Knee Injury and Osteoarthritis Outcome Score (~00~)--Development of a Self-Administered Outcome Measure

Transcription

1 Knee Injury and Osteoarthritis Outcome Score (~00~)--Development of a Self-Administered Outcome Measure Ewa M. Roos, PT, MSc ' Harald P. Roos, MD, PhD * L. Stefan Lohmander, MD, PhD ' Charlotte Ekdahl, PhD, PT Bruce D. Beynnon, PhD 0 utcomes following injury of ligaments or menisci of the knee have usually been evaluated in terms of laxity, clinical examination, or radie logic findings (8,12,14,24). Patients, however, are usually more concerned with symptoms and function. Although symptoms and function have been assessed for these patients, the measures to do so have not always been proven reliable, valid for the subjects at interest, or responsive to clinical change. Furthermore, instruments used for assessing patient-relevant outcomes have often been administered by the surgeon who performed the operation, creating the opportunity for observer bias. In a critical analysis of knee ligament rating systems, Sgaglione et a1 (27) pointed out that three commonly used instruments-the Hospital for Special Surgery (36), the Lysholm (32), and the Cincinnati Knee Ligament (1) rating systems-yield different results and may reflect different underlying constructs. The Hospital for Special Surgery rating scale aggregates findings from a clinical examination and information regarding symptoms and function into one score, whereas the Lysholm scaring scale assesses symptoms and func- There is broad consensus that good outcome measures are needed to distinguish interventions that are effective from those that are not. This task requires standardized, patientcentered measures that can be administered at a low cost. We developed a questionnaire to assess short- and long-term patient-relevant outcomes following knee injury, based on the WOMAC Osteoarthritis Index, a literature review, an expert panel, and a pilot study. The Knee Injury and Osteoarthritis Outcome Score (KOOS) is self-administered and assesses five outcomes: pain, symptoms, activities of daily living, sport and recreation function, and knee-related quality of life. In this clinical study, the KOOS proved reliable, responsive to surgery and physical therapy, and valid for patients undergoing anterior cruciate ligament reconstruction. The KOOS meets basic criteria of outcome measures and can be used to evaluate the course of knee injury and treatment outcome. Key Words: outcome, knee injury, validity ' Graduate Student, Institute of Musculoskeletal Disease, Department of Physical Therapy, Box 5 134, University of lund, S lund Sweden. The KOOS questionnaire and scoring manual can be obtained from Ms. Roos at this address. Assistant Professor, lnstitute of Musculoskeletal Disease, Department of Orthopaedics, lund University Hospital, lund, Sweden ' Professor, lnstitute of Musculoskeletal Disease, Department of Orthopaedics, lund University Hospital, 1 und, Sweden Associate Professor, lnstitute of Musculoskeletal Disease, Department of Physical Therapy, University of Lund, lund, Sweden Director of Research, McClure Musculoskeletal Research Center, Department of Orthopaedics and Rehabilitation, University of Vermont, Burlington, VT Funding was received from the Zdga Foundation, The Swedish Sports Confederation (Sports Research Council), Wenner-Cren Foundation, The Swedish Foundation for Health Care Sciences and Allergy Research, Medical Research Council, lund University and Hospital, Astra lakemedel, Sweden. tions in one score. The Cincinnati Knee Ligament rating system includes physical examination, instrumented testing, and assessment of symptoms, function, and activity level in separate scores. None of these systems are meant to be self-administered. The International Knee Documentation Committee knee ligament rating form (ll), another instrument used to evaluate knee ligament injuries, categorizes subjective assessment, symptoms, range of motion, and ligament examination into four catege ries: normal, nearly normal, abnor- Volume 78 Number 2 Aupst 19% JOSPT

2 - 7 -." -. RESEARCH STUDY.. mal, and severely abnormal. Again, this system is meant to be used by an observer, and the aim is to aid the doctor in his/her assessment of the patient. To avoid observer bias, the patient should assess patient-relevant outcomes such as symptoms and function (13). Also, the questions and answer options should reflect the patient's opinion. A way of dealing with this problem, making questions and answer options patient-relevant, is to use visual analog scales or Likert boxes to record how much difficulty the patient experiences with each item (3). A self-administered questionnaire, comprising a set of visual analog scales to assess symptoms and function of anterior cruciate ligament (ACL) injury, was introduced by Flandry et al in 1991 (9) and compared with other common ACL rating systems. The authors concluded that the format was user-friendly, minimized observer bias, and