INTRODUCTION

This chapter reviews commonly used scales and outcome instruments that rate athletic and daily functional activities for a variety of knee injuries and disorders. The subjects in the investigations typically represent young active patients, with the majority treated for anterior cruciate ligament (ACL) ruptures. The instruments are analyzed according to their strengths and potential biases as well as reliability, validity, and responsiveness test findings.

Outcome measures are available for older or sedentary populations with osteoarthritis such as the Western Ontario and MacMaster Universities (WOMAC),¹ the Knee Pain Scale,¹⁸ and the Knee Society Clinical Rating System.⁸ There are also several generic health status instruments, such as the Medical Outcomes Study Short Form-36 Health Survey (SF-36) and the Medical Outcomes Study Short Form-12 (SF-12). It is not the purpose of this chapter to review these instruments. Scales from the Cincinnati Knee Rating System (CKRS) and the International Documentations Committee Rating System are not included, because they are described in detail in Chapters 44 and 45, respectively.

It is important to stress that, in the authors’ opinion, knee rating systems must measure not only symptoms and athletic and daily functional activities but objective physical findings as well. The assessment of patient outcome using subjective questionnaire data only does not provide an accurate understanding of the ability of the treatment protocol to restore normal knee function. Factors such as range of knee motion, joint effusion, tibiofemoral and patellofemoral crepitus, knee ligament subluxations, radiographic findings, and limb symmetry on functional testing should all be included in clinical investigations.

SPORTS ACTIVITY SCALES

7 Competitive sports   Recreational sports 6 Recreational sports 5 Work Heavy labor (e.g., building, forestry)   Competitive sports Cycling     Cross-country skiing   Recreational sports Jogging on uneven ground at least twice/wk 4 Work Moderately heavy labor (e.g., truck driving, heavy domestic work)   Recreational sports 3 2 Light labor 1 Sedentary work 0 Sick leave or disability pension because of knee problems  

From Tegner, Y.; Lysholm, J.: Rating systems in the evaluation of knee ligament injuries. Clin Orthop Relat Res 198:43–49, 1985.

Several problems are incurred with this activity rating instrument. First, work activities are rated within the same scale as sports activities. Patients who work in heavy-labor occupations are awarded only a level 5 (out of a possible 10), but the analysis of the stress on the lower limb in some of these occupations would probably show that these knees are functioning at a level equivalent to that of competitive athletes. They should not be awarded a lower level simply because they are not athletes or did not return to highly competitive sports. Athletics and occupational activities should be measured on separate rating scales (see Chapter 44).

Second, this scale does not separate various sports levels according to frequency of participation or the intensity of the sport, which is determined by accounting for the forces placed on the lower extremity. For instance, only national and international elite soccer players are listed on level 10, whereas basketball is listed on a level 7. In the United States, it could be argued that competitive high school, collegiate, or professional basketball players are asked to place similar demands on the knee joint and lower extremity as elite soccer players. For patients who play or return to sports not listed on the scale, problems are incurred in trying to determine exactly which level accurately defines their sport. Third, this scale does not allow an assessment of a change in athletic participation between time periods owing to a change in lifestyle (e.g., graduated from school and no longer participate in a league). In addition, one cannot detect knee abusers in either the sport or the work levels.

Two independent investigations assessed the reliability, validity, and responsiveness of the Tegner scale. Briggs and coworkers³ determined its reliability, validity, and responsiveness in 122 patients with meniscus injuries. The scale was found to have adequate reliability (intraclass correlation coefficient [ICC], 0.817), content validity (no ceiling or floor effects), criterion validity with the generic SF-12 scale, and construct validity. However, only moderate effect sizes and standardized response mean values were reported. The investigators concluded that the scale measures only moderate changes in activity levels and noted that patients in the United States may have difficultly completing the scale because it was designed for sports commonly played in Europe. Paxton and associates¹⁷ assessed the reliability and validity of this scale in 153 patients followed 2 to 5 years after an acute patellar dislocation. The scale was found to have adequate reliability (ICC, 0.92) and content validity.

