Environmental Risk Factors in Han and Uyghur Children with Dyslexia: A Comparative Study

Background Several studies have been conducted to explore risk factors for dyslexia. However, most studies examining dyslexia have been skewed toward Western countries, and few have considered two nationalities simultaneously. This study focused on differences in dyslexia prevalence and potential environmental risk factors between Han and Uyghur children. Methods A cross-sectional study was conducted in Kashgar and Aksu, cities in Xinjiang province, China. A two-stage sampling strategy was used to recruit 2,854 students in grades 3–6 from 5 primary schools in 5 districts; 2,348 valid student questionnaires were included in the analysis. Dyslexia checklists for Chinese and Uyghur children and pupil rating scales were used to identify children with dyslexia. Questions related to the home literacy environment and reading ability were used to evaluate potential environmental risk factors. Single factor analysis and multivariate logistic regression were used to examine prevalence and risk factors for dyslexia. Results Dyslexia prevalence differed significantly between Han (3.9%) and Uyghur (7.0%) children (P < 0.05), and the boy-to-girl diagnosis ratio was almost 2:1. Multiple logistic regression analysis showed that ethnic differences in dyslexia prevalence between Han and Uyghur children could have occurred because of factors such as mother’s occupation (P = 0.02, OR = 0.04, 95% CI = 0.01–0.68) and the frequency with which parents told stories (P = 0.00, OR = 4.50, 95% CI = 1.67–12.11). Conclusions The prevalence of dyslexia was high in all children, particularly those in the Uyghur group. Environmental factors could have been responsible for some of the differences observed. The results contribute to the early identification and management of dyslexia in children from these two groups and research examining developmental dyslexia and differences in racial genetics.


Introduction
Developmental dyslexia (DD), or reading disability, is characterized by impaired reading performance despite adequate motivational, educational, and intellectual opportunities and the absence of sensory or neurological disability [1]. The disorder affects recognition speed, language processing, and understanding. In Western countries, the prevalence of DD is estimated at 5-12% in school-age children whose native language is English [2,3]. In China, research into DD began relatively late and focused mainly on Han children, for whom the incidence of DD was 3.9-8% [4,5]. DD is an important and controversial concern worldwide. Research has shown that DD initially manifests in the early years of childhood. Although its etiology and pathology remain elusive, there is abundant evidence suggesting that both genetic and environmental factors contribute to DD. Genetic association studies have identified several candidate genes associated with DD, including DCDC2, DYX1C1, ROBO1, and KIAA0319 [1,2,6]. These genes can change the structure of protein, which may lead to changes in its functional properties. However, studies in different languages have yielded conflicting results [6][7][8][9][10].
Moreover, cross-sectional studies have shown that some risk factors, such as maternal educational levels, home literacy environment (HLE), and socioeconomic status (SES), were associated with the treatment of children with DD [4,11,12]. However, the results of several other studies have indicated that SES factors play a minimal role in DD [5]. Moreover, risk factors associated with DD could differ across language, race, and culture, resulting in differences between children from different ethnic backgrounds. Xinjiang lies in the far northwest of China, where Central and Western Asian cultures converge. The region is populated by various ethnic groups and predominantly by Uyghur and Han. Because of differences in the nature of the languages spoken and the specificity of the cultures, DD prevalences and risk factors differ between Han and Uyghur groups.
With respect to the nature of the languages spoken, Uyghur children are typically educated in both Uyghur and Chinese. The home language (Uyghur) is used until children enter school, but once they reach the third grade, there is a gradual change, whereby Chinese is used in Chinese lessons and Uyghur in others. In contrast, Chinese children are educated only in Chinese. Uyghur belongs to the Turkic language group, which is a branch of the Altaic language family, and uses a phonetic writing system. It involves a completely transparent script and has predictable, bidirectional mapping between orthography and phonology, while Chinese has an ideographic writing system, whereby symbols are used to express words or morphemes in the language, with low transparent orthography [13]. Therefore, children could experience difficulty in one or both languages. To date, knowledge of DD in Uyghur children has been limited relative to that of DD in children acquiring English (low transparent orthography with inflectional morphology). Results of studies focusing on DD in the Han Chinese language (low transparent orthography with isolation or noninflectional morphology) cannot be generalized to Uyghur (high transparent orthography with agglutination in morphology).
In addition, cultural differences could lead to greater diversity in genetic makeup and lower SES in Uyghur people relative to that of the Han population. After the People's Republic of China was founded, many Han people migrated to Xinjiang for military reclamation purposes.
For decades, Han distribution was largely in northern Xinjiang, in the Tianshan veins, whereas Uyghur people gathered in the southern area of the province [14]. In addition, Uyghur people follow the Muslim religion, which prohibits marriage to non-Muslims [14], and formed an isolated genetic group. From a wider perspective, Xinjiang remains relatively undeveloped economically, socially, and educationally, and this is particularly evident in the Uyghur group. Uyghur people typically have low household incomes, and the implementation of a more liberal childbirth policy for this group would exacerbate this situation. The latest statistics indicate that there are approximately 1.9 million students in the 3,533 primary schools in Xinjiang province. Although less well researched, DD also exists in school-age children of Uyghur ethnicity. Zuo et al. [15] noted that the prevalence of DD in Uyghur children was approximately 6.8%. In addition, previous research [4,5,15] has indicated that DD prevalence differs between the Han and Uyghur groups. We conducted this study to examine differences in DD prevalence and potential environmental risk factors between Han and Uyghur children. We also explored racial differences in DD in an attempt to obtain a more accurate pathogenesis of the disorder and provide evidence for further intervention.

