Relationship between Anemia and Academic Performance in Chinese Primary School Students: Evidence from a Large National Survey

,


Introduction
Anemia is a worldwide public health problem and especially afects people in developing countries.Te most susceptible populations are children and women [1,2], with relatively few studies focusing on children with anemia.According to WHO estimates, the prevalence of anemia among children under 5 years old in China was 18.8% in 2019 [3], which is lower than the global prevalence of 39.8% for children under 5 [4].However, this rate is lower compared to developing countries like South Africa (44.4%) and India (53.4%), but higher compared to developed countries like the United States (6.1%), Germany (15.1%), and Japan (16.7%).Social and economic development in China has brought about a decrease in the prevalence of anemia in recent years [5]; however, there are regional diferences among the population.Survey results indicate that the prevalence of anemia among children is higher in Western China and in the female population [6].Te prevalence of anemia in children decreases with age before puberty and increases during puberty [5].Schoolgoing age is a key period of physical and mental development for children and an important period for acquiring new knowledge.Anemia may cause irreversible impairment of learning ability in children.As estimated by the World Health Organization, iron defciency anemia is the primary cause of loss of disability-adjusted life years among 10-14-year-old children [7].
However, childhood anemia, which is common, has not received enough attention in China, particularly in rural areas.With the recent slowdown in China's economic development and the particularly challenging job market for young people, exacerbated by the COVID-19 pandemic, the pressure on education and the rapid devaluation of educational qualifcations has extended to primary schools and even kindergartens.Tis has intensifed parents' focus on students' academic performance and investments in education.Te main objective of this research is to draw more attention from parents to the potential health issue of childhood anemia in China through examining its relationship with academic performance.It also aims to provide an explanation and possible solutions for children with poor academic performance due to anemia.By addressing the issue of childhood anemia and its impact on academic performance, we strive to support efective interventions and strategies that can improve both the health and educational outcomes of afected children.In the long run, it is also benefcial for the accumulation of human resources and the promotion of sustainable socioeconomic development.
Tere is a long history of academic performance research among children, but only a limited number of studies have been devoted to investigating the relationship between anemia and academic performance.Based on the literature published in the past, although there were difering results regarding the relationship between anemia and academic performance, more studies tended to lean towards an association between the two, especially in recent research.For instance, a recent study systematically reviewed the relationship between iron status, anemia, iron interventions, cognition, and academic performance in adolescents aged 10-19 years [8].Tis review included 50 relevant publications that were published prior to 2020, consisting of 26 cross-sectional observational studies and 24 iron intervention studies.It indicated that in certain contexts, iron status and anemia were related to academic performance, and iron supplementation during adolescence might improve academic performance and attention.However, the review found that the quality rating of all cross-sectional studies was poor, and almost all supplementation trials showed a bias towards medium to high risk.Additionally, it was also found that there was no evidence of a relationship between iron status or anemia and attention, intelligence, or memory among adolescents.On the other hand, a diferent conclusion was reached in a more recent meta-analysis [9].Based on 13 randomized controlled experiments, this study found that iron supplementation signifcantly improved intelligence, attention, and memory in children aged 6-12 years, though no relationship was observed between iron supplementation and academic performance.Given the higher prevalence of anemia in middle-and low-income countries, researchers conducted a meta-analysis focusing on these settings [10].Based on 9 studies from 5 countries, this analysis demonstrated that increasing iron supplementation dosage could improve intelligence test scores, but had no efect on attention, short-term memory, long-term memory, or academic performance.Moreover, there were two large-scale observational studies, one crosssectional and one longitudinal, which supported an association between anemia and academic performance in adolescents [11,12].
Regarding studies conducted in China on the relationship between anemia and academic performance, the literature was relatively scarce.A survey conducted in the 1980s involved 478 middle school students aged 13-15 in Shanghai, and it did not fnd a relationship between iron defciency anemia and IQ or academic performance [13].However, recent studies have shown positive associations.One study included 2,646 students aged 7-14 in Taiyuan, and the results showed a positive correlation between academic performance and hemoglobin levels [14].Another study conducted in Gansu province in Western China included 45,549 students from grades 1-5.Using instrumental variable estimation, the study found a signifcant negative impact of iron defciency on children's academic performance [15].Furthermore, a study conducted in Henan province, located in central China, surveyed 493 fourthgrade students and found that anemic students were more likely to have poor scores on standard math exams [16].Te few studies published to date regarding anemia and academic performance in children in China have been limited by (1) relatively small sample sizes [13,16] that were concentrated in specifc regions of China [13][14][15][16].Tis resulted in low statistical power, and the results could not be extrapolated to the general children population throughout China; (2) some surveys using an age range for children that was too wide, with relatively small sample sizes for certain ranges.Tis provided insufcient data to explain the relationship between anemia and academic performance among children of specifc ages [16]; and (3) some studies focusing on the relationship between anemia and academic performance or cognition, without exploring the mechanism or mediation analysis between them [13][14][15][16].Exploring the relationship between anemia status and academic performance would provide valuable insight.
Tis study involved four mediating variables, namely cognition, BMI, sleep duration, and CESD score.Currently, there was no literature reporting the mediating efects of these variables on the relationship between anemia and academic performance.Te notion of mediating variables was more commonly utilized in the felds of sociology or psychology, while epidemiological research predominantly focused on establishing the connection between anemia and academic performance, with little exploration of potential mechanisms.However, based on existing published literature, these four variables had the potential to play signifcant roles in the relationship between anemia and academic performance.In short, frstly, there had been numerous reports, especially recently, both domestically and internationally, on the relationship between anemia and cognition [17][18][19].Moreover, there was a general consensus that iron supplementation improved cognitive levels [9,10].Furthermore, many studies investigating the relationship 2 Health & Social Care in the Community between anemia and academic performance also simultaneously discussed cognitive levels [9,10], and the association between cognition and academic performance was evident.Terefore, cognition level was the frst mediating variable to be considered.Secondly, there was evidence from metaanalyses demonstrating the relationship between childhood obesity and dual burden of micronutrient defciency, as well as the relationship between obesity and anemia [20].Similarly, meta-analyses had also shown a negative correlation between BMI and academic performance [21].Terefore, BMI might have exerted a mediating efect.Tirdly, as primary school students, sleep duration played a crucial role as a measure of their health behaviors.Previous population studies indicated a relationship between anemia and sleep duration in adults [22,23], and studies on children found an association between sleep and academic performance [24].Tis suggested that sleep duration might have had a mediating efect on the association between anemia and academic performance.Lastly, among the main symptoms caused by anemia were fatigue, lethargy, and depression [25].Te relationship between depression and academic performance had been reported in cross-sectional studies [26], and it had been found that this association could persist long term [27].
Terefore, depression might have also acted as a mediating variable.In light of this, the hypotheses to be tested in this study were as follows: (1) Anemia was negatively correlated with academic performance.(2) Anemia was negatively correlated with cognitive levels, and cognition mediated the relationship between anemia and academic performance.(3) Anemia was positively correlated with BMI, and BMI mediated the relationship between anemia and academic performance.(4) Anemia was positively correlated with sleep duration, and sleep duration mediated the relationship between anemia and academic performance.( 5) Anemia was positively correlated with depression, and depression mediated the relationship between anemia and academic performance (see Figure 1).Tis study used the baseline data of primary school students who completed the China Education Panel Survey in 2018 to analyze the relationship between anemia and academic performance and further explored the mediation efect of health-related factors.Te goal was to provide scientifc evidence for the relationship between anemia and academic performance and the possible factors mediating this association.

