Clinical library · Outcome measures
KOOS
Five subscales spanning the whole range from acute knee injury to established osteoarthritis. Its strength is that you can choose the subscale that fits your patient. Its weakness is that the minimal detectable change is large — 14 to 20 points.
In one line. Forty-two items in five separately scored subscales — Pain, Symptoms, Activities of Daily Living, Sport and Recreation, and Knee-related Quality of Life — each transformed to 0 (worst) to 100 (best). The subscales are reported individually; there is no total score.
It was built as an extension of the WOMAC to cover younger, more active patients, and that is exactly where it earns its place. The design decision that matters clinically is which subscale you make primary.
Choose the subscale before you collect it. The systematic review is explicit: the ADL subscale has better content validity for older patients, Sport and Recreation for younger patients with knee injuries, and Pain for painful knee conditions. It recommends a suggested order of subscales for different knee conditions to be used in hierarchical testing of endpoints. [1] Reporting all five and highlighting whichever moved is not a neutral act.
The numbers you actually need
| Property | Value | Source and caveat |
|---|---|---|
| Minimal detectable change | 14.3 to 19.6 (younger adults); 20 or more (older adults) | The review notes measurement error requires further evaluation [1] |
| Reliability, meniscal injury | ICC 0.84 to 0.89; alpha 0.72 to 0.95 | Standard error of measurement 7.0 to 12.6, smallest detectable difference 19.4 to 35.0 across subscales [2] |
| MCID, tibial tubercle osteotomy for patellofemoral pain | Pain 10.2 · Symptoms 10.1 · ADL 9.9 · Sport 14.2 · QoL 14.2 | 77 patients, 82 knees, median follow-up 1.7 years [4] |
| MCID, same procedure for patellar instability | Pain 10.1 · Symptoms 10.6 · ADL 10.2 · Sport 16.0 · QoL 13.2 | Same study — the indication changes the threshold [4] |
| Responsiveness after ACL reconstruction | KOOS QoL effect size 2.07, the second highest of 34 measures | 108 articles, 9,034 patients. KOOS QoL had greater relative efficiency than IKDC, KOOS Pain, Symptoms, ADL and Sport [3] |
| Floor and ceiling effects, meniscal injury | Floor: Sport and QoL. Ceiling: ADL | Within the smallest detectable difference of the extreme score [2] |
| Healthy population values | Vary significantly by age, sex and BMI | 714 subjects. Women scored lower than men on all subscales except QoL [5] |
What it measures
Knee-related symptoms and their consequences, across a wider functional range than the WOMAC covers. The Sport and Recreation and Quality of Life subscales are the additions that make it usable in athletic and younger populations, where a WOMAC would show a ceiling.
The review found adequate internal consistency, test-retest reliability and construct validity in both young and old adults with knee injury or osteoarthritis, with responsiveness demonstrated by larger effect sizes after surgical than non-surgical interventions. [1] That last clause is worth noticing: the instrument is more responsive to surgery than to rehabilitation, which shapes what a null result in a physiotherapy trial means.
Where it misleads
1. The minimal detectable change is bigger than most reported treatment effects
At 14.3 to 19.6 points in younger individuals and 20 or above in older ones, [1] the measurement error for an individual patient exceeds the size of many published between-group differences. In the meniscal injury study the smallest detectable difference ran from 19.4 to 35.0 across subscales. [2] A patient improving by 10 points has, statistically, not been shown to have improved.
2. Its five-factor structure is not established
The review states plainly that the five-factor structure of the original KOOS is unclear, that there is only some evidence the subscales are sufficiently unidimensional, and that this requires confirmation. Structural validity, cross-cultural validity and measurement error all require further evaluation. [1] The subscales are used as though they were five clean constructs; that has not been demonstrated.
3. Subscale choice can determine the result
After ACL reconstruction, KOOS QoL had an effect size of 2.07 and greater relative efficiency than KOOS Pain, Symptoms, ADL and Sport — against KOOS ADL the relative efficiency was 3.03. [3] Two investigators studying the same intervention with different primary subscales will reach different conclusions about its size. Pre-specify.
4. MCIDs shift with the indication, not just the procedure
In one study of a single operation, the MCID for the Sport subscale was 14.2 when the indication was patellofemoral pain and 16.0 when it was patellar instability; QoL moved the other way, 14.2 against 13.2. [4] The surgery was the same. Population-specific means more specific than most people assume.
5. In meniscal tears specifically, it was outperformed
A head-to-head comparison found the IKDC performed best on all measurement properties and concluded it should be used in preference to KOOS or WOMAC for meniscal tears. Floor effects were found for KOOS Sport and QoL, and a ceiling effect for KOOS ADL. [2]
What the evidence supports — and what it does not
Supported
- Adequate reliability, content validity and construct validity for age- and condition-relevant subscales. [1]
- Choosing the subscale by population — ADL for older, Sport for younger injured, Pain for painful conditions. [1]
- KOOS QoL as a responsive endpoint after ACL reconstruction. [3]
- Procedure- and indication-specific MCIDs where derived. [4]
- Age-, sex- and BMI-adjusted normative comparison. [5]
Not supported
- Interpreting individual change below about 15 to 20 points. [1][2]
- Treating the five-factor structure as established. [1]
- Reporting all five subscales and emphasising the one that moved. Relative efficiency differs up to threefold. [3]
- Transferring an MCID between indications for the same operation. [4]
- Preferring it to the IKDC in meniscal injury. [2]
How certain is this?
