Brain & Development xxx (2015) xxx–xxx www.elsevier.com/locate/braindev
Original article
Verification of the reliability and validity of a Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J) Eri Moriguchi a,b,⇑, Mikiko Ito b, Toshisaburo Nagai b b
a Department of Nursing, Kyoto Koka Women’s University, Japan Division of Health Sciences, Osaka University Graduate School of Medicine, Japan
Received 2 December 2014; received in revised form 23 February 2015; accepted 8 April 2015
Abstract Objective: A Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J) was developed using international guidelines as a QOL scale for childhood epilepsy; its reliability and validity were examined, focusing on Japanese pediatric epilepsy patients applicability. Methods: A pilot test questionnaire survey was conducted; involving parents of pediatric epilepsy patients aged 4–15 undergoing outpatient treatment. 278 responses were obtained and analyzed. Results: Internal consistency for the 16 QOLCE-J subscales, except for , was sufficient, and a high overall coefficient a was obtained. The intraclass correlation coefficient was also high, supporting the test–retest reliability of this version. Associations among the subscales, high correlations of r > 0.7 were observed among , , and , representing cognitive and behavioral aspects, and among these and . In contrast, correlations among others were moderate or weaker. Furthermore, correlations of r > 0.35 were observed among the subscales of the SDQ (Strength and Difficulties Questionnaire) used as an external criterion and the QOLCE-J, confirming the criterion validity of the study version. Analysis of associations between the total QOLCE-J score and pathology of epilepsy, found significant correlation with age of onset and frequency of seizures, ADL, and antiepileptics side effects’ symptoms. QOLCE has mostly been used in treatment resistant pediatric patients, the influence of interictal period presently observed, like antiepileptic side effects’ symptoms; suggest usefulness for pediatric patients with seizures under control. Conclusions: The QOLCE-J with sufficient reliability and validity may be applicable as a QOL scale for Japanese children with epilepsy. Ó 2015 The Japanese Society of Child Neurology. Published by Elsevier B.V. All rights reserved.
Keywords: Children; Epilepsy; Quality of life; Japan; QOLCE-J
1. Introduction Children with epilepsy suffer from various limitations in their daily activities due to epileptic seizures, related ⇑ Corresponding author at: Department of Nursing, Kyoto Koka Women’s University, 38 Kadono-cho, Nishikyogoku, Ukyo-ku, Kyoto, Japan. Tel.: +81 75 366 2669; fax: +81 75 366 2652. E-mail address: [email protected] (E. Moriguchi).
accidents, and functional impairment derived from the side effects of antiepileptics. The presence of seizures also interferes with their learning activities and leads to learning difficulties, a limited range of activities, and passivity. Furthermore, seizure-related accidents prevent them from freely participating in school or leisure activities. The historical background, such as regarding epilepsy as a mental disorder, has been reported to be responsible for the persistent discrimination or prejudice
http://dx.doi.org/10.1016/j.braindev.2015.04.005 0387-7604/Ó 2015 The Japanese Society of Child Neurology. Published by Elsevier B.V. All rights reserved.
Please cite this article in press as: Moriguchi E et al. Verification of the reliability and validity of a Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J). Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.04.005
2
E. Moriguchi et al. / Brain & Development xxx (2015) xxx–xxx
against those with the disease [1]; therefore, when examining appropriate methods to improve the QOL of those with childhood epilepsy, it is also necessary to take this factor into consideration. While the Health-related Quality of Life in Children with Epilepsy developed by Ronen et al. [2] and QOLIE-AD-48 for young adults developed by Cramer et al. [3] are scales to assess the QOL of those with childhood epilepsy, the Kid-KINDL and Kiddo-KINDL developed by Bullinger et al. [4] are QOL scales for all children with or without a disease. With the 31 epilepsy-specific items added to the 25 original ones, the latter are also applicable as epilepsy-related QOL scales. Multiple epilepsy-specific health-related QOL scales have already been developed in other countries [5]. In contrast, in Japan, a Japanese version of the QOL Scale for Children with Epilepsy is currently under development, but its applicability has not yet been confirmed. Although a Japanese version of QOLIE-AD-48 has already been tested, and is currently being applied in epilepsy treatment and care, it is limited to those aged 11–18, and not applicable to infants or school-age children with epilepsy. In the case of the Kid-KINDL [6] and Kiddo-KINDL [7], their Japanese versions as QOL scales target 2 age groups: elementary and junior high school students. They consist of items applicable to children generally, and, unlike their original versions, do not contain those specific to the disease. In short, childhood epilepsy-related QOL scales are not available in Japan at present. Considering this situation, the author focused on the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE) [8–11], developed by Sabaz et al. in Australia to assess the QOL of children with refractory epilepsy aged 4–18, as a health-related scale for children with epilepsy. In addition to the original version, US [12], Polish [13], and Korean [14] versions are also available, allowing international comparisons of the QOL of children with epilepsy. In the present study, a Japanese version of the QOLCE (QOLCE-J) was developed, and its reliability and validity were examined, with a view to applying it to Japanese pediatric epilepsy patients. 2. Methods The QOLCE is a QOL-related section of the CEQ-P (Child Epilepsy Questionnaire – Parent Form). As shown in Table 1, the CEQ-P consists of 2 parts: part I: CSP (Child Seizure Profile) and part II: QOLCE. While the former profiles epileptic seizures, such as their occurrence and the side effects of antiepileptics, the latter serves as a QOL scale for children with epilepsy. The CSP items are based on 2 scales to assess the severity of seizures and side effects, respectively: the Hague Seizure Severity (SS) Scale and the Hague Side Effect
Table 1 Structure of the CEQ-P.
