honestly they help you see what your days looked like before a bad one, which is a different and more useful question than "how bad is it right now".

What is the McGill pain scale?

Ronald Melzack published it in Pain in 1975. It is not a number you pick — it is a vocabulary.

The questionnaire contains 78 words describing pain, grouped into 20 subclasses across four dimensions: sensory (throbbing, shooting, burning), affective (exhausting, sickening, fearful), evaluative (annoying, unbearable) and miscellaneous. You choose the words that fit. Each word carries a rank value by its position in its group, and the sum is your Pain Rating Index.

The maximums: sensory 42, affective 14, evaluative 5, miscellaneous 17. Total 78.

The point was never a score. It was that pain has qualities, and that "burning and exhausting" is clinically different information from "throbbing and sickening" even when both are rated severe.

A short form followed in 1987 — 15 descriptors, maximum 45 — and a revised version in 2009 that rates 22 descriptors on a 0–10 scale and reports means rather than sums.

Where does the "out of 50" pain chart come from?

Here is the part that surprised us.

Melzack did publish one figure comparing conditions. It appears in his 1984 John J. Bonica Lecture, The myth of painless childbirth, and it plots mean Pain Rating Index values for groups of patients seen in hospital clinics and an emergency department.

The Y-axis of that figure runs to 50 — because no group mean went higher, not because 50 is the instrument's ceiling. That axis appears to be where the famous "out of 50" came from.

What is actually on Melzack's figure: causalgia, digit amputation, childbirth in first-time and experienced mothers, chronic back pain, non-terminal cancer pain, phantom limb pain, postherpetic neuralgia, toothache, arthritis, bruise, laceration, sprain, fracture, cut.

Cluster headache is not on it. Kidney stones are not on it. Both are signature entries on the chart that circulates.

CRPS is there, under its older name causalgia — as one plotted group mean, not a rank.

And the specific numbers everyone quotes — CRPS 42, cluster headache in the forties, childbirth 35, fracture 20 — appear in no publication we could find. Not in Melzack 1975, not in 1987, not in the 2009 revision, not in the IASP terminology, not in the NIH materials. Treat the chart as an internet reconstruction descended from one forty-year-old graph.

What do the real numbers look like?

They exist, and they make a different point than a ranking does.

A 1990 meta-analysis pooled 51 studies using the McGill questionnaire, covering 3,624 people.

Look at the right-hand column rather than the left. Cancer pain ranges from 5.4 to 44.4 across studies. The spread inside a single condition is far larger than the gap between conditions, and every band overlaps every other band.

That is the empirical answer to "which hurts most": the question does not survive contact with the data.

Why can't pain be ranked between people?

Because of what pain is. The International Association for the Study of Pain revised its definition in 2020: an unpleasant sensory and emotional experience associated with, or resembling that associated with, actual or potential tissue damage.

The notes that accompany it matter as much as the definition. Pain is always a personal experience. Pain and nociception are different things, and pain cannot be inferred from nerve activity alone. People learn the concept of pain through their lives. A person's report of pain should be respected.

So when you say seven, you are placing your experience on your own internal scale, built from everything you have felt. When someone else says seven, they are doing the same thing on a different scale. The numbers are valid within a person over time — your seven today against your four last week — and not between people.

This is why a ranking of conditions is not a slightly imperfect tool. It is a category error.

Which pain scale do doctors actually use?

Not the McGill, usually. The McGill takes time and is mostly a research and specialist-clinic instrument.

The 0–10 numeric rating scale is the practical standard. Zero is no pain, ten is the worst you can imagine. A 2011 review of 54 papers found better patient compliance with it than with the alternatives in most comparisons.

Two figures worth knowing. The conventional cut-points, derived from how much pain interferes with function, are mild 1–4, moderate 5–6, severe 7–10. And a change counts as clinically meaningful at roughly two points or about 30% — which is the number to have in mind when judging whether a treatment did anything.