Modified Hospital for Special Surgery Sports Scale

The modified Hospital for Special Surgery (HSS) knee assessment instrument (see Table 46-2) was developed for an investigation of 48 patients with ACL-deficient knees who underwent arthroscopic partial meniscectomy.⁶ The instrument includes a scale composed of four levels of various sports ranked “very stressful,” “moderately stressful,” “mildly stressful,” and “no sports activities.” Examples of sports are provided in categories A, B, and C to assist with the ratings. The advantages of this scale include the assessment of sports activities only (separately from occupations) and the attempt to sort sports participation according to the level of intensity. The problems with this scale are that ambiguous terms are used to rate the sports categories (as “very stressful” may have different meanings to different patients), an assessment of reasons for change in athletics owing to non–knee-related reasons is not provided, and the detection of knee abusers is not possible. The authors did not conduct reliability, validity, or responsiveness testing of their outcome instrument.

Seto Sports Participation Survey

Seto and colleagues²⁴ developed a sports participation survey in order to determine activity levels in 19 patients who had undergone ACL reconstruction. In this survey, subjects rated their activity level for 13 sports, which are divided into cutting and noncutting categories. Each sports activity receives 0 to 4 points, based on the ability to participate with no symptoms (4 points), participate with “recurring mild episodes of pain, swelling, or instability” (3 points), limited participation owing to “moderate or severe episodes of pain, swelling, or instability” (2 points), or not able to participate owing to either knee-related (1 point) or non–knee-related (0 points) reasons.

Patients also rate their sports participation according to frequency and duration as either competitive (6–7 times/wk or regular participation in organized competition, 3 points), recreational (3–5 times/wk, or 2 times/wk and more than 2 hr/session, or seasonal sports 5 days or more/mo, 2 points), or weekend (1–2 times/wk for 1 hr or participation in seasonal sports ≤4 days/mo, 1 point).

This scale allows the analysis of athletic participation according to frequency of play and the detection of changes in activity levels between evaluations. Knee abusers may be identified from the categories of “mild” or “moderate” symptoms occurring with activity. Because the activity level is rated per sport, reporting of data can be difficult if a population is composed of individuals who play a wide variety of sports. Reliability, validity, and responsiveness testing were not conducted on this survey.

Straub and Hunter’s Sports Performance Index

Straub and Hunter²⁶ devised a sports performance index to determine the outcome of ACL reconstruction. The index is based on scores from two factors: a participation level rating and an intensity level rating. The participation level rating is accomplished on a six-level gradient that ranges from professional (5 points); intercollegiate (4 points); intramural (college), varsity (high school) or recreational (intense) (3 points); casual (2 points); junior high school (1 point); to no sports (0 points). The intensity rating scale sorts various sports into categories according to the authors’ perception of the intensity of those activities on the lower extremity. The most intense activities receive 5 points and the least intense receive 1 point. The two rating scores are summed to produce a sports performance index.

This scale does not allow for the rating of sports that are not included in the intensity rating scale, because the rationale for the classification of activities into the various categories is unknown. The frequency and duration of activities is not included in the sports level assessment. Knee abusers cannot be detected. The authors did not assess reliability, validity, or responsiveness of this instrument.

Daniel Sports Activity Rating

Daniel and coworkers⁴ described a sports rating scheme that was used to determine outcome in both ACL-deficient and ACL-reconstructed knees. The sports activity rating assessed frequency of participation (number of hours played per year) and intensity of the activities. Frequency of participation is determined in a retrospective manner; patients are asked to recall the two most frequent sports they played in before their ACL rupture and to provide the number of weeks per year and the number of hours per week that they played each sport. The number of hours per year of participation in the two sports is summed. A four-level intensity rating is also included in which level I comprises jumping, pivoting, and hard cutting activities; level II, lateral motion but less jumping or hard cutting than level I; level III, “other sports”; and level IV, no sports. A few examples of sports are provided for levels I, II, and III.

The retrospective analysis of hours of participation can be problematic, especially in long-term clinical studies in which patients may be asked to recall hours of participation from many years ago. The scale does not allow for the detection of knee abusers or patients who decreased their activity level or stopped athletics owing to reasons not related to their knee condition. Reliability, validity, and responsiveness testing were not conducted by the investigators.