Part I (CSP) Part II (QOLCE)
Section
Number of items
1 2 3 4 5 6 7 8 9
56 31 12 19 23 12 23 1 1
Seizure description AED Side Effects Your child’s physical activities Wellbeing Cognition Your child’s social activities Your child’s behavior General health Quality of life
(SE) Scale [15]. The QOLCE is a questionnaire consisting of 7 sections of questions, which are classified into the following 16 subscales on assessment: , , , , , , , , , , , , , , , and . Adopting a proxy assessment style, parents are asked questions about their children’s conditions in the last 4 weeks. Mean scores for each subscale and the entire scale are calculated and evaluated. 2.1. Development of a Japanese version of the QOLCE With the permission of the developers, Ann M.E. Bye et al., the QOLCE was translated, following the guidelines for the IQOLA (International Quality of Life Assessment) Project [16,17]. The IQOLA Project aims to promote the activity of translating the SF-36 [18], a health-related QOL scale used the most frequently throughout the world, and examine its validity from cultural and psychometric viewpoints in each country. It has already been translated into the languages of a number of countries, generally following the procedures specified in the guidelines for the project. In the present study, a forward translation procedure to create a Japanese version was implemented by 3 Japanese engaged in childhood epilepsy research in consideration of consistency with the construct of the original version, as well as Japan’s cultural background. In relation to cultural validity, items related to physical activities, such as ‘go swimming’ and ‘staying out overnight with friends or other family members’, were excluded from the Japanese version, as they do not accurately reflect the daily habits or activities of Japanese people. On the other hand, ‘taking a bath’ was added as an item representing Japanese culture and lifestyle; in fact, in long-term care guidance for children with epilepsy, taking precautions against epileptic seizure-related accidents when bathing is generally recommended.
Please cite this article in press as: Moriguchi E et al. Verification of the reliability and validity of a Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J). Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.04.005
E. Moriguchi et al. / Brain & Development xxx (2015) xxx–xxx
Subsequently, a back translation procedure was implemented by a person with thorough knowledge of English, who was not involved in the forward translation procedure. After back translation, consistency with the original scale was confirmed by a native English speaker to complete this as a pilot Japanese version. 2.2. Pilot test The pilot Japanese version was tested, involving 13 parents of pediatric epilepsy patients aged 4–15 undergoing outpatient treatment at a university hospital located in Osaka Prefecture within the period between May and August 2011. After completing the questionnaire sheet during outpatient consultation, they were individually interviewed for approximately 20 min by the researcher to clarify the comprehensibility of each question and accuracy of interpretation. After reviewing and correcting the Japanese translation of 1 item, for which there were questions and opinions about its meaning and interpretation, the Japanese version of the QOLCE (QOLCE-J) was completed.
3
2.3.2. Methods At the outpatient departments of the above-mentioned facilities, doctors in charge explained the outline of the study to the parents of relevant pediatric patients within the period between January and April 2012, with explanatory documents specifying the study objective, methods, and ethical considerations, questionnaire sheets for 2 rounds (test–retest), and a return envelope. It was also explained with a written document that those who were also able to cooperate with the retest should complete the second-round sheet 1–2 weeks after the first round, and the procedure from completing the first-round sheet to returning was shown using a flowchart. The respondents returned the completed firstround sheet only or first- and second sheets together in a return envelope. 2.3.3. Study items 2.3.3.1. Characteristics of pediatric patients. Part I of the questionnaire CEQ-P consisted of questions to clarify the pediatric patients’ basic attributes, such as age, sex, and educational facility, in addition to disease-related items, including frequency of seizures, ADL, and side effects of antiepileptics.
