International Journal of Rheumatic Diseases 2015
ORIGINAL ARTICLE
Psychometric assessment of the Japanese version of the Revised Fibromyalgia Impact Questionnaire: reliability and validity Tatsuya ISOMURA,1,2 Ikuro NAKAMURA,3 Mika KAWAGUCHI,2 Emiko SATO,2 Kyoko INUZUKA,2 Kenichi OSADA,4 Kenya NISHIOKA5 and Kazuhiro HAYAKAWA6 1
Clinical Study Support Inc., Nagoya, 2Institute of Medical Science, Tokyo Medical University, Tokyo, 3Research Center for Rheumatic Disease, Toranomonkai Medical Corporation, Tokyo, 4Department of Neuropsychiatry, St. Marianna University School of Medicine, Kawasaki, 5Department of Neurology, Juntendo University School of Medicine, Tokyo, and 6Pfizer Japan Inc., Tokyo, Japan
Abstract Aim: To assess the psychometric properties of the Japanese version of the Revised Fibromyalgia Impact Questionnaire (JFIQR) in fibromyalgia (FM) patients. Method: The reliability and validity of the JFIQR were assessed using online data collected from Japanese FM patients. Reliability was evaluated based on test-retest reliability results and internal consistency; validity was evaluated on the basis of concurrent and known-group validity. Results: A total of 105 patients completed the online questionnaire. Intra-class correlation coefficients for testretest were 0.91 for the JFIQR total score with a range of 0.84–0.90 in three domains: function, overall impact and symptoms. Internal consistency results indicated a Cronbach’s alpha of 0.90 for the total score with a range of 0.83 and 0.85 for the domains. Concurrent validity results showed that the total score was correlated to all external criteria (Japanese version of the Fibromyalgia Impact Questionnaire, Fibromyalgia Activity Scale-31, Medical Outcomes Study 36-item Short-Form health survey) from a moderate to strong degree with most indicating a strong correlation. Results of known-group validity showed that the JFIQR total score is capable of discriminating between FM and the other groups, such as rheumatic arthritis and no chronic pain (P < 0.0001 for all pairwise comparisons). Conclusion: The current psychometric assessment of the JFIQR demonstrated that it is a reliable and valid questionnaire in Japanese patients with FM. Usefulness of the JFIQR in clinical studies and medical practice for Japanese-speaking populations is expected. Key words: fibromyalgia, Japanese version, psychometric assessment, reliability, revised fibromyalgia impact questionnaire, validity.
INTRODUCTION Fibromyalgia (FM) is characterized by chronic widespread body pain as a core symptom and includes variCorrespondence: Tatsuya Isomura, M.Sc., Clinical Study Support, Inc., 2-22-8 Chikusa, Chikusa-ku, Nagoya 464-0858, Japan. Email: [email protected]
ous other symptoms such as fatigue, sleep disturbance and depression.1 The worldwide prevalence of FM in the general population ranges between 0.5% and 5%;2,3 similarly, the prevalence in Japan is projected to be approximately 2%, or 2 million individuals.4–7 Due to the negative impact of FM symptoms on one’s function and quality of life (QoL),8 a comprehensive assessment is vital in the treatment of FM.
© 2015 Asia Pacific League of Associations for Rheumatology and Wiley Publishing Asia Pty Ltd
T. Isomura et al.
The Revised Fibromyalgia Impact Questionnaire (FIQR©) is a self-reported questionnaire enabling the comprehensive assessment of FM in multiple aspects.9 The questionnaire consists of 21 items measuring three domains: “function” (nine items); “overall impact” of FM on functioning and overall impact symptom severity (two items); and “symptoms” (10 items). Each item is rated on a numeric rating scale with a score of 0–10. After scores for each domain are summed, the summed score for the function (range of 0–90), overall impact (range of 0–20), and symptoms (range of 0–100) domains are normalized by dividing by 3, 1 and 2, respectively. The total FIQR score is the simple sum of the three normalized domain scores with a range of 0–100: a lower score indicates better QoL. The Fibromyalgia Impact Questionnaire (FIQ),10 predecessor of the FIQR, has been translated and utilized for research worldwide, and the Japanese version of the FIQ (JFIQ)11 is not an exception. However, the FIQ has some limitations in perplexing scoring and lack of symptoms commonly experienced in FM. Additionally, the FIQ was originally intended for women with a certain standard of living in reasonably affluent countries. To address such limitations, the FIQR was developed with the essential properties of the original questionnaire retained. The FIQR, originally developed in English, has good psychometric properties.9 Its score was found to be comparable to the original questionnaire. Moreover, quick administration and simple scoring are beneficial for use in a clinical setting. The FIQR has been translated and used worldwide in such countries as Brazil, Morocco, Spain and Turkey.12–15 To utilize the FIQR for Japanese, the English version was translated and linguistically validated as JFIQR,16 but psychometric validation has not yet been conducted. The purpose of this study is to assess the psychometric properties of the JFIQR in FM patients.
