Total ankle arthroplasty (TAA) has evolved during the past decades. High failure rates and discouraging clinical outcomes of first-generation implants in the 1970s resulted in restricted use but also led to the development of modern three-component implants [15, 17]. These designs allow for flexion and extension and for rotational and sliding movements, resulting in improved congruency, reduced shear forces at the bone-implant interface, and less bone removal during implantation . As a consequence, clinical studies suggest improved outcomes and likely as a result apparently increased interest in this procedure [15, 17]. Considering this evolution, the number of outcome reports will increase and quality outcome measures should be used to reflect high-quality research.
Numerous instruments, clinician-generated and self-reported, are available to assess outcomes after foot and ankle surgery. In 2004, Button and Pinney  identified 49 rating scales of which 18 were used more than once, but the authors stated none of these measures had demonstrable reliability, validity, and responsiveness in patients with various foot and ankle disorders. Similarly, Parker et al.  noted the fundamental problem of existing instruments used to evaluate foot and ankle surgery is their limited exploration and evaluation of what patients perceive to be most important in their outcomes. They concluded none of the existing measures could claim to be valid for patient perceptions of outcome . More recently, Martin and Irrgang  comprehensively surveyed self-reported outcome instruments for various foot and ankle disorders and identified 14 different instruments. Five of these measures had some evidence to support their use relative to content validity, construct validity, reliability, and responsiveness . These reports, however, focused on available scores to assess the outcomes of foot and ankle surgery in general or focused only on self-reported questionnaires and did not comprehensively consider outcome instruments in TAA.
Our systematic review, therefore, addressed the following questions: (1) Which are the most frequently used outcome instruments in studies reporting on TAA? (2) Does the literature provide evidence to support their use in terms of validity, reliability, responsiveness, and interpretability?
Search Strategies and Criteria
Initially, we wrote a protocol defining the objectives, search terms, inclusion and exclusion criteria, and the methods of documentation based on the method described by Wright et al. . According to the protocol, the electronic databases MEDLINE through PubMed, EMBASE, and Cochrane (all until May 2009) were searched using the following search terms: “ankle arthroplasty” OR “ankle replacement” AND “outcome” OR “results” OR “score” OR “questionnaire”. The reference lists of all included articles and recent reviews were checked manually for additional relevant studies. We included all articles that (1) reported on TAA; (2) used a specifically defined outcome instrument (either clinician-generated or self-reported, eg, American Orthopaedic Foot and Ankle Society [AOFAS] hindfoot score  or Foot Function Index [FFI] ); and (3) were published in English, German, French, Italian, or Spanish. We excluded articles that (1) reported on ankle fusion or conversion of TAA to fusion; (2) single or a few cases; (3) used no specific outcome instrument (eg, simple outcome rating as good, fair, or poor); and (4) were published in a language other than one of those previously mentioned.
After removing duplicates, we obtained 763 citations from the searches (Fig. 1). Two independent reviewers (FDN, FMI) then screened titles, abstracts, and full texts against the inclusion and exclusion criteria. Disagreements during each step of the review process were discussed and resolved. Four hundred twenty-six citations appeared to be irrelevant by title, and 221 citations by abstract. The full texts of the remaining 116 references were further analyzed. Forty-five studies did not match the selection criteria and were removed. Eight studies were added after reviewing reference lists and recent reviews. Finally, we included 79 original articles referring to 15 different instruments (Fig. 1). We did not intend to calculate cumulative score values as part of a meta-analysis.
In the next step, the reference lists of the included articles were checked for references of studies that developed or first described the identified outcome instrument or that investigated its psychometric properties in patients with ankle disorders. Additionally, we searched the above-mentioned electronic databases for the following search terms: “name of identified instrument” AND “ankle” AND “validity” OR “validation” OR “reliability” OR “responsiveness” OR “interpretability”. The reference lists of review articles also were analyzed. According to established quality criteria to examine the measurement properties of outcome instruments [40-42], we retrieved information on reproducibility, internal consistency, content validity, criterion validity, construct validity, responsiveness, and interpretability of the identified region- and disease-specific outcome instruments.
