Measures of Adult Knee Function.

Abstract

Patient-reported outcome measures (PROMs) are important to fulfill both clinical and research purposes with regard to assessing knee function in patients with a variety of knee conditions associated with injury, osteoarthritis, or rheumatological disorders. For inclusion in this review, measures of knee function were required to be pertinent to rheumatology, orthopedics, and sports medicine specialties. We identified measures published with scientific analysis and included dimensions that were most important to patients, including pain, quality of life (QOL), and activity level. A 2011 review of nine tools was published and focused on many of these issues as they related to rheumatology and orthopedic surgery. Based on the aforementioned criterion and the goals of this review, we used the same nine measures developed specifically for patient-reported knee function and perceptions: the International Knee Documentation Committee (IKDC) Subjective Knee Evaluation Form, the Knee Injury and Osteoarthritis Outcome Score (KOOS), the KOOS Physical Function Short Form (KOOS-PS), the Knee Outcome Survey Activities of Daily Living Scale (KOS-ADLS), the Lysholm Knee Scoring Scale (LKS), the Oxford Knee Score (OKS), the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), the Activity Rating Scale (ARS), and the Tegner Activity Score (TAS). This updated review was conducted by doing a thorough search of new publications for each measure between January 1, 2010, and March 1, 2020. We also included the Patient-Reported Outcomes Measurement Information System Physical Function (PROMIS-PF) measure based on its rising popularity and the amount of research dedicated for its use in a variety of knee conditions. A basic summary of the properties of the different measures is displayed in Table 1. Psychometric data pertaining to the floor and ceiling effects, validity, reliability, responsiveness, and minimum clinically important difference (MCID) of each patient-reported outcome are displayed in Table 2. Floor and ceiling effects were considered to be absent if no participants scored the bottom or top score, respectively, and to be acceptable if less than 15% of the cohort scored the bottom or top score, respectively. Validity was measured by assessing content, face, and construct validity. Content validity was present if patients were involved in development. Face validity was present if expert reviewers made a similar assessment and considered the measured items adequate. Construct validity was considered adequate if expected correlations were found with existing measures that assess similar (convergent construct validity) and dissimilar (divergent construct validity) constructs. Internal consistency was considered adequate if Cronbach's α was at least 0.7 1, and test-retest (intrarater) reliability was adequate if the intraclass correlation coefficient was at least 0.8 for groups and 0.9 for individuals. Responsiveness was determined with a measure of ability to detect change over a period of time or intervention. MCID is the amount of change in a patient-reported outcome that represents a meaningful change to the patient. Since 2011, there have been numerous studies evaluating the psychometric properties of the above measures. There has also been ample research assessing the utility and psychometric properties of the PROMIS-PF function. Extensive work has been performed to add available translations and culturally adapted versions of the above measures. Our review summarizes the available information about how these measures perform for different patient populations in different settings. The IKDC detects improvement or deterioration in symptoms, function, and sports activities due to knee impairment caused by a variety of knee conditions, including ligament injuries, meniscal injuries, articular cartilage lesions, and patellofemoral pain 2. The IKDC was formed in 1987 to develop a standardized international documentation system for knee conditions. The IKDC Standard Knee Evaluation Form, which was designed for knee ligament injuries, was subsequently published in 1993 3 and revised in 1994 4. The IKDC Subjective Knee Evaluation Form was developed as a revision of the Standard Knee Evaluation Form in 1997. It has undergone subsequent minor revisions since its publication in 2001. The IKDC Subjective Knee Evaluation Form contains the following three domains: 1) symptoms, including pain, stiffness, swelling, locking/catching, and giving way; 2) sports and daily activities; and 3) current knee function and knee function prior to knee injury (not included in the total score) 2. The IKDC Subjective Knee Evaluation Form consists of 18 items (7 items for symptoms, 1 item for sport participation, 9 items for daily activities, and 1 item for current knee function). Response options vary for each item. Item 6 dichotomizes response into yes/no; items 1, 4, 5, 7, 8, and 9 use five-point Likert