Survey Introduction.
Tis study was based on primary school student cohort data from the China Education Panel Survey.Te feldwork of the cohort was conducted by the National Survey Research Center at the Renmin University of China from September 2018 to June 2019.Te survey's main goal was to elucidate the efects of family, school, community, and social structure on educational attainment, as well as educational infuence on a child's life course.A probability proportional to the size sampling method in four stages was used to sample students.Te baseline survey sample covered all fourth-grade students in 160 primary schools located across 40 counties within 20 provinces throughout China.A paper-based, self-administered survey with instruction from a supervisor was conducted for all respondents; respondents included students, their fathers or mothers, head teachers, main teachers, and school masters.Te student questionnaire included information on demographics, physical and mental health status, life experience, parents' involvement, Hukou and immigration, learning at school, after-school activities, parents' educational expectations, relationships with teachers and classmates, social behavior, and family members' information.Te survey tested students' cognition and collected the results of the previous mid-term and fnal examinations.Te physical parameters measured included height, weight, visual acuity, and hemoglobin levels.All physical measurements were performed by medical staf trained by the Chinese Center for Disease Control and Prevention.
Student assent and parental consent were obtained before conducting the survey.Informed consent was provided and the strict confdentiality of information was guaranteed.Te study was approved ethically by the Biomedical Ethics Review Board of the Chinese Center for Disease Control and Prevention and conducted in accordance with the Declaration of Helsinki and its later amendments.

