Validation of the Core Outcome Measures Index in Patients With Femoroacetabular Impingement Franco M. Impellizzeri, Ph.D., Anne F. Mannion, Ph.D., Florian D. Naal, M.D., and Michael Leunig, M.D.

Purpose: To evaluate a short, hip-oriented outcome instrument, based on the Core Outcome Measures Index (COMI), in patients undergoing surgery for femoroacetabular impingement (FAI). Methods: The following full-length questionnaires were completed preoperatively and at 6 and 12 months postoperatively by 159 consecutive FAI patients: Hip Outcome Score (HOS); Oxford Hip Score; Western Ontario and McMaster Universities Arthritis Index; Short Form 12 Health Survey; World Health Organization Quality of Life questionnaire, short version; and EuroQoleFive Dimensional index. The scores for the 6 hip-oriented Core Outcome Measures Index (COMI-Hip) itemsdaddressing pain, function, symptom-specific well-being, quality of life, and disabilitydwere extracted from established full-length questionnaires, and their performance as an index was compared with that of the full-length instruments. Results: Scores for the single items of the COMI-Hip questionnaire correlated well with the scores for the corresponding full-length instruments from which they were extracted (r ¼ 0.89 to 0.62, P < .001). The COMI-Hip sum score also correlated well with the Oxford Hip Score and the Western Ontario and McMaster Universities Arthritis Index pain and function scores (r ¼ 0.85 to 0.70, P < .001), as well as with the HOS (r ¼ 0.72 to 0.60, P < .001), an instrument specifically developed for assessing FAI patients. Internal responsiveness (Cohen d for effect size) of the COMI-Hip sum score from preoperatively to 12 months postoperatively was similar to that of the HOS activitieseofedaily living subscale (d ¼ 0.76 and d ¼ 0.68, respectively; P < .001). Significant correlations were found between the change scores of the COMI-Hip sum score and those of the HOS activitieseofedaily living and sport subscales at 6 months (r ¼ 0.62 and r ¼ 0.60, respectively; P < .001) and 12 months (r ¼ 0.69 and r ¼ 0.61, respectively; P < .001), showing the external responsiveness of the COMI-Hip. Conclusions: The COMI-Hip is a simple yet valid and responsive outcome instrument for the efficient assessment of patients undergoing surgery for FAI. It performs at least as well as the current reference instrument for FAI, the HOS, and can therefore be considered a potentially valuable instrument for routine use in both research and clinical practice. Level of Evidence: Level II, development of diagnostic criteria based on consecutive patients (with universally applied reference gold standard).

F

emoroacetabular impingement (FAI) is a recently characterized hip disorder associated with pain and dysfunction in young, active adults.1 From a clinical point of view, the aim of surgical treatment is to restore a more normal morphology and create impingement-free

From the Department of Research and Development, Schulthess Clinic, Zurich, Switzerland. The study was funded by Schulthess Klinik Research Funds. The authors report the following potential conflict of interest or source of funding: F.D.N. receives support from Smith & Nephew, consultant fees for academic hip courses. Deutsche Arthrose-Hilfe e.V, grant for research project in arthroplasty. M.L. receives support from Smith & Nephew. Received April 22, 2014; accepted January 16, 2015. Address correspondence to Franco M. Impellizzeri, Ph.D., Department of Research and Development, Schulthess Clinic, Lengghalde 2, 8008 Zurich, Switzerland. E-mail: [email protected] Ó 2015 by the Arthroscopy Association of North America 0749-8063/14333/$36.00 http://dx.doi.org/10.1016/j.arthro.2015.01.014

motion to prevent or delay the progression of degenerative changes of the hip joint.2 For the patient, treatment should relieve pain and improve daily functioning, work capacity, and quality of life.3 For this reason, it is common to use patient-related outcome measures to evaluate the success of treatment. Both generic and disease-specific instruments should be used for a comprehensive evaluation,4 but the use of several lengthy questionnaires can increase the burden on both the respondent and the administrative system, thus limiting the feasibility of large questionnaire sets in the routine, systematic evaluation of patient outcome. For some orthopaedic disorders, this problem has been addressed by developing and validating short, multidimensional instruments covering the most relevant outcome domains. The Core Outcome Measures Index (COMI), originally developed for patients with back pain,5,6 is an example of such a questionnaire (6 items), subsequently used with success in several

Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol

-,

No

-

(Month), 2015: pp 1-9

1

2

F. M. IMPELLIZZERI ET AL.

studies and international registries.7,8 A recent study has shown that an adaptation of the COMIdthe hip-oriented Core Outcome Measures Index (COMIHip)dis a valid and responsive instrument for assessing pain, function, quality of life, and disability in patients undergoing hip replacement.9 Because these core outcome domains are also those of greatest importance in assessing the outcome of treatment in patients with FAI, the purpose of this study was to evaluate a short, hip-oriented outcome instrument, based on the COMI, in patients undergoing surgery for FAI. We hypothesized that the COMI-Hip is a valid instrument for assessing outcome in patients with FAI (i.e., that there would be large correlations between the COMI-Hip and the corresponding full instruments) and that the responsiveness is comparable with the reference questionnaire (Hip Outcome Score [HOS]). The validity and responsiveness of the COMI-Hip items were evaluated in relation to (1) longer jointspecific hip disability questionnaires, (2) a reference questionnaire specifically developed for FAI patients (HOS),10,11 and (3) generic quality-of-life questionnaires.

Table 1. Baseline Patient Sociodemographic Data Age, mean (SD), yr BMI, mean (SD), kg/m2 Gender Male Female ASA class I II Living condition Rural City Education Primary school High school Professional school University Civil status Married Single Divorced/separated Comorbidities None 1 2 3

Data 35.3 (11.5) 23.6 (4.1) 48% 52% 57% 43% 56% 44% 15% 4% 67% 14% 40% 55% 5% 60% 28% 8% 4%

ASA, American Society of Anesthesiologists; BMI, body mass index.

Methods This was a single-center retrospective analysis of prospectively collected data, involving a total of 159 consecutive patients with FAI who were undergoing either arthroscopic surgery with labral preservation (54%) or limited anterolateral open surgery with labral resection (46%). The inclusion criteria were as follows: cam, localized pincer, or mild to moderate mixed impingement in hips with at most early osteoarthritis ( 0.50 (large correlation). The internal responsiveness was determined with the Cohen d, for which a sample size greater than 90 was necessary to detect at least a medium effect size (0.5). Measures of centrality and dispersion include means and standard deviations unless otherwise stated. Convergent (construct) validity and external responsiveness were examined with Pearson product moment correlations. Strength of association was interpreted using the Cohen benchmarks as follows: less than 0.10, trivial; 0.10 to 0.30, small; 0.30 to 0.50, moderate; and greater than 0.50, large. Internal responsiveness was given by Cohen effect sizes (Cohen d ¼ [Post-test meanePretest mean]/Standard deviation baseline) and

presented with the corresponding 95% confidence intervals. External responsiveness was evaluated by examining the correlation between the change scores of the COMI-Hip sum score and those of the reference instrument (HOS). Changes over time were examined using 1way analysis of variance, and pre-post differences were presented as mean differences with the corresponding 95% confidence intervals. Floor and ceiling effects were calculated as the percentage of patients showing the worst and best values for the instrument, respectively. All analyses were carried out using SPSS software (version 17; SPSS, Chicago, IL).

Results Convergent and Construct Validity The correlation coefficients between the scores for the single COMI-Hip items and the corresponding fulllength instruments from which the items were extracted were large (r ¼ 0.89 to 0.62, P < .001) (Table 3). Large correlations were also found between a given COMI-Hip item and other questionnaire subscales measuring a similar construct, for example, r ¼ 0.70 to 0.65 for the correlation between the COMI-Hip pain item and WOMAC pain subscale and r ¼ 0.79 to 0.60 for the correlation between the COMI-Hip function item and either the OHS (a measure of pain and function) or WOMAC physical function subscale (P < .001). The coefficients describing the strength of the correlation between the COMI-Hip sum score and all fulllength condition-specific instruments were (with the exception of the WOMAC stiffness subscale) consistently large, ranging from 0.85 to 0.70 (P < .001) (Table 4). Similarly, the correlation coefficients between the COMI-Hip sum score and the EuroQoleFive Dimensional index, EuroQoleVisual Analogue Scale, and physical component scores of the generic questionnaires were large (r ¼ 0.83 to 0.67, P < .001). Lower correlation coefficients were found for the relation between the COMI-Hip sum score and the