MATERIALS AND METHODS Patients who were aged 20 years or older diagnosed with FM and were capable of completing questionnaires by themselves were included in the analysis. Participants were recruited online utilizing an online research company, Rakuten Research (Tokyo, Japan), where over 2.2 million individuals nation-wide voluntarily registered. An invitation was electronically sent out to eligible participants who previously reported that they had
2
experienced or were experiencing FM regardless of their present treatment status. For those interested, a screening questionnaire was given to verify ones with FM: participants who had ever been diagnosed as having FM and who had widespread pain lasting for 3 months or longer. Participants responded to the first questionnaire administration after the screening. Those who completed the first questionnaire administration were then asked to respond to the second questionnaire administration after an interval of 1 week. Questionnaire administrations were implemented in September 2013 after ethics approval was granted by the ethics committee at Kitamachi Clinic (Tokyo, Japan). All the participants consented to the questionnaire administration and were compensated. The first questionnaire included questions on demographic and clinical characteristics, JFIQR, JFIQ, Fibromyalgia Activity Scale-31 (FAS-31),17 and the Medical Outcomes Study (MOS) 36-Item Short-Form Health Survey version 2 (SF-36v2â).18,19 The second questionnaire contained questions on whether the FM conditions had changed since the first questionnaire administration, JFIQR and FAS-31. The JFIQ is a self-administered questionnaire containing 10 items assessing a FM patient’s physical function (10 sub-items), feeling good, work missed, ability to do job, pain, fatigue, morning tiredness, stiffness, anxiety and depression during the previous week.11 The score of each item is standardized from 0 to 10 with a total score of ranging from 0 to 100. A higher score indicates a worse condition. A Japanese version of the JFIQ, demonstrated to be reliable and valid as the original FIQ, was used in the present study.11 The FAS-31 is a self-administered questionnaire based on the 2010 American College of Rheumatology (ACR) criteria.17 The questionnaire consists of two domains: widespread pain index (WPI) (one item) and symptom severity (SS) (two items). The score of the WPI is the amount of pain in 19 areas that a FM patient experienced during the previous week. Areas were explained with names of 19 body locations and a diagram indicating the locations: the jaw (left, right), shoulder girdle (left, right), upper arm (left, right), lower arm (left, right), chest, abdomen, upper leg (left, right), lower leg (left, right), neck, back (upper, lower) and hip (left, right). Two items of the SS are current symptoms relating to FM (three sub-items) and 41 symptoms that the participant experienced during the previous week (one sub-item). The former part of each sub-item score ranged from 0 to 3 based on severity. The latter part of subitem scores are based on a total number of symptoms:
International Journal of Rheumatic Diseases 2015
Psychometric assessment of the JFIQR
if 1–5 symptoms, a score of 1; if 6–20 symptoms, a score of 2; and if 21–41 symptoms, a score of 3. A total score is obtained by adding the scores of WPI and SS with a range of 1–31. The SF-36v2 is a questionnaire containing 36 items for the assessment of generic health-related QoL.18,19 The questionnaire is categorized into eight domains: physical function, role limitation-physical, vitality, general health perception, bodily pain, social function, role limitations-emotional and mental health. The score of each domain ranges from 0 to 100, with a higher score indicating a better QoL. A Japanese version of the SF36v2, which has demonstrated good reliability and validity, was utilized in the study.18,19
Statistical analysis Descriptive analyses on demographic and clinical characteristics were performed. Item analyses on the JFIQR domain score and total score were conducted to assess missing data and floor or ceiling effects (> 60%). The scoring was carried out in accordance with the scoring of the English version of FIQR.9 To assess the psychometric properties of the JFIQR, the reliability and validity were evaluated. The reliability of the JFIQR was evaluated by assessing test-retest reliability and internal consistency. For the test-retest reliability, the extent to which responses to the JFIQR between two time points agree among participants with FM whose symptoms remained stable was examined using the intraclass correlation coefficient (ICC). A coefficient of ≥ 0.7 was considered sufficient to determine test-retest reliability.20 For internal consistency, the extent to which a set of subscales as a group was interrelated was computed using Cronbach’s alpha. A Cronbach’s alpha of 0.7 or greater was considered to demonstrate good internal subscale consistency.20 Validity of the JFIQR was evaluated using concurrent validity and known-group validity. For concurrent validity, the degree of correlation with the external criteria (JFIQ, FAS-31, SF-36v2) was assessed using Pearson’s correlation coefficient. Scales measuring similar concepts were expected to be moderately to highly correlated, whereas those measuring different concepts were expected to be poorly correlated. The correlation coefficient was interpreted as follows: 0.1 is considered weak, 0.3 moderate and 0.5 strong.21 Known-group validity was examined to test the discriminating ability of the JFIQR across varying disease groups. For the assessment of known-group validity, patients with rheumatic arthritis (RA) and individuals
