This cross-sectional study was undertaken to translate a validated set of 6 hierarchically arranged questions, that comprise the Gozal sleep questionnaire (6Q) (Fig. 1), to assess the prevalence of SDB among Saudi pre-school children. [11, 15] The questions use Likert-type responses for the preceding 6 months with scores as following: “never” or “mildly quite” (0), “rarely” or “medium loud” (1), “occasionally” or “loud” (2), “frequently” or “very loud” (3), and “almost always” or “extremely loud” (4). A cumulative score is then calculated using the following formula where Q1 = raw score to question1, Q2 = raw score to question 2 and so on:
A= (Q1 + Q2)/2; B= (A + Q3)/2; C= (B + Q4)/2; D= (C + Q5)/2; and the final score = (D + Q6)/2
Participants were recruited from 8 pre-schools located in different areas in Jeddah, Saudi Arabia between October 2017 and April 2018. Two pre-schools were selected randomly from each of the northern, southern, eastern, and western areas. Of the 8 pre-schools, 4 were public and 4 were private. A description of the study and consent forms were placed in the backpacks of 1783 enrolled students. Study inclusion criteria for the children were:1) Saudi or a permanent resident of Saudi Arabia, 2) age between 3 and 5 years old, 3) resides with his/her mother, 4) healthy with no serious medical problems or history of food allergies, and 5) mother is an Arabic speaker. Two hundred and nine mothers returned the signed consent forms and were contacted by research assistants over the telephone to answer the questionnaire. This study was part of a larger scale study to evaluate eating behaviors and weight status among Saudi pre-school children.[26] The protocol was approved by the research ethics committee at King Abdulaziz University (#366 − 16). All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and later versions. Written informed consent was obtained from all parents or legal guardians. This project was funded by the Deanship of Scientific Research (DSR) at King Abdulaziz University, Jeddah (grant number J-734-142-38).
The study included three stages: (1) translating the English questionnaire into Arabic, (2) testing the validity and reliability of the translated Arabic questionnaire, and (3) Using the translated Arabic questionnaire to obtain the prevalence of SDB among pre-school children in Jeddah, Saudi Arabia.
Stage 1. Translating The English Questionnaire Into Arabic:
To translate the questionnaire, forward-translation and back-translation methods refined by the WHO were adopted [27]. To produce a final translation 6 steps were followed:
a. Forward-translation:
Forward-translation was performed by the primary investigator (PI), an American trained health professional. The PI’s mother tongue is Arabic but is familiar with the terminology in the field of sleep having obtained a PhD in neuroscience and sleep from the United States. The PI emphasized conceptual rather than literal translation of words. Translations were also made to be concise avoiding long sentences with many clauses. Finally, translations were aimed to be clear and simple to consider non-working mothers as the typical respondent. This resulted in an initial Arabic forward translation (Fig. 2).
b. Expert panel discussion of initial forward-translation:
A bilingual (Arabic and English) expert panel was convened by the PI to discuss the initial forward-translation. The panel included 5 American trained professionals. There were 4 health professionals with either a PhD or a Doctorate degree, and an engineer with a PhD. The PI provided the panel with reading material that familiarized them with SDB, snoring, sleep apnea, diagnosis and screening methods, and sleep questionnaires. The panel was also informed of the target audience, the way in which the questionnaire will be administered, and the conceptual framework.
The expert panel reviewed each question to identify and resolve inadequate concepts of translation, and inadequate expressions used in the translated version to ensure cultural sensitivity.[28] The panel also reviewed the questions to ensure consistency between the original and translated versions and identify discrepancies between of the questionnaires. This was carried out by questioning some expressions and words and suggesting alternatives until finally agreeing on a modified Arabic forward translation (Fig. 2).
c. Back-translation:
Back-translation to English was performed by another health professional. The translator’s mother tongue is English but is fluent in the Arabic language having studied Arabic for 14 years in an Arabic school in Jeddah, Saudi Arabia. The translator was provided with the same reading material and information provided to the expert panel. The same concepts used by the PI for forward translation were used to create a back-translation (Fig. 2).
d. Expert panel discussion of back-translation:
The back-translation was discussed by the same expert panel and in the same manner in which the forward-translation was discussed. The panel compared the original English version with the back-translated version to make sure the content and concept remained the same. Changes were made to the modified forward-translation and a pretest Arabic translation was composed (Fig. 2).
