Search PubMed⌕ Search

Biomedical subjects

S Greenland

Publications and source records attributed to S Greenland.

At least 37 records · Page 2Linked to original sources
PubMed2000

Multilevel modeling in epidemiology with GLIMMIX.

Previous work has shown that multilevel modeling can be a valuable technique for epidemiologic analysis. The complexity of using this approach, however, continues to restrict its general application. A critical factor is the lack of flexible and appropriate software for multilevel modeling. SAS provides a macro, GLIMMIX, that can be used for multilevel modeling, but that is not sufficient for a complete epidemiologic analysis. We here provide additional code to obtain epidemiologic output from GLIMMIX, illustrated with new data on diet and breast cancer from the European Community Multicenter Study on Antioxidants, Myocardial Infarction, and Breast Cancer (EURAMIC). Our results give epidemiologists an easily used tool for fitting multilevel models.

Breast Neoplasms↗
PubMed2000

Side effects and time course of response in a placebo-controlled trial of fluoxetine for the treatment of geriatric depression.

A high rate of improvement among patients who receive placebo in controlled trials of antidepressants can complicate the evaluation of true drug effect. Placebo response may be a reaction to the psychosocial factors of study participation or a function of changes in the natural course of depression. Drug side effects may also influence patients' expectations, and they should be distinguished from the somatic symptoms associated with major depression. The authors reanalyzed data from a large, multicenter, placebo-controlled clinical trial of fluoxetine treatment of geriatric depression to evaluate similarities and differences between responders and nonresponders in both treatment groups. Specifically, the authors examined weekly somatic complaints as possible predictors of response and of dropout, as well as the time course and onset of response. Fluoxetine was superior to placebo on all outcome measures. Among somatic complaints associated with fluoxetine response, headache before and after randomization was associated with a good response and anxiety after randomization was associated with a poor response. Somnolence before and after randomization was associated with a good placebo response. Early and persistent improvement occurred among similar proportions of responders in both groups. The difference between fluoxetine and placebo seemed to be a persistent response beginning during the 4th week. Pretreatment somnolence was associated with early, persistent improvement in both groups and may serve as a marker for placebo response.

Aged↗
PubMed2000

When should epidemiologic regressions use random coefficients?

Regression models with random coefficients arise naturally in both frequentist and Bayesian approaches to estimation problems. They are becoming widely available in standard computer packages under the headings of generalized linear mixed models, hierarchical models, and multilevel models. I here argue that such models offer a more scientifically defensible framework for epidemiologic analysis than the fixed-effects models now prevalent in epidemiology. The argument invokes an antiparsimony principle attributed to L. J. Savage, which is that models should be rich enough to reflect the complexity of the relations under study. It also invokes the countervailing principle that you cannot estimate anything if you try to estimate everything (often used to justify parsimony). Regression with random coefficients offers a rational compromise between these principles as well as an alternative to analyses based on standard variable-selection algorithms and their attendant distortion of uncertainty assessments. These points are illustrated with an analysis of data on diet, nutrition, and breast cancer.

Breast Neoplasms↗
PubMed2000

Empirical Bayes adjustments for multiple results in hypothesis-generating or surveillance studies.

Traditional methods of adjustment for multiple comparisons (e.g., Bonferroni adjustments) have fallen into disuse in epidemiological studies. However, alternative kinds of adjustment for data with multiple comparisons may sometimes be advisable. When a large number of comparisons are made, and when there is a high cost to investigating false positive leads, empirical or semi-Bayes adjustments may help in the selection of the most promising leads. Here we offer an example of such adjustments in a large surveillance data set of occupation and cancer in Nordic countries, in which we used empirical Bayes (EB) adjustments to evaluate standardized incidence ratios (SIRs) for cancer and occupation among craftsmen and laborers. For men, there were 642 SIRs, of which 138 (21%) had a P < 0.05 (13% positive with SIR > 1.0 and 8% negative with SIR < or = 1.0) when testing the null hypothesis of no cancer/occupation association; some of these were probably due to confounding by nonoccupational risk factors (e.g., smoking). After EB adjustments, there were 95 (15%) SIRs with P < 0.05 (10% positive and 5% negative). For women, there were 373 SIRs, of which 37 (10%) had P < 0.05 before adjustment (6% positive and 4% negative) and 13 (3%) had P < 0.05 after adjustment (2% positive and 1% negative). Several known associations were confirmed after EB adjustment (e.g., pleural cancer among plumbers, original SIR 3.2 (95% confidence interval, 2.5-4.1), adjusted SIR 2.0 (95% confidence interval, 1.6-2.4). EB can produce more accurate estimates of relative risk by shrinking imprecise outliers toward the mean, which may reduce the number of false positives otherwise flagged for further investigation. For example, liver cancer among chimney sweepers was reduced from an original SIR of 2.2 (range, 1.1-4.4) to an adjusted SIR of 1.1 (range, 0.9-1.4). A potentially important future application for EB is studies of gene-environment-disease interactions, in which hundreds of polymorphisms may be evaluated with dozens of environmental risk factors in large cohort studies, producing thousands of associations.