allowed for standardized assessment of patient perceptions. The correlation to the other instruments was positive, indicating validity, but no test-retest data or responsiveness data were pre vided. No outcome measures specific to meniscus injury and meniscus symp toms have been developed, and rating systems specific to the ACL have often been used instead (lo,l4,24). The MOS 36item short-form health survey (SF-36) (35). a self-administered, widely used health status instrument that measures physical function among other constructs, has been used to assess outcomes after meniscectomy ( 15). Patients with ACL or meniscus injury have a higher than average risk of developing osteoarthritis, and about 50% have radiological signs of osteoarthritis years after injury (25). Ideally, any outcome measure to be used for long-term follow-up of patients with knee injury should address problems associated with osteoarthritis as well as those associated with the injury. For the elderly popu- lation with knee osteoarthritis, wellvalidated outcome measures exist (5, 18). However, no outcome measure has been validated for use in younger or physically more active subjects with knee osteoarthritis. It may not be appropriate to use outcome measures assessing function as activities of daily living alone in younger and more active subjects with posttraumatic osteoarthritis. The present paper reports on the development of a self-administered questionnaire, the Knee Injury and Osteoarthritis Outcome Score (KOOS) for assessing pain; symp toms, such as swelling and restricted range of motion; activities of daily living; sport and recreation function; and knee-related quality of life in young and middle-aged subjects with ACL injury, meniscus injury, or posttraumatic osteoarthritis. The instrument was evaluated for test-retest reliability, construct validity, and responsiveness to clinical change. MATERIALS AND METHODS Development of the Questionnaire The concepts of reliability, validity, and responsiveness follow those described by Bellamy (3) and Liang and Jette and Liang et al (19,20). Content Validity 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. The literature indicates three principal areas of patient-relevant outcomes: symptoms, functional status, and satisfaction (3,27,3O). An expert panel comprised of patients referred to physical therapy because of knee injuries, orthopaedic surgeons, and physical therapists from both Sweden and the United States was asked to identify short- and longterm symptoms and functional disabilities resulting from a meniscus or ACL injury. Seven factors were identified by the panel: pain, early diseasespecific symptoms, late diseasespecific symptoms (eg., symptoms of osteoarthritis), function, quality of life, activity level, and satisfaction. A pilot study was conducted to identify the subjectively most relevant factors among patients with posttraumatic osteoarthritis. Seventy-five 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 (X = 56) and showed radiological signs of knee osteoarthritis, defined as joint space narrowing and osteophytes (17). One of the questionnaires was constructed to assess symptoms of ACL injury (9) and the other for assessing symptoms of osteoarthritis (18). 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. KOOS (Knee Injury and Osteoarthritis Outcome Score) The Knee Injury and Osteoarthritis Outcome Score (KOOS) was constructed on the basis of the literature review, expert panel, and pilot study described above. The KOOS covers five dimensions that are reported sep arately: pain, symptoms, activities of daily living, sport and recreation function, and knee-related quality of life (Table 1). To ensure content validity for the older population with osteoarthritis, the questions from the Western Ontario and MacMaster Universities (WOMAC) Osteoarthritis Index (an outcome measure covering pain, stiffness, and function) (5) were included in their full and original form in the KOOS questionnaire [with permission (2)]. The WOMAC scores can thus be calculated from the KOOS questionnaire. The KOOS JOSPT Volume 78 Number 2 August 1998

3 RESEARCH STUDY PI. How often is your knee painful? What degree of pain have you experienced the last week when...? P2. twistinglpivoting on your knee P3. straightening knee fully P4. bending knee fully P5. walking on flat surface P6. going up or down stain P7. at night while in bed P8. sitting or lying P9. standing upright Symptoms Syl. How severe is your knee stiffness after tint wakening in the morning? Sy2. How severe is your knee stiffness after sitting, lying, or resting later in the day? Sy3. Do you have swelling in your knee? Sy4. Do you feel grinding, hear clicking, or any other type of noise when your knee moves? Sy5. Does your knee catch or hang up when moving? Sy6. Can you straighten your knee fully? Sy7. Can you bend your knee fully? Activities of daily living What difficulty have you experienced the last week...? At. Descending