KOOS FAQs

Content:

What is the KOOS?

The Knee injury and Osteoarthritis Outcome Score (KOOS) is a patient-reported outcome measurement instrument, developed to assess the patient’s opinion about their knee and associated problems. The KOOS evaluates both short-term and long-term consequences of knee injury and also consequences of primary osteoarthritis (OA). It holds 42 items in five separately scored subscales: KOOS Pain, KOOS Symptoms, Function in daily living (KOOS ADL), Function in Sport and Recreation (KOOS Sport/Rec), and knee-related Quality of Life (KOOS QOL) (Roos and Lohmander 2003).

Why assess health-related Quality of Life with the KOOS?

The main reason for developing a single instrument with the purpose of covering several types of knee injury and including osteoarthritis (OA) was that traumatic knee injuries often cause concomitant damage to multiple structures (ligaments, menisci, cartilage, etc.) and frequently lead to the subsequent development of OA. To be able to follow patients after a trauma and to gain insight into the change in symptoms, function, etc. over time, a questionnaire, which covers both the short-term and long-term consequences is needed. Prior instruments such as the Lysholm knee scoring scale have focused only on the short-term consequences, and instruments such as the WOMAC Osteoarthritis Index only on the long-term consequences. An instrument intended for use in follow-up evaluation of these patients needs to adequately monitor both the acute injury consequences in the physically active and younger patients, as well as the chronic outcomes in the elderly (Roos and Lohmander 2003).

What are the applications of the KOOS?

The KOOS is intended to be used over short and long time intervals; to assess changes from week to week induced by treatment (medication, surgery, physical therapy) or over the years due to a primary knee injury, posttraumatic OA or primary OA (Roos and Lohmander 2003).

In which populations can the KOOS be used?

KOOS has been used in patients 13-79 years of age.
KOOS is intended to be used for knee injury that can result in posttraumatic osteoarthritis (OA); i.e. knee ligament injury (ACL, posterior cruciate ligament [PCL], medial collateral ligament [MCL]), meniscal tears, knee cartilage lesions, knee OA, and osteochondritis dissecans, etc.
KOOS is meant to be used over short and long time intervals; to assess changes from week to week induced by treatment (examples include: ligament reconstruction (ACL, PCL, MCL), meniscectomy, microfracture, osteochondral autografts, tibial osteotomy, total knee replacement (TKR), exercise (land-based, aquatic-based), intra-articular sodium hyaluronate injection, pharmacologic therapy, and glucosamine supplementation) or over years due to the primary injury, post traumatic OA or primary OA (Roos and Lohmander 2003).

How was the KOOS developed?

KOOS was developed by Ewa Roos and co-authors in the 1990s as an instrument to assess the patient’s opinion about their knee and associated problems. Items were selected based on: (1) the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), version 3.0; (2) a literature review; (3) an expert panel (patients referred to physical therapy for knee injuries, orthopedic surgeons, and physical therapists from Sweden and the US); and, (4) a pilot study of two questionnaires (one for symptoms of ACL injury, one for symptoms of OA) in individuals with posttraumatic OA.

What are the KOOS subscales and what areas of health do they measure?

The KOOS collects data on five knee-specific patient-centered outcomes: (1) KOOS Pain; (2) KOOS Symptoms: Other symptoms such as swelling, restricted range of motion and mechanical symptoms; (3) KOOS ADL: Disability on the level of daily activities; (4) KOOS Sport/Rec: Disability on a level physically more demanding than activities of daily living; (5) KOOS QOL: Quality of life, mental and social aspects such as awareness and lifestyle changes.

KOOS translations

What translations are available?

The English and Swedish versions were developed concurrently. Translated versions are: Austria-German, Chinese, Croatian, Czech, Danish, Dutch, Estonian, French, German, Greek, Hindi (India), Italian, Japanese, Korean, Latvian, Lithuanian, Norwegian, Persian, Polish, Portuguese, Russian, Singapore English, Slovakian, Slovenian, Spanish (Peru), Spanish (US), Thai, Turkish, and Ukrainian.