Methods Participants
Our study was conducted in the Kashgar and Aksu administrative offices located in the south of Xinjiang province, China, which has a population of 6,685,800 and spans 264,723 square kilometers. In 2013, there were approximately 633,142 students in the 1,552 primary schools in the area. Based on population and economic status, these two regions are classed as mediumsized cities [16]. Therefore, the results of the current study could be generalized to other medium-sized cities in Xinjiang province. As there are numerous primary schools in each region, we used cluster sampling to select 42 primary schools, with a total of 7,452 students in grades 3-6. Of these, we chose 5 Uyghur-Chinese bilingual primary schools, with a total of 2,854 students in grades 3-6, for inclusion in the study.

Materials
In our cross-sectional survey, the questionnaire included the following: 1) The Pupil Rating Scale Revised Screening for Learning Disability (PRS) [17] is a convenient instrument that has been widely used to identify reading disability in China. The questionnaire contains 24 items divided into 5 categories including listening comprehension and memory, language, time and location judgment, athletic ability, and social behavior. It was completed by the head teacher and based on students' performance at school. The mean score and standard deviation of PRS in the present study were 76.77 and 10.21 respectively. 2) The Dyslexia Checklist for Chinese Children (DCCC) [18] is a Chinese rating scale for DD, with well-established reliability and validity. The questionnaire contains 57 items divided into 8 factors: oral ability, written expression, reading habits, attention, visual perception, writing, hearing, and comprehension. The scale uses 5-point scoring criteria, with 1 indicating that the behavior is never observed, and 5 indicating that it always occurs. The participants' Chinese parents (either one or both) completed the questionnaire based on the children's daily performance. Higher scores indicated that the child's ability was considered poorer. Referring to our norm of DCCC, the mean score and standard deviation were 118.87 and 39.75 respectively. 3) The Dyslexia Checklist for Uyghur Children (DCUC) was based on the DCCC and developed using forward and back translation. The questionnaire was found to have good reliability (retest correlation coefficients from 0.67% to 0.76%) and fair validity (criterion validity correlation coefficients from −0.39 to −0.54). The participants' Uyghur parents (either one or both) completed the questionnaire based on the children's daily performance. Referring to our norm of DCUC, the mean score and standard deviation were 130.39 and 42.66 respectively. 4) The Home Literacy Environment and Reading Ability Survey Scale (HLE-RA) [4] contains three factors: general information, home literacy environment, and children's reading habits. The participants' guardians completed the questionnaire to evaluate the children's family background and reading environment. 5) The China-Wechsler Intelligence Scale for Children (C-WISC) [19] classifies children into three groups: those with low intelligence (intelligence quotient [IQ] of <70), critical state (IQ from 70 to 80), and normal intelligence (IQ of >80). The mean score and standard deviation of C-WISC in the present study were 94.83 and 11.47 respectively.