Marx Sports Activity Scale

Marx and associates¹⁴ developed a sports activity rating scale to use as a general research tool to determine outcome for a variety of knee injuries and operations. This scale takes into account frequency of participation and intensity of the activity. Four separate activities are rated: running, cutting, decelerating, and pivoting. Frequency of participation for each activity is classified as either “none,” “1 time a month,” “1 time a week,” “2 or 3 times a week,” and “4 or more times a week.”

Reliability and construct validity were assessed in 40 volunteers with no history of knee problems. The scale showed acceptable reliability (ICC, 0.97) and correlated with the Tegner, Cincinnati, and Daniel sports activity scales. Responsiveness testing was not completed. Although this scale is useful in rating activities that involve the specific motions selected by the authors, it cannot rate low-impact activities such as swimming, bicycling, or low-impact aerobics. If a patient swam or bicycled several times a week, he or she would receive 0 out of a possible 16 points on this scale.

ACTIVITIES OF DAILY LIVING SCALES

Support Locking Instability Pain Swelling Stair-climbing Squatting

The symptoms of pain, swelling, and instability are graded according to the activity in which they occur. For example, for the factor of instability, patients may receive 15 of a possible 25 points if they experience instability frequently during athletics or heavy exertion (or are unable to do these activities). Independent studies have demonstrated that the Lysholm system “may inflate results, particularly if raw scores are combined to obtain an overall final rating”²⁵ and is “less sensitive for ACL patients … than three other diagnostic categories.”² Risberg and colleagues¹⁹ found poor sensitivity in detecting changes over time (responsiveness) and concluded, “The Lysholm score was not sensitive to detect changes over time, and therefore it is not recommended as an outcome measurement after ACL reconstruction.”¹⁹

Other authors have reported adequate reliability and responsiveness of this scale when it was tested in combined samples of patients with a variety of knee disorders including ligament and meniscus injuries, patellofemoral pain, and degenerative joint disease. Irrgang and coworkers⁹ assessed internal consistency, criterion validity, and responsiveness of the Lysholm scale in a group of 397 patients who had a wide variety of knee disorders. Data collected at baseline and at 1, 4, and 8 weeks later were compared with those obtained with the authors’ Activities of Daily Living (ADL) Scale. The Lysholm scale demonstrated adequate criterion validity with the ADL scale (Pearson product-moment correlation coefficients ranged from R = 0.78 at baseline to R = 0.86 at 8 wk postbaseline) and with the patients’ global rating of the overall knee condition. Large effect sizes (>0.80), indicating adequate responsiveness, were noted at 4 and 8 weeks postbaseline testing. Internal consistency of the Lysholm scale was lower than that of the ADL scale, resulting in higher standard error of measurement for all time periods of data collection.

Marx and associates¹³ determined reliability, validity, and responsiveness of the Lysholm scale along with the subjective components of the CKRS, the American Academy of Orthopaedic Surgeons (AAOS) sports knee-rating scale, and the ADL scale in a population of patients with a wide variety of knee disorders. All of the scales demonstrated high reliability (ICCs > 0.80), adequate face and content validity, and responsiveness (standardized response means > 0.80).

Paxton and associates¹⁷ measured the reliability and validity of the overall Lysholm score in 153 patients followed 2 to 5 years after an acute patellar dislocation. The overall score demonstrated adequate reliability (ICC, 0.88), content validity, and construct validity. Kocher and colleagues¹¹ assessed the Lysholm scale in a population of 1657 patients with traumatic and degenerative chondral lesions. Adequate reliability was found for the overall score and all individual components of the scale except for pain and stair-climbing. Adequate content validity was found for the overall score. However, unacceptably high ceiling effects were found for the domains of limp, instability, support, and locking and a high floor effect was found for the domain of squatting. Criterion validity was acceptable, with significant correlations between the overall score and the SF-12, Western Ontario, and Western Ontario and MacMaster Universities (WOMAC) scales. Construct validity was proved in the nine hypotheses generated by the investigators. Adequate responsiveness was found for the overall score and all of the individual components except for instability.