Translation work is currently ongoing into: Arabic (Egypt), Arabic (Saudi Arabia), Bengali (India), Brazilian Portuguese, Gujarati (India), Icelandic, Kannada (India), Malayalam (India), Marathi (India), Romanian, Tamil (India), Telugu (India) and Vietnamese.

All the available language versions can be downloaded as pdf files here.

Do you want to translate the KOOS into another language?

If you want to translate the KOOS into a new language, please contact us to be sure no other translation work in your intended language is currently underway. We recommend that you follow the guidelines for cross-cultural validation proposed by Beaton et al. (Beaton, Bombardier et al. 2000).
As soon as the linguistic process (and if applicable, the clinical validation study) is ready, we offer to publish the new language version on our website, www.koos.nu. We will assist in the appropriate formatting of the questionnaire. An abstract describing the linguistic process (and if available, the results from a clinical validation study) should accompany the questionnaire. The reason we offer to publish the new language version after the linguistic process is completed, is to minimize the risk of multiple versions being created by different translators simultaneously. This policy also enables other authors to perform clinical validation studies of the already linguistically-validated version. The stage of validation (linguistic and/or psychometric) is always evident from the abstract accompanying all the different language versions available at www.koos.nu.

In what settings have the measurement properties of the KOOS been evaluated?

In what settings have test-retest reliability of the KOOS been evaluated?

Test-retest reliability is critical because the measurement precision of an attribute is inversely related to the size of the measurement error (Roos, Engelhart et al. 2011).
In studies of patients with knee injury, intraclass correlation coefficients (ICCs) for KOOS Pain range from 0.85-0.95, KOOS Symptoms from 0.75-0.91, KOOS ADL from 0.75-0.91, KOOS Sport/Rec from 0.61-0.89 and KOOS QOL from 0.83-0.95. The minimal detectable changes were for KOOS Pain 6-6.1, for KOOS Symptoms 5-8.5, for KOOS ADL 7-8, for KOOS Sport/Rec 5.8-12, and for KOOS QOL 7-7.2 (Collins, Misra et al. 2011).
In studies of patients with knee OA, ICCs for KOOS Pain range from 0.8-0.97, KOOS Symptoms from 0.74-0.94, KOOS ADL from 0.84-0.94, KOOS Sport/Rec from 0.65-0.92 and KOOS QOL from 0.6-0.91 (Collins, Misra et al. 2011).

In what settings have KOOS’ internal consistency (Cronbach’s α) been evaluated?

In studies of patients with knee injuries, the internal consistency range (Cronbach’s α) for KOOS Pain was 0.84–0.91, for KOOS Symptoms: 0.25–0.75, for KOOS ADL: 0.94–0.96, for KOOS Sport/Rec: 0.85–0.89, and for KOOS QOL: 0.64–0.9 (Collins, Misra et al. 2011).
In studies of patients with knee OA the internal consistency range for KOOS Pain was 0.65–0.94, for KOOS Symptoms: 0.56–0.83, for KOOS ADL: 0.78–0.97, for KOOS Sport/Rec: 0.84–0.98 and for KOOS QOL: 0.71–0.85 (Collins, Misra et al. 2011).

How was face and content validity of the KOOS assured?

Content validity of the KOOS has been confirmed in patients with a complete rupture of the ACL requiring surgical reconstruction, patients with an isolated meniscal tear of the knee requiring arthroscopy, and patients with mild to moderate OA (Kellgren and Lawrence grades I-III) of the knee who were treated both non-operatively and operatively by Tanner et al. (Tanner, Dainty et al. 2007). They conclude that: “Of general knee instruments studied, the International Knee Documentation Committee Standard Evaluation Form and the Knee injury and Osteoarthritis Outcome Score contain the most items important to patients.”