You will also meet the visual analogue scale, a 100mm line you mark, used mainly in research; and the Wong-Baker FACES scale, six faces from 0 to 10. The faces measure pain, not mood — the point is which face matches your pain, not which matches your feelings about it.

Which conditions really are at the extreme?

Some conditions do have published severity data. It just does not come as a league table.

Cluster headache has the strongest numbers. A 2021 survey of 1,604 people with cluster headache found a mean attack intensity of 9.7 out of 10, with 72% rating it a flat 10. In the same survey, people who had experienced them rated childbirth at 7.2, pancreatitis at 7.0 and kidney stones at 6.9 by comparison — though those are recollections by people living with cluster headache, not head-to-head measurement.

Cluster headache also carries documented suicidality: in a 2019 study of 175 patients, during attacks 64% reported passive suicidal ideation and 36% active ideation, falling to near zero between bouts. It is nicknamed the suicide headache for that reason.

CRPS, formerly causalgia, is described by the NIH as intense and often long-lasting pain following an injury that may have seemed mild, with light touch becoming painful and most people still in some pain a year later. No authoritative body ranks it as the most painful condition — that framing is not from the clinical literature.

Kidney stones at emergency presentation averaged 8.57 on a 0–10 scale in one study of 183 patients.

How bad is migraine pain, really?

Migraine sits oddly in this conversation. It rarely appears on the shareable charts, and it is one of the most disabling conditions on earth.

In the Global Burden of Disease 2019 analysis, migraine ranked second among all causes of years lived with disability worldwide — and first among women aged 15 to 49. Headache disorders account for 5.4% of all years lived with disability globally, and migraine is 88% of that.

The diagnostic criteria are more modest than the burden suggests. Migraine requires two of four features — one-sided, pulsating, moderate or severe intensity, worsened by routine activity — so severity is one optional characteristic among four, not a requirement. An attack can meet the criteria without being the worst pain in the room.

The disability does not come from peak intensity. It comes from duration (4 to 72 hours untreated), frequency, and the fact that the attack does not begin with the pain.

What happens before a migraine attack?

This is the part worth knowing, and the part that changes what is worth tracking.

The prodrome is a symptomatic phase that begins up to 48 hours before the pain. The diagnostic classification lists fatigue, mood change in either direction, food cravings, repetitive yawning and neck stiffness among its features.

How common is it? Honestly, the estimates vary enormously. A 2016 study of 2,714 people found 77% of those with migraine reported premonitory symptoms, averaging three of them, with yawning the single most common at 34%. A 2022 meta-analysis pooled 29% in population studies and 66% in clinic studies, with the authors concluding the evidence is too heterogeneous for a reliable single figure.

Individual symptom frequencies from that meta-analysis: fatigue 49%, neck stiffness 46%, mood change 37%, difficulty concentrating 30%, nausea 29%, light sensitivity 29%, sound sensitivity 26%, yawning 22%.

This is not psychological. Imaging in the pain-free premonitory phase shows activation in the posterolateral hypothalamus, midbrain, periaqueductal grey and dorsal pons. A 2016 study scanning one patient daily for 30 days across three untreated attacks found altered hypothalamic responses in the 24 hours before pain onset. The attack has already started before it hurts.

And people can learn to read it. In a landmark electronic diary study, 97 people recorded symptoms for three months on devices that would not let them edit entries retrospectively. They correctly predicted a migraine headache from 72% of the diary entries that recorded premonitory symptoms. The honest caveat: the participants were selected because they already believed they could predict their attacks, so 72% is a ceiling rather than a population figure.

One reframe falls out of this and it is worth the whole section. Food craving is a prodromal symptom. So the chocolate eaten the evening before a migraine may not be the trigger at all — it may be the first sign that the attack had already begun. Experimental work supports the scepticism: when 27 people who self-reported bright light or exertion as triggers were deliberately exposed to them, only three developed a migraine with aura, and photic stimulation produced none at all.

How do you find your own pattern?

Since pain scores do not transfer between people, the only comparison that means anything is you against yourself. That makes tracking a genuinely different activity from scoring.

