Brazilian Version of the Foot Health Status Questionnaire (FHSQ-Br): Cross-Cultural Adaptation and Evaluation of Measurement Properties

OBJECTIVE To conduct a cross-cultural adaptation of the Foot Health Status Questionnaire into Brazilian-Portuguese and to assess its measurement properties. INTRODUCTION This instrument is an outcome measure with 10 domains with scores ranging from 0–100, worst to best, respectively. The translated instrument will improve the examinations and foot care of rheumatoid arthritis patients. METHODS The questions were translated, back-translated, evaluated by a multidisciplinary committee and pre-tested (n = 40 rheumatoid arthritis subjects). The new version was submitted to a field test (n = 65) to evaluate measurement properties such as test-retest reliability, internal consistency and construct validity. The Health Assessment Questionnaire, Numeric Rating Scale for foot pain and Sharp/van der Heijde scores for foot X-rays were used to test the construct validity. RESULTS The cross-cultural adaptation was completed with minor wording adaptations from the original instrument. The evaluation of measurement properties showed high reliability with low variation coefficients between interviews. The α-Cronbach coefficients varied from 0.468 to 0.855, while correlation to the Health Assessment Questionnaire and Numeric Rating Scale was statistically significant for five out of eight domains. DISCUSSION Intra- and inter-observer correlations showed high reliability. Internal consistency coefficients were high for all domains, revealing higher values for less subjective domains. As for construct validity, each domain revealed correlations with a specific group of parameters according to what the domains intended to measure. CONCLUSION The FHSQ was cross-culturally adapted, generating a reliable, consistent, and valid instrument that is useful for evaluating foot health in patients with rheumatoid arthritis.


INTRODUCTION
Foot pro�lems, particularly foot pain and foot deformity, are the cause of numerous visits to physicians. � A patient with a painful foot or foot deformity is more suscepti�le to falls due to impaired �alance and may have marked functional deterioration, reducing his or her health-related quality of life and independence; thus, foot pro�lems are a pu�lic health issue. [2][3][4][5] The prevalence of foot pro�lems in patients with rheumatoid arthritis (RA) is more than 50%, with one study reporting the proportion to �e as high as 93 out of 99 patients. 6 Clinical studies suggest that foot pain may �e pro�lematic in a�out one-third of patients in early disease with more frequent involvement of the metatarsophalangeal joints (34%) than the midtarsal (4%) and ankle (20%) regions. 7 A severe negative impact on mo�ility and functional capacity is o�served when the patient's feet �egin to �e affected, 8 and studies have shown that feet are affected earlier than hands in the course of the disease. 9-�� Furthermore, Priolo et al. (�997) stated that foot involvement is indicative of a more aggressive disease. �� The development of clinics specialized in treating feet of patients with RA has provided health professionals with an opportunity to gain a greater understanding of the impact of RA on feet and ankles. Treatment of foot-related conditions often requires more than systemic drug therapy, such as podiatry, physiotherapy and orthopedic services. 8,�0 The RA �urden and the efficacy of multidisciplinary foot-health care in RA are measura�le using several generic 3,�2,�3 instruments and a single RA-specific 4 foothealth instrument, which are used for �oth routine clinical purposes and research.
The Foot Health Status Questionnaire (FHSQ) was developed and validated in Australia as a patient outcome measure of foot health, and has now �een used in nearly ten countries. �4 It was primarily intended to assess the outcome of surgical treatment, �ut it was validated across pathologies such as skin, nail, neurological, orthopedic and musculoskeletal disorders, among other conditions. �5-�9 The FHSQ has also �een used to evaluate the effectiveness of foot orthoses. 20 This instrument is divided into three sections. Section I evaluates foot health in four domains: Foot Pain, Foot Function, Footwear, and General Foot Health. Section II evaluates general health in four domains: General Health, Physical Activity, Social Capacity and Vigor. Sections I and II are composed of questions with response options presented in phrases and corresponding num�ers. Section III collects general demographic data of the patients. This structure makes the FHSQ a complete instrument, which has �een shown to have good responsiveness and relia�ility. �4,�5 The scores for the FHSQ may �e easily o�tained using software (The Foot Health Status Questionnaire, Version �.03). The answers to the questions are entered, and the software provides a score for each domain ranging from 0 to �00 (worst and �est conditions, respectively). �4,�5 The cross-cultural adaptation of this instrument will �e of great use �ecause until now, Brazil has lacked instruments for researchers focusing on foot health status. As Beaton et al. (2000) point out, translated versions of questionnaires may �e useful and may provide researchers with an instrument that allows international comparisons of scores. 2� The purpose of this study was to conduct a cross-cultural adaptation of the FHSQ into Brazilian-Portuguese and evaluate measurement properties of the Brazilian version of the FHSQ (FHSQ-Br).