stairs A2. Ascending stairs A3. Rising from sitting A4. Standing A.5. Bending to floorlpi;k up an object A6. Walking on flat surface A7. Getting idout of car A8. Going shopping A9. Putting on sockdstockings A1 0. Rising from bed A1 1. Taking off sockdstockings A12. Lying in bed (turning over, maintaining knee position) A13. Getting idout of bath A14. Sitting A1 5. Getting odoff toilet Al6. Heavy domestic duties (shoveling snow, scrubbing floors, etc.) A1 7. Light domestic duties (cooking, dusting, etc.) Sport and remation function What difficulty have you experienced the last week...? Spl. Squatting Sp2. Running Sp3. Jumping Sp4. Turning/hvisting on your injured knee Sp5. Kneeling Knee-relatd quality of life Q1. How often are you aware of your knee problems? Never, monthly, weekly, daily, always None, mild, moderate, severe, extreme None, mild, moderate, severe, extreme Never, rarely, sometimes, often, always Always, often, sometimes, rarely, never None, mild, moderate, severe, extreme None, mild, moderate, severe, extreme Nwer, monthly, weekly, daily, always Q2. Have you modified your lifestyle to avoid potentially damaging activities to your knee? Not at all, mildly, moderately, severely, totally 43. How troubled are you with lack of confidence in your knee? Not at all, mildly, moderately, severely, extremely Q4. In general, how much difficulty do you have with your knee? None, mild, moderate, severe, extreme TABLE 1. The 42 Knee Injun~ and Osteoarthritis Outcome Score (KOOSI items arranged in the iive subscaks: pain, svmptoms, activities oidailv living, sport and recreation iunction, and knee-related qualitv of liie. Each item can be answered bv a 3-point likert scale. The five answer options are given aner each item. In case several following items have identical answer options, the answer options are only given after the first item. 90 Volume 78 Number 2 August 1998 JOSPT

4 , > ,* RESEARCH STUDY dimension of activities of daily living is equivalent to that of function in the WOMAC Osteoarthritis Index. Questions included in the subscales, sport and recreational function and knee-related quality of life, were adopted as they were originally written or with some modification from other outcome measures used to assess ACL injury (9,22). Satisfaction and activity level, two dimensions also considered relevant by the panel of experts, were not included in the KOOS, since it did not seem possible to agree on wording that would be applicable for all situations. An American and a Swedish version of the KOOS were developed simultaneously. Both feature the standardized, user-friendly format of the WOMAC Osteoarthritis Index and five-item Likert scales. The questionnaire is selfexplanatory and takes about 10 minutes to complete. KOOS Score Calculation The five dimensions of KOOS were scored separately: pain (nine items) ; symptoms (seven items) ; activities of daily life function (17 items); sport and recreation function (five items); and knee-related quality of life (four items). All items were scored from 0 to 4, and each of the five scores was calculated as the sum of the items included, in accordance with score calculations of the WOMAC Osteoarthritis Index (5). Scores were then transformed to a scale, with zero representing extreme knee problems and 100 rep resenting no knee problems, as common in orthopaedic scales (1,32,36). Scores between 0 and 100 represent the percentage of total possible score achieved. Table 2 provides a transformation formula and information necessary to apply the formula to each scale. The separate scores of the five dimensions can be visualized as a profile (Figure). An aggregate score was not calculated since it was re- ' - more than two items were omitted, the response was considered invalid. Clinical Study A clinical study was designed to assess reliability, construct validity, and responsiveness. The American version of the KOOS was used, and - the study was conducted at the De- Forn1ul,1 md cwn~plc for trmiormt~on of r w ~cde partment of Orthopaedics and Rehascores to a zero- 100 scale (0 = extreme knee problems, 100 = no knee problemsj. bilitation, University of Vermont, Bur-... lington, - VT. The study. protocol. was Transformed scale = reviewed and approved by the Institu- Actual raw score x 100 loo - Possible raw score range Example: A pain raw score of 16 would be transiormed as follows: TABLE 2. Formulas for scoring and transforming the Knee Injury and Osteoarthritis Outcome Score (KOOSj Scales. garded desirable to analyze and interpret the five dimensions separately. If a mark was placed outside a box, the closest box was used. If two boxes were marked, the box which indicated the more severe problems was chosen. Missing data were treated in agreement with the SF-36 (33), substituting missing values with the average value for the dimension. If KOOS Profile tional Review Board. Subjects Twenty-one subjects (nine men and 12 women) about to undergo ACL reconstruction were enrolled