2.3. Psychometric analysis 2.3.1. Subjects Among the targeted children aged 4–15 undergoing outpatient antiepileptic therapy, those who were able to communicate appropriately with their parents were included, and their parents responded to the questionnaire as proxies. In the surveys conducted for the development of the original version, only pediatric patients with refractory epilepsy requiring surgical treatment, aged 4–18, were studied. In contrast, in the present study, those in whom seizures were well controlled were also included, in addition to pediatric patients with refractory epilepsy, focusing on the influences of the frequency and severity of seizures. Furthermore, considering the developmental aspect of psychosociology and parents’ difficulty in accurately recognizing the delicate emotions of their children as senior high school students with established identity, only junior high school students or younger children aged 4–15 were targeted. While participants were basically recruited in facilities where the targeted children were undergoing treatment, requests for cooperation with this study were made through medical institutions specializing in childhood epilepsy throughout Japan, excluding those where the majority of children were likely to have difficulty in appropriately communicating with their parents, such as schools for special needs education or severely retarded children, with written explanations regarding the study objective, methods, and ethical considerations. Cooperation was obtained from 43 facilities.
2.3.3.2. Epilepsy-related QOL scale. The QOLCE-J after pilot testing was used as an epilepsy-related QOL scale. 2.3.3.3. Strength and Difficulties Questionnaire (SDQ). In the surveys conducted for the development of the QOLCE, the CHQ (Child Health Questionnaire) and CBCL (Child Behavior Checklist) were used to confirm its criterion validity; the reliability and validity of these 2 scales had already been confirmed. In the present study, there were concerns over the accuracy of responses, as a Japanese version of the CHQ was not available in Japan, and the large numbers of QOLCE and CBCL (113) items were likely to cause fatigue in respondents; therefore, the Strength and Difficulties Questionnaire (SDQ) was adopted to examine the criterion validity of the study version. The SDQ was developed by Goodman et al. [19], and has been translated throughout the world in 75 versions. In Japan, Matsuishi et al. [20] developed a Japanese version, the reliability and validity of which have been confirmed. The SDQ is a scale to screen the behavioral characteristics of infants to school-age children, which has been widely used for developmental disability and psychological health assessments. It has also been reported to be correlated with the CBCL [21]. The scale consists of 25 items, classified into 5 subscales: , , , , and . Each question has 3 answer options: ‘Yes’, ‘Relatively yes’, and ‘No’, and, on the basis of scores for each subscale, it is possible to determine the level of difficulty or
Please cite this article in press as: Moriguchi E et al. Verification of the reliability and validity of a Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J). Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.04.005
4
E. Moriguchi et al. / Brain & Development xxx (2015) xxx–xxx
need for support as low, some, or high. A higher score for each of the 4 subscales, , , , and , indicates a higher level of need, and their sum is adopted as a TDS (total difficulties score). In contrast, a higher score for indicates a lower level of need. 2.3.4. Analysis 2.3.4.1. Descriptive statistics. All data were analyzed using statistical analysis software SPSS Ver. 22. The mean score, standard deviation, and range for each QOLCE-J subscale were calculated. 2.3.4.2. Reliability. Among the 16 QOLCE-J subscales, those with 3 or more questions were examined by calculating Cronbach’s (coefficient) a to confirm internal consistency. Those with 2 questions were examined by calculating the single correlation coefficient between the questions. The reproducibility of the QOLCE-J was examined by calculating the correlation coefficient using the test– retest method. The retest as the second questionnaire round was conducted, as described in the Methods section, and 260 responses (response rate: 30.4%) were obtained, 235 of which were regarded as valid responses (valid response rate: 90.4%). 2.3.4.3. Validity. 2.3.4.3.1. Discriminant and convergent validity. Discriminant and convergent validity was examined by comparing correlations among the 16 QOLCE-J subscales. 2.3.4.3.2. Criterion validity. As previously mentioned, criterion validity was examined using the SDQ. 2.3.4.3.3. Factorial validity. To clarify the influences of age, sex, educational facility, and ADL, as well as the pathology of epilepsy under treatment, on the QOLCE-J, multiple regression analysis was performed, with the QOLCE-J as a dependent variable and the age of onset, frequency, type of seizures, and symptoms due to the side effects of antiepileptics as independent variables. The symptoms due to the side effects of antiepileptics were classified into 4 pathological categories: : drowsiness, headaches, malaise, depression, and bradykinesia; : nausea, constipation, diarrhea, and loss of appetite; : dizziness, walking difficulty, falls, tremors, speech difficulty, and double or blurred vision; and : hyperactivity, dysphoria, aggressiveness, declined academic performance, and abnormal behavior. On analysis of internal consistency among these 4 categories, the coefficient a for was low. Therefore, this category was excluded (Table 2). Higher scores for independent variables were considered to represent the less frequent occurrence of seizures, higher ADL independence levels, and milder side effects of antiepileptics. 2.3.5. Ethical considerations The study was conducted with approval of the research ethics committee of the rerated organization. The study objective and ethical considerations were explained to the parents by the doctor in charge to obtain their consent to cooperate with the study, and returning the completed sheets was regarded as their consent to cooperate with the study. 3. Results A total of 854 questionnaire sheets were distributed, 312 (response rate: 36.5%) were completed and returned, and 278 (valid response rate: 89.1%) were studied as valid responses. Among the respondents who returned valid responses, the numbers of the mothers, fathers, and grandparents of pediatric patients were 262 (94.2%), 15 (5.4%), and 1 (0.4%), respectively. 3.1. Characteristics of pediatric patients 3.1.1. Attributes (Table 3) The numbers of male and female pediatric patients were 146 (52.5%) and 130 (46.8%), respectively; the sex of two patients was not clarified. The age ranged from 4 to 15, with a mean of 9.8 (SD: ±2.9). The educational facilities they belonged to included: kindergartens/nursery schools (31; 11.9%); elementary schools (179; 64.4%); and junior high schools (66; 23.7%).
