Search PubMed⌕ Search

Biomedical subjects

K F Yu

Publications and source records attributed to K F Yu.

14 recordsLinked to original sources
PubMed1998

What's the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes.

Logistic regression is used frequently in cohort studies and clinical trials. When the incidence of an outcome of interest is common in the study population (>10%), the adjusted odds ratio derived from the logistic regression can no longer approximate the risk ratio. The more frequent the outcome, the more the odds ratio overestimates the risk ratio when it is more than 1 or underestimates it when it is less than 1. We propose a simple method to approximate a risk ratio from the adjusted odds ratio and derive an estimate of an association or treatment effect that better represents the true relative risk.

Cohort Studies↗
PubMed1997

Adolescent age at first pregnancy and subsequent obesity.

Adolescent pregnancy has been associated with subsequent obesity. This paper examines the patterns of obesity for second and third pregnancies among women who had their first singleton pregnancy as teenagers. We used maternally-linked data from 1978 to 1990 among 43,160 Missouri resident women. Age, parity, interpregnancy interval and prior body mass index were significantly associated with subsequent obesity among adolescents. Race, marital status and smoking had significant interactions with age. Among older women, being African-American and never having married was associated with an increased probability of obesity, and smoking had a greater effect on obesity at higher maternal age. Race and marital status did not have significant effects on obesity among younger women. The most important predictor of obesity was prior body mass index. Body mass index before the first pregnancy had a greater effect on subsequent obesity if the intervening interpregnancy weight gains were large. Excessive weight gain during pregnancy presents the health care provider with a dilemma. An increase in birthweight associated with high prenatal weight gains may diminish the risk of infant mortality and morbidity in an index pregnancy, but subsequent obesity may increase perinatal mortality rates, the rates of obstetric problems and neural tube defects.

Adolescent↗
PubMed1997

Does changing paternity contribute to the risk of intrauterine growth retardation?

An immune reaction initiated by paternal antigens may be necessary for healthy placental development, pregnancy maintenance and infant growth. An inadequate immune response may result in intrauterine growth retardation (IUGR). We hypothesised that a change in paternity may interfere with the immune response and cause poor placentation with resultant IUGR. In this paper we examine the risk of IUGR associated with changes in paternity. We used the Utah Successive Pregnancies Data Set that contains information on women across their pregnancies. We restricted the analysis to 141,817 women with two or three pregnancies. Women who did not have an IUGR infant in the previous pregnancy were at a 20-30% greater risk of developing IUGR if they had changed partners. Women who had a previous IUGR infant were at no increased risk for IUGR after a change in paternity. These results may point to an immune mechanism or may be as a result of residual confounding of unmeasured risk factors for IUGR. Further studies with data that contain more sociodemographic and biological risk factors for IUGR are necessary to exclude residual confounding.

Birth Order↗
PubMed1996

Evaluation of the effectiveness of a community-based enriched model prenatal intervention project in the District of Columbia.

OBJECTIVE: To evaluate an enriched prenatal intervention program designed to reduce the risk of low birth weight. STUDY SETTING: Freestanding community-based prenatal intervention project located in a poor inner-city community, serving mostly African American women. STUDY DESIGN: All women less than 29 weeks pregnant were eligible to participate. They were compared to women who lived in neighborhoods with similar rates of poverty. DATA COLLECTION: The birth certificate was the source of data on maternal age, education, marital status, timing and frequency of prenatal care attendance, parity, gravidity, prior pregnancy terminations, fetal and child deaths, and birth weight. PRINCIPAL FINDINGS: Thirty-eight percent of the women who delivered live-born infants in the study area participated in the program. There were no differences in low- and very low birthweight rates in the study and comparison groups. In a secondary analysis comparing participants and nonparticipants in the study census tracts, participants were at higher risk for low and very low birth weight, and they adhered more closely to the schedule of prenatal visits than nonparticipants. Low- and very low birthweight rates were lower among participants than among nonparticipants and comparison women. CONCLUSION: The Better Babies Project did not have an effect on the overall low- and very low birthweight rates in the study census tracts. This was probably due to the low participation rates and the high population mobility.

Adult↗
PubMed1994

A note on testing the Hardy-Weinberg law across strata.

The problem of testing the Hardy-Weinberg law when the data are stratified in K strata is considered. Previous methods lose power when the departure from the law is irregular from stratum to stratum. Two methods based on the squared distance are proposed to overcome this problem. Simulations show that the new methods can have a dramatic improvement over the previous methods. The methods are applied to red cell glyoxalase genotype data from populations in India.

Erythrocytes↗
PubMed1993

Stopping a clinical trial very early because of toxicity: summarizing the evidence.

When a trial is stopped early because of a specific toxicity, it may be important to summarize the statistical evidence for stopping. Such a summary needs to take into account the sequential nature of the stopping rule. We address some practical issues involved in analyzing such toxicity data coming from a trial that was stopped after the fourth patient was evaluated.

Bayes Theorem↗
PubMed1993

Birth weight, gestational age and perinatal mortality: biological heterogeneity and measurement error.