International Journal of Rheumatic Diseases 2015
with no chronic pain (CP) were recruited from individuals voluntarily registered with Rakuten Research. RA patients were those who previously reported they had experienced or were experiencing RA, whereas individuals with no CP were those who previously reported they had experienced no particular diseases. A screening questionnaire was given for verification purposes: participants with RA were those who had a history of diagnosis with RA, continuous swelling in three or more joints for a while, and morning stiffness. Participants with no CP were those who had no history of diagnosis with particular diseases and who had no pains lasting for 3 months or longer. The mean JFIQR scores were compared across three groups (FM, RA and no CP). If there were any significant differences as a result of the analysis of variance (ANOVA), pairwise multiple comparisons, Tukey’s honestly significant difference (HSD) test, were performed to clarify where the statistically significant difference in least-squares means were. Knowngroup validity was assessed using the JFIQ total score in the same manner. All the statistical tests were two-sided with a significance level of 5%. All the analyses were performed using SAS software version 9.3 (SAS Institute, Inc., Cary, NC, USA).
RESULTS A total of 105 participants with FM took part in the first questionnaire administration. Demographic and clinical characteristics of the participants are summarized in Table 1. Mean age of participants with FM was 46.7 years, and 72.4% of those were female. Mean duration of experiencing FM was 5.2 years. The participants with FM experienced pain during the previous week in 10.5 areas in average. Results of item analysis are shown in Table 2. No data were missing and no ceiling or flooring effect was found in either the JFIQR domain or total scores.
Table 1 Demographic and clinical characteristics of participants with FM (n = 105) Characteristics Age, years, mean (SD) Sex, female, n (%) Duration of FM symptoms, years, mean (SD) Number of pain areas, mean (SD)
46.7 (9.4) 76 (72.4) 5.2 (3.8) 10.5 (4.6)
FM, fibromyalgia; SD, standard deviation. Number of pain area was asked in the Fibromyalgia Activity Scale-31 (FAS-31).
3
T. Isomura et al.
Reliability For the analysis of the test-retest reliability, 65 participants whose FM symptoms between the first and second questionnaire administrations remained stable on a self-reporting basis were selected. The ICC for the JFIQR total score was 0.91, with the ICCs of 0.84 for function, 0.85 for overall impact and 0.90 for symptoms domain. The internal consistency was evaluated using the data collected from participants of the first questionnaire administrations. The Cronbach’s alpha for the total score was 0.90. The values within the domains were 0.83 for function, 0.85 for overall impact and 0.85 for symptoms.
Validity For the assessment of the concurrent validity, the correlation coefficients between the JFIQR and external criteria (JFIQ, FAS-31 and SF-36v2) were calculated (Table 3). The JFIQR total score was correlated to all the external criteria, including the item or domain and total scores from a moderate to strong degree. Most of those indicated a strong correlation of 0.5 or above (or 0.5 or below for the SF-36v2). Of those, the strongest correlation was found between the JFIQR total score and JFIQ total score (0.89). The JFIQR domains were also correlated to the external criteria from a moderate to strong degree with few exceptions (i.e., correlations between all the domains and “feel good” in the JFIQ). To examine the known-group validity, responses from 69 participants with RA and 45 participants with no CP were included in the analysis. The JFIQR total score between the FM, RA and no CP groups was highest in the FM group, followed by RA and no CP groups as shown in Table 4. A statistically significant difference was in the JFIQR total score across the three groups (FM, RA and no CP) (F[2, 216] = 107.36, P < 0.0001). Furthermore, the difference in the JFIQR total score between any pairs of the three groups was statistically
significant (Tukey’s HSD test P < 0.0001 for all three comparisons). Similar results were found in the known-group validity using the JFIQ total score. The highest JFIQ total score was observed in the FM group, followed by the RA and no CP groups (Table 4). The difference in the total score between the three groups was statistically significant (F[2, 216] = 145.27, P < 0.0001). A statistically significant difference in the JFIQ total score was found in all pairs of the three groups (Tukey’s HSD test P < 0.0001 for all three comparisons).