Table 2 Symptoms due to the side effects of antiepileptics. Category
Manifested symptoms
Cronbach’s a
Suppressed neural activity Gastrointestinal symptoms Specific neurological symptoms Higher brain dysfunction
Drowsiness, headaches, malaise, depression, and bradykinesia Nausea, constipation, diarrhea, and loss of appetite Dizziness, walking difficulty, falls, tremors, speech difficulty, and double or blurred vision Hyperactivity, dysphoria, aggressiveness, declined academic performance, and abnormal behavior
0.72 0.63 0.77 0.84
Please cite this article in press as: Moriguchi E et al. Verification of the reliability and validity of a Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J). Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.04.005
E. Moriguchi et al. / Brain & Development xxx (2015) xxx–xxx Table 3 Pediatric patients’ attributes (n = 278). Number of patients
(%)
Sex Male Female No answer
146 130 2
(52.5) (46.8) (0.7)
Age 4–6 7–9 10–12 13–15
38 93 94 53
(13.7) (33.5) (33.8) (19.1)
Educational facility Kindergarten/nursery school Elementary school Junior high school
33 179 66
(11.9) (64.4) (23.7)
Age of onset 0–3 4–6 7–9 10–12 13–15 No answer
100 81 56 30 4 7
(36.0) (29.1) (20.1) (10.8) (1.4) (2.5)
Frequency of seizures in the past year Almost daily Once a week or more frequently Several times a month Several times a year Not occurred
28 14 30 69 137
(10.1) (5.0) (10.8) (24.8) (49.3)
ADL independence level Independent Partially independent Requiring assistance No answer
223 40 14 1
(80.5) (14.4) (5.1) (0.4)
The age of onset of epileptic seizures ranged from 0 to 14, with a mean of 5.18 (SD: ±3.42). Regarding the frequency of seizures in the past year, 28 (10.1%) respondents answered , 14 (5.0%) answered , 30 (10.8%) answered , 69 (24.8%) answered , and 137 (49.3%) answered . Regarding ADL independence, 223 (80.2%) answered , 40 (14.4%) answered , and 14 (5.0%) answered . There was no answer from 1.
5
3.1.2. Symptoms due to the side effects of antiepileptics The incidence of symptoms due to the side effects of antiepileptics listed in Table 2 was scored as follows: : 1; : 2; : 3; and : 4. Scores were totaled to obtain a mean for each symptom, and the obtained means were generally high: >3.5 (Table 4). 3.2. Scores for the QOLCE-J items In the original QOLCE, each question has 5 answer options: ‘Very Often’, ‘Fairly Often’, ‘Sometimes’, ‘Almost Never’, and ‘Never’. The most positive answer is scored as 100, followed by 75, 50, and 25, and the most negative answer is scored as 0. In short, if ‘Very Often’ is the most positive answer to a question, it is scored as 100. Similarly, if ‘Almost Never’ is the most positive answer to another (reverse) question, it is also scored as 100. It was difficult, however, to adopt this 5-grade answering method comprehensively in the QOLCE-J, as there might be slight differences in the impression of each frequency between the original and translated versions, possibly preventing Japanese respondents from answering accurately. Therefore, in the latter, a 4-grade answering method was adopted: ‘Often’, ‘Sometimes’, ‘Rarely’, and ‘Never’. Subsequently, methods to score these 4 answer options were examined, discussing the appropriateness of scoring them at 100, 75, 25, and 0 or 100, 70, 30, and 0. Both patterns were applied, and principal component analysis was performed, revealing no marked differences between them. Therefore, the 4 answer options were scored as 100, 70, 30, and 0 as a less polarized answering method. 3.3. Examination of each subscale On item-total correlation analysis for each subscale to examine the reliability of the original version, a total of 14 questions showed a weaker correlation (
Original article
Verification of the reliability and validity of a Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J) Eri Moriguchi a,b,⇑, Mikiko Ito b, Toshisaburo Nagai b b
a Department of Nursing, Kyoto Koka Women’s University, Japan Division of Health Sciences, Osaka University Graduate School of Medicine, Japan
Received 2 December 2014; received in revised form 23 February 2015; accepted 8 April 2015
Abstract Objective: A Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J) was developed using international guidelines as a QOL scale for childhood epilepsy; its reliability and validity were examined, focusing on Japanese pediatric epilepsy patients applicability. Methods: A pilot test questionnaire survey was conducted; involving parents of pediatric epilepsy patients aged 4–15 undergoing outpatient treatment. 