To investigate the methodologic quality and results of these clinimetric studies, we used the checklist described by van der Leeden et al.  to assign different levels of evidence (Levels 1 to 3) for the following criteria. Reproducibility refers to the degree to which repeated measurements in clinically stable individuals (test-retest) provide similar results. A distinction can be made between reliability and agreement . The intraclass correlation coefficient (ICC) and kappa statistics are adequate measures to assess reliability . Similar to van der Leeden et al. , we assigned a Level 1 rating if an ICC or kappa value was reported for a sample size of at least 50 patients. A positive rating was assigned if these values were greater than 0.70, and a negative for lower values. We assigned a Level 2 rating if a Pearson correlation coefficient was reported for a sample size of also at least 50 patients. Coefficients greater than 0.80 were rated positively, and lower values negatively. A Level 3 evidence was assigned if an ICC, a kappa value, or a Pearson coefficient was reported for a sample size of less than 50 patients. ICC or kappa values greater than 0.80 or Pearson coefficients greater than 0.90 were assigned a positive rating, and lower values a negative rating . Agreement refers to the absolute measurement error of an instrument and therewith describes the precision of this instrument . The standard error of the measurement (SEM), Bland and Altman's limits of agreement, or the smallest detectable change (SDC) are adequate measures to assess agreement . The absolute measurement error should be smaller than the minimal clinically important difference (MCID) of score values in individuals with time . Therefore, the MCID of an instrument or subscale should be defined. We assigned a Level 1 rating for agreement if the limits of agreement, SEMs, or SDCs were determined in a sample size of at least 50 patients, and if the MCID was reported. A Level 2 rating was assigned if the sample size was less than 50 patients. Values for the limits of agreement, SEM, or SDC below the MCID were rated positively, and values above the MCID were rated negatively. A Level 3 rating was assigned if the MCID was not defined .
Internal consistency describes the homogeneity of an instrument or its (sub)scales. It is an important quality criterion for instruments that intend to measure one concept or construct . A Level 1 evidence was designated if factor analysis was performed in a sample size of at least seven times the number of items and in a minimum of 100 patients. Additionally, the Cronbach's alpha (CA) had to be reported for each of the instruments' (sub)scales. A Level 1 rating also was assigned if Rasch analysis was used and the methodology was completely defined. A Level 2 evidence was assigned if the descriptive information on the Rasch methodology was incomplete. A Level 2 rating also was assigned if factor analysis was used in a sample size of at least four patients per item and greater than 50 patients in total, and a Level 3 rating needed four patients per item and a total sample size less than 50 patients. A CA greater than 0.70 relates to a positive rating, and lower CA values to a negative rating .
Content validity examines the extent to which the concept of interest is measured or represented by the items of an instrument . In the present context, items must reflect areas of importance of patients with ankle osteoarthritis. A positive rating therefore was assigned if patients were involved during the item generation and selection .
Criterion validity examines the extent to which an instrument is related to a gold standard . In the present context, we are not aware of any instrument that really can be considered a gold standard related to ankle osteoarthritis and ankle arthroplasty. Considering gold standards used in other scientific areas (eg, doubly labeled water or total oxygen uptake for the quantification of physical activity), we defined no specific instrument (such as the SF-36) as the gold standard in this review, but physical performance tests closely related to ankle function (eg, single heel lifts). A Level 1 rating was assigned if hypotheses concerning expected relationships between the instrument and the gold standard were specified in advance, and these relationships were investigated in a sample of at least 50 patients. A Level 2 rating was assigned if no specific hypotheses were specified beforehand. A Level 3 evidence was designated if plausible relations were found in a sample of less than 50 patients. A positive rating needed a correlation with the gold standard greater than 0.70, and weaker correlations received a negative rating .
Construct validity refers to the extent an instrument correlates with other measures in a manner consistent with theoretically derived hypotheses concerning the concepts being measured . Predefined hypotheses regarding these relationships should be specified as precisely as possible. Construct validity can be determined by convergent and divergent (or discriminant) validity. Evidence of convergent validity is provided by moderate to high correlations with other instruments measuring the same construct. In contrast, there should be no or only weak associations with instruments measuring different constructs (divergent validity) . A Level 1 rating was assigned if hypotheses concerning expected correlations with other instruments were specified in advance, and these relationships were investigated in a sample of at least 50 patients. A positive rating needed confirmation of at least 75% of these hypotheses. A Level 2 rating was assigned if no specific hypotheses were specified beforehand and the sample size was greater than 50. A Level 3 evidence was designated if plausible relations with other measures were found in a sample of less than 50 patients .