DISCUSSION The JFIQR was linguistically validated on earlier research. The present study assessed psychometric properties of the JFIQR using data collected from Japanese patients with FM. Results of the assessment overall suggested the JFIQR has good reliability and validity. Both test-retest reliability and internal consistency demonstrated a good reliability of the JFIQR; both concurrent validity and known-group validity indicated a good validity. In test-retest reliability, the ICCs for the JFIQR domains (function, overall impact, symptom) satisfied the sufficient level of 0.7 with the highest of symptoms (0.90); the lowest of function (0.84). The ICC for the total score (0.91) also satisfied the level. Given that the ICC required for clinical practice is 0.9 or higher,20 the ICC for the present study suggests a sufficient level for meeting this requirement. Cronbach’s alpha coefficient evaluating internal consistency indicated an acceptable level of internal subscale consistency. Cronbach’s alpha of 0.9 is considered as the appropriate level for individual patient assessment.20,21 The value suggests a good internal consistency of meeting the level for use in individual patient assessments. The correlation coefficients for the concurrent validity assessment showed a moderate to strong agreement
Table 2 Descriptive statistics and reliability for the JFIQR in FM patients (n = 105) JFIQR item
Mean
SD
Function Overall impact Symptoms Total score
15.1 12.6 31.8 59.5
8.6 6.3 10.7 23.9
Median
Min—Max
Floor effect (%)
Ceiling effect (%)
15.3 14.0 33.5 61.0
0.0–30.0 0.0–20.0 6.5–49.0 7.5–98.0
1.9 1.9 0.0 0.0
1.0 24.8 0.0 0.0
JFIQR, Japanese version of the Revised Fibromyalgia Impact Questionnaire; FM, fibromyalgia; SD, standard deviation; Min, minimum; Max, maximum.
4
International Journal of Rheumatic Diseases 2015
Psychometric assessment of the JFIQR
Table 3 Correlations† between the JFIQR and external criteria (a) JFIQ, (b) FAS-31 and SF-36v2 (n = 105) JFIQR
(a) Function Overall impact Symptom Total score
JFIQ Physical function
Feel good
Work missed
Ability to do work
Pain
Fatigue
Morning tiredness
Stiffness
Anxiety
Depression
Total score
0.74 0.64 0.61 0.71
0.14 0.20 0.21 0.20
0.62 0.58 0.56 0.63
0.71 0.80 0.74 0.80
0.73 0.79 0.85 0.85
0.64 0.69 0.80 0.77
0.52 0.59 0.70 0.65
0.50 0.46 0.56 0.55
0.44 0.66 0.72 0.65
0.53 0.72 0.77 0.72
0.77 0.84 0.89 0.89
JFIQR
(b) Function Overall impact Symptom Total score
FAS-31
SF-36v2
WPI
SS
Total score
PF
RP
BP
GH
VT
SF
RE
MH
0.44 0.32 0.37 0.41
0.67 0.67 0.78 0.76
0.59 0.51 0.58 0.61
0.81 0.60 0.65 0.74
0.60 0.67 0.60 0.66
0.68 0.74 0.71 0.76
0.42 0.44 0.49 0.49
0.55 0.61 0.62 0.63
0.59 0.67 0.59 0.66
0.53 0.58 0.59 0.61
0.39 0.56 0.60 0.56
JFIQR, Japanese version of the Revised Fibromyalgia Impact Questionnaire; JFIQ, Japanese version of the Fibromyalgia Impact Questionnaire; FAS31, Fibromyalgia Activity Scale-31; WPI, widespread pain index; SS, symptom severity; SF-36v2, Medical Outcomes Study 36-Item Short-Form Health Survey version 2; PF, physical function; RP, role limitations-physical; VT, vitality; GH, general health perception; BP, bodily pain; SF, social function; RE, role limitations-emotional; MH, mental health; †Pearson’s correlation. All the results are the P-value of
ORIGINAL ARTICLE
Psychometric assessment of the Japanese version of the Revised Fibromyalgia Impact Questionnaire: reliability and validity Tatsuya ISOMURA,1,2 Ikuro NAKAMURA,3 Mika KAWAGUCHI,2 Emiko SATO,2 Kyoko INUZUKA,2 Kenichi OSADA,4 Kenya NISHIOKA5 and Kazuhiro HAYAKAWA6 1
Clinical Study Support Inc., Nagoya, 2Institute of Medical Science, Tokyo Medical University, Tokyo, 3Research Center for Rheumatic Disease, Toranomonkai Medical Corporation, Tokyo, 4Department of Neuropsychiatry, St. Marianna University School of Medicine, Kawasaki, 5Department of Neurology, Juntendo University School of Medicine, Tokyo, and 6Pfizer Japan Inc., Tokyo, Japan