in the study. Ten of the subjects had a combined meniscus injury. Postoperatively, all subjects underwent controlled rehabilitation under supervision of a physical therapist. Participants ranged in age from 18 to 46 years, with a mean of 32 years. Thirteen of the 21 subjects had sustained their knee injury less than 3 months prior to surgery and 18 of the 21 had sustained it less than 6 months prior to surgery. Ten subjects were competing in sports, nine were recreational athletes, and two were not active in A 12 months post-op, n=14 A 6 months post-op, n=l9 0 3 months post-op. "=it3-1.i 10 0 Pain Syrnptans ADL Sport/Rec QOL FIGURE. The Knee lnjury and Osteoarthritis Score (KOOS) profile. Mean KOOS scores preoperatively, 3, 6, and 12 months postoperatively. ADL = Activities of dailv living, SporVRec = Sport and recreation function, QOL = Knee-related quality of life. JOSPT Volume 78 Number 2 August 19%

5 RESEARCH STUDY sports. All participants gave their informed consent to participate in the study. Test-Retest Reliability To assess reliability, the KOOS was administered twice preoperatively. Reliability coefficients between corresponding dimensions (pain, symptoms, activites of daily living, sport and recreation function, kneerelated quality of life) were calculated. Thirteen of the 21 participants completed the questionnaire twice within a 9day time interval. Elapsed time between assessment 1 and 2 averaged days (X 2 SD). Construct Validity Participants completed the SF-36 preoperatively at the same time that they completed the KOOS. For sub jects completing the KOOS twice preoperatively, the SF-36 was completed at the second administration of the KOOS. The SF-36 is a widely used measure of general health status which comprises eight subscales: physical functioning, role-physical, bodily pain, general health, vitality, social functioning, roleemotional, and mental health. The SF-36 physical functioning measures limitations in the ability to perform physical activities, a construct the activities of daily living and sport and recreation scales of the KOOS are intended to measure. The SF-36 bodily pain measures pain/ache and disturbances in normal activities, a construct similar to knee pain which the KOOS pain scale is designed to measure. We expected the highest correlations when comparing the scales that are intended to measure the same or similar constructs. Further, the eight subscales of the SF-36 have been shown to produce valid indices of physical health and mental health (34). Since the KOOS is designed to measure physical health rather than mental health, we expected to observe higher correlations between the KOOS subscales and the SF-36 sub scales of physical function, bodily pain, and role-physical (convergent construct validity) than between KOOS subscales and the SF-36 sub scales of mental health, vitality, roleemotional, social functioning, and general health (divergent construct validity). Responsiveness The KOOS and the SF-36 were administered at 3 and 6 months postoperatively. We expected that surgery and rehabilitation would induce a change in the patient's perceptions of symptoms and function that could be measured by the questionnaires. Full recovery after ACL surgery can be expected at the earliest of 6 months postoperatively. Three months postoperatively, most patients had not started agility training and were still undergoing physical therapy. It was desirable to follow this normal course of ACL rehabilitation with the KOOS scores. One subject was excluded from the study directly after surgery, leaving 20 subjects. Three subjects attended only the 3 or the &month follow-up, leaving 18 of the remaining 20 subjects for the Smonth follow-up and 19 of the remaining 20 subjects for the &month follow-up. Twelve-month follow-up data are available for 14 of the sub jects. Statistical Analyses The data obtained from questionnaires like the KOOS are ordinal rather than continous, which implies the use of nonparametric statistics. However, means and standard deviations are usually calculated instead of medians and interquartile range for questionnaire data (4,20,32,33,37). The test-retest reliability has been calculated using the random effects intraclass correlation coefficient (ICC) (29). An intraclass correlation coefficient of less than 0.4 is considered poor; between 0.4 and 0.75, fair; and greater than 0.75, excellent (26). The Spearman's correlation coefficient was used to assess construct validity. Postoperative change across all times was assessed by Friedman's test, and postoperative change at 3 and 6 months was assessed by Wilcoxon's signed rank test. Responsiveness was calculated by effect size. Effect size was defined as mean score change divided by the standard deviation of the preoperative score (16) In general, effect sizes of 0.2 are considered small, of 0.5 are considered moderate, and those of at least 0.8 are considered large (7). The 6-month follow-up data were used for calculations of