| ObjectiveShort stature in children is identified as the important issue among global health.The aim of this study is analyzing the prevalence and influencing factors for short stature,constructing a nomogram prediction model and risk classification system for short stature,as well as providing screening and assessment of short stature individually.MethodsThis study was a cross-sectional study which was including 12504 children aged 6-14 from 13 primary and secondary schools in Pingshan district,Shenzhen.Physical examination was used to measure the height and weight of children.Questionnaires were used to obtain the information about children,including gender,age,family environment,social environment,maternal conditions during pregnancy,born and fed,and life style.SPSS 25.0 statistical software was used to analyze the characteristics of height distribution and the prevalence of short stature in Pingshan district.Adjust age confounding variables via 1:1 Propensity Score Matching(PSM)analysis and 1076 children were selected in the influencing factor analysis.Chi-square test and multivariate logistic regression analysis were used to screen the influencing factors for short stature.R studio statistical analysis software was utilized for establishing a clinical prediction model and risk classification system for short stature.Results1.For children in Pingshan district,Shenzhen,the average height of boys is higher than girls at 6-9 years old and 12-14 years old,and the average height of girls is higher than boys at 10-11 years old.2.For the same age and gender,there is no significant difference between the height of children in Pingshan district compared with the height of 2005 Chinese data.3.The prevalence of short stature in children from 6 to 14 years old is 4.3%in Pingshan district.The prevalence of short stature in boys is 4.1%,the prevalence of short stature in girls is 4.5%.The prevalence of short stature in girls is slightly higher than that of boys,but there is no significant difference in the prevalence of short stature between different sexes.The prevalence of short stature in children peaks at the age of 9,and then gradually decreases with age.When the age past 10,the decline is significant,and there is a significant difference between ages among short stature.4.Multivariate logistic regression model shows that the influencing factors of short stature are:father’s height,mother’s height,annual family income,father’s education level,and concern about their children’s height in future by parents(P<0.05).5.Based on the multivariate logistic regression model of short stature,a nomogram prediction model of short stature is constructed for visualizing the influencing factors for short stature.Evaluate the degree of discrimination of the nomogram by ROC curve.The model presents a good degree of discrimination because the area under the ROC curve(AUC)value is 0.748.According to the calibration curve,the Hosmer-Lemesio test value is 0.917 and the model is accuracy.6.Based on a risk classification system which stemmed from nomogram prediction model for short stature,the total score of the nomogram is 127.5 and regards as the cut-off point,which divides all children into low-risk group and high-risk group.ConclusionThis research reveals the prevalence of short stature in children from Pingshan district,Shenzhen via a cross-sectional survey.Because of limitation of preliminary assessment methods for short stature,this study supplements the existing preliminary assessment methods for short stature,by constructing a nomogram prediction model and risk classification system based on the influencing factors of short stature. |
PDF Full Text Request