Abstract Aim: To assess the psychometric properties of the Japanese version of the Revised Fibromyalgia Impact Questionnaire (JFIQR) in fibromyalgia (FM) patients. Method: The reliability and validity of the JFIQR were assessed using online data collected from Japanese FM patients. Reliability was evaluated based on test-retest reliability results and internal consistency; validity was evaluated on the basis of concurrent and known-group validity. Results: A total of 105 patients completed the online questionnaire. Intra-class correlation coefficients for testretest were 0.91 for the JFIQR total score with a range of 0.84–0.90 in three domains: function, overall impact and symptoms. Internal consistency results indicated a Cronbach’s alpha of 0.90 for the total score with a range of 0.83 and 0.85 for the domains. Concurrent validity results showed that the total score was correlated to all external criteria (Japanese version of the Fibromyalgia Impact Questionnaire, Fibromyalgia Activity Scale-31, Medical Outcomes Study 36-item Short-Form health survey) from a moderate to strong degree with most indicating a strong correlation. Results of known-group validity showed that the JFIQR total score is capable of discriminating between FM and the other groups, such as rheumatic arthritis and no chronic pain (P < 0.0001 for all pairwise comparisons). Conclusion: The current psychometric assessment of the JFIQR demonstrated that it is a reliable and valid questionnaire in Japanese patients with FM. Usefulness of the JFIQR in clinical studies and medical practice for Japanese-speaking populations is expected. Key words: fibromyalgia, Japanese version, psychometric assessment, reliability, revised fibromyalgia impact questionnaire, validity.
INTRODUCTION Fibromyalgia (FM) is characterized by chronic widespread body pain as a core symptom and includes variCorrespondence: Tatsuya Isomura, M.Sc., Clinical Study Support, Inc., 2-22-8 Chikusa, Chikusa-ku, Nagoya 464-0858, Japan. Email: [email protected]
ous other symptoms such as fatigue, sleep disturbance and depression.1 The worldwide prevalence of FM in the general population ranges between 0.5% and 5%;2,3 similarly, the prevalence in Japan is projected to be approximately 2%, or 2 million individuals.4–7 Due to the negative impact of FM symptoms on one’s function and quality of life (QoL),8 a comprehensive assessment is vital in the treatment of FM.
© 2015 Asia Pacific League of Associations for Rheumatology and Wiley Publishing Asia Pty Ltd
T. Isomura et al.
The Revised Fibromyalgia Impact Questionnaire (FIQR©) is a self-reported questionnaire enabling the comprehensive assessment of FM in multiple aspects.9 The questionnaire consists of 21 items measuring three domains: “function” (nine items); “overall impact” of FM on functioning and overall impact symptom severity (two items); and “symptoms” (10 items). Each item is rated on a numeric rating scale with a score of 0–10. After scores for each domain are summed, the summed score for the function (range of 0–90), overall impact (range of 0–20), and symptoms (range of 0–100) domains are normalized by dividing by 3, 1 and 2, respectively. The total FIQR score is the simple sum of the three normalized domain scores with a range of 0–100: a lower score indicates better QoL. The Fibromyalgia Impact Questionnaire (FIQ),10 predecessor of the FIQR, has been translated and utilized for research worldwide, and the Japanese version of the FIQ (JFIQ)11 is not an exception. However, the FIQ has some limitations in perplexing scoring and lack of symptoms commonly experienced in FM. Additionally, the FIQ was originally intended for women with a certain standard of living in reasonably affluent countries. To address such limitations, the FIQR was developed with the essential properties of the original questionnaire retained. The FIQR, originally developed in English, has good psychometric properties.9 Its score was found to be comparable to the original questionnaire. Moreover, quick administration and simple scoring are beneficial for use in a clinical setting. The FIQR has been translated and used worldwide in such countries as Brazil, Morocco, Spain and Turkey.12–15 To utilize the FIQR for Japanese, the English version was translated and linguistically validated as JFIQR,16 but psychometric validation has not yet been conducted. The purpose of this study is to assess the psychometric properties of the JFIQR in FM patients.
