Search PubMed⌕ Search

Biomedical subjects

N M Laird

Publications and source records attributed to N M Laird.

At least 73 records · Page 4Linked to original sources
PubMed1989

A comparison of statistical methods for combining event rates from clinical trials.

We compare two statistical methods for combining event rates from several studies. Both methods treat each study as a separate stratum. The Peto-modified Mantel-Haenszel (Peto) method estimates a combined odds ratio assuming homogeneity across strata and provides a test for heterogeneity. The DerSimonian and Laird modified Cochran method (D&L) produces a weighted average of rate differences, where the weights allow for among-study variability. We analyse 22 meta-analyses from ten reports by both methods. The pooled estimates are divided by their standard errors to produce a Z-statistic. A t-test comparing Z-statistics from all 22 studies suggests that the D&L method tends to be more conservative [d(Peto - D&L) = 0.29, t = 2.53, p = 0.02]. For a subset of 14 non-heterogeneous studies, the difference is smaller and non-significant (d = 0.09, t = 0.72, p = 0.49). The results from the methods correlate well (r = 0.66 for all 22 studies, r = 0.95 for 14 non-heterogeneous studies). Thus, the presence of heterogeneity influences our conclusion. We discuss the statistical and scientific implications of these findings.

Clinical Trials as Topic↗
PubMed1987

Inter-laboratory variability in Ames assay results.

The Ames test is widely used in the screening of chemicals and compounds for potential carcinogenic effect. There is, however, considerable inter-laboratory variability in results from this assay. Using data from the RTI Collaborative Study of the EPA Ames Test Protocol, we show that their reported standard errors of estimates of mutagenicity fall far short of capturing day-to-day or laboratory-to-laboratory variation. We estimate the factors by which the standard errors must be inflated to account for these sources of variation. The laboratory protocol and previous studies suggest that much of this variation may be caused by factors that are relatively constant within days (e.g. technician, incubation temperature, S9 liver homogenate preparation) but vary over days and across laboratories. Therefore, such variation might be reduced through use of a reference compound tested on the same day and under the same conditions as the test chemical. This conjecture was, however, not supported by analyses that considered the positive control compound and a pure chemical as possible reference assays.

Laboratories↗
PubMed1987

Thyroid cancer risk from exposure to ionizing radiation: a case study in the comparative potency model.

Considerable controversy exists about the relative risk of thyroid cancer following exposure to external radiation compared to the risk after exposure to internally deposited 131I. The human epidemiological data are equivocal, and studies are not directly comparable owing to differing ages at exposure, dose ranges, and periods of follow-up. Limited experimental data at low dose ranges support the hypothesis of equal potency in animals. This report utilizes a relative potency model to reconcile data from different sources, and to provide an estimate of thyroid cancer risk following human exposure to 131I. We utilize data from epidemiological studies of external radiation and 131I exposure in humans and data from an experimental animal study. This analysis shows that the data provide no compelling evidence to suggest that the risks accompanying external radiation or 131I exposure are different.

Animals↗
PubMed1986

Kartagener's syndrome. A blinded, controlled study of cilia ultrastructure.

We investigated respiratory mucosa cilia ultrastructure in patients homozygous for the gene for Kartagener's syndrome (KS) and patients apparently phenotypic for KS who had bronchiectasis and sinusitis but without situs inversus. Parents, as obligate carriers of the recessive KS gene, were also evaluated among other control groups. The four patients with KS had significantly fewer cilia outer dynein arms than normal subjects or parents of patients with KS. Two of five patients apparently phenotypic for KS demonstrated distinctive ultrastructural changes. No other subjects demonstrated explicit ultrastructural abnormalities. Internal control specimens showed that the number of outer dynein arms was consistent within a subject compared with variation between subjects. The outer dynein arm serves as a dependable ultrastructural marker. Carriers of KS do not demonstrate distinctive morphologic cilia abnormalities. Not every patient with chronic bronchiectasis and sinusitis demonstrates abnormal cilia ultrastructure.

Adolescent↗
PubMed1985

Aging and generational effects on drinking behaviors in men: results from the normative aging study.