Study Variables
(1) Te dependent variables were the fnal examination results from the previous semester for the subjects Chinese, Mathematics, and English; the average score of all the three subjects was calculated.To adjust for diferent test contents in diferent schools, the standardized scores for each student were calculated by school and the standardized scores were normalized back in the fnal data analysis.(2) Te main independent variable was anemia.Te existence of anemia was determined by measuring hemoglobin levels in blood samples taken from the fngerstick blood test, adjusted by local altitude [28].Studies Depression Scale that included fve questions regarding physical symptoms, fve regarding depression, and three regarding positive emotions [30].All questions had four options: none (0 points), a little (1 point), some (2 points), and quite a lot (3 points).Te total score was the sum of the 13 questions, in which the scores for the questions regarding positive emotions was reversed as none (3 points), a little (2 point), some (1 points), and quite a lot (0 points): a higher score indicates a more severe state of depression.Te cognition test-characterized by international comparability and national standardization [31]-focused on logical thinking and problem-solving capacity.Graphic reasoning was tested with 30 questions that had to be answered within 25 minutes.Ten questions relating to an expanded graph had to be answered within 15 minutes.According to the number of correct responses, the score ranged from 0 to 80, with higher values indicating better cognition.

Statistical Analysis.
For descriptive analyses, continuous variables were expressed as means and standard deviations.Te t-test, analysis of variance, or nonparametric test was used to compare students with anemia and students without anemia.Categorized variables were expressed as percent (%) and the Chi-square test was administered to compare between the two groups.
Ordinary least squares (OLS) regression was used to analyze the relationship between anemia and academic performance, and multivariate analysis was adjusted by sex, age, Hukou type, only child or not, family socioeconomic status, the highest educational attainment of parents, number of books in the family, parents' educational expectations, learning pressure, student-student relationships, and teacherstudent relationships.Te fxed efect of school was considered in the regression analysis to control for the school efect after the academic performance was standardized.Analyses stratifed by sex and age grouping were likewise performed in the multivariable OLS regression analyses.
Mediation analysis was used to explore the possible health-related factors mediating the association between anemia and academic performance [32].Te mediators included the cognition score, BMI, sleep duration, and degree of depression.Briefy, there were several steps for the mediation analysis.Firstly, the previous statistical relationship analysis (model 3) between each test score and anemia was the basis for further mediation exploration.Secondly, models 13 were constructed based on model 3 by adding four mediators, respectively.Te mediation efect was examined using "sgmediation" code in Stata.Te bootstrap method was used to calculate the confdence intervals of the associated products of the mediation analyses.Finally, the direct and indirect efect sizes were calculated.When examining the mediating efects in the analysis, sex, age, Hukou type, only child or not, self-assessed economic  Health & Social Care in the Community status, the highest educational attainment of parents, number of books in the family, parents' educational expectations, learning pressure scores, relationships with other students, teacher-student relationship scores, and school fxed efect were controlled.All statistical analyses were performed using Stata statistical software (version 16.0, College Station, TX, USA).

Social Demographic and Health-Related Characteristics.
Te study involved a total of 36,826 respondents, but 18,543 students did not participate in the hemoglobin measurement.After excluding 588 students who were not 10 or 11 years old, a total of 17,695 fourth-grade students were ultimately used for this study (see Figure 2).Table 1 presents the social demographic and health-related characteristics of the study population.Te prevalence of anemia was 6.52%, with no diference in prevalence between males and females (P > 0.05).Anemia was more prevalent among rural Hukou children.Higher family economic status, larger number of books in the family, and higher parents' educational expectations were associated with a lower prevalence of anemia in children, as were better student-student or teacherstudent relationships.High learning pressure was associated with a high prevalence of anemia.Children with upper quantile cognition scores or body mass indices had the lowest prevalence of anemia.Longer sleep duration and higher CESD score were associated with a higher prevalence of anemia (all P < 0.05).