278 responses were obtained and analyzed. Results: Internal consistency for the 16 QOLCE-J subscales, except for , was sufficient, and a high overall coefficient a was obtained. The intraclass correlation coefficient was also high, supporting the test–retest reliability of this version. Associations among the subscales, high correlations of r > 0.7 were observed among , , and , representing cognitive and behavioral aspects, and among these and . In contrast, correlations among others were moderate or weaker. Furthermore, correlations of r > 0.35 were observed among the subscales of the SDQ (Strength and Difficulties Questionnaire) used as an external criterion and the QOLCE-J, confirming the criterion validity of the study version. Analysis of associations between the total QOLCE-J score and pathology of epilepsy, found significant correlation with age of onset and frequency of seizures, ADL, and antiepileptics side effects’ symptoms. QOLCE has mostly been used in treatment resistant pediatric patients, the influence of interictal period presently observed, like antiepileptic side effects’ symptoms; suggest usefulness for pediatric patients with seizures under control. Conclusions: The QOLCE-J with sufficient reliability and validity may be applicable as a QOL scale for Japanese children with epilepsy. Ó 2015 The Japanese Society of Child Neurology. Published by Elsevier B.V. All rights reserved.
Keywords: Children; Epilepsy; Quality of life; Japan; QOLCE-J
1. Introduction Children with epilepsy suffer from various limitations in their daily activities due to epileptic seizures, related ⇑ Corresponding author at: Department of Nursing, Kyoto Koka Women’s University, 38 Kadono-cho, Nishikyogoku, Ukyo-ku, Kyoto, Japan. Tel.: +81 75 366 2669; fax: +81 75 366 2652. E-mail address: [email protected] (E. Moriguchi).
accidents, and functional impairment derived from the side effects of antiepileptics. The presence of seizures also interferes with their learning activities and leads to learning difficulties, a limited range of activities, and passivity. Furthermore, seizure-related accidents prevent them from freely participating in school or leisure activities. The historical background, such as regarding epilepsy as a mental disorder, has been reported to be responsible for the persistent discrimination or prejudice
http://dx.doi.org/10.1016/j.braindev.2015.04.005 0387-7604/Ó 2015 The Japanese Society of Child Neurology. Published by Elsevier B.V. All rights reserved.
Please cite this article in press as: Moriguchi E et al. Verification of the reliability and validity of a Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J). Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.04.005
2
E. Moriguchi et al. / Brain & Development xxx (2015) xxx–xxx
against those with the disease [1]; therefore, when examining appropriate methods to improve the QOL of those with childhood epilepsy, it is also necessary to take this factor into consideration. While the Health-related Quality of Life in Children with Epilepsy developed by Ronen et al. [2] and QOLIE-AD-48 for young adults developed by Cramer et al. [3] are scales to assess the QOL of those with childhood epilepsy, the Kid-KINDL and Kiddo-KINDL developed by Bullinger et al. [4] are QOL scales for all children with or without a disease. With the 31 epilepsy-specific items added to the 25 original ones, the latter are also applicable as epilepsy-related QOL scales. Multiple epilepsy-specific health-related QOL scales have already been developed in other countries [5]. In contrast, in Japan, a Japanese version of the QOL Scale for Children with Epilepsy is currently under development, but its applicability has not yet been confirmed. Although a Japanese version of QOLIE-AD-48 has already been tested, and is currently being applied in epilepsy treatment and care, it is limited to those aged 11–18, and not applicable to infants or school-age children with epilepsy. In the case of the Kid-KINDL [6] and Kiddo-KINDL [7], their Japanese versions as QOL scales target 2 age groups: elementary and junior high school students. They consist of items applicable to children generally, and, unlike their original versions, do not contain those specific to the disease. In short, childhood epilepsy-related QOL scales are not available in Japan at present. Considering this situation, the author focused on the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE) [8–11], developed by Sabaz et al. in Australia to assess the QOL of children with refractory epilepsy aged 4–18, as a health-related scale for children with epilepsy. In addition to the original version, US [12], Polish [13], and Korean [14] versions are also available, allowing international comparisons of the QOL of children with epilepsy. In the present study, a Japanese version of the QOLCE (QOLCE-J) was developed, and its reliability and validity were examined, with a view to applying it to Japanese pediatric epilepsy patients. 2. Methods The QOLCE is a QOL-related section of the CEQ-P (Child Epilepsy Questionnaire – Parent Form). As shown in Table 1, the CEQ-P consists of 2 parts: part I: CSP (Child Seizure Profile) and part II: QOLCE. While the former profiles epileptic seizures, such as their occurrence and the side effects of antiepileptics, the latter serves as a QOL scale for children with epilepsy. The CSP items are based on 2 scales to assess the severity of seizures and side effects, respectively: the Hague Seizure Severity (SS) Scale and the Hague Side Effect
Table 1 Structure of the CEQ-P.