Procedure
This research was conducted over two months, November and December 2014, by researchers who were familiar with the scales and experienced in epidemiology. The head teachers gathered all of the parents (either one or both) in classrooms and explained the aim of the study and demands of the questionnaire. The PRS scale was completed by the head teachers, while the DCCC (for Han children), DCUC (for Uyghur children), HLE-RA scales were completed by parents (either one or both). Results of the C-WISC test indicated that all participants had IQ scores in the typical range. Based on the definition of DD in the International Classification of Diseases 10 (ICD-10) and DD diagnostic criteria in the Diagnostic and Statistical Manual of Mental Disorders fourth Edition (DSM-IV), we adopted a step-by-step screening method for children with DD. The diagnosis of DD was based on the following criteria: 1) PRS score of <65 points; 2) The score of DCCC was two standard deviations higher than the mean scores of Han Chinese children; DCUC score was two standard deviations higher than the mean scores of Uyghur children; 3) IQ of >80 assessed via the C-WISC test; and 4) no visual and/or auditory disorders or psychiatric diseases.

Analysis
Following double data entry, the SPSS12.0 package was used to perform the analyses. Descriptive analysis was performed to obtain means ± standard deviations (Mean ± SD) for quantitative variables and frequencies for qualitative variables. We used logistic regression analysis to examine the relationship between environmental factors and DD diagnosis in different groups. The difference in potential environmental risk factors between Han and Uyghur DD children was examined. Variables found to be statistically significant in the univariate analysis were included in a multivariate logistic regression model. All P values were two-tailed, with a significance level of 0.05.

Ethics statement
Prior to initiation of the research, written informed consent was obtained from all participants and their guardians. Moreover, the study was approved by the ethical committee at the Medical Association of Shihezi University.

Participant characteristics
Overall, 2,854 students from grades 3-6 in the selected schools agreed to participate in the study. However, only 2,511 students returned the questionnaires, with 2,348 completed appropriately (response rate: 93.5%). The participants' sociodemographic characteristics are shown in Table 1. Of the 2,348 participants, 1,156 (49.2%) were Han children, 1,192 (50.8%) were

Factors influencing DD diagnosis in Han and Uyghur children
We analyzed the potential risk factors for DD in Han and Uyghur children in the logistic regression analysis. Factors such as sex, mother's occupation, parental educational level, and most HLE factors differed significantly between children with and without DD in both groups (P < 0.05). However, this difference was not observed for age in either group. In addition, family income, paternal occupation, and maternal educational level were specifically associated with DD in Uyghur children (P < 0.05), and child's educational grade and book-buying expenses were associated with DD in Han children (P < 0.05; Table 2). We then removed the data for children who had not been diagnosed with DD. Table 2 shows the differences in environmental risk factors between Han and Uyghur children with DD, based on univariate analysis. The results indicate that the prevalence rates for DD were 5.5%, 5.0%, 2.4%, and 1.3% in grades 3, 4, 5, and 6, respectively, in Han children with DD; these rates were significantly lower relative to those of Uyghur children with DD (P < 0.05). There was significant difference in family income between the Han and Uyghur children with DD (P < 0.05). With respect to income, 4.3%, 4.3%, 3.3%, 3.7% of families of Han children with DD earned incomes of >3000 RMB, 2000-3000 RMB, 1000-2000 RMB, and <1000 RMB, respectively. These proportions were lower relative to those of families of Uyghur children with DD, of which 3.3%, 7.1%, 9.9%, and 7.6% earned incomes of >3000 RMB, 2000-3000 RMB, 1000-2000 RMB, and <1000 RMB, respectively. Regarding maternal occupation, 26.0% and 1.1% of mothers of Uyghur and Han children with DD, respectively, were service staff members or manual workers. In addition, when mothers were educated to junior college level or above, DD prevalence rates differed significantly between the two groups (P < 0.05).
Moreover, there was significant association between the prevalence difference of DD in this two groups and two HLE factors (e.g., Frequency of parents telling stories (P = 0.01, OR = 2.54) and Children have regular reading time (P = 0.04, OR = 0.46)).