Test Score Distribution.
Te average test scores of the three subjects was 54.34 ± 0.18 and 51.04 ± 0.71 in the nonanemia and anemia groups, respectively, as shown in Table 2. On average, children in the nonanemia group had higher average scores than those in the anemia group within categories of sex, age, Hukou type, only child.As shown in Figure 3, there was a left-skewed distribution with the nonanemia group having a higher proportion of higher scores than the anemia group.
Te test scores for Chinese, Mathematics, and English, were higher for children in the nonanemia group than for children in the anemia group by subtotal, sex, age, Hukou type, and only child (see Figure 4 and Table 2); the distribution was almost similar to the average test score of the three subjects (see Figures S1-S3 in the supplementary material).3, univariate and multivariate OLS regression analysis found that anemia status was negatively associated with the average test score of the three subjects after they were adjusted by sex, age, Hukou type, only child or not, self-assessed economic status, the highest educational attainment of parents, number of books in the family, parents' educational expectations, learning pressure scores, relationships with other students, and teacher-student relationship scores.Tis negative association was also found with each of the three subjects.

Multivariate Analysis. As shown in Table
Univariate and multivariate OLS regression analyses were conducted on the association between anemia status and academic performance according to sex and age grouping.Anemia status was associated with a decrease of average, Chinese and Mathematics test score in female students within multivariate OLS regression analyses (P < 0.01), but not in male students.After stratifying by age group, the association between anemia status and average, mathematics and English test score remained in the subpopulations aged 11 years (P < 0.05), but lack of association was found in students aged 10 years.4 used the average test score as the dependent variable and added the independent variables of cognition score, BMI, sleep duration, and CESD score to explore the direct efect of anemia and the indirect efect of the four independent variables.Te results showed that anemia had a direct efect on the average test score (P < 0.05), and it afected the average test score indirectly by decreasing the cognition score (P < 0.05); the indirect efect accounted for 19.9% of the total efect.No indirect efect on the average test score was found for BMI, sleep duration, and CESD score.In addition, the mediation models were conducted for test scores of Chinese, Mathematics, and English respectively, and the results were almost similar to the average test score of the three subjects (see Tables S1-S3 in the supplementary material).

Discussion
Tis study utilized a large-scale nationally representative dataset to analyze the prevalence of anemia among fourthgrade primary school students in China in 2018.It revealed the current situation of anemia among fourth-grade students in China and reported diferences in anemia prevalence among diferent socioeconomic statuses.Additionally, this study provided empirical evidence from a large population survey for the hypothesis that anemia afects academic performance, which has not only direct but also indirect efects through cognition score.Tis study not only provides fundamental data for researching the prevalence of anemia among children in China but also ofers valuable insights into the factors infuencing children's academic performance.It serves as a valuable reference for future policymaking related to the prevention and control of anemia among children.
Tis study showed that the prevalence of anemia is 6.52% in Chinese Grade 4 students, which is similar to the results of other current national surveys on the prevalence of anemia in China.Te China Nutrition and Health Surveillance data showed that the prevalence of anemia was 4.01% and 3.25% in children aged 10 and 11 years, respectively [33].Data from the Chinese National Survey on Student Constitution and Health revealed that the prevalence of anemia in children aged 9 and 12 years was 6.8% and 9.6%, respectively [34].Te reasons for the diferent results may be associated with diferent survey sampling and methods for testing for anemia, as well as diagnostic criteria.In addition, surveys Health & Social Care in the Community conducted in the poorer or western regions of China reported a higher prevalence of anemia than other regions.A 2019 nutritional improvement program survey for compulsory education for students in rural areas showed that the prevalence of anemia was 8.3% among 11 to 13-year-old children [35].Te prevalence of anemia among girls was higher than that among boys.According to data from the 2015-2017 China Nutrition and Health Surveillance, the prevalence of anemia was 8.5% among girls aged 10-17 years old [36].
Te current study found that the average test score of children without anemia was higher than that of children with anemia, and each individual subject had similar results (all P < 0.05).Anemia was negatively associated with average test score adjusted by sex, age, Hukou type, only child, selfassessed family economic status, highest educational attainment of parents, number of books in the family, parents' educational expectations, learning pressure scores, and good relationships with other students.On average, the average test score of children with anemia was 1.96 points lower than that of children without anemia.A negative association between anemia and test scores was also found individually for the subjects Chinese, Mathematics, and English.Tese results indicated we detected total efects of anemia status on academic performance.Tese total efects were found only in female or subpopulation aged 11 years, which meant anemia status may mainly afect the girls and children in puberty.Our fndings were similar to those of other studies in China and abroad [15,16,37] and provided new evidence for the relationship between anemia and academic performance.
According to current studies, the negative impact of anemia on academic performance may be associated with brain dysfunction.Anemia can decrease the oxygen saturation of blood in the brain [38] and cause silent infarcts [39] as well as cognitive dysfunction, which manifests as pallor, listlessness, behavioral disturbances, impaired performance in some cognitive tasks, et cetera [40].All these syndromes afect academic performance in children.It is worth mentioning that the current study mainly focused on children aged 10 and 11 years, but the age at which onset of anemia occurred was difcult to determine.Onset of iron defciency anemia at a young age increases the possibility of mental retardation in later life [41].
To explore the mechanism underlying the relationship between anemia and academic performance, this study analyzed the mediating efects of cognition, BMI, sleep duration, and CESD score.After mediation analysis, anemia was shown to afect academic performance directly (P < 0.05) and, by decreasing the cognition score, indirectly (P < 0.05).Te indirect efect was 19.9%.Te results provide statistical support for the causal chain of anemia-cognition dysfunctiondecreasing academic performance.Tis result was similar to other studies in China and abroad [42][43][44]; however, additional high quality research to confrm it is required.
Te study had several advantages: (1) it was the frst nationwide study to explore the association between anemia and academic performance in China.Except for age, the study population had a high level of heterogeneity and thus the results are generalizable to the entire Chinese population.Te heterogeneity also provided the study with stronger evidence with higher statistical power; (2) it was based on a survey that focused on Grade 4 students, which narrowed the age range of the study population.Tis narrow range was valuable in removing any interference that age may have imposed on the results, thus beneftting the exploration of the association between anemia and academic performance; and (3) it frst explored the mediating efect of cognition on the relationship between anemia and academic performance based on nationwide survey data.Although the results were still statistically signifcant, this provided a possible explanation for the relationship between anemia and academic performance.6 Health & Social Care in the Community   Health & Social Care in the Community (2) the analysis used baseline data from the China Education Panel Survey for primary school students, so the association between anemia and academic performance shows a cross-sectional relationship, not a causal efect.More data from a follow-up survey are needed to provide stronger evidence for this relationship; and (3) it only explored health-related mediation factors such as cognition score, BMI, sleeping duration, and CESD score.Other factors might have been missed that also could have mediated the relationship between anemia and academic performance.