responsiveness. Test-Retest Reliability The random effects intrac1a.s correlation coefficients could all be regarded a. high and were 0.85 for pain, 0.93 for symptoms, 0.75 for activities of daily living, 0.81 for sport and recreation function, and 0.86 for knee-related quality of life. Construct Validity As expected, the highest correlation occurred between the SF-36 scales and the KOOS scales that are intended to measure similar constructs (physical function vs. activities of daily living, r = 0.57, physical function vs. sport and recreation function, r = 0.47, bodily pain vs. pain, r = 0.46). Generally, higher correlations were seen when comparing KOOS scales to SF-36 scales with a high ability to measure physical health (convergent construct validity), and lower correlations were seen when comparing KOOS scales to SF-36 scales with a high ability to measure mental health (divergent construct validity), as shown in Table 3. Volume 78 Number 2 August 1998 JOSPT

6 Physical Mental KOOS KOOS KOOS KOOS KOOS SF-36 Subscale Health* Health* Pain Symptoms ADL SwrVRec QOL Physical function Strong Weak Role-physical Strong Weak Bodily pain Strong Weak Dimension E M Size KOOS pain WOMAC pain SF-36 pain KOOS symptoms WOMAC stiffness General health Moderate Moderate KOOS ADL Vitality Moderate Moderate Social Moderate Strong functioning - KOOS sportlrec WOMAC function SF-36 physical function KOOS QOL Role-emotional Weak Strong KOOS = Knrr Inpn, and Ottcoarthr~t~s Outcome Score; ADL = Actwites oi dailv living; SporURec = Mental health Weak Strong Sport and recreation function; WOMAC = Western Ontario and MacMaster Universities Osteoarthritis In- * SF-36 subscales abilitv to measure phvsical health vs. mental hcalth rn1. dex; QOL = Knee-related quality of life. KOOS = Knee Injury and Osreoarthritis Outcome Score; ADL = Activites oidailv living; SporURec = Sport and TABLE 5. sizes 6 months postoperatively for recreation function; QOL = Knee-related quality of life. KOOS five subscales; SF-36 subscales, bodily pain TABLE 3. Spearman's correlation coefficients (1,) determined when comparing KOOS' five dimensions to the and physical function; and WOMAC ~steod'rthritis SF-36 eight different subscales. Significant correlations (p < 0.05) in bold figures, N = 20. Index, N = 19. Responsiveness The scores of all subscales improved significantly from the preop erative to the postoperative visits (Friedman's test, p < 0.02). There were significant changes in the scores for pain, activites of daily living, and knee-related quality of life from the preoperative to the.?-month postop erative assessment. By the &month follow-up, all scores had improved significantly (Table 4). By the 12- month follow-up, further improve- ment was seen compared with the &month follow-up in the sport and recreation function and knee-related quality of life subscales (Figure). All mean subscores but knee-related quality of life were greater than 92 at 12 months postoperatively. The mean knee-related quality of life score was 75. Six months after surgery, the effect sizes could all be regarded as high (>0.8), ranging from 0.84 to 0.94 for pain, symptoms, and activi- Preoperative 3 Months Posttop KOOS Dimension (N = 21) 6 ~ O & patop value Pain Symptoms IN = 18) IN = 191 Activites of daily living ties of daily living and from 1.16 to 1.65 for sport and recreation function and knee-related quality of life. The sport and recreation function and knee-related quality of life subscales were the most responsive. Effect sizes for corresponding dimensions of the WOMAC and SF-36 were similar (Table 5). No comparisons could be made to the KOOS kneerelated quality of life, the most responsive subscale, because neither the WOMAC nor the SF36 measure this dimension. DISCUSSION Symptoms and functional disabilities are the principle reasons why patient5 seek treatment. These outcomes should, therefore, be considered paramount in assessing response to surgery, physical therapy, or other treatment. Thus, the KOOS was developed as a measure of patient-relevant out- Sport and recreation function comes to be used in studies of the treatment of ACL and meniscus in- Knee-related quality jury. Special emphasis was given to of life ascertaining validity for young and KOOS = Knee lnjurv and Osteoarthritis Outcomr Score. TABLE 4. Preoperative and postoperative mean scores and standard deviations of the five KOOS dimensions. p values received when assessing postoperative change (Wikcoxon's signed rank test). middle-aged patients with osteoarthritis so that long-term follow-up could be conducted. Studies of knee injury JOSPT Volume 78 Number 2 August