Part I (CSP) Part II (QOLCE)
Section
Number of items
1 2 3 4 5 6 7 8 9
56 31 12 19 23 12 23 1 1
Seizure description AED Side Effects Your child’s physical activities Wellbeing Cognition Your child’s social activities Your child’s behavior General health Quality of life
(SE) Scale [15]. The QOLCE is a questionnaire consisting of 7 sections of questions, which are classified into the following 16 subscales on assessment: , , , , , , , , , , , , , , , and . Adopting a proxy assessment style, parents are asked questions about their children’s conditions in the last 4 weeks. Mean scores for each subscale and the entire scale are calculated and evaluated. 2.1. Development of a Japanese version of the QOLCE With the permission of the developers, Ann M.E. Bye et al., the QOLCE was translated, following the guidelines for the IQOLA (International Quality of Life Assessment) Project [16,17]. The IQOLA Project aims to promote the activity of translating the SF-36 [18], a health-related QOL scale used the most frequently throughout the world, and examine its validity from cultural and psychometric viewpoints in each country. It has already been translated into the languages of a number of countries, generally following the procedures specified in the guidelines for the project. In the present study, a forward translation procedure to create a Japanese version was implemented by 3 Japanese engaged in childhood epilepsy research in consideration of consistency with the construct of the original version, as well as Japan’s cultural background. In relation to cultural validity, items related to physical activities, such as ‘go swimming’ and ‘staying out overnight with friends or other family members’, were excluded from the Japanese version, as they do not accurately reflect the daily habits or activities of Japanese people. On the other hand, ‘taking a bath’ was added as an item representing Japanese culture and lifestyle; in fact, in long-term care guidance for children with epilepsy, taking precautions against epileptic seizure-related accidents when bathing is generally recommended.
Please cite this article in press as: Moriguchi E et al. Verification of the reliability and validity of a Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J). Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.04.005
E. Moriguchi et al. / Brain & Development xxx (2015) xxx–xxx
Subsequently, a back translation procedure was implemented by a person with thorough knowledge of English, who was not involved in the forward translation procedure. After back translation, consistency with the original scale was confirmed by a native English speaker to complete this as a pilot Japanese version. 2.2. Pilot test The pilot Japanese version was tested, involving 13 parents of pediatric epilepsy patients aged 4–15 undergoing outpatient treatment at a university hospital located in Osaka Prefecture within the period between May and August 2011. After completing the questionnaire sheet during outpatient consultation, they were individually interviewed for approximately 20 min by the researcher to clarify the comprehensibility of each question and accuracy of interpretation. After reviewing and correcting the Japanese translation of 1 item, for which there were questions and opinions about its meaning and interpretation, the Japanese version of the QOLCE (QOLCE-J) was completed.
3
2.3.2. Methods At the outpatient departments of the above-mentioned facilities, doctors in charge explained the outline of the study to the parents of relevant pediatric patients within the period between January and April 2012, with explanatory documents specifying the study objective, methods, and ethical considerations, questionnaire sheets for 2 rounds (test–retest), and a return envelope. It was also explained with a written document that those who were also able to cooperate with the retest should complete the second-round sheet 1–2 weeks after the first round, and the procedure from completing the first-round sheet to returning was shown using a flowchart. The respondents returned the completed firstround sheet only or first- and second sheets together in a return envelope. 2.3.3. Study items 2.3.3.1. Characteristics of pediatric patients. Part I of the questionnaire CEQ-P consisted of questions to clarify the pediatric patients’ basic attributes, such as age, sex, and educational facility, in addition to disease-related items, including frequency of seizures, ADL, and side effects of antiepileptics.