Conclusions
Tis study analyzed the data of primary school students who participated in the China Education Panel Survey.Te fndings showed that the prevalence of anemia in children aged 10 and 11 years was 6.52%, and the prevalence was diferent among students with diferent characteristics.Tis indicates that anemia is a very important health issue among Grade 4 students in China, even in heterogeneous populations.Analyses showed that, in the study population, anemia was associated with academic performance, and that it afects academic performance both directly and indirectly.More nutritionrelated interventions are needed to reduce the prevalence of anemia among children to ensure a healthy population base and guarantee human capital accumulation in China [45].

Figure 1 :
Figure 1: Hypothetical model diagram of the study.

Figure 2 :
Figure 2: Sample selection process diagram for analysis in this study (China, 2018).

Figure 4 :
Figure 4: Comparison of the average and individual test score of three subjects between the nonanemia and anemia groups (China, 2018).

Table 1 :
Social demographic and health-related characteristics of the enrolled participants (China, 2018).

Table 2 :
Comparison of test score distribution between nonanemia and anemia groups by social and demographic characteristics (China, 2018).

Table 3 :
Univariate and multivariate ordinary least squares (OLS) analysis regarding the association between anemia and academic performance by total, sex and age group (China, 2018).Group Health & Social Care in the Community Te study's limitations are that (1) the sample population was Grade 4 students of 10 to 11 years old, so the fndings can only be representative of children in this age group and cannot be extrapolated to other age groups;

Table 4 :
Mediation efect analysis of factors mediating the association between anemia and average test score (China, 2018).(−2.93, −0.22) a −1.92 (−3.33, −0.50) a −2.00 (−3.43, −0.58) b −2.17 (−3.59, −0.75) b 13-16 were based on the dependent variable of average test score, adjusted for factors such as sex, age, Hukou type, only child, self-assessed economic status, highest educational qualifcation of parents, number of books in the family, parents' educational expectations, learning pressure score, good relationships with other students, teacher-student relationship score, and fxed efects at the school level.Tis table only reported the coefcients of the variables related to anemia and mediation.a P < 0.05; b P < 0.01; c P < 0.001.Central Universities and the research funds of Renmin University of China.Program name: China Education Panel Survey. Models