Responsiveness refers to the ability of an instrument to detect clinical changes with time. It can be considered an aspect of longitudinal validity . The effect size (ES), standardized response mean (SRM), and area under the receiver operating characteristics curve (AUC) are adequate measures to assess responsiveness . As responsiveness refers to longitudinal validity, predefined hypotheses also should be specified . The levels of evidence, therefore, were assigned similarly to construct validity . A positive rating was given if the ES or SRM was greater than 0.8 (high responsiveness) or the AUC was greater than 0.70 .
Floor and ceiling effects occur when patients score the lowest or highest score on an instrument, respectively. As a consequence, clinical deterioration or improvement cannot be assessed, and patients scoring lowest or highest possible cannot be distinguished from each other. Floor and ceiling effects were considered present if greater than 15% of patients achieved the lowest or highest score, respectively . A positive rating was assigned for the absence of floor or ceiling effects .
Interpretability refers to the degree to which one can assign qualitative meaning to quantitative scores . Interpretability, therefore, is related to the MCID, which should be determined in a sample size of at least 50 patients to receive a positive rating .
We identified 15 distinct outcome instruments used to determine the clinical outcome of TAA in 79 studies (Fig. 2). Eight of these measures were region specific, one was disease specific, three were generic, two were related to physical activity, and one could not be classified. The most commonly used instruments were the AOFAS hindfoot score (n = 41), the Kofoed ankle score (n = 21) [22, 23], the visual analog scale (VAS) assessing pain (n = 15), and the generic SF-36 (n = 6) . Eight additional instruments were used only once or twice.
We identified 13 articles providing information on quality criteria for the previously identified region- and disease-specific questionnaires (Tables 1, 2). Except for the AOFAS hindfoot score, we found no information on quality criteria for all other clinician-based instruments. Five studies provided information on properties of the AOFAS hindfoot score, eight studies on the FFI, and one study on the Ankle Osteoarthritis Scale (AOS)  (Tables 1, 2). Evidence in terms of the different quality criteria was low for the AOFAS hindfoot score and the AOS, and we could assign moderate ratings for the FFI (Table 3). None of the instruments provided evidence in terms of interpretability attributable to unknown MCIDs.
TAA has evolved during the past decades and the improvements in survivorship and clinical outcomes have led to the development of various new implants [15, 17]. Assessment of outcomes using appropriate outcome measures is the cornerstone of successful clinical research and allows for comparisons of patients' function and different treatment modalities or implants. Considering the growing popularity of TAA and therewith the increasing need for quality outcome research, this review addressed the following questions: (1) Which are the most frequently used outcome instruments in studies reporting on TAA? (2) Does the literature provide evidence to support their use in terms of validity, reliability, responsiveness, and interpretability?
Some limitations and aspects must be considered before interpreting our results. First, we did not include all electronic databases in the systematic search. Therefore, some instruments used in studies reporting on TAA outcomes may have been missed. Second, the definitions of criterion validity and construct validity have been used rather confusingly in the different studies. Budiman-Mak et al. [3, 4], for example, defined the “50 feet walking time test” as a measure for criterion validity in their first paper on the FFI  but as a measure for construct validity in their last paper on the FFI . The SF-36 and the WOMAC  also have been used to measure criterion validity , whereas the SF-36 was used to evaluate construct validity in another study . We doubt if any generic tool, such as the SF-36, Musculoskeletal Function Assessment (MFA) , or Quality Adjusted Life Year (QUALY) , could really be considered a gold standard for patients undergoing TAA. Similarly, the WOMAC might be considered the gold standard instrument for osteoarthritis, but only for patients having hip and knee arthroplasties. In this review, we, therefore, considered these different instruments and questionnaires as being related to construct validity, and we defined physical performance tests closely linked to ankle function as being related to criterion validity. The VAS pain was the third most frequently used instrument in original studies reporting on TAA, but it is neither foot nor ankle specific and represents only a single-item construct. The SF-36 , MFA , and Health Assessment Questionnaire  are generic tools assessing general health or health-related quality of life. These measures are self-reported, but they cannot be considered to adequately determine the region-specific health state of patients with ankle osteoarthritis or after TAA. It is beyond question these measures, similar to the physical activity-assessing University of California at Los Angeles activity scale or International Physical Activity Questionnaire , offer important additional information about the patients' health state. Nevertheless, the following discussion focuses on the identified region- and disease-specific outcome instruments.