The KOOS was developed in 1994-95. At that time, it was not common to report the developmental stage in the same detail as it is today (2012). What has been reported about the content validity is the following (Roos, Roos et al. 1998): “To ensure content validity for subjects with ACL injury, meniscus injury, and early osteoarthritis, we reviewed the literature, consulted an expert panel and conducted a pilot study.”… “An expert panel comprised of patients referred to physical therapy because of knee injuries, orthopedic surgeons, and physical therapists from both Sweden and the United States was asked to identify short-term and long-term symptoms and functional disabilities resulting from a meniscus or ACL injury.”… “A pilot study was conducted to identify the subjectively most relevant factors among patients with posttraumatic osteoarthritis. 75 individuals who had meniscus surgery 20 years previously were asked to respond to two questionnaires, both self-administered. The participants ranged in age from 35 to 76 (mean 56) and showed radiological signs of knee osteoarthritis, defined as joint space narrowing and osteophytes. One of the questionnaires was constructed to assess symptoms of ACL injury (Flandry, Hunt et al. 1991) and the other for assessment of OA (Lequesne, Mery et al. 1987). Questions that most frequently received high responses and were thus considered to reflect the most predominant symptoms included those relating to pain, swelling, stiffness, and the ability to run, jump, kneel, and squat.”

Additional information (unpublished): Semi-structured interviews were conducted individually with 30+ patients with knee injury or OA during the developmental phase. The interview guide included questions such as: What are the most important symptoms you experience from your knee injury (OA)? What are the effects of your knee injury (OA) on your function and quality of life? Additionally, patients were shown already existing questionnaires (Flandry, Lequesne, WOMAC) and were asked to comment on the relevance and importance of the items. These patients were, as reported, treated with physical therapy or surgery in Sweden or the US, and had different types of knee injuries (including ACL, meniscus, and cartilage) or OA.
Additionally, two focus groups (4-5 patients in each group) were held with patients on a waiting list for TKR to confirm content validity for this group. When tested in a sample of 105 patients having TKR (Roos and Toksvig-Larsen 2003), “…over 90% reported that improvement in the four subscales of Pain, Symptoms, ADL, and QOL was extremely or very important when deciding to have their knee operated on. Also, 51% reported that improvement in functions included in the subscale Sport and Recreation function such as squatting, kneeling, jumping, turning/twisting and running was extremely or very important when deciding to have their knee operated on. The group who reported that items related to Sport and Recreation function were extremely or very important had a greater proportion of men but was not biased with regard to age and preoperative ADL function” (Roos and Toksvig-Larsen 2003).

In what settings have KOOS’ construct validity been evaluated?

Multiple studies in patients with knee injury and knee OA report that the KOOS demonstrates expected convergent and divergent construct validity, with the KOOS more strongly correlated with subscales of the ShortForm- 36 (SF-36) that measure similar constructs (e.g., KOOS ADL with physical function, KOOS Sport/Rec with physical function, KOOS Pain with bodily pain), and less strongly with SF-36 subscales that measure mental health (Collins, Misra et al. 2011). Rasch analysis conducted using patient data 20 weeks post–ACL reconstruction showed that only the KOOS Sport/Rec and KOOS QOL subscales exhibited unidimensionality, not the three subscales that were based on the WOMAC. A more recent study reported that the KOOS subscales had acceptable dimensionality in this patient group (Salavati, Akhbari et al. 2011).

In what settings is the KOOS responsive to change?

KOOS’ responsiveness has been reported following surgical procedures including ACL reconstruction, meniscectomy, cartilage repair procedures, tibial osteotomy, total knee replacement as well as physical therapy, nutritional and pharmaceutical interventions. Following orthopedic surgery, including total knee replacement, KOOS QOL is usually as responsive, or more responsive, than KOOS Pain and KOOS ADL (Collins and Roos 2012).

What is the Minimal Detectable Change of the KOOS in different settings?