7 RESEARCH STUDY - treatment should incorporate a range of outcome measures, including o h jective assessment of impairment, complications, and patient-oriented measures, such as the KOOS scale. The KOOS is meant to supplement, not replace, the measures usually o b tained in clinical studies. The critical properties of outcome measures in general are patient-relevancy, user-friendliness, reliability, validity, and responsiveness to clinical change (19). Typically in o h The critical properties of outcome measures in general are pa tien t-relevancy, user- friendliness, reliability, validity, and responsiveness to clinical change. server-administered scores, the questions and the grading of the answer options are based on how knee function and reduction in knee function are defined by the constructors of the questionnaires. However, depending on actual and desired activity level of a specific subject, the same increased knee instability or decreased knee flexion might be regarded as more or less relevant by the patient. By asking the patient how much difficulty they experience with each task, instead of using answer options like ">90 of knee flexion" or "6-10 squats," the answers become patient-relevant. A daycare teacher or a weight lifter might experience extreme problems with only a minor limitation in knee flexion because their work/sport requires a lot of squatting. A salesperson or longdistance runner might only experience minor problems although the actual limitation in ability to squat is the same and would give the same score reduction in most observeradministered instruments. The KOOS is self-administered, requires approximately 10 minutes to complete, and can be administered in the waiting room or by mail; thus, the KOOS imposes only minimal bias and a negligible burden on patients and investigators. The KOOS showed high test-retest reliability, indicating adequate standardization and adaptation to patients. By way of comparison, intrarater agreement of experienced radiologists on the presence of osteoarthritis has been reported to range between (weighted kappa statistics) (31 ). Thus, while occasionally designated as so-called soft outcomes, questionnaires such as the KOOS provide highly reproducible data in comparison with other measures considered objective, even when these are applied by trained and experienced professionals. Validity is difficult to assess since there is no universally accepted gold standard for measuring patient-relevant outcomes. Therefore, there is no standard against which the KOOS can be compared. If there were, the KOOS might well be unnecessary. It is generally accepted that convergent construct validity is demonstrated if the correlation between scores on the same health components, a. measured by two different instruments, is positive and appreciably above zero (3). McDowell and Newell have noted, in a review of rating scales and questionnaires, that correlation coefficients for convergent construct validity often fall between 0.20 and 0.60 and rarely are above 0.70 (21). It is also necessary that the level of correlation is predicted n pion' and not determined post hoc. For these reasons, it is obvious that the instrument used for validation comparisons must be well-defined regarding construct validity. The SF-36 is such a well-validated instrument and has been used to assess outcomes after knee injury (15,28). The Lysholm scoring scale is frequently used to assess functional outcomes after knee injury, but its validity when used for patients with ACL injury has been questioned. Risberg and Ekeland found the Lysholm score was not sensitive to problems with sports function (23). Similarly, Bengtsson et a1 found the sensitivity of the Lysholm score being better for subjects with meniscus injury, patellofemoral pain syndrome, and lateral ankle sprain than for subjects with injury to the ACL (6). These findings indicate the Lysholm score not being sensitive to symptoms of ACL injury. For comparative reasons, the patients in the current study were also assessed with the Lysholm knee scoring scale. No single score was below 75 postoperatively, which leaves little room for improvement. Additionally, scores of 100 were found at all postoperative follow-ups, indicating the Lysholm score was not sensitive enough to symptoms and functional disabilities following ACL reconstruction. Responsiveness to clinical change is an important characteristic of outcome measures. High responsiveness indicates that fewer subjects are needed to demonstrate a statistically significant difference, making possible more managable studies. Cohen has suggested that, in comparative studies, examples of small, medium, and large effect sizes might have values of 0.2, 0.5, and 0.8, respectively (7). For the KOOS instrument, the effect sizes could be regarded as large 6 months after surgery, being largest for sport (effect size = 1.1 6) and knee-related quality of life (effect size = 1.65). The participants had only minor problems with pain and activities of daily living preoperatively; thus, there was little room for improvement. The pain score and the activites of daily living score, equivalent to the WOMAC function score, are believed to be important to ensure content validity for subjects with meniscus injury and osteoarthritis. A concomitant meniscus tear was found in 10 of the 2 1 subjects with ACL in- 94 Volume 78 Number 2 August 1998 JOSPT