Our review suggests the AOFAS hindfoot score  and the Kofoed ankle score [22, 23] are the two most frequently used outcome instruments in studies reporting on TAA. Both are clinician-based, region-specific 100-point scores, with 100 points reflecting the best clinical state. Their structure slightly differs in that the AOFAS attributes 40 points to the pain component, 50 points to the function component (including 16 points for hindfoot motion), and 10 points to hindfoot alignment. The Kofoed, in contrast, attributes 50 points to the pain component, only 30 points to the function component, and 20 points to range of motion (ROM). These differences illustrate, despite being 100-point scores, the absolute values cannot simply be compared between studies. The less frequently used instruments, ie, that of Evanski and Waugh  (pain 40, function 50, ROM 10), the New Jersey Orthopaedic Hospital ankle score  (pain 40, function 40, ROM 15, deformity 5), that of Takakura et al.  (40 pain, 40 function, 20 ROM), and that of Mazur et al.  (pain 50, function 40, ROM 10), are similarly clinician-based, region-specific 100-point rating systems, also with differences in pain, function, and ROM weightings.
However, for all these clinician-generated measures, except the AOFAS, the literature provides no evidence of validity, reliability, responsiveness, or interpretability of these scores. Also, no details on development strategies or the rationale of their structures could be identified, either in the original reports or in following studies. The lack of a clear theoretical framework behind these instruments also makes it difficult to interpret any evidence of content or construct validity as it is not clear what these instruments were supposed to measure. The World Health Organization, with its International Classification of Functioning, Disability and Health (ICF), proposed a conceptual model according to which items of a measure can be categorized . The ICF identified three levels of human functioning: (1) body or body part, (2) whole person, and (3) whole person in a social context. Disability involves dysfunctioning at one or more of these three levels: impairments (problems in body function or structures), activity limitation (difficulties in executing activities), and participation restriction (problems in involvement in a life situation) . Symptoms and clinical signs (ie, pain, ROM, alignment, etc) are related to the impairment domain, whereas activities of daily living such as self-care or sports are related to the activity limitation and participation restriction domains. As there is evidence that these different domains are not necessarily dependent on or correlated with each other, combining items of these domains into one score is questionable. This occurs, however, with all of the above-mentioned clinician-based outcome measures but also with the self-reported AOS  and FFI [3, 4].
The AOFAS hindfoot score has been the subject of concern before. Guyton  described several conceptual limitations of the AOFAS using Monte Carlo modeling. He pointed out, in addition to other drawbacks, the small number of answer categories in several subscales of the score is a major confounding factor leading to skewed data. He concluded the AOFAS cannot produce reliable data and score values obtained by parametric statistics must be interpreted with care . SooHoo et al.  correlated the AOFAS score with the SF-36 and found only weak associations, suggesting poor construct validity of this instrument. Two other studies investigating the association between the AOFAS and the generic questionnaires QUALY and MFA also found only low correlations between these instruments [26, 34]. A greater association was found between the subjective part of the AOFAS and the FFI . There are two other possible weaknesses of the AOFAS, as with the other clinician-based outcome measures. First, including a clinical examination in a score always introduces a possible confounder as different examiners might measure different things. Intrarater or interrater relations have so far not been determined for the AOFAS or the other clinician-determined instruments. Second, these rating systems might be not specific enough to measure TAA outcomes. The inclusion of objectively measured ROM represents a problem for patients who have, for example, an additional subtalar arthrodesis. Such patients may be completely satisfied with a well-functioning TAA, but they lose 8 points on the AOFAS owing to their fused subtalar joint. However, because of its wide use in the literature, AOFAS score values still offer the best comparison between different studies.