The minimal detectable change (MDC) is dependent on the test-retest reliability of an instrument. The more reliable a measure is, the smaller the difference that can be detected longitudinally. The test-retest reliability is, in turn, dependent on factors such as patient characteristics and time between assessments.
In studies of patients with knee injury, the MDC ranges for KOOS Pain were 6-6.1, for KOOS Symptoms: 5-8.5, for KOOS ADL: 7-8, for KOOS Sport/Rec: 5.8-12, and for KOOS QOL: 7-7.2 (Collins, Misra et al. 2011).
In one study including patients with knee OA, the MDCs for KOOS Pain were 13.4, for KOOS Symptoms: 15.5, for KOOS ADL: 15.4, for KOOS Sport/Rec: 19.6, and for KOOS QOL: 21.1 (Collins, Misra et al. 2011).

What is the Minimal clinically Important Change in the KOOS in different settings

The minimal clinically important change (MIC) can be defined as the smallest change score needed for the effect to be considered clinically relevant (de Vet, Ostelo et al. 2007). MIC can be calculated in many different ways using anchor-based and distribution-based methods. Recent publications highlight that there is probably no such thing as one MIC for a specific patient-reported outcome measure, such as the KOOS. Instead MIC seems to be dependent on context factors including patient group, intervention, time to follow-up, etc. Therefore, it is more probable that in the future, a range of MICs will be established for an instrument. Research is ongoing to establish the MIC for the KOOS in many different contexts. Until then, a MIC of 8-10 is considered appropriate for the KOOS (Roos and Lohmander 2003). For performing a sample size calculation, usually an SD of 15 can be used. However, the variation within the group to be studied will depend on the characteristics of the group. Usually elderly or more impaired groups have greater within-group variation than younger or healthier groups.

What is the Patient Acceptable Symptom State of the KOOS in different settings?

The Patient Acceptable Symptom State (PASS) can be defined as the value beyond which the patients can consider them well. The PASS is an absolute value (not a change value) and can be used as a patient centered and clinically relevant treatment target (the proportion of patients reaching a satisfactory state) (Tubach, Ravaud et al. 2005). Similar to the MIC, research is ongoing within this area to establish PASS for KOOS when used in different contexts.

Where can I find normative data/reference values?

KOOS reference values are available from several groups including a population-based sample (Paradowski, Bergman et al. 2006), a group of active soccer players (Frobell, Svensson et al. 2008), a group of healthy female soccer players (Östenberg, Roos et al. 2000), male and female patients following ACL reconstruction (Ageberg, Forssblad et al. 2010), and from another population-based group of 50 participants (mean 53 years, 37-79) who had no previous and no current clinical signs of injury to the ACL or menisci and no radiographic signs of OA (Roos, Roos et al. 1999).

KOOS administration

Who answers the questionnaire?

The KOOS is self-administered and completed by the patient. No interview or phone formats are available.

How long does it take to complete the KOOS?

The KOOS questionnaire takes about 10 minutes to fill out.

How is the KOOS administered?

The KOOS is self-explanatory and can be administered in the waiting room prior to a clinical visit or used in a postal mail survey. KOOS has not been validated for interview administration (Collins, Misra et al. 2011).

Do we need permission to use it?

Permission is not required to use the KOOS.

Are there any costs associated with using the KOOS?

The KOOS is free of charge and all the available language versions can be downloaded as pdf files.

Is there an electronic format available?

KOOS is available in paper format only from koos.nu. Excel scoring files are also available from koos.nu. There are however a range of electronic (commercial and non-commercial) software applications where the KOOS is included. Since KOOS is in the public domain, you are free to include the KOOS in your own software application. You may however not change the format or the text compared to the paper-based version. A recent study found an electronic version of KOOS to be comparable to the paper version (Gudbergsen, Bartels et al. 2011). As a courtesy, you are asked to inform the developer if you include KOOS in your own software. You can find contact information here.

KOOS Scores

How is the KOOS scored?