Multivariate logistic regression analysis of the factors influencing DD diagnosis in Uyghur and Han children
Multivariate logistic regression analysis was performed to examine the effect sizes for potential risk factors detected in the univariate analysis (Table 3). For Han children, two risk factors (paternal educational level and frequency with which parents encouraged children to read extracurricular books) were entered into the equation. Similarly, five risk factors (sex; paternal occupation; maternal occupation; frequency with which parents tell stories; and child's daily TV viewing duration) were entered into the equation for Uyghur children. In addition, we included DD prevalence in Uyghur and Han children as a dependent variable and the six potential risk factors as independent variables in the analysis of the sizes of the effects of these factors on DD prevalence rates. The results showed that four risk factors (maternal occupation and frequency with which parents tell stories) were significantly associated with DD prevalence differences between the Han and Uyghur groups. Relative to the differences in the proportions of Han and Uyghur children with DD whose mothers were employed as professional or   technical staff, the difference between proportions of Han and Uyghur with DD children whose mothers were employed as business staff was greatest (P = 0.02, OR = 0.04, 95% CI = 0.01-0.68). This indicates that in those whose mothers were employed as business staff, Uyghur children were 0.04 times more likely to have DD relative to Han children. For parents with occasionally telling stories, the OR to be the prevalence differences of DD was 4.5 times as high as those parents only telling story sometimes.

Discussion
To our knowledge, this study was the first to consider the prevalence of DD and associated risk factors in a large sample of Han and Uyghur children. We also explored the environmental risk factors affecting DD prevalence differences between the Han and Uyghur groups. Our results showed that DD prevalence differed between Han (3.9%) and Uyghur (7.0%) children, and the boy-to-girl diagnosis ratio was almost 2:1. In addition, SES and HLE factors were significantly associated with DD in both groups. Factors such as grade, family income, maternal occupation, maternal educational level, and two HLE factors (maternal occupation and frequency with which parents tell stories) could exert an impact on differences in DD prevalence between the More than 3 hours 0.02 2.66 1.14-6.20 All DD Subjects Environmental Risk Factors for Dyslexia two groups; this was particularly likely for maternal occupation and frequency with which parents tell stories. Because of linguistic and genetic differences between Uyghur, Chinese, and other populations, the DD prevalence varies between different studies. In European populations, the prevalence of DD is 5-12% [2,3]. The corresponding figure for China has been reported as 3.9-8.0% [4]. Zuo et al. showed that the DD prevalence in a Uyghur population was estimated at 6.9% [15]. Similarly, our finding that DD prevalence in the Han group (3.9%) was significantly lower relative to that of the Uyghur group (7.0%) was consistent with those of previous studies. There are a number of possible reasons for this result. For instance, Uighur children receive bilingual Uyghur and Chinese education, and Uyghur is an alphabetical language with a linear onedimensional alphabet [20] that differs from Chinese (ideograph language) in linguistic characteristics [21]. In addition, previous studies have shown that the cerebral regions activated by Uyghur and Chinese language are not identical and found that activation in the left anterior cingulate gyrus could be associated with the Uyghur language [20]. Moreover, research focusing on bilingual Chinese and English individuals showed that when children are educated in two languages, the second language is likely to be influenced by the linguistic and cultural conventions of the first and vice versa [22]. In addition, molecular studies have found that the Uyghur population is of typical mixed genetic origin, with 60% and 40% of the population of European and East Asian ancestry, respectively; this may explain, at least in part, the genetic influence on studied DD children of Uyghur whose prevalence rate is high than Chinese children but lower than European children [23]. Further, the Uyghur population formed an isolated genetic group, because it was not strongly influenced by recent migration [23], and is overwhelmingly of Muslim faith, which prohibits marriage to non-Muslims [14].