8 RESEARCH STUDY jury in this study, and it was associated with more pain and more symptoms preoperatively. Different diagnoses and different combinations of diagnoses are currently being studied in the continued validation work of the KOOS. When comparing the five different KOOS scales, the subjects experienced more problems with sport and recreational activities and kneerelated quality of life than with pain, symptoms, and activities of daily living. Sport and recreation function and knee-related quality of life are also the dimensions that change the most after surgery. Presenting outcome as five separate scores, assessing five separate dimensions, makes interpretation easier than if symptoms and function are weighted and combined into one score. If a total KOOS score was calculated, an improvement could be observed, but it would not be possible to follow the phases of rehabilitation as closely as with separate scores. In a clinical setting, the KOOS profile of the five separate scores will help clarify the major concerns of the patient and therefore may be a tool for optimizing rehabilitation. In addition, the statistical effectiveness would be lower for a total score than for the most responsive subscales. This supports the perspective of Sgaglione et al that aggregated scores should neither be calculated, nor converted into the categories of excellent, good, fair, or poor (27). It is important that the changes of an outcome score are not only statistically significant but also reflect clinically relevant change. The mean KOOS scores at 3, 6, and 12 months postreconstruction reflected clinical experience of ACL rehabilitation. Three months postoperatively, the subjects experienced some pain, swelling, and restriction of range of motion and had not pushed their knee during sporting activities. This is reflected by nonsignificant changes in pain, symptoms, and sports over this time interval. Six months postop eratively, subjects were back at sports activities and had few symptoms, reflected by significant changes in all subscores. However, 6 months postsurgery, the majority of subjects were still aware of their knee problems daily and had, to some extent, modified their lifestyle to avoid potentially damaging activities to their knee, reflected by a mean knee-related quality of life score of 61. Twelve months postoperatively, additional improvement can be seen in the sport and knee-related quality of life scores. Pain and activities of daily living scores are close to normal. Limited knee flexion is the most common reason for still having residual symp toms (mean score go), and problems with kneeling is the most common reason for a mean sport score of 90. A mean knee-related quality of life score of 75 reflects that all subjects are still aware of their knee, at least monthly and at most daily. Five have made no lifestyle changes while nine have made minor changes in lifestyle to protect their knee. Similarly, five have full confidence in their knee, while nine experienced mild lack of confidence in their knee. There is broad consensus that good outcome measures are needed to distinguish interventions that are effective from those that are not. This task requires standardized, patient-centered measures that can be administered at low cost. These measures should be reliable, valid, and responsive and amenable to cross-study comparisons. The Knee Injury and Osteoarthritis Outcome Score (KOOS) meets these criteria and can be used to evaluate the course of knee injury and the effects of treatment. However, further methodological studies of the KOOS, engaging larger numbers of patients with knee injuries and posttraumatic osteoarthritis, are ongoing and needed to confirm the usefulness of the KOOS. SUMMARY The Knee Injury and Osteoarthritis Outcome Score proved to be a measure of sufficient reliability, validity, and responsiveness for surgery and physical therapy after reconstruction of the ACL. Five outcomes (pain, other disease-specific symp toms, activities of daily living, sport and recreation function, and kneerelated quality of life) are assessed in separate scores. The result is presented as an outcome profile. The most sensitive and responsive subscales were sport and recreation function and knee-related quality of life. JOSPT ACKNOWLEDGMENT The authors would like to acknowledge Dr. Robert Johnson, Department of Orthopaedics and Rehabilitation, University of Vermont, Burlington, VT, for sharing his knowledge and encouraging us in our work on the KOOS. REFERENCES 1. Barber S, Noyes F, Mangine R, McCloskey J, Hartman W: Quantitative assessment of functional limitations in normal and anterior cruciate deficient knees. Clin Orthop 255: , Bellamy N: London Health Sciences Centre, Suite 303, 375 South Street, London, Ontario N6A 4G5, Canada (personal communication), Bellamy N: Musculoskeletal Clinical Metrology, Dordrecht: Kluwer Academic Publishers, Bellamy N: WOMAC Osteoarthritis User's Guide, London, Ontario, Canada: Victoria Hospital, Bellamv N, Watson Buchanan W, ~oldshith CH, Campbell J, Stitt LW: Validation study of WOMAC: A health status instrument for measuring clinically important patient-relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or the knee. J Rheumatol 15: , 1988 Bengtsson J, Mollborg J, Werner S: A study for testing the sensitivity and reliability of the Lysholm knee scoring scale. Knee Surg Sports Traumatol Arthroscopy 4:27-3 7, Cohen J: Statistical Power Analysis for the Behavioural Sciences, New York: Academic Press, JOSPT Volume 78 Number 2 August 1998

