Home About us Contact
|
Home About us Contact | ||
|
|
||
Fractional Polynomials (fractional + polynomial)
Selected AbstractsEstimating herd-specific force of infection by using random-effects models for clustered binary data and monotone fractional polynomialsJOURNAL OF THE ROYAL STATISTICAL SOCIETY: SERIES C (APPLIED STATISTICS), Issue 5 2006Christel Faes Summary., In veterinary epidemiology, we are often confronted with hierarchical or clustered data. Typically animals are grouped within herds, and consequently we cannot ignore the possibility of animals within herds being more alike than between herds. Based on a serological survey of bovine herpes virus type 1 in cattle, we describe a method for the estimation of herd-specific rates at which susceptible animals acquire the infection at different ages. In contrast with the population-averaged force of infection, this method allows us to model the herd-specific force of infection, allowing investigation of the variability between herds. A random-effects approach is used to account for the correlation in the data, allowing us to study both population-averaged and herd-specific force of infection. In contrast, generalized estimating equations can be used when interest is only in the population-averaged force of infection. Further, a flexible predictor model is needed to describe the dependence of covariates appropriately. Fractional polynomials as proposed by Royston and Altman offer such flexibility. However, the flexibility of this model should be restricted, since only positive forces of infection have a meaningful interpretation. [source] Maternal blood glucose in diabetic pregnancies and cognitive performance in offspring in young adulthood: a Danish cohort studyDIABETIC MEDICINE, Issue 7 2010G. L. Nielsen Diabet. Med. 27, 786,790 (2010) Abstract Aims, Maternal diabetes is a known risk factor for perinatal complications, but there are little data on consequences for long-term intellectual outcome in offspring. We assess cognitive performance in military conscripts according to maternal blood glucose levels during pregnancy. Methods, We identified a cohort of 60 Danish male offspring of insulin-treated diabetic mothers born between 1976 and 1984 and followed this cohort to military conscription. From medical records, we extracted data on all available values of maternal blood glucose categorized as fasting and non-fasting and by day in pregnancy, together with maternal White class, smoking habits and socio-economic status. The main outcome was cognitive performance at conscription measured with a validated intelligence test. The association between maternal blood glucose level and cognitive performance was assessed by multivariate linear regression and a fitted fractional polynomial. Results, Median fasting blood glucose values in the second half of pregnancy was negatively associated with cognitive scores at conscription [adjusted coefficient ,1.7; 95% confidence interval (CI) ,3.0; ,0.4]. Restriction to only first-born sibling slightly strengthened the association (coefficient ,1.9; 95% CI ,3.3; ,0.5), but after exclusion of two pregnancies with the blood glucose > 10 mmol/l the association became insignificant (coefficient ,0.6; 95% CI ,2.6; 1.4). Conclusions, Maternal blood glucose level during diabetic pregnancy is negatively associated with cognitive performance in offspring at military conscription. In pregnancies with fasting blood glucose levels below 10 mmol/l, the association is weak and considered to be without clinical relevance. [source] Alcohol and hypertension: gender differences in dose,response relationships determined through systematic review and meta-analysisADDICTION, Issue 12 2009Benjamin Taylor ABSTRACT Aims To analyze the dose,response relationship between average daily alcohol consumption and the risk of hypertension via systematic review and meta-analysis. Design A computer-assisted search was completed for 10 databases, followed by hand searches of relevant articles. Only studies with longitudinal design, quantitative measurement of alcohol consumption and biological measurement of outcome were included. Dose,response relationships were assessed by determining the best-fitting model via first- and second-degree fractional polynomials. Various tests for heterogeneity and publication bias were conducted. Findings A total of 12 cohort studies were identified from the literature from the United States, Japan and Korea. A linear dose,response relationship with a relative risk of 1.57 at 50 g pure alcohol per day and 2.47 at 100 g per day was seen for men. Among women, the meta-analysis indicated a more modest protective effect than reported previously: a significant protective effect was reported for consumption at or below about 5 g per day, after which a linear dose,response relationship was found with a relative risk of 1.81 at 50 g per day