MATERIALS AND METHODS Patients who were aged 20 years or older diagnosed with FM and were capable of completing questionnaires by themselves were included in the analysis. Participants were recruited online utilizing an online research company, Rakuten Research (Tokyo, Japan), where over 2.2 million individuals nation-wide voluntarily registered. An invitation was electronically sent out to eligible participants who previously reported that they had
2
experienced or were experiencing FM regardless of their present treatment status. For those interested, a screening questionnaire was given to verify ones with FM: participants who had ever been diagnosed as having FM and who had widespread pain lasting for 3 months or longer. Participants responded to the first questionnaire administration after the screening. Those who completed the first questionnaire administration were then asked to respond to the second questionnaire administration after an interval of 1 week. Questionnaire administrations were implemented in September 2013 after ethics approval was granted by the ethics committee at Kitamachi Clinic (Tokyo, Japan). All the participants consented to the questionnaire administration and were compensated. The first questionnaire included questions on demographic and clinical characteristics, JFIQR, JFIQ, Fibromyalgia Activity Scale-31 (FAS-31),17 and the Medical Outcomes Study (MOS) 36-Item Short-Form Health Survey version 2 (SF-36v2â).18,19 The second questionnaire contained questions on whether the FM conditions had changed since the first questionnaire administration, JFIQR and FAS-31. The JFIQ is a self-administered questionnaire containing 10 items assessing a FM patient’s physical function (10 sub-items), feeling good, work missed, ability to do job, pain, fatigue, morning tiredness, stiffness, anxiety and depression during the previous week.11 The score of each item is standardized from 0 to 10 with a total score of ranging from 0 to 100. A higher score indicates a worse condition. A Japanese version of the JFIQ, demonstrated to be reliable and valid as the original FIQ, was used in the present study.11 The FAS-31 is a self-administered questionnaire based on the 2010 American College of Rheumatology (ACR) criteria.17 The questionnaire consists of two domains: widespread pain index (WPI) (one item) and symptom severity (SS) (two items). The score of the WPI is the amount of pain in 19 areas that a FM patient experienced during the previous week. Areas were explained with names of 19 body locations and a diagram indicating the locations: the jaw (left, right), shoulder girdle (left, right), upper arm (left, right), lower arm (left, right), chest, abdomen, upper leg (left, right), lower leg (left, right), neck, back (upper, lower) and hip (left, right). Two items of the SS are current symptoms relating to FM (three sub-items) and 41 symptoms that the participant experienced during the previous week (one sub-item). The former part of each sub-item score ranged from 0 to 3 based on severity. The latter part of subitem scores are based on a total number of symptoms:
International Journal of Rheumatic Diseases 2015
Psychometric assessment of the JFIQR
if 1–5 symptoms, a score of 1; if 6–20 symptoms, a score of 2; and if 21–41 symptoms, a score of 3. A total score is obtained by adding the scores of WPI and SS with a range of 1–31. The SF-36v2 is a questionnaire containing 36 items for the assessment of generic health-related QoL.18,19 The questionnaire is categorized into eight domains: physical function, role limitation-physical, vitality, general health perception, bodily pain, social function, role limitations-emotional and mental health. The score of each domain ranges from 0 to 100, with a higher score indicating a better QoL. A Japanese version of the SF36v2, which has demonstrated good reliability and validity, was utilized in the study.18,19
Statistical analysis Descriptive analyses on demographic and clinical characteristics were performed. Item analyses on the JFIQR domain score and total score were conducted to assess missing data and floor or ceiling effects (> 60%). The scoring was carried out in accordance with the scoring of the English version of FIQR.9 To assess the psychometric properties of the JFIQR, the reliability and validity were evaluated. The reliability of the JFIQR was evaluated by assessing test-retest reliability and internal consistency. For the test-retest reliability, the extent to which responses to the JFIQR between two time points agree among participants with FM whose symptoms remained stable was examined using the intraclass correlation coefficient (ICC). A coefficient of ≥ 0.7 was considered sufficient to determine test-retest reliability.20 For internal consistency, the extent to which a set of subscales as a group was interrelated was computed using Cronbach’s alpha. A Cronbach’s alpha of 0.7 or greater was considered to demonstrate good internal subscale consistency.20 Validity of the JFIQR was evaluated using concurrent validity and known-group validity. For concurrent validity, the degree of correlation with the external criteria (JFIQ, FAS-31, SF-36v2) was assessed using Pearson’s correlation coefficient. Scales measuring similar concepts were expected to be moderately to highly correlated, whereas those measuring different concepts were expected to be poorly correlated. The correlation coefficient was interpreted as follows: 0.1 is considered weak, 0.3 moderate and 0.5 strong.21 Known-group validity was examined to test the discriminating ability of the JFIQR across varying disease groups. For the assessment of known-group validity, patients with rheumatic arthritis (RA) and individuals