The five patient-relevant subscales of KOOS are scored separately: KOOS Pain (9 items); KOOS Symptoms (7 items); KOOS ADL (17 items); KOOS Sport/Rec (5 items); KOOS QOL (4 items). A Likert scale is used and all items have five possible answer options scored from 0 (No Problems) to 4 (Extreme Problems) and each of the five scores is calculated as the sum of the items included. Scores are transformed to a 0–100 scale, with zero representing extreme knee problems and 100 representing no knee problems as is common in orthopaedic assessment scales and generic measures. Scores between 0 and 100 represent the percentage of total possible score achieved. An aggregate total score is not calculated since it is regarded desirable to analyze and interpret the five dimensions separately. Please find more extensive instructions on score calculations in the scoring section of the KOOS User’s Guide.
For statistical purposes, when used as the primary outcome in an RCT, a single score can be constructed. In this case, the five individual subscale scores should be secondary outcomes. Please see reference (Roos, Engelhart et al. 2011) for more information on this procedure.

How are the KOOS scores interpreted?

The score is a percentage score from 0 to 100, with 0 representing extreme problems and 100 representing no problems. This scoring direction, 100 indicating no problems, is common in orthopaedic instruments and generic measures like the SF-36. In measures developed by rheumatologists, like the WOMAC, 100 usually represents the worst possible result.

How should the patient be instructed on scoring activities he/she is not able to perform?

It is important to determine whether or not each subscale is relevant at the time point chosen, considering the specific study population. For example, difficulty with Sport/Rec function may not be relevant to assess 2 weeks post-operatively.
The following guideline for study staff is available:
KOOS Pain and KOOS ADL subscales: If a subject avoids an activity (e.g., twisting/pivoting or going up or down stairs) due to doctor's orders or because the subject has chosen to avoid the activity, the subject should be instructed to choose “(4) Extreme” for those items.
KOOS Sport/Rec subscale: The same as above. Also, if a subject does not normally engage in an activity (e.g., running or jumping), the subject should be instructed to leave the item blank.

How do I handle missing items?

Missing data: If a mark is placed outside a box, the closest box is chosen. If two boxes are marked, that which indicates the more severe problem is chosen.
As long as at least 50% of the subscale items are answered for each subscale, a mean score can be calculated. If more than 50% of the subscale items are omitted, the response is considered invalid and no subscale score should be calculated. For KOOS Pain, this means that 5 items must be answered; for KOOS Symptoms, 4 items; for KOOS ADL, 9 items; for KOOS Sport/Rec, 3 items; and for KOOS QOL, 2 items in order to calculate a subscale score. Subscale scores are independent and can be reported for any number of the individual subscales, i.e. if a particular subscale is not considered valid (for example, KOOS Sport/Rec 2 weeks after total knee replacement), the results from the other subscales can be reported at this time-point.

Can a total score be calculated?

An aggregate total score is not calculated since it is regarded desirable to analyze and interpret the five dimensions separately. For statistical purposes, when used as the primary outcome in an RCT, a single score can be constructed. In this case, the five individual subscale scores should be secondary outcomes. For more information on this procedure, see reference (Roos, Engelhart et al. 2011) or the FAQs section on statistics.

Is there automated administration or scoring software?

Scoring software is available in Microsoft Excel-format and can be downloaded here.
NB: The excel Scoring sheet available for download does not automatically take the number of missing items into account.

How do I present my data?

The results of the five subscales can be plotted as an outcome profile, preferably in a graph with scores from 0-100 on the y-axis and the five subscales on the x-axis (an example is presented in the KOOS User’s Guide, under the heading of KOOS Profile). The order of the subscales from left to right should be: KOOS Pain, KOOS Symptoms, KOOS ADL, KOOS Sport/Rec and KOOS QOL. This way profiles can be compared visually across studies. Changing this order makes visual comparison more difficult.

How do I use KOOS as the primary outcome in an RCT?