2.3. Psychometric analysis 2.3.1. Subjects Among the targeted children aged 4–15 undergoing outpatient antiepileptic therapy, those who were able to communicate appropriately with their parents were included, and their parents responded to the questionnaire as proxies. In the surveys conducted for the development of the original version, only pediatric patients with refractory epilepsy requiring surgical treatment, aged 4–18, were studied. In contrast, in the present study, those in whom seizures were well controlled were also included, in addition to pediatric patients with refractory epilepsy, focusing on the influences of the frequency and severity of seizures. Furthermore, considering the developmental aspect of psychosociology and parents’ difficulty in accurately recognizing the delicate emotions of their children as senior high school students with established identity, only junior high school students or younger children aged 4–15 were targeted. While participants were basically recruited in facilities where the targeted children were undergoing treatment, requests for cooperation with this study were made through medical institutions specializing in childhood epilepsy throughout Japan, excluding those where the majority of children were likely to have difficulty in appropriately communicating with their parents, such as schools for special needs education or severely retarded children, with written explanations regarding the study objective, methods, and ethical considerations. Cooperation was obtained from 43 facilities.
2.3.3.2. Epilepsy-related QOL scale. The QOLCE-J after pilot testing was used as an epilepsy-related QOL scale. 2.3.3.3. Strength and Difficulties Questionnaire (SDQ). In the surveys conducted for the development of the QOLCE, the CHQ (Child Health Questionnaire) and CBCL (Child Behavior Checklist) were used to confirm its criterion validity; the reliability and validity of these 2 scales had already been confirmed. In the present study, there were concerns over the accuracy of responses, as a Japanese version of the CHQ was not available in Japan, and the large numbers of QOLCE and CBCL (113) items were likely to cause fatigue in respondents; therefore, the Strength and Difficulties Questionnaire (SDQ) was adopted to examine the criterion validity of the study version. The SDQ was developed by Goodman et al. [19], and has been translated throughout the world in 75 versions. In Japan, Matsuishi et al. [20] developed a Japanese version, the reliability and validity of which have been confirmed. The SDQ is a scale to screen the behavioral characteristics of infants to school-age children, which has been widely used for developmental disability and psychological health assessments. It has also been reported to be correlated with the CBCL [21]. The scale consists of 25 items, classified into 5 subscales: , , , , and . Each question has 3 answer options: ‘Yes’, ‘Relatively yes’, and ‘No’, and, on the basis of scores for each subscale, it is possible to determine the level of difficulty or
Please cite this article in press as: Moriguchi E et al. Verification of the reliability and validity of a Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J). Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.04.005
4
E. Moriguchi et al. / Brain & Development xxx (2015) xxx–xxx
need for support as low, some, or high. A higher score for each of the 4 subscales, , , , and , indicates a higher level of need, and their sum is adopted as a TDS (total difficulties score). In contrast, a higher score for indicates a lower level of need. 2.3.4. Analysis 2.3.4.1. Descriptive statistics. All data were analyzed using statistical analysis software SPSS Ver. 22. The mean score, standard deviation, and range for each QOLCE-J subscale were calculated. 2.3.4.2. Reliability. Among the 16 QOLCE-J subscales, those with 3 or more questions were examined by calculating Cronbach’s (coefficient) a to confirm internal consistency. Those with 2 questions were examined by calculating the single correlation coefficient between the questions. The reproducibility of the QOLCE-J was examined by calculating the correlation coefficient using the test– retest method. The retest as the second questionnaire round was conducted, as described in the Methods section, and 260 responses (response rate: 30.4%) were obtained, 235 of which were regarded as valid responses (valid response rate: 90.4%). 2.3.4.3. Validity. 2.3.4.3.1. Discriminant and convergent validity. Discriminant and convergent validity was examined by comparing correlations among the 16 QOLCE-J subscales. 2.3.4.3.2. Criterion validity. As previously mentioned, criterion validity was examined using the SDQ. 2.3.4.3.3. Factorial validity. To clarify the influences of age, sex, educational facility, and ADL, as well as the pathology of epilepsy under treatment, on the QOLCE-J, multiple regression analysis was performed, with the QOLCE-J as a dependent variable and the age of onset, frequency, type of seizures, and symptoms due to the side effects of antiepileptics as independent variables. The symptoms due to the side effects of antiepileptics were classified into 4 pathological categories: : drowsiness, headaches, malaise, depression, and bradykinesia; : nausea, constipation, diarrhea, and loss of appetite; : dizziness, walking difficulty, falls, tremors, speech difficulty, and double or blurred vision; and : hyperactivity, dysphoria, aggressiveness, declined academic performance, and abnormal behavior. On analysis of internal consistency among these 4 categories, the coefficient a for was low. Therefore, this category was excluded (Table 2). Higher scores for independent variables were considered to represent the less frequent occurrence of seizures, higher ADL independence levels, and milder side effects of antiepileptics. 2.3.5. Ethical considerations The study was conducted with approval of the research ethics committee of the rerated organization. The study objective and ethical considerations were explained to the parents by the doctor in charge to obtain their consent to cooperate with the study, and returning the completed sheets was regarded as their consent to cooperate with the study. 3. Results A total of 854 questionnaire sheets were distributed, 312 (response rate: 36.5%) were completed and returned, and 278 (valid response rate: 89.1%) were studied as valid responses. Among the respondents who returned valid responses, the numbers of the mothers, fathers, and grandparents of pediatric patients were 262 (94.2%), 15 (5.4%), and 1 (0.4%), respectively. 3.1. Characteristics of pediatric patients 3.1.1. Attributes (Table 3) The numbers of male and female pediatric patients were 146 (52.5%) and 130 (46.8%), respectively; the sex of two patients was not clarified. The age ranged from 4 to 15, with a mean of 9.8 (SD: ±2.9). The educational facilities they belonged to included: kindergartens/nursery schools (31; 11.9%); elementary schools (179; 64.4%); and junior high schools (66; 23.7%).