Self-reported questionnaires, in contrast, might more adequately reflect the patients' perspective. What we have learned from outcome research in fields other than foot and ankle surgery is that self-reported outcome instruments allow for a more complete estimation of the patients' health status and of issues relevant to the patients. The only self-reported region- or disease-specific measures used in studies on TAA, however, were the FFI , AOS , and “Oxford” . The “Oxford” is put in quotation marks because this instrument is a not validated, self-developed questionnaire modeling to the original Oxford Hip Score . The original FFI also was a result of an expert panel, initially developed for patients with rheumatoid arthritis . Although, in the meantime, numerous studies broadened its use to the entire spectrum of foot and ankle disorders, included patient perceptions, and adapted the instrument for use in different languages [1, 24, 30, 36, 38, 46], several limitations of this instrument have been highlighted, resulting in a recently performed extensive revision of this questionnaire based on Rasch analysis . Although we found reasonable ratings in terms of the different quality criteria for the FFI, the above-mentioned studies resulted in at least five different FFI versions (FFI original, FFI-R long, FFI-R short, FFI-D, FFI-5pt) and its use in patients having TAA, therefore, can be recommended only cautiously. Considering the AOS is based on the original FFI , it can be concluded no quality region- or disease-specific tools have been used regularly in TAA studies until now. Recognizing the recent literature, it is interesting that very well-developed self-reported instruments, such as the Foot and Ankle Ability Measure (FAAM), have not yet been used in studies reporting on TAA outcomes . The FAAM has shown evidence of validity, reliability, and responsiveness in patients with a broad spectrum of foot complaints, including ankle osteoarthritis .
For all the instruments identified in our review, no studies have supplied enough information to understand the interpretability of the results; in particular, no MCIDs have been reported yet. Interpretability is defined by the Scientific Advisory Committee as “the degree to which one can assign easily understood meaning to an instrument's quantitative score” . To facilitate interpretability, various kinds of information are needed, eg, norm values, differences between subgroups expected to differ in scores, and MCIDs [40, 42]. Unfortunately, most of this information is not available for the instruments used in TAA studies.
Several different outcome instruments have been used in studies reporting on TAA, with the AOFAS hindfoot score and the Kofoed ankle score being the most common. However, there is no or only limited evidence to support their use in terms of patient relevance, validity, reliability, responsiveness, and interpretability. Self-reported questionnaires to assess TAA outcomes are rather uncommon until now, and considerable research is required to broaden the knowledge regarding the existing measures and to develop and investigate new measures that validly and reliably assess outcomes in this target population.
1. Agel, J., Beskin, JL., Brage, M., Guyton, GP., Kadel, NJ., Saltzman, CL., Sands, AK., Sangeorzan, BJ., SooHoo, NF., Stroud, CC. and Thordarson, DB. Reliability of the Foot Function Index: a report of the AOFAS Outcomes Committee. Foot Ankle Int.
2005; 26: 962-967.
2. Bellamy, N., Buchanan, WW., Goldsmith, CH., Campbell, J. and 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 knee. J Rheumatol.
1988; 15: 1833-1840.
3. Budiman-Mak, E., Conrad, K., Stuck, R. and Matters, M. Theoretical model and Rasch analysis to develop a revised Foot Function Index. Foot Ankle Int.
2006; 27: 519-527.
4. Budiman-Mak, E., Conrad, KJ. and Roach, KE. The Foot Function Index: a measure of foot pain and disability. J Clin Epidemiol.
1991; 44: 561-570. 10.1016/0895-4356(91)90220-4
5. Buechel, FF., Pappas, MJ. and Iorio, LJ. New Jersey low contact stress total ankle replacement: biomechanical rationale and review of 23 cementless cases. Foot Ankle.
1988; 8: 279-290.
6. Button, G. and Pinney, S. A meta-analysis of outcome rating scales in foot and ankle surgery: is there a valid, reliable, and responsive system? Foot Ankle Int.
2004; 25: 521-525.
7. Dawson, J., Fitzpatrick, R., Carr, A. and Murray, D. Questionnaire on the perceptions of patients about total hip replacement. J Bone Joint Surg Br.
1996; 78: 185-190.
8. Vet, HC., Terwee, CB., Knol, DL. and Bouter, LM. When to use agreement versus reliability measures. J Clin Epidemiol.
2006; 59: 1033-1039. 10.1016/j.jclinepi.2005.10.015
9. Domsic, RT. and Saltzman, CL. Ankle osteoarthritis scale. Foot Ankle Int.
1998; 19: 466-471.
10. Engelberg, R., Martin, DP., Agel, J., Obremsky, W., Coronado, G. and Swiontkowski, MF. Musculoskeletal Function Assessment instrument: criterion and construct validity. J Orthop Res.
