Beyond the pain score: how progress is really measured.

Why isn't a 0–10 pain score enough on its own?

Because it answers a narrow question, how much does it hurt right now, and chronic pain rehabilitation is trying to change something wider. The 0–10 numeric rating scale was built for acute pain, where intensity is the whole clinical picture. In pain that has lasted months or years, intensity is one piece of a picture that also includes sleep, mood, work, movement, and the beliefs that govern all of them.

Three problems show up when a program relies on the pain score alone.

• It lags. People usually regain activity, sleep, and confidence before their pain intensity drops. A program judged on pain alone in week six can look like it is doing nothing while the patient is quietly getting their life back.
• It drifts. After years of pain, a person's internal scale recalibrates. A reading that would once have been an 8 becomes 'a normal day'. Two readings months apart may not be measured against the same anchor.
• It hides the mechanism. If fear of movement is the thing limiting someone, measuring only pain will miss the thing that needs to change, and miss it when it changes.

None of this means the pain score should be dropped. It is quick, validated, and patients care about it. It means the score should sit in a row with other measures rather than standing alone. The wider family of pain scales is covered in chronic pain scales.

What outcomes should a chronic pain program measure?

The clearest answer comes from the IMMPACT consensus, a group of pain researchers, regulators, and patient representatives who agreed in 2005 on what every chronic pain trial should report. Their six core domains are pain, physical functioning, emotional functioning, the patient's global rating of improvement and satisfaction, symptoms and adverse events, and participant disposition, meaning who stayed in treatment and who dropped out. A clinical program is not a trial, but the same domains are a reasonable yardstick for what a good one should be tracking.

No program measures everything on this list, and a questionnaire burden that drives people away is its own harm. A sensible minimum is pain intensity, one function measure, one belief measure, and the patient's global rating, taken at baseline and again at fixed points. For how to ask a remote program what it tracks, see how to evaluate a remote pain rehabilitation program.

What counts as a meaningful change?

A change big enough for the patient to notice and value. Researchers call this the minimal clinically important difference (MCID), and it is a different thing from statistical significance. A large study can find a statistically significant change too small for any individual to feel. The MCID asks the opposite question: how much does a score need to move before the person behind it says 'yes, that helped'?

Worked example: the Oswestry Disability Index

The ODI scores back-related disability from 0% to 100%. Suppose someone starts a program at 46%, which sits in the 'severe' band, and finishes at 34%, which is 'moderate'. Is a 12-point drop real progress or noise? The ODI literature puts the smallest change patients consider worthwhile at around 10 points on the 0–100 scale, so a 12-point improvement clears that bar. A drop from 46% to 42% would not, and should be reported as 'no meaningful change yet' rather than dressed up as progress. The scoring bands and the MCID discussion are on the Oswestry Disability Index page.

Pain intensity

For the 0–10 scale, the best-known anchor comes from a study that compared changes in pain score against what patients said on a global impression question. A reduction of roughly 2 points, or about 30%, corresponded to patients describing themselves as 'much improved'. A 6 becoming a 4 is a meaningful result by that standard, even though it is not a zero.

Two cautions. Thresholds depend on the starting point: a 10-point ODI drop means something different at 70% than at 22%, which is why some researchers prefer a relative change, such as 30% from the person's own baseline. And thresholds describe what one person would notice in their own score over time. They do not license comparing one patient's 34% against another's.

Can function improve even if pain stays the same?

Yes, and in chronic pain it is the usual order of events rather than the exception. The sequence is predictable once you look at what rehabilitation actually changes first.

Consider a delivery driver two years after a back injury. She rates her pain a 6. She has stopped lifting her daughter, sits only on hard chairs, and sleeps about four hours. Over twelve weeks of graded exposure she starts lifting her again, sits through a film, and sleeps six hours. Her pain is still a 6. On the pain score nothing has happened. On the Oswestry she has moved two bands, her FABQ has fallen, and her PGIC says 'much improved'.

This happens because much of chronic pain disability is driven by avoidance rather than by the sensation itself. When fear drops, activity expands, and the activity expands before the nervous system has finished turning the alarm down. The fear-avoidance model describes this loop in detail, and the mechanism behind the slow pain change is covered in central sensitization.