scales; and items 2, 3, and 10 use 11-point numerical rating scales. The recall period is not specified for items 1, 3, 5, 7, 8, and 9; it is 4 weeks for items 2, 4, and 6. Item 10a refers to function prior to knee injury, and item 10b refers to current function. Free to use. Cost of administration and information storage was not assessed and varies for each practice. See https://www.sport​smed.org/aossm​imis/Stagi​ng/Resea​rch/IKDC_Forms.aspx. The IKDC Subjective Knee Evaluation Form is a patient-completed questionnaire. The form has not been validated for administration by interview, either in person or via telephone. The response to each item is scored using an ordinal method (ie, 0 for responses that represent the highest level of symptoms or lowest level of function). The most recent version has assigned scores for each possible response printed on the questionnaire. Scores for each item, excluding item 10a, are summed to give a total score. The total score is calculated as the sum of items divided by the maximum possible score multiplied by 100, to give a total score out of 100. An online scoring sheet is available (www.sport​smed.org/tabs/research/ikdc.aspx) that provides a patient's raw score and percentile score (relative to age- and sex-based norms). The item regarding knee function prior to knee injury is not included in the total score. The revised scoring method states that in cases in which patients have up to two missing values (ie, responses have been provided for at least 16 items), the total score is calculated as the sum of completed items divided by the maximum possible sum of completed items multiplied by 100. Possible scores range from 0 to 100, in which 100 = no limitation with daily or sporting activities and the absence of symptoms. Normative data are available from the general US population stratified for age, sex, and current/prior knee problems 5. The IKDC Subjective Knee Evaluation Form takes 10 minutes to complete 6. It uses simple language that is suitable for patients. The IKDC Subjective Knee Evaluation Form takes approximately 5 minutes to score. Training is not necessary. Manual scoring can be performed easily using the scoring instructions supplied with the questionnaire. The form is available in English, Arabic, Brazilian Portuguese, traditional Chinese (Taiwan and Hong Kong), simplified Chinese (China and Singapore), Czech, Dutch, French, German, Greek, Italian, Japanese, Korean, Norwegian, Polish, Spanish, Swedish, Thai, and Turkish. Cross-cultural adaptations have been conducted for the Arabic 7, Brazilian 8, Chinese 9-11, Danish 12, Dutch 13, German 14, Greek 15, Italian 6, Korean 16, Romanian 17, Thai 18, and Turkish 19 translations. The initial set of items was developed by the IKDC by considering questions from the Standard Knee Evaluation Form, the Musculoskeletal Outcomes Data Evaluation and Management Systems Lower Limb Instrument, and the Activities of Daily Living (ADLs) and Sports Activity Scales of the Knee Outcome Survey. Pilot testing of the initial version (n = 144) resulted in revision or deletion of existing items and the addition of new items. Testing of the second version (n = 222) resulted in further revisions and deletions (based on missing data), producing a final version. Item-response theory was used to create the scoring system. Patients were not involved in development; rather, the items were selected by the IKDC, a committee of international orthopedic surgeons 2. Following development, validation, and implementation of the IKDC Subjective Knee Evaluation Form, a pediatric form was developed (the Pedi-IKDC), which has been tested for psychometric properties and normative data as well as electronic use 20-22. Studies consistently report no floor or ceiling effects (ie, no participants scored the lowest or highest score) 2, 6, 8, 13, 15, 23, 24. Internal consistency is adequate for patients with knee injuries and mixed knee pathologies (Table 1). Test-retest reliability is adequate for groups of patients with knee injuries and mixed pathologies and individuals with knee injuries. It has also been shown to be adequate in pediatric populations 20. The test-retest reliability is slightly below adequate for individuals who fall into a broader category of knee pathologies. However, studies have shown superior reliability over other measurement forms. The Chinese IKDC has better reliability than the Chinese KOOS 25. The Dutch IKDC had better reliability than the WOMAC and KOOS for meniscal injury 26. The domains covered by the IKDC Subjective Knee Evaluation Form appear to represent elements that are likely to be important to patients. However, the lack of patient contribution to the selection and revision of items in the IKDC Subjective Knee Evaluation Form means that content validity cannot necessarily be assumed. There are consistent reports of high convergent and divergent construct validity, with the IKDC Subjective Knee Evaluation Form more strongly