Evidence grade: Moderate.
The core psychometric picture comes from a systematic review with meta-analysis that is careful about what it has and has not established — adequate on reliability and construct validity, unclear on factor structure, incomplete on measurement error. [1] That candour is itself a marker of quality.
The responsiveness evidence after ACL reconstruction is large: 108 articles and 9,034 patients. [3] It measures relative efficiency between instruments rather than absolute clinical meaning.
The MCID values are from a single centre with 77 patients across two indications, [4] and the meniscal comparison from one Dutch-language validation study. [2] Both are useful and neither is definitive. The normative data come from 714 people in a single study. [5]
What would change the grade: confirmation of subscale unidimensionality, and MCID derivation in non-surgical rehabilitation populations, which is where most physiotherapy use sits and where almost none of this evidence was generated.
Common questions
Which subscale should I make my primary outcome?
Decide before collecting, and match it to the patient. The review recommends ADL for older patients, Sport and Recreation for younger patients with knee injuries, and Pain for painful knee conditions, and proposes a subscale order for hierarchical endpoint testing. [1] After ACL reconstruction specifically, KOOS QoL was the most responsive KOOS subscale. [3]
How much change is real in one patient?
More than you would like. Minimal detectable change is 14.3 to 19.6 points in younger individuals and 20 or above in older ones, [1] and in meniscal injury the smallest detectable difference ran from 19.4 to 35.0 depending on subscale. [2] Below that, an individual's change cannot be distinguished from measurement error.
Is KOOS better than WOMAC?
For younger and more active patients, it has more headroom — that is what the Sport and QoL subscales are for. For meniscal tears specifically, a head-to-head study found the IKDC better than both. [2] See the WOMAC page for that instrument's own limitations, which include a pain subscale with no divergent validity.
What is a normal KOOS in someone without knee problems?
Not 100. Across 714 healthy subjects, ADL and Sport scores fell with age while Symptoms and QoL scores rose, and women scored lower than men on every subscale except QoL. [5] Read a patient's score against their own age and sex band rather than against the maximum.
Why do my rehabilitation patients show smaller changes than surgical ones?
Partly a real difference and partly the instrument. The review found that evidence of responsiveness comes from larger effect sizes following surgical — especially total knee replacement — than non-surgical interventions. [1] Almost all the MCID work has been done in surgical cohorts, which is a genuine gap for physiotherapy.
References
- Collins NJ, Prinsen CA, Christensen R, et al. Knee Injury and Osteoarthritis Outcome Score (KOOS): systematic review and meta-analysis of measurement properties. Osteoarthritis and Cartilage. 2016 Aug;24(8):1317–29. doi:10.1016/j.joca.2016.03.010 PMID 27012756 Systematic review and meta-analysis
- van de Graaf VA, Wolterbeek N, Scholtes VA, et al. Reliability and Validity of the IKDC, KOOS, and WOMAC for Patients With Meniscal Injuries. American Journal of Sports Medicine. 2014 Jun;42(6):1408–16. doi:10.1177/0363546514524698 PMID 24618098 Measurement properties study
- Abed V, Kapp S, Nichols M, et al. Lysholm and KOOS QoL Demonstrate High Responsiveness in Patients Undergoing Anterior Cruciate Ligament Reconstruction: A Systematic Review and Meta-analysis of Randomized Clinical Trials. American Journal of Sports Medicine. 2024 Oct;52(12):3161–3166. doi:10.1177/03635465231219966 PMID 38352999 Systematic review and meta-analysis
- Retzky JS, Shah AK, Neijna AG, et al. Defining the minimal clinically important difference for IKDC and KOOS scores for patients undergoing tibial tubercle osteotomy for patellofemoral pain or instability. Journal of Experimental Orthopaedics. 2024 Jul;11(3):e12115. doi:10.1002/jeo2.12115 PMID 39076849 Cohort study
- Marot V, Murgier J, Carrozzo A, et al. Determination of normal KOOS and WOMAC values in a healthy population. Knee Surgery, Sports Traumatology, Arthroscopy. 2019 Feb;27(2):541–548. doi:10.1007/s00167-018-5153-6 PMID 30251101 Cross-sectional normative study
About this resource
Using this in clinic
Every figure here is traceable to its source.
Every threshold on this page names its subscale, its population and its indication, because all three move the number and most sources quote only one of them. Where a value could not be verified against the paper it came from, it is not on this page, and the omission is stated rather than filled with a number from a secondary source.