Log the day before, not just the bad day. The prodrome window is up to 48 hours. If you only record attacks, you record the ending and miss the beginning.

Write down the ordinary things. Yawning, a stiff neck, an odd mood, wanting a specific food, struggling to concentrate. These are the documented premonitory symptoms, and individually they look like nothing — which is exactly why memory loses them.

Give it weeks, not days. Patterns need repetitions. Two to three weeks of consistent noting is roughly where signals start separating from noise.

Record the average, not the peak. One frightening number tells a clinician less than "this happens about twice a month, usually starting on Thursday evenings".

Be sceptical about triggers. As above, provocation studies mostly fail to reproduce self-reported triggers, and some apparent triggers are early symptoms. A pattern you observe repeatedly in your own log is better evidence than a list from the internet.

A word on what the evidence does not support: no consumer device reliably forecasts a migraine. The published attempts are small — studies of 18, 23, a handful of participants — and rely on models trained on one person at a time. And the assumption that short sleep triggers migraine was tested prospectively in 98 people over 4,406 days and was not supported. Anyone promising prediction is ahead of the evidence.

How to bring this up with your doctor

Bring frequency and duration before intensity. "Three attacks a month, each lasting a day and a half, and I lose those days" is more actionable than a number out of ten. Disability is what treatment decisions turn on.

Use the 0–10 scale the way it works. Compare yourself to yourself: "usually a 6, but the last three were 8s." A two-point change is the threshold clinicians treat as meaningful.

Bring the day before. If you have noticed yawning, neck stiffness or mood change preceding attacks, say so — it is a recognised prodrome, not a coincidence, and it affects what can be tried.

If you are not taken seriously, the disability figures are legitimate to cite. Migraine is second globally for years lived with disability and first in women aged 15 to 49.

When not to wait. A sudden severe headache unlike any before, a headache with fever and a stiff neck, one following a head injury, or one with new weakness, numbness, confusion or vision loss needs urgent assessment rather than a diary entry.

How Welltory fits

Welltory does not measure pain, and no app does. What it does is hold the context around it.

Its My Patterns feature works on exactly the logic this article arrives at. When it detects a stress or recovery episode, it asks what was happening, and you answer in your own words — "bad night", "deadline", "neck stiff since morning". Those become tags. One tag is nothing. After a few weeks the tags start showing which things recur before your worst days, which reliably help, and what time of day and week they cluster. There is also a section for the rare tags that hit hard, so they do not vanish among the frequent ones.

It needs about 7 tagged occurrences in a month, with the previous month to compare against, before it says anything — roughly two to three weeks of noting.

If your attacks might be migraine, Welltory's 2-minute check-in shows the stress-and-recovery pattern in your nervous system behind them. It does not diagnose migraine and it does not predict attacks — the evidence does not support anyone claiming that. What it does is make the buildup visible, which is the thing memory is worst at holding.

Our limits, plainly: Welltory holds no regulatory clearance, it measures physiological signals rather than pain, and nothing here replaces a clinician.

How we made it

Every figure here comes from the original instrument, a named peer-reviewed study, or a professional body — the International Association for the Study of Pain, the International Headache Society, the NIH and the Global Burden of Disease analyses — retrieved on 23 September 2026. Where a widely repeated claim could not be traced to any publication, as with the ranked pain index chart, that is stated rather than repeated. Where estimates disagree wildly, as with how many people experience a migraine prodrome, the range is given instead of a single figure.

Made with AI tools, then edited and fact-checked by people.

Data analysis by Jane Smorodnikova — the founder and CEO of Welltory. A recognized tech leader with two Master's degrees and experience at MIT, she has scaled Welltory to over 17 million users.

Written by Tatsiana Yashyna — Deputy COO at Welltory. With a background in medicine and years of working with health data, she translates research and real physiological signals — sleep, stress, heart rate, and hormones — into clear, evidence-based explanations that help people understand what their bodies are telling them.

Discounts for blog readers: up to 36% off

See what affects your energy, stress, sleep, and daily state with Welltory

Explore your data