Analysis of Variance↗
PubMed1999

Random-effects meta-analyses are not always conservative.

It is widely held that random-effects summary effect estimates are more conservative than fixed-effects summaries in epidemiologic meta-analysis. This view is based on the fact that random-effects summaries have higher estimated variances and, consequently, wider confidence intervals than fixed-effects summaries when there is evidence of appreciable heterogeneity among the results from the individual studies. In such instances, however, the random-effects point estimates are not invariably closer to the null value nor are their p values invariably larger than those of fixed-effects summaries. Thus, random-effects summaries are not predictably conservative according to either of these two connotations of the term. The authors give an example from a meta-analysis of water chlorination and cancer in which the random-effects summaries are less conservative in both of these alternative senses and possibly more biased than the fixed-effects summaries. The discussion of when to use random effects and when to use fixed effects in computing summary estimates should be replaced by a discussion of whether summary estimates should be computed at all when the studies are not methodologically comparable, when their results are discernibly heterogeneous, or when there is evidence of publication bias.

Bias↗
PubMed1999

Presenting statistical uncertainty in trends and dose-response relations.

When one estimates the effects of a polytomous exposure, it is common practice to express all effects relative to a baseline or reference level. Certain authors have challenged this practice and proposed alternatives, which we review here. One alternative, the "floating absolute risk" method, can supply useful statistics and trend graphs, but it does not yield valid confidence intervals for relative risks. All categorical methods have further shortcomings when the exposure is continuous, however. These shortcomings can be addressed by plotting or tabulating confidence limits for points on a flexible curve fitted to the uncategorized data.

Birth Weight↗
PubMed1999

A unified approach to the analysis of case-distribution (case-only) studies.

A number of new study designs have appeared in which the exposure distribution of a case series is compared to an exposure distribution representing a complete theoretical population or distribution. These designs include the case-genotype study, the case-cross-over study, and the case-specular study. This paper describes a unified likelihood-based approach to the analysis of such studies, and discusses extensions of these methods when a control group is available. The approach clarifies certain assumptions implicit in the methods, and helps contrast these assumptions to those underlying ordinary case-control studies. There are several reasons to expect discrepancies between ordinary case-control estimates and case-distribution estimates; for example, case-distribution estimates can be more sensitive to exposure misclassification. Some discrepancies are illustrated in an application to case-specular data on wire codes and childhood cancer.

Case-Control Studies↗
PubMed1999

Application of the case-specular method to two studies of wire codes and childhood cancers.

This paper presents the results of applying the case-specular method to two earlier studies of wire codes and childhood cancers (DA Savitz et al, Am J Epidemiol 1988;128:21-38, and SJ London et al, Am J Epidemiol 1991;9:923-937). The method compares the wire codes of case residences with the wire codes of specular residences constructed by switching the location of the case residence across the center of the street. The method was designed to discriminate between the magnetic field hypothesis, which postulates that childhood cancer is affected by magnetic fields and that wire codes are a proxy for magnetic fields, and the neighborhood hypothesis, which postulates that childhood cancer is affected by some characteristics of the neighborhood other than magnetic fields and that wire codes are a proxy for those characteristics. Although the results from the two applications of the method have limited precision, they support the results originally reported (odds ratios of around 2 for very high current configuration residences and childhood cancers) and do not support suggestions that the associations are due to confounding by socio-economic and neighborhood factors. The results leave open the question of whether or not control selection bias could have influenced the original associations, because there was no convincing evidence that the control-specular matrices were symmetric.