At low birth weight the variance of last menstrual period based gestational age is wide and the distribution is positively skewed toward higher values. In this study the variance of gestational age decreases rapidly as birth weight increases, skewness decreases and kurtosis increases in approaching the mean of the birth weight distribution. Some of the wider variance and positive skewness of gestational age at low birth weight appears to reflect heterogeneity of intrauterine growth, in which infants with high values of gestational age are growth retarded. We show by partitioning each birth weight group into two groups of infants with different gestational age distributions, that at low birth weight, infants with low gestational ages have higher neonatal mortality rates but lower fetal mortality rates than infants with a higher gestational age for birth weight. The differences in mortality described between small infants at different gestational ages suggest that infants with a high LMP-based gestational age have experienced a slower rate of intrauterine growth. Some authors interpret the distributional characteristics as indications of systematic error in last menstrual period based assessment of gestational age. It appears from this study that the extent of systematic error in the estimation of LMP based gestational age may have been overstated in the past.

Birth Weight↗
PubMed1992

Maternal vitamin levels during pregnancies producing infants with neural tube defects.

Women at very high risk for having a child with a neural tube defect (NTD) because they had previously delivered affected children significantly reduced their recurrence rate by taking folate supplements before conception. To clarify how these results might apply to a lower-risk general obstetric population, we measured folate, vitamin B12, and retinol levels in maternal serum drawn early in 89 pregnancies resulting in NTD offspring and 178 control pregnancies identified from the Finnish Registry of Congenital Malformations. In 86.5% of the subjects, specimens were collected within 8 weeks after neural tube closure. In the NTD case mothers the mean (+/- SD) levels were not significantly lower than in control mothers: folate, 4.13 +/- 2.36 versus 4.28 +/- 2.52 ng/ml; vitamin B12, 482.8 +/- 161.1 versus 520.3 +/- 191.9 pg/ml; and retinol, 51.2 +/- 17.0 versus 50.5 +/- 16.9 micrograms/dl. After adjustment for age of the specimen, gestational age at which the specimen was drawn, maternal age, and maternal employment status, the odds ratios for being a case mother were 1.00 (95% confidence interval (CI) 0.91 to 1.10) for folate, 1.05 (95% CI 0.92 to 1.19) for vitamin B12, and 0.99 (95% CI 0.88 to 1.10) for retinol. Excluding NTD cases with known or suspected causes unrelated to vitamins, restricting the analyses to interviewed subjects, and excluding subjects whose specimens were collected after 15 gestational weeks confirmed that NTD case and control vitamin levels did not differ significantly. This population-based investigation in a low rate area demonstrated no relationship between maternal serum folate, vitamin B12, or retinol levels during pregnancy and the risk of NTDs.

Adult↗
PubMed1992

On estimating standardized risk differences from odds ratios.

An estimator proposed by Greenland and Holland (1991, Biometrics 47, 319-322) for a standardized risk difference parameter is shown to be a maximum likelihood estimator if the consistent estimator of the common odds ratio is appropriately chosen. The statistical problem under consideration is reparameterized. Likelihood equations are derived.

Epidemiologic Methods↗
PubMed1982

Breath-to-breath variations in rate and depth of ventilation in sleeping infants.

Ventilatory measurements were made noninvasively over 2- to 3-h periods during sleep in each of nine normal infants at 1 mo of age. To assess the changes that occur in ventilation on a breath-to-breath basis, we 1) examined the variations of each of tidal volume (VT), respiratory cycle time (Ttot), expiratory time (TE), and inspiratory time (TI) and 2) studied their interrelationships. We found that the variations of VT, Ttot, and TE but not of TI were significantly greater in rapid-eye-movement (REM) than in quiet sleep. In addition, on a breath-to-breath basis, VT had a positive linear relationship and strong correlation with TI; however, the correlation between VT and TE was weak in both sleep states. VT/Ttot was found to be moderately and negatively correlated with Ttot in both REM and quiet sleep. VT was weakly correlated with Ttot in REM sleep and was, on the average, more correlated with Ttot in quiet sleep. We suggest that in infants 1) on a breath-to-breath basis, VT/Ttot is likely to drop if respiratory frequency is decreased and 2) VT is nonlinearly related to Ttot during sleep; this lack of linearity depends on the lack of constancy of VT/Ttot, which is in turn closely related to the variability of the "on-switching" of inspiratory activity.

Humans↗
PubMedDate not supplied

Statistical methods of computing reference values for side-to-side differences in nerve conduction studies.

This study uses theoretic derivations, statistical simulations, and empirical estimations to compare two ways of deriving reference values for side-to-side differences in nerve conduction parameters of healthy subjects. The two methods involve using the side-to-side differences (STSD) and the absolute values of the STSD (AVSTSD). The theoretic derivations showed that the population reference value of the AVSTSD is greater than the STSD reference values by 0.18%. Simulation studies showed that the AVSTSD yields greater sampling errors than the STSD method when establishing the reference values for nerve conduction parameters. However, the sampling variability is substantially reduced by using study samples of greater than 50, and the differences between the two methods in sampling errors are trivial as sample size approaches 100. Using H reflex (HR) and extensor digitorum brevis reflex (EDBR) clinical data, the two methods were compared. In contrast to the small theoretic differences in reference values, the AVSTSD method overestimated the reference value by 0.5 ms for the EDBR and 0.1 ms for the H reflex, when using data from a population sample, increasing the type II error (reducing sensitivity) for the EDBR. The STSD method is recommended for establishment of normal values and for clinical comparisons.

Electromyography↗