correlated with the Short Form 36 (SF-36) physical subscales and component summary than with the mental subscales and component summary 2, 8, 11, 13, 23, 24, 27. Construct validity is acceptable in the pediatric form 20 and improved over the KOOS-Child form 28. Studies have shown the IKDC Subjective Knee Evaluation Form score to be highly correlated with the Cincinnati Knee Rating System, pain visual analog scale (VAS), Oxford 12 Questionnaire, WOMAC, Lysholm score, and SF-36 physical component, physical function, and bodily pain subscales 8, 13, 29. The IKDC Subjective Knee Evaluation Form has been shown to be adequately responsive 24. In a study comparing responsiveness of the IKDC Subjective Knee Evaluation Form with that of the KOOS for anterior cruciate ligament (ACL) injuries, the IKDC Subjective Knee Evaluation Form was found to be adequately responsive, but the KOOS was not 30. The same finding was found in a Chinese study comparing the IKDC Subjective Knee Evaluation Form with the KOOS 25. Further testing has specifically shown its adequate responsiveness for meniscal injury 31. The Pedi-IKDC has also been shown to have acceptable responsiveness 20. When directly compared with the KOOS-Child, it has superior responsiveness 28. The minimal detectable change has been reported to be between 8.8 and 15.6, and the SEM has been reported to be between 3.2 and 5.6. Few studies have shown the minimal important changes (MICs). One study shows the MIC to be 10.9 for meniscal injuries 31. Another study showed it to be 9.8 in the Chinese population 25, and another showed 12.0 for pediatric populations 12. The MCID has been reported to be 6.3 at 6 months and 16.7 at 12 months following cartilage repair 32 and 11.5 to 20.5 (range 6-28 months) in those who have undergone various surgical procedures for mixed (various) knee pathologies 33. The patient-acceptable symptom state (PASS) has not been determined. At face value, the domains covered by the IKDC Subjective Knee Evaluation Form appear to represent elements that are likely to be important to patients. It shows adequate internal consistency and has no floor or ceiling effects across mixed groups of patients with knee conditions. The IKDC Subjective Knee Evaluation Form has been shown to be responsive to change following surgical interventions, highlighting its usefulness in this patient population. It has particularly been shown to be a stronger measure for ACL injuries and meniscal injuries. It has also been shown to be a strong measure in the pediatric population. Despite demonstrating face validity, the lack of patient contribution to item selection indicates that content validity cannot necessarily be assumed and has not been thoroughly investigated. The relatively long recall period associated with three of the items may be a problem for some patients. The use of one aggregate score to represent symptoms, activities, and function may mask deficits in one domain. Psychometric testing is lacking for patients with knee osteoarthritis as an isolated group, as is responsiveness following nonsurgical management, highlighting areas for future studies. The IKDC Subjective Knee Evaluation Form involves minimal administrative and respondent burden and can be easily scored in the clinic using the online scoring sheet. However, clinicians using the online scoring system need to keep in mind that the normative data provided are from a particular population and may not be representative of their individual patient's population. Psychometric evaluation supports the use of the IKDC Subjective Knee Evaluation Form in research for a variety of knee conditions. Because some versions of the measure published online contain subtle differences in the wording of instructions and items, researchers should ensure that they utilize the version published as a component of the 2000 IKDC Knee Forms to ensure that findings of psychometric properties still apply and that comparisons can be made with previous studies. Administrative and respondent burden would not limit research use, although researchers should be diligent in checking for missing data. To measure the opinions of young, middle-aged, and elderly patients with posttraumatic osteoarthritis (OA) and other injuries leading to OA, regarding their knees and associated problems over short- and long-term follow-up 34. Examples of conditions include knee ligament injury (ACL, posterior cruciate ligament [PCL], or medial collateral ligament [MCL]), meniscal tears, knee cartilage lesions, knee OA, and osteochondritis dissecans. Interventions include ligament reconstruction (ACL, PCL, or MCL), meniscectomy, microfracture, osteochondral autografts, tibial osteotomy, total knee replacement (TKR), exercise (land-based or aquatic), intra-articular sodium hyaluronate injection, pharmacologic therapy, and glucosamine supplementation. The KOOS consists of the following five domains: 1) pain frequency and severity during functional activities; 