This study consists of two phases: (1) translation of the QOLCE-16 to Chinese, followed by a cross-cultural adaptation among the parents with children with epilepsy; (2) testing of the psychometric properties of the QOLCE-16-C, including reliability and validity assessments.
Phase 1: Translation and cross-cultural adaptation procedure
The goal of phase 1 was to produce a semantically equivalent, grammatically fluent, culturally compatible, and readily comprehensible QOLCE-16-C. Due to the lack of gold translation standard[17], we adopt the 10 steps of Principles of Good Practices for translation and cultural adaptation of measures established by the International Society for Pharmacoeconomics and Outcomes Research (ISPOR)[18], which has been commonly used in large international researches including those of Patient-reported outcome measurement information system (PROMIS)[19], the World Health Organization Quality of Life (WHOQOL)[20], the Neuro-quality of life (Neuro-QoL)[21], and others[22, 23]. Figure 1 shows the specific translation steps.
Forward and backward translation
After receiving permission for translating from the development group of QOLCE-16 (Professor Kathy N. Speechley) through E-mails, the forward translation team (1 pediatric neurology nurse, 2 master students) translated the instrument independently. All three of them were bilingual and all have experience living in English-speaking countries for at least one year. The three forward translated version were then reconciled into one through discussion among the three translators as well as the leader of the project. After that, another three bilingual translators (1 child neurologist, 1 child health doctor, 1 English-speaking translator) who were blind to the study objects and original QOLCE-16 conducted the back-translation. After discussion, an integrated and back-translated English version was obtained.
Harmonization
A 5-member harmonization team (1 linguistic specialist, 2 medical school teachers with experience in scale translation research, 2 pediatric nursing doctoral students) conducted an evaluation and comparison between the original scale, forward version and back-translated version to produce an agreement on a draft version.
Culture adaption
To evaluate the equivalence of the draft version with the original version, the cognitive interviewing technique was used in 5 parents whose child had epilepsy. Prior to the cognitive interview, an interview outline was developed and evaluated by experts to make sure its rationality. According to the interview outline, parents were asked about their understanding of the whole scale and the expression of specific items. A 10-point scale was used to measure how well parents understood the items. A rating of “1” meant that the item was difficult to answer, while a rating of “10” meant that the item was easy to answer. General probing and paraphrasing techniques were used for some of the items[24, 25].
Expert review
The Delphi Method was used during this procedure[26]. 9 experts (1 child neurologist, 1 child neurosurgeon, 1 child health doctor, 1 child rehabilitation doctor, 1 pediatrician in public health, 2 professor of nursing, 1 head nurse of neurosurgery, and 1 head nurse of neurology) were invited to make comparison among the above original scale, forward version, back-translated version, and cognitive interview integrated version through E-mail, and to form the final integrated version, the QOLCE-16-C.
Phase 2: QOLCE-16-C psychometric testing procedure
Study subjects and settings
Parents with children who had epilepsy from Shanghai Children’s hospital were invited to take part in the study. A convenience sampling method was adopted to recruit the eligible participants in Neurology department inpatient and outpatient clinics. The inclusion criteria were children who were (1) aged 4 to 18 years, (2) with a diagnosis of epilepsy, (3) with parents able to speak and read Chinese, and (4) willing to participate in this study and have their parent’s permission. Parents who have cognitive or mental impairment and refuse to participate in the investigation were excluded. When the children and their parents came to the clinic, the researchers explained to the participants the purpose of the study and asked them if they would be willing to participate in the survey. After obtaining the written informed consent of the participants, the data were collected through self-filled questionnaires which were filled out in the doctor’s office. All the questionnaires will be returned on the spot after they had been filled out. A total of 430 questionnaires were distributed from June to November 2020, and 414 were effectively returned, with an effective response rate of 96.3%. To assess the test-retest reliability, 40 parents were randomly selected to answer the QOLCE-16-C again 2–3 weeks after their first completement. The test-retest data was collected in two ways: paper questionnaires and electronic questionnaires filled out on the WeChat platform. Finally, 32 valid retest questionnaires were recovered.