There are two ways to apply the KOOS as the primary outcome in an RCT. Either way, the primary endpoint should be specified a priori. You can either (1) choose the single subscale you consider is the most likely to measure the change from the intervention of interest, or you can (2) create a composite score, referred to as KOOS4 or KOOS5, depending on the number of subscales included in this composite score. Please note that this is very different from calculating a total score! Calculating an average score from subscale scores ensures similar weight from all subscales in the composite score. Calculating a total score is never recommended. In either case, all KOOS subscales should be included as secondary endpoints to enable clinical interpretation of the results.
As examples, since exercise is aimed at improving physical function, it may be appropriate to apply the KOOS ADL subscale as the primary endpoint in an older group of OA patients, and the KOOS Sport/Rec subscale as the primary endpoint in a younger group of knee-injured patients in a trial of an exercise intervention. However, in pharmacological trials, the subscale KOOS Pain may be more appropriate as the primary endpoint since pharmacological agents often are supposed to relieve pain. When two different interventions are included (for example a surgical and a non-surgical) or when the intervention is supposed to affect more than one subscale to a similar degree, a composite score may be most appropriate as the primary outcome. Below, an example is given where KOOS4 was applied. You can also read more about “Single Subscale versus Overall score” and “Multiplicity” in the review by Roos et al (Roos, Engelhart et al. 2011).

What is KOOS4?

KOOS4 is an example of how the KOOS was used as the primary endpoint in an RCT on treatment of acute ACL injury. KOOS4 was first used in the paper by Frobell et al, published in the New England Journal of Medicine in 2010 (Frobell, Roos et al. 2010). In this paper, studying the effects of surgical reconstruction and exercise as treatments of acute ACL injury, a single outcome was preferred for the RCT. After a discussion within the author group, based on knowledge about previously published KOOS results, it was decided to construct an average score for four out of the five KOOS subscale scores. The reason for not including all five subscale scores was that in this group of patients, difficulties with ADL function is at most mild if existing at all. Including the subscale ADL would then add unwanted ‘noise’ to the constructed outcome.
KOOS4 was calculated as the average score of the four subscale scores for KOOS Pain, KOOS Symptoms, KOOS Sport/Rec and KOOS QOL. Please note that this is very different from calculating a total score! Calculating an average score from subscale scores ensures similar weight from all subscales in the primary outcome (KOOS4).Calculating a total score is never recommended. Doing so would significantly increase the influence of ADL function on the total score since the subscale KOOS ADL holds 17 items to be compared with KOOS Pain (9 items), KOOS Symptoms (7 items), KOOS Sport/Rec (5 items) and KOOS QOL (4 items).

How do I estimate the sample size needed in my study?

When using the KOOS, a change of 8-10 is most commonly considered the minimally important change (MIC) and the standard variation is set to 15. These numbers are recommendations but may of course vary depending on your setting. You are always responsible for performing your own sample size estimation and make sure that the data you enter into the calculation is appropriate for your study design and patient group. Please refer to the MIC section of this FAQs for more complete information.

How is the KOOS used in individual patients?

KOOS is also widely used for clinical purposes. In the clinic, KOOS is used to monitor groups and individuals over time. When completed prior to a consultation, due to its comprehensiveness, KOOS can be used to guide the consultation as to the symptoms and difficulties experienced by the patient.

Contact information

Do you have any other questions regarding KOOS?

If you have any queries, first contact the KOOS web manager. The KOOS web manager also handles requests to translate the KOOS for use in other countries:
Email: webmanager@koos.nu
If necessary, the KOOS web manager will contact the developer:
Professor Ewa Roos, Department of Sports Science and Clinical Biomechanics, University of Southern Denmark, Odense, Denmark
Email: eroos@health.sdu.dk

References

Ageberg, E., M. Forssblad, et al. (2010). "Sex differences in patient-reported outcomes after anterior cruciate ligament reconstruction: data from the Swedish knee ligament register." Am J Sports Med 38(7): 1334-1342.

Beaton, D. E., C. Bombardier, et al. (2000). "Guidelines for the process of cross-cultural adaptation of self-report measures." Spine (Phila Pa 1976) 25(24): 3186-3191.

Collins, N. J., D. Misra, et al. (2011). "Measures of knee function: International Knee Documentation Committee (IKDC) Subjective Knee Evaluation Form, Knee Injury and Osteoarthritis Outcome Score (KOOS), Knee Injury and Osteoarthritis Outcome Score Physical Function Short Form (KOOS-PS), Knee Outcome Survey Activities of Daily Living Scale (KOS-ADL), Lysholm Knee Scoring Scale, Oxford Knee Score (OKS), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), Activity Rating Scale (ARS), and Tegner Activity Score (TAS)." Arthritis care & research 63 Suppl 11: S208-228.