2) symptoms such as the severity of knee stiffness and the presence of swelling, grinding or clicking, catching, and range of motion restriction; 3) difficulty experienced during ADLs; 4) difficulty experienced with sport and recreational activities; and 5) knee-related QOL 34. The original KOOS remains unchanged, although there have been other subscales developed, including the KOOS-12 Short Form, the KOOS-Joint Replacement Form, and the KOOS-Child Form The KOOS contains 42 items across five subscales. All items are rated on a five-point Likert scale (0-4) specific to each item. The recall period is the previous week for the pain, symptoms, ADL, and sport/recreation subscales. It is not defined for the QOL subscale. Free of charge. The costs of distribution, collection, and data storage are not assessed. The KOOS is available with associated documentation at www.koos.nu. The KOOS is a patient-completed, in-person questionnaire. It can be administered on paper, tablet, or computer. Scoring sheets (manual and computer spreadsheets) are provided on the website. Each item is scored from 0 to 4. The five dimensions are scored separately as the sum of all corresponding items. Scores are then transformed to a 0 to 100 scale (percentage of total possible score achieved) 34. If a mark is placed outside a box, the closest box is chosen. If two boxes are marked, the box that indicates more severe problems is chosen. One or two missing values within a subscale are substituted with the average value for that subscale. If two or more items are missing, the response is considered invalid, and a subscale score is not calculated. A score of 0 is equivalent to the most severe knee problems, and a score of 100 is representative of no knee problems. Population-based normative data stratified by age and sex are available 35. The KOOS takes 10 minutes to complete 34. It uses simple language and similar one-word responses for each item. The KOOS takes approximately 5 minutes to score if using the scoring spreadsheet. It can be automatically calculated with any type of data management software. The KOOS is available in Egyptian Arabic, Saudi Arabian Arabic, Austrian German, Bengali (India), Czech, Hong Kong Chinese, Singapore Chinese, Croatian, Danish, Dutch, Estonian, English, Finnish, Filipino (Philippines), French, German, Greek, Hindi (India), Icelandic, Italian, Japanese, Kannada (India), Korean, Latvian, Lithuanian, Malayalam (India), Malay, Marathi (India), Norwegian, Persian, Portuguese, Brazilian Portuguese, Polish, Romanian, Russian, Singapore, Slovakian, Slovenian, Spanish, US Spanish, Peruvian Spanish, Swedish, Tamil (India), Telugu (India), Thai, Turkish, Ukrainian, Urdu (India), Vietnamese, Welsh, and Zulu. Validation of the cross-cultural adaptations have been conducted and found adequate in the following languages 36, 37: mainland Chinese 38, 39, Singapore Chinese 40, Greek 41, Icelandic 42, Spanish 43, 44, Dutch 45, French 46, Saudi Arabian Arabic 47, Hong Kong Chinese 48, Japanese 49, Persian 50, Portuguese 51, Russian, Singapore English 40, Thai 52, and Turkish 53. Studies consistently report no floor or ceiling effects or acceptable floor or ceiling effects in knee injury cohorts 36, 50, 54 and in patients with mild or moderate knee OA 37, 40, 46, 51. In those with severe OA awaiting TKR 37, 40, 45, 46, 51, there are consistent reports of floor effects for the sport/recreation subscale (16%-73.3% scored the lowest score), and ceiling effects have been reported for the pain (15%-22%), sport/recreation (16%), and QOL (17%) subscales for up to 12 months following TKR 37. Comparatively, it has been shown to have lower ceiling effects in all categories except for pain against the Knee Society Function score 55. Studies have shown that the original KOOS was not well understood by children, and subsequently the KOOS-Child was formed 56. The KOOS-Child was found to have no floor or ceiling effects 57. For patients with knee injuries, the pain, ADL, and sport/recreation subscales have adequate internal consistency in all reports, whereas the symptom and QOL subscales have had reports of lower or adequate internal consistency (Table 1). In patients with knee OA, the ADL, sport/recreation, and QOL subscales have adequate internal consistency, whereas the pain and symptoms subscales have reports of lower or adequate internal consistency. Test-retest reliability is adequate for group evaluation in all reports on the pain, symptoms, and QOL subscales for patients with knee injuries, whereas there are reports of lower and adequate reliability, respectively, for the ADL and sport/recreation subscales. Recent meta-analysis has shown adequate test-retest reliability for age- and condition-relevant subscales 58. Across the five subscales, the minimal detectable change ranges from 6 to 12 for knee injuries and from 13.4 to 21.1 for knee OA. For the five KOOS subscales, the pooled smallest detectable change (SDC) for individuals ranged