International Journal of Rheumatic Diseases 2015
with no chronic pain (CP) were recruited from individuals voluntarily registered with Rakuten Research. RA patients were those who previously reported they had experienced or were experiencing RA, whereas individuals with no CP were those who previously reported they had experienced no particular diseases. A screening questionnaire was given for verification purposes: participants with RA were those who had a history of diagnosis with RA, continuous swelling in three or more joints for a while, and morning stiffness. Participants with no CP were those who had no history of diagnosis with particular diseases and who had no pains lasting for 3 months or longer. The mean JFIQR scores were compared across three groups (FM, RA and no CP). If there were any significant differences as a result of the analysis of variance (ANOVA), pairwise multiple comparisons, Tukey’s honestly significant difference (HSD) test, were performed to clarify where the statistically significant difference in least-squares means were. Knowngroup validity was assessed using the JFIQ total score in the same manner. All the statistical tests were two-sided with a significance level of 5%. All the analyses were performed using SAS software version 9.3 (SAS Institute, Inc., Cary, NC, USA).
RESULTS A total of 105 participants with FM took part in the first questionnaire administration. Demographic and clinical characteristics of the participants are summarized in Table 1. Mean age of participants with FM was 46.7 years, and 72.4% of those were female. Mean duration of experiencing FM was 5.2 years. The participants with FM experienced pain during the previous week in 10.5 areas in average. Results of item analysis are shown in Table 2. No data were missing and no ceiling or flooring effect was found in either the JFIQR domain or total scores.
Table 1 Demographic and clinical characteristics of participants with FM (n = 105) Characteristics Age, years, mean (SD) Sex, female, n (%) Duration of FM symptoms, years, mean (SD) Number of pain areas, mean (SD)
46.7 (9.4) 76 (72.4) 5.2 (3.8) 10.5 (4.6)
FM, fibromyalgia; SD, standard deviation. Number of pain area was asked in the Fibromyalgia Activity Scale-31 (FAS-31).
3
T. Isomura et al.
Reliability For the analysis of the test-retest reliability, 65 participants whose FM symptoms between the first and second questionnaire administrations remained stable on a self-reporting basis were selected. The ICC for the JFIQR total score was 0.91, with the ICCs of 0.84 for function, 0.85 for overall impact and 0.90 for symptoms domain. The internal consistency was evaluated using the data collected from participants of the first questionnaire administrations. The Cronbach’s alpha for the total score was 0.90. The values within the domains were 0.83 for function, 0.85 for overall impact and 0.85 for symptoms.
Validity For the assessment of the concurrent validity, the correlation coefficients between the JFIQR and external criteria (JFIQ, FAS-31 and SF-36v2) were calculated (Table 3). The JFIQR total score was correlated to all the external criteria, including the item or domain and total scores from a moderate to strong degree. Most of those indicated a strong correlation of 0.5 or above (or 0.5 or below for the SF-36v2). Of those, the strongest correlation was found between the JFIQR total score and JFIQ total score (0.89). The JFIQR domains were also correlated to the external criteria from a moderate to strong degree with few exceptions (i.e., correlations between all the domains and “feel good” in the JFIQ). To examine the known-group validity, responses from 69 participants with RA and 45 participants with no CP were included in the analysis. The JFIQR total score between the FM, RA and no CP groups was highest in the FM group, followed by RA and no CP groups as shown in Table 4. A statistically significant difference was in the JFIQR total score across the three groups (FM, RA and no CP) (F[2, 216] = 107.36, P < 0.0001). Furthermore, the difference in the JFIQR total score between any pairs of the three groups was statistically
significant (Tukey’s HSD test P < 0.0001 for all three comparisons). Similar results were found in the known-group validity using the JFIQ total score. The highest JFIQ total score was observed in the FM group, followed by the RA and no CP groups (Table 4). The difference in the total score between the three groups was statistically significant (F[2, 216] = 145.27, P < 0.0001). A statistically significant difference in the JFIQ total score was found in all pairs of the three groups (Tukey’s HSD test P < 0.0001 for all three comparisons).