e. Pretesting and cognitive interviewing:
The pretest Arabic translation was pretested on a group of 10 Saudi mothers visiting a private dental practice in Jeddah. Care was taken to include mothers of different socioeconomic levels. After the mothers filled the questionnaires, each mother was interviewed separately to discuss their responses. The mothers were asked whether they thought each question was clear, what they thought each question was asking, and to repeat each question in their own words. Additionally, the mothers were asked about certain words or phrases which the expert panel thought might be problematic, and alternative options were presented to the mothers to choose from. Finally, the mothers were asked if there were any words or phrases, other than the ones selected by the panel, that they did not understand or that could be misunderstood. A written report of all answers and problems arising during these interviews was prepared.
f. Expert panel discussion of pretesting results:
The written report that was formulated as a result of pretesting and cognitive interviewing was presented to the same expert panel for discussion. The panel discussed all answers and issues that arose during the interviews in addition to alternative options to problematic words or phrases. Discrepancies were discussed and resolved through consensus among the expert panel to derive a final Arabic translation (Fig. 2).
Stage 2. Testing the validity and reliability of the translated Arabic questionnaire:
a. Validity:
Face validity was determined subjectively using a dichotomous scale by asking the mothers to rate each question as “clear” or “not clear”. Face validity was assessed during pretesting when cognitive interviewing was performed.
Content validity was undertaken by the expert panel review which was ongoing during the translation process. Content validity was assessed at three timepoints: after forward translation, after back-translation, and after pretesting. After forward translation to ensure that adequate concepts and expression were used and consistency with the original question was maintained. After back-translation to ensure the back-translated version maintained the same concept and content as the original English version. Finally, after pretesting, each member of the panel rated each question independently in terms of “consistency with the English version” and “clarity” on a four-point scale (Table 1).[29] The item-level content validity index (I-CVI) for each question (proportion of experts giving a question a score of either 3 or 4), and the scale-level content validity index (S-CVI) (proportion of questions on a questionnaire that achieved a score of 3 or 4 by all the experts), were calculated as will be described in the statistics section.[30]
Table 1
Criteria for scoring questions for content validity
Score | Consistency | Clarity |
1 | Not consistent with English version | Not clear |
2 | Item needs some revision | Item needs some revision |
3 | Consistent but needs minor revision | Clear but needs minor revision |
4 | Very consistent with the English version | Very clear |
b. Reliability:
Reliability was conducted after content validity. The study team randomly selected 34 mothers out of the total sample of 209 to examine internal consistency and test-retest reliability. The mothers were contacted by phone to complete the questionnaire. The questionnaire had 6 questions assessing 3 constructs: apnea (questions 1 and 2), breathing difficulty (questions 3 and 4), and snoring (questions 5 and 6). To evaluate how reliably questions that were designed to measure the same construct actually did so, internal consistency was assessed. This was accomplished by determining how highly questions within the same construct were correlated and how well they predicted each other using Cronbach's alpha. In order to examine test-retest reliability, the same 34 mothers were contacted by phone approximately 2 weeks after the initial call to complete the questionnaire for a second time. The cumulative scores from the first attempt were correlated with scores from the second attempt using Interclass correlation coefficient (ICC).
Stage 3. Using the translated Arabic questionnaire to obtain prevalence of SDB among pre-school children in Jeddah, Saudi Arabia:
As previously described, the 209 mothers who returned the signed consent forms were contacted by research assistants over the telephone to complete the questionnaire. Answers were tabulated and a cumulative score for each questionnaire was calculated.
Statistical analysis:
Statistical analysis was performed using IBM SPSS version 20. Face validity was assessed for each question separately. This was carried out by adding up the number of mothers rating each question as “clear” out of the 10 mothers the translated questionnaire was pretested on. Prior to the calculation of CVI, the consistency and clarity from the expert rating was recorded as 1 (score of 3 or 4) or 0 (score of 1 or 2).[30] To calculate the I-CVI for each question, the number of experts giving a score of either 3 or 4 was divided by the total number of experts. To calculate the S-CVI, the total number of questions with a score of either 3 or 4 from all experts combined was divided by the total number of questions (S-CVI with universal agreement). [30]
Cronbach’s alpha was used to assess internal consistency of the raw scores within each construct and among all questions combined. Cronbach’s alpha values of at least 0.60 are considered “good”, while values of at least 0.70 are considered “favorable” [31, 32] The cumulative scores of the sleep questionnaires were calculated using the formula described previously. Interclass correlation coefficient (ICC) of the questionnaire cumulative scores from the 34 mothers who answered the questionnaire twice was used to examine test-retest reliability. Statistical significance for all analysis was noted at α = 0.05.
To obtain the prevalence of SDB, the questionnaire cumulative scores from all 209 questionnaires were tested for normality and symmetry using Shapiro-Wilk test for normality and the average and spread were calculated accordingly. The prevalence of SDB was calculated using cumulative score cut-off values of ≥ 1.5, ≥ 2, and ≥ 2.5 to define different SDB severities.