MATERIALS AND METHODS
The author of the original instrument was contacted, and he provided the FHSQ software and his consent to the translation and evaluation of the instrument's measurement properties.
The cross-cultural adaptation was divided into four phases, according to Guillemin et al. (�993): (i) translation of the FHSQ into Brazilian-Portuguese, (ii) �ack-translation to evaluate the initial translation, (iii) committee analysis to identify cultural differences and language errors in the new instrument and (iv) a pre-test to assess cultural equivalence, considering the patients' impressions a�out the instrument. A written report including every suggestion and decision was made for each phase. 2� To assess the measurement properties of the FHSQ-Br, we conducted three phases: (i) test-retest relia�ility (inter and intra-o�servers), (ii) internal consistency and (iii) construct validity. Construct validity was evaluated �y comparing patient answers to the FHSQ-Br to clinical parameters commonly used in RA follow-up. The study was approved �y the Ethics Committee of the São Paulo University Medical School, and all patients signed an informed consent document.

Cross-cultural adaptation
i) Translation. The FHSQ was translated into Brazilian-Portuguese �y two �ilingual Brazilian translators with different profiles, creating versions T� and T2. Only one of the translators was aware of the concepts explored in the questionnaire. The different profiles of the translators ensured the use of unexpected words in the translation. 2�-25 Versions T� and T2 were discussed with �oth translators and with a team coordinator to synthesize the translations and to create Version �.
ii) Back-translation. Version � was �ack-translated into English �y two �ilingual native English-speaking teachers, neither of whom was aware of the concepts explored and neither had any medical �ackground. 2�-25 The two �acktranslations (BT� and BT2) were discussed with �oth �acktranslators and with the team coordinator to check if the translated version truly reflected the original instrument and to explore unexpected meanings of the translated items. 2�-25 Based on these adaptations, Version 2 was created and was ready to �e su�jected to committee analysis.
iii) Committee Analysis. The committee consisted of two rheumatologists, two physiotherapists, two translators, two �ack-translators and the team coordinator. The purpose of this phase was to achieve semantic, idiomatic, experiential and conceptual equivalence of the translated version. Semantic equivalence refers to the actual meaning of the words; idiomatic equivalence refers to attention to colloquialisms; experiential or cultural equivalence questions if the activities explored are common in that culture; and conceptual equivalence questions the cultural importance given to the situation �eing explored. 2�-23,26 Version 3 was created only after every item of the FHSQ had �een thoroughly discussed and agreed upon �y the committee.
iv) Pre-test. The pre-test group was composed of 40 RA patients from the Rheumatology Outpatient Division/ University of São Paulo. They were asked to answer the items on Version 3, commenting on the instructions, questions, answers and what they found confusing a�out the questionnaire. An answer option stating "not applica�le" was added to every question, to �e marked if the respondent did not understand the question. Any item of the instrument marked "non-applica�le" �y more than �0% of the patients was re-written and re-tested until all items were accepted �y more than 90% of the patients. Version 4 was then created and named FHSQ-Br.

Evaluation of the measurement properties
i) Test-retest Relia�ility. The study group was composed of sixty-five RA patients. None of them had participated in the pre-test phase. The patients were diagnosed according to the American College of Rheumatology revised criteria (�987). 27 Patients who had undergone foot surgery or therapeutic alterations or who were designated as functional class IV were excluded. 28 Each patient was asked to answer the FHSQ-Br three times in face-to-face interviews, always in the afternoon. Although the original version is a self-administered instrument, due to the educational level of the patients in this study, the authors chose to administer the instrument using an interview format. Interviewers were instructed to simply read the questions to the patients and tick their answers, providing no explanations to the questions to avoid interference. Patients who did not have a clear understanding of the questions were excluded at that moment.
The first interview was conducted �y interviewer A, the second interview was conducted 45 minutes after the first �y interviewer B and the third and final interview was conducted �5 days later, again �y interviewer A. During these �5 days, the patients did not have any modifications to their treatment and were asked to maintain their regular activities. Such instructions were intended to ensure the sta�ility of the clinical status to test the instrument relia�ility. 24 ii) Internal Consistency. The scores o�tained from the interviews were used to determine correlations �etween different questions on the same domain.
iii) Construct Validity. The Health Assessment Questionnaire (HAQ), Numeric Rating Scale (NRS) and foot X-rays were correlated to the FHSQ-Br scores, which ranged from 0 to �00 (worst and �est conditions, respectively), to assess the construct validity. 24 The HAQ is a disease-specific instrument used worldwide for the evaluation of RA. 25 The scores of the HAQ range from 0 to 3, from the �est to worst condition, respectively. 25 The NRS for foot pain was used to score patient "foot pain in the last few days." These scores range from 0 to �0, in which 0 is equivalent to no pain at all and �0 is the worst pain imagina�le. Foot X-rays were scored using the Sharp/van der Heijde method, which assesses joint surface erosion (0 to �20) and joint space narrowing (0 to 48) of 6 joints, resulting in a score that ranges from 0 to �68 for �oth feet. 29 Statistical Analysis. The variation coefficients of the scores from the three interviews were used to test the inter-and intra-o�server relia�ility. The α-Cron�ach coefficient was applied to test the internal consistency of the domains. Spearman Correlation was used to determine the construct validity.