4

F. M. IMPELLIZZERI ET AL.

Table 3. Correlation Coefficients for Relation Between COMI Items and Full Instruments (Generic and Condition Specific) Correlation Coefficient for Relation Between COMI Item and Full Instrument COMI pain item (from OHS) OHS* WOMAC pain subscale COMI function item (from SF-12) SF-12 PCS* OHS WOMAC function subscale COMI QoL item (from WHO) WHO physical health subscale* WHO psychological health subscale WHO social subscale WHO environment subscale SF-12 MCS EQ-VAS EQ-5D index COMI social disability item (from SF-12) SF-12 MCS* SF-12 PCS WHO psychological health subscale WHO social subscale WHO environment subscale COMI work disability item (from OHS) OHS* SF-12 MCS SF-12 PCS WHO psychological health subscale WHO social subscale WHO environment subscale

Baseline

6 mo

12 mo

0.73* 0.65

0.72* 0.66

0.62* 0.70

0.86* 0.73 0.65

0.84* 0.69 0.60

0.89* 0.79 0.78

0.70* 0.56 0.30 0.43 0.37 0.70 0.63

0.71* 0.51 0.28 0.32 0.45 0.77 0.74

0.66* 0.52 0.32 0.48 0.32 0.65 0.62

0.66* 0.41 0.51 0.45 0.41

0.73* 0.32 0.54 0.33 0.27

0.74* 0.27 0.50 0.34 0.49

0.74* 0.24 0.60 0.24 0.14 0.25

0.83* 0.24 0.71 0.18 0.01 0.27

0.79* 0.15 0.68 0.27 0.13 0.35

COMI, Core Outcome Measures Index; EQ-5D, EuroQoleFive Dimensional; EQ-VAS, EuroQoleVisual Analogue Scale; MCS, mental component summary; OHS, Oxford Hip Score; PCS, physical component summary; QoL, quality of life; SF-12, Short Form 12 Health Survey; WHO, World Health Organization Quality of Life questionnaire, short version; WOMAC, Western Ontario and McMaster Universities Arthritis Index. *Correlation coefficients on the first line of each section indicate the correlation between the given hip-oriented Core Outcome Measures Index (COMI-Hip) item and the full-length instrument from which the item was extracted.

Table 4. Correlation Coefficients for Relation Between COMI Sum Scores and Full Instrument Scores (Generic and Condition Specific) Correlation Coefficient for Relation Between COMI Sum Score and Full Instrument Score Generic instruments EQ-VAS EQ-5D index WHO physical health subscale WHO psychological health subscale WHO social subscale WHO environment subscale SF-12 PCS SF-12 MCS Condition-specific instruments HOS ADL subscale HOS sport subscale OHS WOMAC total WOMAC pain subscale WOMAC stiffness subscale WOMAC function subscale

Baseline

6 mo

12 mo

0.72 0.76 0.82 0.50 0.28 0.38 0.79 0.40

0.73 0.80 0.83 0.36 0.14 0.30 0.80 0.42

0.67 0.73 0.78 0.41 0.26 0.40 0.79 0.20

0.71 0.70 0.82 0.71 0.70 0.44 0.71

0.72 0.70 0.85 0.73 0.73 0.51 0.73

0.72 0.63 0.85 0.77 0.79 0.49 0.74

ADL, activities of daily living; COMI, Core Outcome Measures Index; EQ-5D, EuroQoleFive Dimensional; EQ-VAS, EuroQoleVisual Analogue Scale; HOS, Hip Outcome Score; MCS, mental component summary; OHS, Oxford Hip Score; PCS, physical component summary; SF-12, Short Form 12 Health Survey; WHO, World Health Organization Quality of Life questionnaire, short version; WOMAC, Western Ontario and McMaster Universities Arthritis Index.

CORE OUTCOME MEASURES INDEX

mental and psychological domains of the generic questionnaires (Table 4). The correlation coefficients describing the relation between the COMI-Hip sum score and the reference instrument (HOS) were large, ranging from 0.72 to 0.63. Responsiveness and Change Scores Table 5 shows the mean scores at each time point, the change scores at each follow-up, and the internal responsiveness (Cohen d) for the individual COMI-Hip items, COMI-Hip sum score, and full questionnaires. All scoresdexcept for the mental component of the Short Form 12 Health Survey; the World Health Organization Quality of Life questionnaire, short version, social and psychological health subscales; and the COMI social disability itemdshowed a significant improvement from baseline to follow-up (P < .001). All full-length questionnaires assessing pain and physical function (i.e., the condition-specific instruments) showed moderate responsiveness, with values ranging from 0.50 (WOMAC total score) to 0.72 (HOS sport subscale) at 6 months and from 0.51 (WOMAC function subscale) to 1.18 (HOS sport subscale) at 12 months. The single items of the COMI-Hip, with the exclusion of social disability, showed values ranging from 1.09 (COMI pain item at 12 months) to 0.32 (COMI quality-of-life item at 6 months). Large correlation coefficients (external responsiveness) were also found for the relation between the change scores of the reference questionnaire (HOS) and those of the COMI-Hip sum score at 6 months (r ¼ 0.62 and r ¼ 0.60 for HOS activitieseofedaily living [ADL] subscale and HOS sport subscale, respectively) and 12 months (r ¼ 0.69 and r ¼ 0.61 for HOS ADL subscale and HOS sport subscale, respectively). At baseline, no floor or ceiling effects (i.e., no proportions >15%) were found for either the HOS or COMI. At 6 and 12 months’ follow-up, there was no floor or ceiling for the COMI whereas there were higher ceiling effects for the HOS (12% for ADL subscale and 5% for sport subscale), though still not exceeding the acceptable level (15%).