and of 2.81 at an average daily consumption of 100 g pure alcohol per day. Among men, Asian populations had higher risks than non-Asian populations. Conclusions The risk for hypertension increases linearly with alcohol consumption, so limiting alcohol intake should be advised for both men and women. [source] Estimating herd-specific force of infection by using random-effects models for clustered binary data and monotone fractional polynomialsJOURNAL OF THE ROYAL STATISTICAL SOCIETY: SERIES C (APPLIED STATISTICS), Issue 5 2006Christel Faes Summary., In veterinary epidemiology, we are often confronted with hierarchical or clustered data. Typically animals are grouped within herds, and consequently we cannot ignore the possibility of animals within herds being more alike than between herds. Based on a serological survey of bovine herpes virus type 1 in cattle, we describe a method for the estimation of herd-specific rates at which susceptible animals acquire the infection at different ages. In contrast with the population-averaged force of infection, this method allows us to model the herd-specific force of infection, allowing investigation of the variability between herds. A random-effects approach is used to account for the correlation in the data, allowing us to study both population-averaged and herd-specific force of infection. In contrast, generalized estimating equations can be used when interest is only in the population-averaged force of infection. Further, a flexible predictor model is needed to describe the dependence of covariates appropriately. Fractional polynomials as proposed by Royston and Altman offer such flexibility. However, the flexibility of this model should be restricted, since only positive forces of infection have a meaningful interpretation. [source] Fetal size charts for the Italian population.PRENATAL DIAGNOSIS, Issue 6 2005Normative curves of head, abdomen, long bones Abstract Objective To describe size charts developed from fetuses of Italian couples. Method Prospective cross-sectional investigation conducted in three referral centers for prenatal diagnosis. The population of the study included fetuses between the 16th and the 40th week of gestation recruited prospectively and examined only once for the purpose of this study. Exclusion criteria comprised all maternal and/or fetal conditions possibly affecting fetal biometry. The following biometric variables were measured: biparietal diameter, head circumference, abdominal circumference, femur, tibia, humerus, ulna and radio length. The statistical procedure recommended for analyzing this type of data set was employed to derive normal ranges and percentiles. Birthweight was also recorded. Our centiles were then compared with results from other studies. Results The best-fitted regression model to describe the relationships between head circumference and abdominal circumference and gestational age was a cubic one, whereas a simple quadratic model fitted BPD, and length of long bones. Models fitting the SD were straight lines or quadratic curves. Neither the use of fractional polynomials (the greatest power of the polynomials being 3) nor the logarithmic transformation improved the fitting of the curves. Conclusion We have established size charts for fetuses from Italian couples using the recommended statistical approach. Since the mean birthweight in this study is not statistically different from the official birthweight reported for the Italian population, these reference intervals, developed according to the currently approved statistical methodology, can be employed during second- and third-trimester obstetric ultrasound of fetuses from Italian couples. Copyright © 2005 John Wiley & Sons, Ltd. [source] Logistic Regression Models for Polymorphic and Antagonistic Pleiotropic Gene Action on Human Aging and LongevityANNALS OF HUMAN GENETICS, Issue 6 2003Qihua Tan Summary In this paper, we apply logistic regression models to measure genetic association with human survival for highly polymorphic and pleiotropic genes. By modelling genotype frequency as a function of age, we introduce a logistic regression model with polytomous responses to handle the polymorphic situation. Genotype and allele-based parameterization can be used to investigate the modes of gene action and to reduce the number of parameters, so that the power is increased while the amount of multiple testing minimized. A binomial logistic regression model with fractional polynomials is used to capture the age-dependent or antagonistic pleiotropic effects. The models are applied to HFE genotype data to assess the effects on human longevity by different alleles and to detect if an age-dependent effect exists. Application has shown that these methods can serve as useful tools in searching for important gene variations that contribute to human aging and longevity. [source] | |||||||||||||