Font Size: a A A

A Study On Determinants And Discriminant Analysis Forecasting Model Of Cleft Lip And Palate

Posted on:2011-07-03Degree:MasterType:Thesis
Country:ChinaCandidate:C H FuFull Text:PDF
GTID:2154360305993968Subject:Child and Adolescent Health and Maternal and Child Health Science
Abstract/Summary:PDF Full Text Request
Objective:The aim of this study was to explore the determinants of cleft lip and palate, and then construct discriminant analysis forecasting model of cleft lip and palate.Methods:A 1:2 matched hospitals-based case-control study was conducted.113 cases and 226 controls were interviewed with questionnaires. Database was established by Epi Data 3.0 and analyzed by SPSS 13.0. Conditional logistic regression models were used to select determinants of cleft lip and palate. According to results of univariate logistic analysis, stepwise discriminant analysis was adopted to construct fisher linear discriminant analysis forecasting model.Results:Results from univariate logistic regression analysis showed that low education level, suffering from cold and pregnancy complications during early pregnancy, contraceptive taking before pregnancy, exposure to hazardous substance before or during pregnancy, housing renovation before or during pregnancy, regular consumption of pickle, pets contacting, husband's exposure to hazardous substance, husband's suffering from chronic diseases, husband's drinking strong tea, low family income, family heredity history, premature delivery and low birth weight were the risk factors of cleft lip and palate. Pre-marital medical examination, regular consumption of meat and eggs, fruits and vegetables, milk and soya bean milk during early pregnancy were protective factors of cleft lip and palate. As shown in multivariable conditional logistic model, suffering from cold in early pregnancy(OR=2.395), exposure to hazardous substance before or during pregnancy(OR=22.018), regular consumption of milk and soya bean milk during early pregnancy(OR=0.167), family heredity history (OR=20.817), husband's suffering from chronic diseases(OR= 17.677), low family income(OR=3.785), low birth weight (OR=25.532)were significantly associated with cleft lip and palate. The variables entered into forecasting model including exposure to hazardous substance before or during pregnancy(X1), pre-marital medical examination(X2), contraceptive taking before pregnancy(X3), housing renovation before or during pregnancy (X4), regular consumption of milk and soya bean milk during early pregnancy (X5), husband's exposure to hazardous substance (X6), family heredity history (X7) and family income (X8). The function of forecasting model was:Z= 0.366 X1-0.237X2+0.239X3+0.494X4-0.297 X5+0.281X6+0.332X7+0.325X8. Self-validating, jackknife-validating and ROC curve analysis were used to evaluate validity of discriminant analysis forecasting model. The accuracy of self-validating of forecasting model was 77.6%, the sensitivity was 62.8%, the specificity was 85.0%; the accuracy of jackknife-validating was 77.3%, the sensitivity was 61.9%, the specificity was 85.0%; area under ROC curve of forecasting model was 0.796. Conclusion:Genetic factor is an important cause of the development of cleft lip and palate. In addition, environmental factors such as family economic condition, maternal exposure to hazardous substance, maternal nutritional status and husband's health status are associated with the development of cleft lip and palate. Discriminant analysis forecasting model can forecast the occurrence of cleft lip and palate in the term of epidemiology.
Keywords/Search Tags:Cleft lip and palate, Case-control study, Determinants, Conditional logistic regression, Discriminant analysis forecasting model
PDF Full Text Request
Related items