The effects of aging on alcohol consumption behaviors are unclear because of confounding with period and cohort effects. In 1973, 1,859 male participants in the Normative Aging Study, born between 1892 and 1945, described their drinking behaviors by responding to a mailed questionnaire. In 1982, 1,713 of the participants in this study responded to a similar questionnaire. We used multivariate techniques, adjusting regression coefficients for the correlations between repeated responses of the same individuals, to assess the effects of birth cohort and aging on mean alcohol consumption level, on the prevalence of problems with drinking, and on the prevalence of averaging three or more drinks per day. Older men drank significantly less than younger men at both times yet there was no tendency for men to decrease their consumption levels over time. Each successively older birth cohort had a prevalence of problems with drinking estimated to be 0.037 lower than the prevalence of the next youngest cohort (95 per cent confidence interval: 0.029-0.045), yet there was no decrease in drinking problems over nine years. Interpretation of these findings requires consideration of the changes in attitudes as well as the increases in per capita consumption occurring in the United States throughout the 1970s. Results suggest that aging is not as important a factor in changes in drinking behaviors as generational or attitudinal changes.

Adult↗
PubMed1982

Random-effects models for longitudinal data.

Models for the analysis of longitudinal data must recognize the relationship between serial observations on the same unit. Multivariate models with general covariance structure are often difficult to apply to highly unbalanced data, whereas two-stage random-effects models can be used easily. In two-stage models, the probability distributions for the response vectors of different individuals belong to a single family, but some random-effects parameters vary across individuals, with a distribution specified at the second stage. A general family of models is discussed, which includes both growth models and repeated-measures models as special cases. A unified approach to fitting these models, based on a combination of empirical Bayes and maximum likelihood estimation of model parameters and using the EM algorithm, is discussed. Two examples are taken from a current epidemiological study of the health effects of air pollution.

Air Pollution↗
PubMed1981

Effect of the hemodialysis prescription on patient morbidity: report from the National Cooperative Dialysis Study.

This report summarizes morbidity in 151 patients in a cooperative trial designed to evaluate the clinical effects of different dialysis prescriptions. Four treatment groups were divided along two dimensions: dialysis treatment time (long or short), and blood urea nitrogen (BUN) concentration averaged with respect to time (TACurea) (high or low). Dietary protein was not restricted. There was no difference in mortality between the groups. Withdrawal of patients from the high-BUN groups for medical reasons was significantly greater than withdrawal from the low-BUN groups. Hospitalization was also greater in the high-BUN groups, but dialysis treatment time had no significant effects. The data indicate that the occurrence of morbid events is affected by the dialysis prescription. Increased morbidity appears to accompany prescriptions associated with a relatively high BUN. Conversely, morbidity may be decreased by prescriptions associated with more efficient removal of urea if the dietary intake of protein and other nutrients is adequate.

Adolescent↗
PubMed1979

Sample-size estimation: a sensitivity analysis in the context of a clinical trial for treatment of mild hypertension.

The effectiveness of treatment for mild hypertension (diastolic pressures of 85 to 105 mm Hg) has not been conclusively demonstrated. Both the costs of a carefully designed clinical trial and the likelihood that it will produce definitive answers will depend importantly on the sample size. This paper presents sample-size estimates under a variety of assumptions regarding the characteristics of the population to be studied, the degree of blood pressure control to be achieved, and the health benefits to be expected. Under a central set of assumptions, the estimated sample size per group is 22,700 with death as an endpoint and 14,000 with morbid events (CHD and stroke) as endpoints. As individual assumptions are varied one at a time, required sample sizes range from 10,900 to 101,100 and from 6,800 to 63,100 for the respective endpoints. Results are most sensitive to the degree of blood pressure control actually achieved to the expected health benefits from blood pressure control. They are also highly sensitive to the sex composition of the population and to expected dropout rates. The choice of sample size will depend on the decision maker's assessment of the likelihood that each assumption will be fulfilled and on the degree of willingness to risk an inconclusive study result. By making explicit the effect of variation in each assumption, decision making is rendered more susceptible to critical examination by outside reviewers.

Adult↗
PubMedDate not supplied

Mixture models for the joint distribution of repeated measures and event times.

Many long-term clinical trials collect both a vector of repeated measurements and an event time on each subject; often, the two outcomes are dependent. One example is the use of surrogate markers to predict disease onset or survival. Another is longitudinal trials which have outcome-related dropout. We describe a mixture model for the joint distribution which accommodates incomplete repeated measures and right-censored event times, and provide methods for full maximum likelihood estimation. The methods are illustrated through analysis of data from a clinical trial for a new schizophrenia therapy; in the trial, dropout time is closely related to outcome, and the dropout process differs between treatments. The parameter estimates from the model are used to make a treatment comparison after adjusting for the effects of dropout. An added benefit of the analysis is that it permits using the repeated measures to increase efficiency of estimates of the event time distribution.

Algorithms↗