Table 2 Symptoms due to the side effects of antiepileptics. Category
Manifested symptoms
Cronbach’s a
Suppressed neural activity Gastrointestinal symptoms Specific neurological symptoms Higher brain dysfunction
Drowsiness, headaches, malaise, depression, and bradykinesia Nausea, constipation, diarrhea, and loss of appetite Dizziness, walking difficulty, falls, tremors, speech difficulty, and double or blurred vision Hyperactivity, dysphoria, aggressiveness, declined academic performance, and abnormal behavior
0.72 0.63 0.77 0.84
Please cite this article in press as: Moriguchi E et al. Verification of the reliability and validity of a Japanese version of the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE-J). Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.04.005
E. Moriguchi et al. / Brain & Development xxx (2015) xxx–xxx Table 3 Pediatric patients’ attributes (n = 278). Number of patients
(%)
Sex Male Female No answer
146 130 2
(52.5) (46.8) (0.7)
Age 4–6 7–9 10–12 13–15
38 93 94 53
(13.7) (33.5) (33.8) (19.1)
Educational facility Kindergarten/nursery school Elementary school Junior high school
33 179 66
(11.9) (64.4) (23.7)
Age of onset 0–3 4–6 7–9 10–12 13–15 No answer
100 81 56 30 4 7
(36.0) (29.1) (20.1) (10.8) (1.4) (2.5)
Frequency of seizures in the past year Almost daily Once a week or more frequently Several times a month Several times a year Not occurred
28 14 30 69 137
(10.1) (5.0) (10.8) (24.8) (49.3)
ADL independence level Independent Partially independent Requiring assistance No answer
223 40 14 1
(80.5) (14.4) (5.1) (0.4)
The age of onset of epileptic seizures ranged from 0 to 14, with a mean of 5.18 (SD: ±3.42). Regarding the frequency of seizures in the past year, 28 (10.1%) respondents answered , 14 (5.0%) answered , 30 (10.8%) answered , 69 (24.8%) answered , and 137 (49.3%) answered . Regarding ADL independence, 223 (80.2%) answered , 40 (14.4%) answered , and 14 (5.0%) answered . There was no answer from 1.
5
3.1.2. Symptoms due to the side effects of antiepileptics The incidence of symptoms due to the side effects of antiepileptics listed in Table 2 was scored as follows: : 1; : 2; : 3; and : 4. Scores were totaled to obtain a mean for each symptom, and the obtained means were generally high: >3.5 (Table 4). 3.2. Scores for the QOLCE-J items In the original QOLCE, each question has 5 answer options: ‘Very Often’, ‘Fairly Often’, ‘Sometimes’, ‘Almost Never’, and ‘Never’. The most positive answer is scored as 100, followed by 75, 50, and 25, and the most negative answer is scored as 0. In short, if ‘Very Often’ is the most positive answer to a question, it is scored as 100. Similarly, if ‘Almost Never’ is the most positive answer to another (reverse) question, it is also scored as 100. It was difficult, however, to adopt this 5-grade answering method comprehensively in the QOLCE-J, as there might be slight differences in the impression of each frequency between the original and translated versions, possibly preventing Japanese respondents from answering accurately. Therefore, in the latter, a 4-grade answering method was adopted: ‘Often’, ‘Sometimes’, ‘Rarely’, and ‘Never’. Subsequently, methods to score these 4 answer options were examined, discussing the appropriateness of scoring them at 100, 75, 25, and 0 or 100, 70, 30, and 0. Both patterns were applied, and principal component analysis was performed, revealing no marked differences between them. Therefore, the 4 answer options were scored as 100, 70, 30, and 0 as a less polarized answering method. 3.3. Examination of each subscale On item-total correlation analysis for each subscale to examine the reliability of the original version, a total of 14 questions showed a weaker correlation (