Collins, N. J. and E. M. Roos (2012). "Patient-reported outcomes for total hip and knee arthroplasty: commonly used instruments and attributes of a "good" measure." Clinics in geriatric medicine 28(3): 367-394.

de Vet, H. C., R. W. Ostelo, et al. (2007). "Minimally important change determined by a visual method integrating an anchor-based and a distribution-based approach." Quality of life research : an international journal of quality of life aspects of treatment, care and rehabilitation 16(1): 131-142.

Flandry, F., J. P. Hunt, et al. (1991). "Analysis of subjective knee complaints using visual analog scales." Am J Sports Med 19(2): 112-118.

Frobell, R. B., E. M. Roos, et al. (2010). "A randomized trial of treatment for acute anterior cruciate ligament tears." The New England journal of medicine 363(4): 331-342.

Frobell, R. B., E. Svensson, et al. (2008). "Self-reported activity level and knee function in amateur football players: the influence of age, gender, history of knee injury and level of competition." Knee Surg Sports Traumatol Arthrosc 16(7): 713-719.

Gudbergsen, H., E. M. Bartels, et al. (2011). "Test-retest of computerized health status questionnaires frequently used in the monitoring of knee osteoarthritis: a randomized crossover trial." BMC musculoskeletal disorders 12: 190.

Lequesne, M. G., C. Mery, et al. (1987). "Indexes of severity for osteoarthritis of the hip and knee. Validation--value in comparison with other assessment tests." Scand J Rheumatol Suppl 65: 85-89.

Östenberg, A., E. M. Roos, et al. (2000). "Physical capacity in Female Soccer Players - Does Age Make a Difference?" Advances in Physiotherapy 2(1): 39-48.

Paradowski, P. T., S. Bergman, et al. (2006). "Knee complaints vary with age and gender in the adult population. Population-based reference data for the Knee injury and Osteoarthritis Outcome Score (KOOS)." BMC Musculoskelet Disord 7: 38.

Roos, E., L. Engelhart, et al. (2011). "ICRS Recommendation Document: Patient-reported outcome instruments for use in patients with articular cartilage defects." Cartilage 2(2): 122-136.

Roos, E. M., H. P. Roos, et al. (1999). "WOMAC Osteoarthritis Index--additional dimensions for use in subjects with post-traumatic osteoarthritis of the knee. Western Ontario and MacMaster Universities." Osteoarthritis and cartilage 7(2): 216-21.

Roos, E. M. and L. S. Lohmander (2003). "Knee injury and Osteoarthritis Outcome Score (KOOS): from joint injury to osteoarthritis." Health Qual Life Outcomes 1: 64.

Roos, E. M., H. P. Roos, et al. (1998). "Knee Injury and Osteoarthritis Outcome Score (KOOS)--development of a self-administered outcome measure." J Orthop Sports Phys Ther 28(2): 88-96.

Roos, E. M. and S. Toksvig-Larsen (2003). "Knee injury and Osteoarthritis Outcome Score (KOOS) - validation and comparison to the WOMAC in total knee replacement." Health Qual Life Outcomes 1(1): 17.

Salavati, M., B. Akhbari, et al. (2011). "Knee injury and Osteoarthritis Outcome Score (KOOS); reliability and validity in competitive athletes after anterior cruciate ligament reconstruction." Osteoarthritis and cartilage / OARS, Osteoarthritis Research Society 19(4): 406-410.

Tanner, S. M., K. N. Dainty, et al. (2007). "Knee-specific quality-of-life instruments: which ones measure symptoms and disabilities most important to patients?" Am J Sports Med 35(9): 1450-1458.

Tubach, F., P. Ravaud, et al. (2005). "Evaluation of clinically relevant states in patient reported outcomes in knee and hip osteoarthritis: the patient acceptable symptom state." Ann Rheum Dis 64(1): 34-37.