Instruments
General demographic scale
The general demographic scale was used to obtain demographic data, including information about children with epilepsy and family information. Data related to children with epilepsy includes gender, age, grade of study and information related to epilepsy disease. Family data includes parents' age, education background, family monthly income, payment methods for medical expenses.
Chinese version of the QOLCE-16
This was a short-form version of the QOLCE-55[14], which consists of 16 items and was divided into 4 dimensions: cognitive, emotional, social, and physical functioning. The internal consistency reliability of the original QOLCE-16 was excellently high (α = 0.90)[16]. A 5-point Likert scale was used to calculate the scores for each domain and the scores for the total scale: 1 = very often, 2 = fairly often, 3 = sometimes, 4 = almost never, 5 = never[16]. Ratings were transformed linearly, presenting total scores take values from 0 (low HRQOL) to 100 (high HRQOL). It is worth noting that the scale also includes the sixth option: 6 = not applicable, but this option does not participate in scoring. If the answer to an item is “not applicable”, the item will be considered as missing data. The translation procedure of the Chinese version of the QOLCE-16 has been reported in phase 1.
Data analysis
IBM SPSS Statistics 23.0 and SPSS Amos24.0 were used for analyzing the psychometric properties.
Item analysis
In item analysis, the mean and standard deviation (‾x ± s) of each and total items in the inventory was reported. The critical ratio (CR) was calculated for every item, with a value above 3 was considered statistically significant[27]. After that, the item-total correlation coefficient was tested.
Reliability
Reliability was tested by internal consistency and test-retest reliability. Cronbach's α coefficient was used to assess internal consistency with a value of α above 0.7 is considered as acceptable[28]. Interclass correlation coefficients (ICC) was used for test-retest reliability, with ICC lower than 0.4 is considered low, 0.4–0.6 considered normal, 0.6–0.8 considered moderate and higher than 0.8 considered high[29].
Content validity
Content validity index (CVI) was used to evaluate the content validity of the scale, including the scale level CVI (S-CVI) and the item level CVI (I-CVI). 9 experts (the 9 experts mentioned in the expert review process) were asked to score the whole scale and each item on clarity and relevancy on a 4-point Likert scale (0 = undesirable, 1 = somewhat desirable, 2 = desirable, 3 = fully desirable). The CVI was calculated by the number of the evaluators gave fully desirable and desirable divided by the total number of questions. The cut-off of the S-CVI should be 0.9 or higher, and the I-CVI should be 0.78 or higher[30].
Construct validity
Although the QOLCE-16 has been proved has good structural validity by its original author, exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were conducted considering that the translated QOLCE-16-C was applied for the first time in Chinese children with epilepsy.
First, exploratory factor analysis (EFA) was performed to obtain the factor structure by using principal components analysis with varimax rotation[31]. Before the EFA, Kaiser-Meyer-Olkin (KMO) measure and Bartlett’s sphericity were chosen for verification, with KMO index greater than 0.50 was considered as eligible to conduct EFA[32]. Criterions used to determine the number of effective factors were: (1) eigenvalues greater than 1.0, (2) the percentage of total explained variance accounted for, and (3) items with loading greater than 40% in absolute value[31]. Second, confirmatory factor analysis (CFA) was conducted to verify the factor structure determined in the exploratory study. The maximum likelihood estimation method was used to estimate the parameters in the model. Items were first loaded onto 4 factors and then integrated into one factor. The model fit indices are deemed acceptable by χ2/df༜3, comparative fit index (CFI)༞0.90, goodness of fit index (GFI)༞0.90 and root-mean-square error of approximation (RMSEA)༜0.08[33]. After that, the convergent validity of each dimension was tested by the values of the average variance extracted (AVE) and discriminant validity was determined by comparing the square root value of AVE of each dimension with the correlation coefficient of a particular dimension with those of other dimensions[34].