9 RESEARCH STUDY Daniel D, Stone M, Dobson B, Fithian D, Rossman D, Kaufman K: Fate of the ACL-injured patient. A prospective outcome study. Am J Sports Med 226): , Flandry F, Hunt JP, Terry GC, Hughston JC: Analysis of subjective knee complaints using visual analog scales. Am J Sports Med 19(2): , Hede A, Larsen E, Sandberg H: The long term outcome of open total and partial meniscectomy related to quantity and site of the meniscus removed. Int Orthop 16: , Hefti F, Muller W, Jakob RP, Stubli HU: Evaluation of knee ligament injuries with the IKDC form. Knee Surg Sports Traumatol Arthroscopy 1(3-4): , Howe JG, Johnson RJ, Kaplan MJ, Fleming B, Jarvinen M: Anterior cruciate ligament reconstruction using quadriceps patellar tendon graft. Part I. Long-term follow-up. Am J Sports Med 19(5): , Hoher J, Bach T, Miinster A, Bouillon B, Tiling T: Does the mode of data collection change results in a subjective knee score? Am 1 Sports Med 25(5): , Jaureguito J, Elliot J, Lietner T, Dixon L, Reider B: The effects of arthroscopic,. partial meniscectomy in an otherwise normal knee: A retrospective review of functional, clinical and radiographic results. Arthroscopy 1 l(1): 29-36, Katz IN, Harris TM, Larson MG, Krushell RJ, Brown CH, Fossel AH, Liang MH: Predictors of functional outcomes after arthroscopic partial meniscectomy. J Rheumatol 19(12): , Kazis L, Anderson I, Meenan R: Effect sizes for interpreting changes in health status. Med Care 27(Suppl):S178- S189, Kellgren JH, Lawrence JS: The Epidemiology of Chronic Rheumatism. Atlas of Standard Radiographs, Oxford: Blackwell Scientific, Lequesne M: Informational indices. Validation of criteria and tests. Scand J Rheumatol80(Suppl): 17-27, Liang M, Jette AM: Measuring functional ability in chronic arthritis. Arthritis Rheum 24(1):80-86, Liang M, Larson M, Cullen K, Schwartz I: Comparative measurement of efxciency and sensitivity of five health status instruments for arthritis research. Arthritis Rheum 28(5): , McDowell I, Newel1 C: Measuring Health: A Guide to Rating Scales and Questionnaires, New York: Oxford University Press, Mohtadi N: Quality of life assessment as an outcome in anterior cruciate ligament reconstructive surgery. In: Jackson D (ed), The Anterior Cruciate Ligament Current and Future Concepts, New York: Raven Press, Risberg MA, Ekeland A: Assessment of functional tests after anterior cruciate ligament surgery. J Orthop Sports Phys Ther 1 9(4): , Rockborn P, Gillquist J: Outcome of arthroscopic meniscectomy. Acta Orthop Scand 66(2): , Roos H, Adalberth T, Dahlberg L, Lohmander LS: Osteoarthrosis of the knee after injury to the cruciate ligament or meniscus: The influence of time and age. Osteoarthritis Cartilage 3: , Rosner B: Fundamentals of Biostatistics, Belmont, CA: Duxbury Press, Sgaglione NA, Del Pizzo W, Fox JM, Friedman MI: Critical analysis of knee ligament rating systems. Am J Sports Med 23(6): , Shapiro ET, Richmond JC, Rockett SE, McGrath MM, Donaldson WR: The use of a generic, patient-based health assessment (SF-36) for evaluation of patients with anterior cruciate ligament injuries. Am J Sports Med 24(2): , Shrout P, Fleiss J: lntraclass correlations: Uses in assessing rater reliability. Psycho1 Bull 86(2): , Small NC, Sledge CB, Katz JN: A conceptual framework for outcome research in arthroscopic meniscectomy: Results of a nominal group process. Arthroscopy 10(5): , Spector T, Dacre JE, Harris FA, Huskisson EC: Radiological progression of osteoarthritis: An 1 1 -year follow-up study of the knee. Ann Rheum Dis 5 1: , Tegner Y, Lysholm J: Rating systems in the evaluation of knee ligament injuries. Clin Orthop 198:43-49, Ware J: How to Score the Revised MOS Short-Form Health Scale (SF- 361, Boston, MA: The Health Institute, New England Medical Center Hospitals, Ware], Snow K, Kosinski M, Gandek B: SF-36 Health Survey Manual and Interpretation Guide, Boston, MA: New England Medical Center, The Health Institute, Ware JE, Sherbourne CD: The MOS 36- item short form health survey (SF-36). 1. Conceptual framework and item selection. Med Care 30: , Windsor R, Insalll, Warren R: The Hospital for Special Surgery knee ligament rating form. Am J Knee Surg 1: , Wright JG, Young NL: The patient-specific index: Asking patients what they want. J Bone Joint Surg 79A(7): , 1997 Volume 78 Number 2 Aupst 1998 s JOSPT