1996; 14: 182-192. 10.1002/jor.1100140204
11. Evanski, PH. and Waugh, TR. Management of arthritis of the ankle: an alternative of arthrodesis. Clin Orthop Relat Res.
1977; 122: 110-115.
12. Fries, JF., Spitz, P., Kraines, RG. and Hotman, HR. Measurement of patient outcome in arthritis. Arthritis Rheum.
1980; 23: 137-145. 10.1002/art.1780230202
13. Guyatt, GH., Deyo, RA., Charlson, M., Levine, MN. and Mitchell, A. Responsiveness and validity in health status measurement: a clarification. J Clin Epidemiol.
1989; 42: 403-408. 10.1016/0895-4356(89)90128-5
14. Guyatt, GH., Feeny, DH. and Patrick, DL. Measuring health-related quality of life. Ann Intern Med.
1993; 118: 622-629.
15. Guyer, AJ. and Richardson, G. Current concepts review: total ankle arthroplasty. Foot Ankle Int.
2008; 29: 256-264. 10.3113/FAI.2008.0256
16. Guyton, GP. Theoretical limitations of the AOFAS scoring systems: an analysis using Monte Carlo modeling. Foot Ankle Int.
2001; 22: 779-787.
17. Haddad, SL., Coetzee, JC., Estok, R., Fahrbach, K., Banel, D. and Nalysnyk, L. Intermediate and long-term outcomes of total ankle arthroplasty and ankle arthrodesis: a systematic review of the literature. J Bone Joint Surg Am.
2007; 89: 1899-1905. 10.2106/JBJS.F.01149
18. Hosman, AH., Mason, RB., Hobbs, T. and Rothwell, AG. A New Zealand national joint registry review of 202 total ankle replacements followed for up to 6 years. Acta Orthop.
2007; 78: 584-591. 10.1080/17453670710014266
19. Ibrahim, T., Beiri, A., Azzabi, M., Best, AJ., Taylor, GJ. and Menon, DK. Reliability and validity of the subjective component of the American Orthopaedic Foot and Ankle Society clinical rating scales. J Foot Ankle Surg.
2007; 46: 65-74. 10.1053/j.jfas.2006.12.002
20. Kirshner, BF. and Guyatt, GH. A methodological framework for assessing health indices. J Chronic Dis.
1985; 38: 27-36. 10.1016/0021-9681(85)90005-0
21. Kitaoka, HB., Alexander, IJ., Adelaar, RS., Nunley, JA., Myerson, MS. and Sanders, M. Clinical rating systems for the ankle-hindfoot, midfoot, hallux, and lesser toes. Foot Ankle Int.
1994; 15: 349-353.
22. Kofoed, H. In: Kossowsky, R. and Kossovsky, V. (eds.), A new total ankle joint prosthesis. Material Sciences and Implant Orthopedic Surgery.
1986: Dordrecht, The Netherlands: Martinus Nijhoff; 75-84.
23. Kofoed, H. Cylindrical cemented ankle arthroplasty: a prospective series with long-term follow-up. Foot Ankle Int.
1995; 16: 474-479.
24. Kuyvenhoven, MM., Gorter, KJ., Zuithoff, P., Budiman-Mak, E., Conrad, KJ. and Post, MW. The foot function index with verbal rating scales (FFI-5pt): a clinimetric evaluation and comparison with the original FFI. J Rheumatol.
2002; 29: 1023-1028.
25. Lohr, KN., Aaronson, NK., Alonso, J., Burnam, MA., Patrick, DL., Perrin, EB. and Roberts, JS. Evaluating quality-of-life and health status instruments: development of scientific review criteria. Clin Ther.
1996; 18: 979-992. 10.1016/S0149-2918(96)80054-3
26. Malviya, A., Makwana, N. and Laing, P. Correlation of the AOFAS scores with a generic health QUALY score in foot and ankle surgery. Foot Ankle Int.
2007; 28: 494-498. 10.3113/FAI.2007.0494
27. Martin, RL. and Irrgang, JJ. A survey of self-reported outcome instruments for the foot and ankle. J Orthop Sports Phys Ther.
2007; 37: 72-84. 10.2519/jospt.2007.2403
28. Martin, RL., Irrgang, JJ., Burdett, RG., Conti, SF. and Swearingen, JM. Evidence of validity for the Foot and Ankle Ability Measure (FAAM). Foot Ankle Int.
2005; 26: 968-983.