RESULTS
Cross-cultural adaptation. Over the translation, �acktranslation, and committee analysis phases, minor adaptations to wording and sentence structure were required in sections I and II for �etter understanding. One of the two items that needed to �e adjusted was Question �2, which stated "I am limited in the num�er of shoes I can wear." In Portuguese the word "num�er" could �e misleading and suggest shoe size, so it was changed to "tipos," meaning "kinds." The other item was Question �5�, which gives examples of moderate activities, one of which is "playing golf." Golf is not a common sport in Brazil, especially in our study group, so considering there were other suita�le examples in the question, it was decided that the elimination of the words "playing golf" would cause no interference.
During the pre-test phase, three items of the instrument were adapted. The first one was in Question �5d; the words "steep hill," when translated into Portuguese, were inadequate for an ur�an scenario and were replaced �y "ladeira." The second item, on Question 25, asked if the patient was "a pensioner or health care cardholder." It was replaced with "Do you have pu�lic health insurance?" ("Você paga INSS ou previdência social?"), which is more comprehensi�le to the Brazilian population. Finally, in Question 29, "…completed a trade certificate or any other educational qualification since leaving school" was replaced with "have you attended any continuing education program after leaving school?" ("Você fez algum aperfeiçoamento ou especialização desde que saiu da escola?"). All questions were considered applica�le after these corrections.

Evaluation of Measurement Properties.
A field test was conducted on a study group of 65 RA patients selected as descri�ed earlier. The FHSQ-Br, which includes demographic and clinical data, was collected in addition to the HAQ, NRS and foot X-rays. This information is presented in Ta�les � through 3.
Test-retest reliability. Out of the 65 patients from the study group who participated in the first interview, 6� participated in the second interview to test inter-o�server relia�ility and 55 in the third interview to test intra-o�server relia�ility. The scores o�tained from the three interviews were analyzed and showed low variation coefficients, representing the homogeneity of the scores from all interviews. The mean FHSQ-Br score of the three interviews and inferior and superior limits of the confidence interval are detailed in Figure �. Construct validity. The HAQ was answered �y 40 patients and the NRS was answered �y all 65 patients; foot X-rays were o�tained from 50 patients in the study group.
The mean FHSQ-Br scores showed a significant correlation to the HAQ scores (-55.6%; p < 0.00�). Individually, five out of eight domains showed significant correlations, with p-values ranging from 0.00� to 0.046. The mean FHSQ-Br   Table 2 -Clinical data characterizing the disease and associated diseases of the study group (n = 65)

DISCUSSION
Although patients with RA complain of foot pain and of foot pro�lems affecting their performance of activities of daily life, physicians sometimes overlook or neglect the feet in routine examinations. 30,3� Disease-related foot-health instruments may help health professionals in measuring the impact on feet related to RA. The Foot Health Status Questionnaire, one of these instruments, was developed and validated in Australia as a patient-outcome measure of foot health. The cross-cultural adaptation of the FHSQ into Brazilian-Portuguese ena�les clinicians to compare results and outcomes of their treatments among different populations with different profiles and lifestyles.
The FHSQ was translated into Brazilian-Portuguese with no difficulty, and it is fully adapted to the culture; very few items needed to �e analyzed in detail through all of the development phases. Furthermore, during the pre-test phase, only three questions had to �e reformulated, and after the necessary changes, every item of the questionnaire was approved �y more than 90% of the patients, concluding the instrument's cultural adaptation.
The FHSQ-Br is a self-administered instrument; �4 however, in Brazil, we chose to administer it with an interviewer, considering the social-economic and educational level of the studied population.
Our study group reflects the RA segment of the Brazilian population �ecause the demographic and clinical data collected are similar to what has �een previously reported for the Brazilian population regarding age, sex distri�ution, disease duration, HAQ scores and percentage of illiterate or pre-school level patients. 32 The scores o�tained from this study group reflect the difficulty that patients have in completing their daily activities as a consequence of their foot disease. Attri�utes like foot pain and foot function reflect their disa�ilities.
The inter-and intra-o�server relia�ility values were in line with those found �y the authors of the original instrument, which varied from 0.740 to 0.9�5. �4 The FHSQ-Br internal consistency analysis showed fair results. General health (0.468) and Physical Activity (0.855) showed the lowest and highest results, respectively. The internal consistency of the original instrument varied from 0.85� to 0.884, depending on the domain �eing analyzed. �4 When compared to other parameters intended to measure disease impact, 25,29,33 the FHSQ-Br showed significant correlations, indicating that the instrument is actually measuring what it is intended to measure. 24 For instance, the HAQ and the NRS scores were significantly correlated with five out of the eight domains of the instrument. The FHSQ-Br is a complete and valid instrument, easy to understand, accepta�le to patients, and a�le to �e administered in �0 to �5 minutes.
Even though RA is a chronic disease, it presents clinical variations with acute periods. This clinical characteristic associated with patient physical limitations constituted the main limitation of our study. Further studies should �e conducted in Brazil with other foot conditions for comparison with our results and proceed with eventual adaptations of the FHSQ for these other conditions. This study shows that the FHSQ-Br will �e of great value in assessments of RA patients in clinical settings and in research.