DISCUSSION The JFIQR was linguistically validated on earlier research. The present study assessed psychometric properties of the JFIQR using data collected from Japanese patients with FM. Results of the assessment overall suggested the JFIQR has good reliability and validity. Both test-retest reliability and internal consistency demonstrated a good reliability of the JFIQR; both concurrent validity and known-group validity indicated a good validity. In test-retest reliability, the ICCs for the JFIQR domains (function, overall impact, symptom) satisfied the sufficient level of 0.7 with the highest of symptoms (0.90); the lowest of function (0.84). The ICC for the total score (0.91) also satisfied the level. Given that the ICC required for clinical practice is 0.9 or higher,20 the ICC for the present study suggests a sufficient level for meeting this requirement. Cronbach’s alpha coefficient evaluating internal consistency indicated an acceptable level of internal subscale consistency. Cronbach’s alpha of 0.9 is considered as the appropriate level for individual patient assessment.20,21 The value suggests a good internal consistency of meeting the level for use in individual patient assessments. The correlation coefficients for the concurrent validity assessment showed a moderate to strong agreement
Table 2 Descriptive statistics and reliability for the JFIQR in FM patients (n = 105) JFIQR item
Mean
SD
Function Overall impact Symptoms Total score
15.1 12.6 31.8 59.5
8.6 6.3 10.7 23.9
Median
Min—Max
Floor effect (%)
Ceiling effect (%)
15.3 14.0 33.5 61.0
0.0–30.0 0.0–20.0 6.5–49.0 7.5–98.0
1.9 1.9 0.0 0.0
1.0 24.8 0.0 0.0
JFIQR, Japanese version of the Revised Fibromyalgia Impact Questionnaire; FM, fibromyalgia; SD, standard deviation; Min, minimum; Max, maximum.
4
International Journal of Rheumatic Diseases 2015
Psychometric assessment of the JFIQR
Table 3 Correlations† between the JFIQR and external criteria (a) JFIQ, (b) FAS-31 and SF-36v2 (n = 105) JFIQR
(a) Function Overall impact Symptom Total score
JFIQ Physical function
Feel good
Work missed
Ability to do work
Pain
Fatigue
Morning tiredness
Stiffness
Anxiety
Depression
Total score
0.74 0.64 0.61 0.71
0.14 0.20 0.21 0.20
0.62 0.58 0.56 0.63
0.71 0.80 0.74 0.80
0.73 0.79 0.85 0.85
0.64 0.69 0.80 0.77
0.52 0.59 0.70 0.65
0.50 0.46 0.56 0.55
0.44 0.66 0.72 0.65
0.53 0.72 0.77 0.72
0.77 0.84 0.89 0.89
JFIQR
(b) Function Overall impact Symptom Total score
FAS-31
SF-36v2
WPI
SS
Total score
PF
RP
BP
GH
VT
SF
RE
MH
0.44 0.32 0.37 0.41
0.67 0.67 0.78 0.76
0.59 0.51 0.58 0.61
0.81 0.60 0.65 0.74
0.60 0.67 0.60 0.66
0.68 0.74 0.71 0.76
0.42 0.44 0.49 0.49
0.55 0.61 0.62 0.63
0.59 0.67 0.59 0.66
0.53 0.58 0.59 0.61
0.39 0.56 0.60 0.56
JFIQR, Japanese version of the Revised Fibromyalgia Impact Questionnaire; JFIQ, Japanese version of the Fibromyalgia Impact Questionnaire; FAS31, Fibromyalgia Activity Scale-31; WPI, widespread pain index; SS, symptom severity; SF-36v2, Medical Outcomes Study 36-Item Short-Form Health Survey version 2; PF, physical function; RP, role limitations-physical; VT, vitality; GH, general health perception; BP, bodily pain; SF, social function; RE, role limitations-emotional; MH, mental health; †Pearson’s correlation. All the results are the P-value of