29. Mazur, JM., Schwartz, E. and Simon, SR. Ankle arthrodesis: long-term follow-up with gait analysis. J Bone Joint Surg Am.
1979; 61: 964-975.
30. Naal, FD., Impellizzeri, FM., Huber, M. and Rippstein, PF. Cross-cultural adaptation and validation of the Foot Function Index for use in German-speaking patients with foot complaints. Foot Ankle Int.
2008; 29: 1222-1228. 10.3113/FAI.2008.1222
31. Naal, FD., Impellizzeri, FM., Loibl, M., Huber, M. and Rippstein, PF. Habitual physical activity and sports participation after total ankle arthroplasty. Am J Sports Med.
2009; 37: 95-102. 10.1177/0363546508323253
32. Neufeld, SK. and Lee, TH. Total ankle arthroplasty: indications, results, and biomechanical rationale. Am J Orthop.
2000; 29: 593-602.
33. Parker, J., Nester, CJ., Long, AF. and Barrie, J. The problem with measuring patient perceptions of outcome with existing outcome measures in foot and ankle surgery. Foot Ankle Int.
2003; 24: 56-60.
34. Pena, F., Agel, J. and Coetzee, JC. Comparison of the MFA to the AOFAS outcome tool in a population undergoing total ankle replacement. Foot Ankle Int.
2007; 28: 788-793. 10.3113/FAI.2006.0788
35. Radford, PJ. In: Pynsent, PB., Fairbank, JC. and Carr, A. (eds.), General outcomes measures. Outcome Measures in Orthopaedics.
1993: London, UK: Butterworth-Heinemann; 59-80.
36. SooHoo, NF., Samimi, DB., Vyas, RM. and Botzler, T. Evaluation of the validity of the Foot Function Index in measuring outcomes in patients with foot and ankle disorders. Foot Ankle Int.
2006; 27: 38-42.
37. SooHoo, NF. and Shuler, M. Fleming LL; American Orthopaedic Foot and Ankle Society. Evaluation of the validity of the AOFAS Clinical Rating Systems by correlation to the SF-36. Foot Ankle Int
2003; 24: 50-55.
38. SooHoo, NF., Vyas, R. and Samimi, D. Responsiveness of the foot function index, AOFAS clinical rating systems, and SF-36 after foot and ankle surgery. Foot Ankle Int.
2006; 27: 930-934.
39. Takakura, Y., Tanaka, Y., Sugimoto, K., Tamai, S. and Masuhara, K. Ankle arthroplasty: a comparative study of cemented metal and uncemented ceramic prostheses. Clin Orthop Relat Res.
1990; 252: 209-216.
40. Terwee, CB., Bot, SD., Boer, MR., Windt, DA., Knol, DL., Dekker, J., Bouter, LM. and Vet, HC. Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol.
2007; 60: 34-42. 10.1016/j.jclinepi.2006.03.012
41. Leeden, M., Steultjens, MP., Terwee, CB., Rosenbaum, D., Turner, D., Woodburn, J. and Dekker, J. A systematic review of instruments measuring foot function, foot pain, and foot-related disability in patients with rheumatoid arthritis. Arthritis Rheum.
2008; 59: 1257-1269. 10.1002/art.24016
42. Veenhof, C., Bijlsma, JW., Ende, CH., Dijk, GM., Pisters, MF. and Dekker, J. Psychometric evaluation of osteoarthritis questionnaires: a systematic review of the literature. Arthritis Rheum.
2006; 55: 480-492. 10.1002/art.22001
43. Ware, JE Jr. and Sherbourne, CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care.
1992; 30: 473-483. 10.1097/00005650-199206000-00002
44. World Health Organization. International Classification of Functioning, Disability and Health (ICF). Geneva, Switzerland: World Health Organization; 2001. Available at: http://www.who.int/classifications/icf/en/
. Accessed June 20, 2009.
45. Wright, RW., Brand, RA., Dunn, W. and Spindler, KP. How to write a systematic review. Clin Orthop Relat Res.
2007; 455: 23-29. 10.1097/BLO.0b013e31802c9098
46. Wu, SH., Liang, HW. and Hou, WH. Reliability and validity of the Taiwan Chinese version of the Foot Function Index. J Formos Med Assoc.
2008; 107: 111-118. 10.1016/S0929-6646(08)60124-2