Briggs and coworkers³ determined reliability, validity, and responsiveness of the Lysholm scale in patients with meniscus injuries. Adequate reliability was reported for the overall score and all individual components except locking. Acceptable content validity was found for the overall score. However, high ceiling effects were reported for the domains of limp, instability, support, and locking and a high floor effect was found for the domain of squatting. Criterion validity was adequate, because the scale correlated with the physical component score of the SF-12. Acceptable construct validity was proved, with all eight hypotheses developed by the authors found to be significant. The overall score had adequate responsiveness, with large effect sizes and standardized response means. However, the individual domains of instability, locking, limping, support, and swelling had only small to moderate effect sizes.

ADL Scale of the Knee Outcome Survey

Irrgang and coworkers⁹ developed the ADL Scale (Table 46-4) to measure limitations in patients with a variety of knee disorders including ligament ruptures, patellofemoral pain, meniscus tears, and articular cartilage damage. The scale rates seven symptoms and 10 functions. The symptoms (pain, crepitus, stiffness, swelling, instability, and weakness) are rated on a six-level gradient as either (1) never occurs with ADL (5 points), (2) occurs but does not affect ADL (4 points), (3) occurs and slightly affects ADL (3 points), (4) occurs and moderately affects ADL (2 points), (5) occurs and severely affects ADL (1 point), or (6) prevents all ADL (0 points).

Functional limitations are determined for walking, walking with a supportive device, limping, ascending and descending stairs, standing, kneeling, squatting, sitting with the knee bent, and rising from a chair. The ratings range from no limitation to the inability to perform the function. The responses to all 17 factors in the scale are summed, divided by 80 (the total number of possible points), and multiplied by 100 to produce a final score that is expressed as a percentage.

The scale was administered to 397 patients at four treatment intervals (over an 8-wk period) to determine its internal consistency, validity, and responsiveness. A subset of 52 patients completed the scale before and after one treatment session for reliability testing. The authors reported acceptable results for all psychometric properties. Because the reliability testing was accomplished during one treatment setting, the authors concluded that further analysis of this factor is required. In addition, ceiling effects may occur in athletes or patients involved with strenuous work activities.

Knee Injury and Osteoarthritis Outcome Score

Roos and coworkers²² developed the Knee Injury and Osteoarthritis Outcome Score (KOOS) scale to assess symptoms and functional limitations with activity for patients with ACL or meniscus injuries or post-traumatic osteoarthritis. The scale is composed of five “dimensions” that are scored separately, including pain (9 questions), symptoms (7 questions), ADL (17 questions), sport and recreational function (5 items, includes squatting and kneeling), and knee-related quality of life (4 questions). The ADL questions are composed of the original 17 WOMAC Osteoarthritis Index Physical Function questions. Therefore, the score of the KOOS ADL dimension is equal to that of the WOMAC Osteoarthritis Index.

All questions are answered on a five-level frequency or severity gradient in which terms such as “none,” “mild,” “moderate,” “severe,” “extreme” or “always,” “often,” “sometimes,” “rarely,” “never” are provided. Two questions assess frequency of pain and are answered as “never,” “monthly,” “weekly,” “daily,” or “always.”

Construct validity was tested in 11 subjects with isolated ACL ruptures and 10 subjects with ACL ruptures and meniscus tears. Reliability was determined from 13 patients who completed the questionnaire twice, an average of 3.6 days apart. Responsiveness was determined 3 months postoperatively in 18 patients, 6 months postoperatively in 19 subjects, and 12 months postoperatively in 14 subjects. Adequate reliability was noted for all five dimensions (ICC ≥ 0.75). Construct validity tested against the SF-36 subscales showed varying correlation coefficients. Responsiveness of the five dimensions 6 months postoperatively was acceptable, because all effect sizes were 0.84 or greater.