There is a practical consequence. A patient who is told 'your pain score hasn't changed, so this isn't working' at week six may abandon a program that is working. The function and belief measures exist to make that early change visible, to the patient as much as to the clinician.

How does Karuna measure and report progress?

Karuna Virtual Embodiment Training (KVET)™ is a 12-week program done at home, with short daily VR sessions on a provided headset, a weekly one-on-one session with a dedicated pain coach, and a clinician who reviews the case before the start and oversees it throughout. Measures are taken before and after the program, and the coach reviews progress with the patient each week and adjusts the plan. The measures Karuna reports are chosen to match the domains above: pain, activity, fear-avoidance, confidence, and sleep, with the FABQ as the fear instrument. Karuna has also used the Oswestry and the Pain Catastrophizing Scale as outcome measures in an institutional review board–approved feasibility study of the technology.

The table below is Karuna's published outcome data, from a real-world evidence analysis of 50 patients. All of them were injured at work, covered under workers' compensation, and in pain for longer than six months. The caveats underneath are part of the result, not a footnote to it.

The pattern is the point of this page. Activity and confidence moved further than pain did: a median of 49% and 71% among those who improved, against 28% for pain. That is the shape you would expect if fear and avoidance, rather than pain intensity, were the main limit on what these patients could do, and it is the shape a pain-only dashboard would have missed entirely.

How should a clinician or case manager read a progress report?

Start with the baseline and the instrument, not with the headline number. A report that cannot answer 'compared with what, measured how' has not reported anything yet.

What to look for

1. A baseline taken before treatment began, on the same instrument used later. A score without a starting point is a description, not a change.
2. The same questionnaire at each time point. Switching from the Oswestry to a different function scale mid-course makes the two scores incomparable.
3. Named instruments with their scale. 'Function improved 40%' means little; 'ODI 46% to 34%' can be checked against a published threshold.
4. More than one domain. Pain, function, and at least one belief measure, plus the patient's global rating. Agreement across domains is far more persuasive than one moving alone.
5. Who completed it and when. Patient self-report at a scheduled interval is the norm. A coach's impression, or a score collected on an unusually good day, is a different kind of evidence.
6. Engagement and dropout. How many sessions were done, and whether the person is still in the program. Outcomes from completers only will always look better than the full picture.
7. The patient's own goals. A Patient-Specific Functional Scale or a plain list ('lift my daughter', 'drive for an hour') ties the numbers to a life.

What a misleading dashboard looks like

• A single large percentage with no n, no instrument, and no definition of 'improved'.
• Pain intensity reported alone, often as 'average pain reduction', with function and work status missing.
• Only responders described. If the report never mentions people who did not improve, assume there were some and ask how many.
• Causal language ('the program reduced pain by X') attached to before-and-after data with no comparison group.

What this means for you

If you are a patient

Ask any program you are considering what it measures, when, and whether you will see your own scores. Then ask for them. Writing down an Oswestry or FABQ score at the start and again at twelve weeks is often the clearest evidence a person has that something is changing, especially during the stretch when pain intensity has not moved yet. Track function weekly rather than pain hourly; close monitoring of pain tends to raise it.

Measurement is for pain that has already been medically evaluated. New weakness or numbness, loss of bowel or bladder control, unexplained weight loss, fever, pain after a major fall or accident, chest pain, or thoughts of harming yourself need prompt medical attention, not a questionnaire.

If you are a clinician

Pair the NRS with one function measure matched to the region (the ODI for low back, a PROMIS short form where no regional tool fits) and one belief measure, and repeat them on a schedule rather than when a patient seems better. Report change against published thresholds and against the patient's own baseline. When a remote program sends you outcomes, apply the checklist above before accepting the summary. A remote or at-home program that cannot show its instruments is asking you to take its word for it.

If you are a case manager or adjuster

Work status is the outcome you are accountable for, and the FABQ work subscale is one of the better predictors of it. Ask for it at intake and at discharge. Treat a report with baseline, named instruments, engagement data, and an honest non-responder count as the standard, and treat a single headline percentage as a request for more information. Karuna takes workers' compensation referrals, and the care team confirms coverage case by case.