Occupation of mother
In the present study, although there was no sex difference observed between groups, the boy-to-girl DD diagnosis ratio was approximately 1:2 in both Han and Uyghur children, and significant differences were observed between sexes in both groups. In Uyghur children, multivariate logistic regression analysis showed that the proportion of girls with DD was 0.51 times that of boys. Our results were consistent with those of Renato, who found that a higher number of boys were diagnosed with DD relative to that of girls [24]. However, Liederman and other researchers considered evidence suggesting a preponderance of boys with DD controversial and weak and posited that it required support from other studies [25,26]. There are several possible reasons for this difference. Previous analyses have indicated that girls enjoy a small advantage in language production until the age of 36 months [27]. Further, in China, particularly in certain ethnic groups such as the Uyghur, many parents view boys as superior; therefore, education and the way the child is raised will depend largely on the child's sex. In terms of academic performance, boys are believed to be good at math, and girls are thought to be good at languages and arts. To some extent, this could reflect sex-related differences linked to the location of the speech-processing center and differences in cognitive development. Our research implicating a important change in intervening DD, education methods should be selected based on children's sex because of the sex difference in DD diagnosis.
Our research also demonstrated that educational grade exerted a significant influence on DD prevalence in Han, but not Uyghur, children. This discrepancy could have occurred because Han children are under greater pressure to enter better middle schools relative to that exerted on Uyghur Children. Moreover, there were significant differences between grades in Han and Uyghur children with DD. Further, DD prevalence showed a downward trend with progression to higher grades. This could be explained by Shaywitz's findings indicating that DD prevalence was lower in higher grades, relative to that in lower grades, because some DD symptoms improve through systematic learning, [28]. However, multiple logistic regression analysis showed that there was no association between grades and risk of DD in Han or Uyghur children. Indeed, the association between grade and DD prevalence remains unclear. In addition, several studies have shown higher DD prevalence in higher grades relative to that observed in lower grades [29]. In order to explore the exact nature of the association between grade and prevalence, further studies should be conducted in the near future.
In previous studies, parental education and occupation and family income represented SES, which has been associated with children's language development [30]. Our results suggest that SES-related factors, such as maternal occupation and paternal educational level, were associated with DD in Han children. In addition, all SES-related factors were associated with DD in Uyghur children. Moreover, in both Han and Uyghur groups, the results showed that higher parental educational levels and family income were related to lower risk of DD. This result could be explained by previous findings indicating that higher SES was related to reduced DD prevalence [4]. Further, DD prevalence in Han children was significantly lower relative to that observed in Uyghur children in family income and maternal education level factors even though SES in Uyghur group are better than that in Han group. This finding could be explained by linguistic and genetic differences between Han and Uyghur populations, which could have led to higher DD prevalence in Uyghur children relative to that observed in Han subjects [20][21][22][23].
Furthermore, results of multiple logistic regression analysis indicated that ethnic differences in DD prevalence between Han and Uyghur children could have occurred because of differences in their mothers' occupations. Relative to the differences in proportions of Uyghur and Han children whose mothers were employed as professional or technical staff, the difference in the proportions of children whose mothers were employed as business staff was greater. This finding indicates that mothers of Han children with DD were employed mainly as business staff, and as such, rarely stayed at home or communicated with their children; therefore, Han children could have received little attention or verbal stimulation, resulting in a developmental lag in reading ability [30]. However, most Uyghur families were involved in farming and animal husbandry, and mothers had more time to take care of their children. Therefore, mothers play an important role in raising children, if mothers remain at home and communicate with their children more frequently, DD prevalence is likely to decrease.
HLE has been found to be one of the most important modifiable risk factors for DD in several studies. Rashid et al. conducted a study involving children with reading disabilities and showed that HLE was significantly associated with children's comprehension and spelling scores [31]. He et al. showed that increasing literacy-related activity and reducing the total time spent using electronic devices was a potentially protective factor for Chinese children at risk of DD [11]. Similarly, the results of the present study highlighted this point, as the risk of DD decreased in children of both groups if their parents provided a better HLE. Further, other factors, such the frequency with which parents tell stories and whether children have regular reading time, were significantly associated with DD prevalence in both Han and Uyghur children. In particular, when parents told stories more frequently, the OR for DD prevalence differences between groups decreased. However, Rashid et al. showed that parents', rather than children's, home literacy activities were significantly associated with several types of academic ability in their children. Therefore, to validate the positive effect of HLE-related factors on DD in both groups, further studies involving Uyghur populations are required.
This study recruited a large sample from two major ethnic groups in the Kashgar and Aksu regions. We collected socioeconomic and environmental variables that were important in examining DD. The results of our study identified risk factors associated with differences in DD prevalence between Han and Uyghur children, and statistical significance for some factors linked to the groups. These findings provide important environmental insight into DD. However, the study was subject to a number of limitations. For example, there are many risk factors associated with DD, and only some of these were analyzed in our study. Details concerning other factors, such as birth history, nursery period, and history of infantile illness, were unknown for the participants in our study. Furthermore, the etiology of DD is complex, and related neurology, genetics, and psychology have been investigated in many studies [2,9,10]. The present study considered only environmental risk factors affecting Han and Uyghur children with DD. Further research is required to understand the molecular mechanisms underlying DD in Uyghur children and whether they differ when compared to other groups.