No patient in the authors’ original series had osteoarthritis or an isolated meniscus tear. However, later studies by Roos and associates²³ provided evidence of satisfactory reliability, validity, and responsiveness of the KOOS for patients who had total knee replacement, meniscectomy, or post-traumatic osteoarthritis.^20,21 The questionnaire format does not provide an analysis of the activity level in which symptoms occur, other than daily activities. Although the scales assess certain sports functions, an activity rating for athletics is not provided, and therefore, it is unknown what activities patients are actually participating in or whether a change in activity level occurred between evaluations. Ambiguous terminology to define gradient levels can be problematic, because patients may interpret terms such as “moderate,” “severe,” and “extreme” differently. Some may assign a grade according to the frequency of the problem, others according to the intensity of the problem, and still others may use a combination of frequency and intensity to assign a rating.

AAOS Lower Limb Scale

The AAOS developed a Lower Limb Core Scale in order to determine pain, stiffness, swelling, and ADL function for a wide variety of knee disorders.¹⁰ The scale is composed of seven questions and designed to be treated as a summative scale to be used either alone for lower limb diagnoses or as a joint-specific scale for hip and knee or foot and ankle problems. The scale was administered to 70 patients with ligament or meniscus injury, patellofemoral pain, and degenerative joint disease. Johanson and associates¹⁰ reported good construct validity, with adequate correlation coefficients of this scale to other instruments such as the SF-36 and the WOMAC. However, there were limitations to this study. Reliability was not assessed with ICC, rather with correlation coefficients that the authors expressed was a potential limitation. Responsiveness was not determined in a large sample size, and floor and ceiling effects were not calculated.

GLOBAL ASSESSMENT SCHEMES

Mohtadi¹⁵ devised a quality of life outcome instrument for ACL-deficient knees. The questionnaire is composed of 31 questions in a visual analog scale (VAS) format and sorted into five domains. There are 5 items in the area of symptoms and physical complaints, 4 items in the area of work, 12 items in the area of sports, 6 items in the area of lifestyle, and 5 items in the area of social and emotion.

The instrument was tested for reliability and responsiveness in 25 patients with chronic ACL-deficient knees. Responsiveness was tested in 25 patients who completed the questionnaire 6 months apart and included patients with ACL-deficient or reconstructed knees. The author reported that the scale had adequate reliability using a paired T-test (instead of ICC) and adequate responsiveness, although effect size statistics were not provided. Face and construct validity was assessed in 50 patients with satisfactory results. Other validity measures such as content (floor and ceiling effects), criterion, and item-discriminant were not assessed.

A subjective knee score comprising 28 questions in a VAS format was originally described by Flandry and colleagues⁵ and later by Hoher and coworkers.⁷ The questionnaire combines symptoms and athletic and daily activity functions into one overall score. The instrument is weighted to an assessment of ADL, because 12 questions (43%) refer to ADL and only 5 (18%) assess sports activities. The questionnaire’s reliability was assessed in 25 volunteers with normal knees and in 21 patients with a variety of knee disorders. The correlation coefficients of the score were 0.86 in the volunteers and 0.92 in the patients. Whereas the score was able to detect a significant difference between normal and injured subjects, only a highly competitive athlete could score the maximum of 100 points. Construct validity, tested against the Lysholm score and an older version of the CKRS,¹⁶ showed adequate correlation coefficients. Other validity analyses were not conducted. Responsiveness effect size data were not provided.

There are many problems with these global VAS questionnaires. VAS formats do not allow for an analysis of symptoms or functional limitations according to activity level. Therefore, results of studies from different clinicians cannot be accurately compared with the knowledge of the percentage of patients who have symptoms with strenuous, moderate, light sports activities or with ADL. Because these instruments combine a variety of factors including symptoms, work, athletics, social, and emotional issues into one final score, there is no effective manner to address these items separately. Mohtadi’s final score¹⁶ is weighted for sports, which is problematic for patients who do not return to sports for reasons not related to the knee condition. Conversely, Hoher and coworkers’s final score⁷ is weighted for ADL and thereby loses sensitivity in defining functional limitations in athletic populations.

The VAS format does not allow for the detection of knee abusers, which can positively bias the final results and conclusions of the effectiveness of a specific treatment. These scales do not produce a measurable sports activity level according to frequency of participation and stress incurred to the lower extremity or allow the detection of a change in activity levels from one time period to another. In addition, overall patient satisfaction and perception of the knee condition are not addressed. These studies did not measure reliability, validity, or responsiveness of either the individual questions or the domains.