Discussion The COMI-Hip sum scores and single-item scores showed strong correlations with the scores on instruments assessing similar constructs, and the responsiveness of the COMI-Hip was comparable with that of the longer questionnaires, including the reference outcome questionnaire for FAI patients, the HOS. The basic psychometric properties of the COMI-Hip were similar to those reported in previous studies in which the COMI was validated for use in patients with spinal disorders,20-23 patients with inguinal hernia,24 and more recently, patients undergoing hip replacement.9 This study suggests that the adapted hip version

5

of the COMI is a potentially useful and valid, parsimonious instrument for the assessment of outcome in patients undergoing surgery for FAI. The original COMI developed for back pain was created by extracting appropriate items from existing, established outcome questionnaires and was validated by examining the relation between the two.6,20 Following a similar validation process, we examined the construct validity of the COMI-Hip by assessing the degree of correlation between the COMI-Hip singleitem scores/sum scores and the scores on the full-length questionnaires from which the COMI-Hip items were extracted. Furthermore, we determined the correlation between the COMI-Hip sum score and the scores on other instruments measuring similar constructs, including a questionnaire specifically developed for the assessment of patients undergoing surgery for FAI (i.e., the HOS). The scores for the single items of the COMIHip showed a strong correlation with the scores for the corresponding full-length questionnaires (r ¼ 0.89 to 0.66) and were comparable with those reported in patients undergoing total hip replacement (r ¼ 0.65 to 0.81).9 We also found strong correlations between the COMI-Hip items and other questionnaire subscales measuring similar constructs. The correlations reported in this study and the previous study on patients undergoing hip replacement13 were similar to the correlations reported in patients with low-back pain between the COMI pain item and other pain scale measures (r ¼ 0.67 to 0.79), as well as between the COMI function item and other measures of function/disability (r ¼ 0.67 to 0.70).6,20 Although the individual COMI items generally showed a strong correlation with the full questionnaires, most of the latter were not specifically developed for use in FAI.25 Therefore, to better evaluate the construct validity of the COMI-Hip, we determined the correlation between its sum scores and the scores on an instrument intentionally developed and validated for use in FAI patients undergoing surgical treatment, the HOS.10,11 Our results showed strong correlations (r > 0.70 in most cases) between the COMI-Hip and the HOS, thus providing evidence of construct validity. These correlations are similar to those found between the COMI sum score and reference instruments in patients with spinal disorders20-23 or inguinal hernia,24 as well as patients undergoing hip replacement.9 In addition to assessing construct validity, which is a prerequisite for a good measurement instrument, we examined the responsiveness of the COMI-Hip, an important attribute for evaluating the performance of an instrument used for evaluative purposes (i.e., longitudinal assessment). “Responsiveness” refers to the sensitivity to change: Internal responsiveness is determined from the distribution of the data (effect sizes), and external responsiveness, from changes in the

6

Table 5. Preoperative, Follow-up, and Change Scores for Different Instruments and COMI Items (Scores Adjusted to 0 to 10 Scale) Preoperative Score

12-mo FU Score

CS, 6 mo e Preoperative

Mean 7.1 5.4 2.9 1.3

SD 2.3 2.1 1.9 1.8

Mean 4.9 4.3 2.3 1.3

SD 2.6 2.1 1.8 1.6

Mean 4.6 4.1 2.0 1.1

SD 2.8 2.1 1.6 1.5

ANOVA (P Value)

Validation of the Core Outcome Measures Index in Patients With Femoroacetabular Impingement.

To evaluate a short, hip-oriented outcome instrument, based on the Core Outcome Measures Index (COMI), in patients undergoing surgery for femoroacetab...
213KB Sizes 0 Downloads 5 Views