from 15.7 (ADL) to 25.1 (sport/recreation). The SDC was greater for older adults and those with knee OA than for younger and ACL cohorts 58. The SEM is reported to be lower for knee injuries than for OA. In addition to exhibiting face validity, the KOOS shows content validity facilitated by direct involvement of patients with knee conditions in the development of the KOOS 34, 37. Multiple studies report that the KOOS demonstrates convergent and divergent construct validity, with the KOOS more strongly correlated with subscales of the SF-36 that measure similar constructs (eg, ADL with physical function, sport/recreation with physical function, and pain with bodily pain) and less strongly with SF-36 subscales that measure mental health 34, 36, 37, 40, 45, 50, 51, 54, 58, 59. Rasch analysis conducted using patient data 20 weeks post–ACL reconstruction showed that only the sport/recreation and QOL subscales exhibited unidimensionality and that the three subscales that were based on the WOMAC did not 60. A more recent study reported that the KOOS subscales had acceptable dimensionality 59. Further meta-analysis more recently found the hypothesis of superior convergent and divergent construct validity were supported when all data were pooled and when data were split by age group and knee condition for the pain, symptoms, ADL, sports/recreation, and QOL subscales 58. They found that further testing was necessary for the short form as well as for structural validity in all categories 58. The KOOS appears to be responsive to change in patients with a variety of conditions that have been treated with nonsurgical and surgical interventions (Table 2). In patients who have undergone partial meniscectomy 3 months previously, large effect sizes are seen on all but the ADL subscale. Large effect sizes are seen in all subscales 6 months after ACL reconstruction. Three years following autologous chondrocyte implantation or microfracture, large effect sizes are seen for the pain, sport/recreation, and QOL subscales, and moderate effects are seen for the symptoms and ADL subscales. In those with knee OA who have undergone physical therapy treatment, large effect sizes are seen at 4 weeks on the pain, symptoms, and ADL subscales, whereas the sport/recreation and QOL subscales show moderate effects. Larger effect sizes are found following TKR than nonoperative treatment 58. Large effect sizes are consistently reported on all subscales 3 to 12 months after TKR, but the effect sizes do not increase over these periods 58. Large effect sizes have been shown to be a strength of the KOOS as opposed to other parameters 61. The MCID of the KOOS Short Form and KOOS QOL has been reported in one study 62. MCID and moderate improvement estimates for the KOOS QOL were 8.0 and 15.6, respectively 62. The KOOS has undergone a substantial amount of psychometric testing. Over the last decade, the creation of subscales paired with psychometric testing has expanded as has cultural adaptation testing. Establishment of the KOOS as a reliable and valid measure across multiple languages highlights its usefulness as a patient-reported measure of knee function for people with knee OA and various combinations of sports- and trauma-related injuries. This has been expanded to include a child form of the test. The use of individual scores for each subscale, rather than an aggregate score, enhances clinical interpretation and, in research, acknowledges the impact of different interventions on different dimensions (eg, exercise therapy is likely to have more impact on ADLs and sports/recreation, whereas pharmacology may have more impact on pain and symptoms) and ensures content validity in groups of different ages and functional activity levels (eg, the sport/recreation subscale is more important in patients with a high physical activity level, whereas the ADL subscale is more important in subjects with a lower physical activity level). The KOOS has not been validated for interview administration, meaning that it may not be appropriate for patients who are unable to read or write or in cases when telephone follow-up is necessary. When administering the KOOS in older or less physically active individuals, higher-level components of the ADL and sport/recreation subscales may not be applicable and could result in missing data. It may be appropriate to leave out the sport/recreation subscale in those with more advanced disease or disability; however, doing so omits the ability to measure improvements seen in these more demanding functions following treatment 37. The MCID has been minimally examined. The KOOS is freely available online. Administration and scoring burden are minimal when online score sheets are utilized. Clinicians should bear in mind that the sport/recreation subscale may not be applicable for less physically active patients and may not have adequate test-retest reliability in individuals with knee injuries. The KOOS fulfills desired criteria for research outcomes, demonstrating adequate reliability for use in groups and validity when used in those with knee injuries and knee OA. The inclusion of the three WOMAC subscales facilitates the comparison of findings with studies that have utilized the WOMAC as a primary measure. The minimal amount of MCID evidence continues to weaken research usability. The purpose of the KOOS-PS is to measure patients’ opinions about the difficulties they experience with physical activity because of their knee problems. The KOOS-PS is a measure of physical function derived from the ADL and sport/recreation subscales of the KOOS 63. Patients rate the degree of difficulty they have experienced over the previous week due to of their knee pain with respect to the following: 1) rising from bed, 2) putting on socks/stockings, 3) rising from sitting, 4) bending to the floor, 5) twisting/pivoting on injured knee, 6) kneeling, and 7) squatting. The KOOS-PS contains seven items. All items are scored on a five-point Likert scale (none, mild, moderate, severe, and extreme) scored from 0 to 4. The KOOS-PS refers to the previous week. Free to use. The cost of administration and data storage is unique to each practice. The KOOS-PS and associated documentation are freely available at www.koos.nu. The KOOS-PS is a patient-completed questionnaire that can be completed in paper form or electronic form. Each question is scored from 0 to 4. The raw score is the sum of the seven items. The interval score from 0 to 100 is obtained using a conversion chart 63. There are no instructions on how to handle missing values. The possible raw score range is 0 to 28. Scores are then transformed to a score from 0 to 100, in which 0 = no difficulty. Not available. Based on findings for the KOOS, the KOOS-PS takes no more than 2 minutes to complete. It uses simple language and the same one-word responses for each of the seven items. Because the items relate to everyday tasks, it is not considered likely that they would have an emotional impact on the individual. The KOOS-PS takes less than 5 minutes to score using the conversion table provided 63. Training is not necessary because the questionnaire and scoring instructions are self-explanatory. The KOOS-PS is available in Arabic, Chinese, Danish, Dutch, English, French, German, Hindi, Italian, Korean, Norwegian, Polish, Portuguese, Russian, Spanish, Swedish, and Turkish. It can easily be compiled by extracting the seven items needed from the full KOOS forms in all languages in which the KOOS is available. Cross-cultural adaptations have been conducted for the French 64, Portuguese 65, and Turkish 66 translations. Rasch analysis was conducted on KOOS and WOMAC data from individuals with knee OA from Sweden, Canada, France, Estonia, and the Netherlands. Patient data from 13 data sets were used (age 26-95 years; male:female ratio 1:1.4). This included community and clinical samples, such as those who had undergone previous meniscectomy, tibial osteotomy, or ACL repair, as well as those scheduled to undergo TKR 63. Rates of missing data have not been reported. Findings of one study indicate no floor or ceiling effects when used in patients with knee OA (ie, no patients had the lowest or highest score, respectively) 64. The KOOS-PS has adequate internal consistency and test-retest reliability for groups of patients with knee OA; however, its reliability is lower than adequate for use in individuals with knee OA (Table 1). The minimal detectable change and SEM have not been reported. Because items are taken directly from the KOOS, which has face and content validity, this can also be assumed for the KOOS-PS, although no studies have evaluated content validity solely for KOOS-PS 58. The KOOS-PS shows evidence of convergent and divergent construct validity. Higher correlations have been shown with the SF-36 physical function, role physical, and bodily pain subscales; WOMAC function subscale (excluding the KOOS-PS items); and the Osteoarthritis Knee and Hip QOL Questionnaire (OAKHQOL) physical activity domain 64, 65, 67. Conversely, lower correlations have been reported with the KOOS pain, symptoms, and QOL subscales; the SF-36 mental health subscales; mental health questionnaires (eg, the Profile of Mood States and the Hospital Anxiety and Depression Scale); and OAKHQOL social support 64, 65, 67. One study found that in patients with knee OA, the KOOS-PS had a unidimensional structure when evaluated using principal component analysis 68. In patients with knee OA, the KOOS-PS shows moderate to large effect sizes following 4 weeks of physical therapy and moderate effects 4 weeks after intra-articular hyaluronic acid injection (Table 2). The KOOS-PS is also able to discriminate between groups of patients based on use of walking aids 65. When compared directly with the WOMAC physical function subscale, the WOMAC physical function subscale was better able to detect changes over time in physical function categories 69. One study