California↗
PubMed1999

The importance of specifying the underlying biologic model in estimating the probability of causation.

There are a number of contexts in which interested parties wish to estimate the probability that an individual's injury was caused by radiation or a toxic chemical. It has been shown, however, that such calculations cannot be made based on epidemiologic data alone, without assumption of a biologic model for the disease process and without a specific definition of causation. To illustrate the relevant theorems, we present a number of examples in which different biologic models produce different values for the probability of causation for individuals from the same population-based epidemiologic data and dose-response curves. As a result of these ambiguities, it is important that anyone attempting to calculate probability of causation for individuals explicitly state the biologic model that has been assumed, as well as state the definition of causation being used. The analyst should test the robustness of the calculations by repeating them for a broad range of underlying biologic models.

Female↗
PubMed1999

Side effects as predictors of drug response in obsessive-compulsive disorder.

Differences between the side effect profiles of clomipramine (CMI) and the selective serotonin reuptake inhibitors may be important factors in both treatment outcome and patient selection in obsessive-compulsive disorder (OCD). Safety and efficacy data from an industry-sponsored, multicenter clinical trial of CMI were analyzed previously using tabular and multiple regression methods. Good response, defined as at least a 35% drop in final scores on the Yale-Brown Obsessive Compulsive Scale (Y-BOCS), was associated with a later age of OCD onset and certain early side effects that may reflect a sensitivity of responders to CMI's serotonergic actions. The authors conducted a similar analysis of data from an industry-sponsored clinical trial of fluoxetine in OCD. Fluoxetine response did not seem to be associated with age of OCD onset. Good response to both drugs was associated with initial nervousness and sexual complaints. The common side effects of fluoxetine (headache, nausea, and gastrointestinal complaints) did not seem to be associated with treatment response. Slight differences in the protocols of the two clinical trials yielded patient populations that were different in factors found to be associated with treatment outcome: subjects in the fluoxetine study had lower scores on the Y-BOCS, higher scores on the Hamilton Rating Scale for Depression, and an earlier age of OCD onset.

Adolescent↗
PubMed1999

Relation of probability of causation to relative risk and doubling dose: a methodologic error that has become a social problem.

Epidemiologists, biostatisticians, and health physicists frequently serve as expert consultants to lawyers, courts, and administrators. One of the most common errors committed by experts is to equate, without qualification, the attributable fraction estimated from epidemiologic data to the probability of causation requested by courts and administrators. This error has become so pervasive that it has been incorporated into judicial precedents and legislation. This commentary provides a brief overview of the error and the context in which it arises.

Causality↗
PubMed1999

A case-control study of colorectal adenomatous polyps and consumption of foods containing partially hydrogenated oils.

The trans fatty acids produced by partially hydrogenating vegetable oils may cause colorectal neoplasia by interfering with cell membrane function or eicosanoid synthesis. This possibility provides a rationale for looking at the relation between colorectal adenomatous polyps and consumption of foods containing partially hydrogenated vegetable oils (PHVOs). A total of 516 cases and 551 controls who underwent screening sigmoidoscopy from 1991-1993 were recruited from a prepaid Los Angeles health plan. Subjects were interviewed and given a self-administered food frequency questionnaire. Food items containing PHVOs were divided into four groups characterized by principal ingredients and preparation methods: sweetened baked goods, candy bars, oils and condiments, and french fries and chips. After adjusting for age, sex, physical activity, body mass index, smoking, total energy, and red meat and vegetable intake, there was a positive association between polyps and sweetened baked goods [350+ versus <50 kcal/day (odds ratio, 2.1; 95% confidence interval, 1.3-3.5)]. No association was found with the other food groups after adjustment for dietary and nondietary covariates. Neither was total dietary trans fatty acid associated with adenomas after adjustment for sweetened baked goods and other covariates. These results do not support the hypothesis that eating foods containing PHVOs increases the risk of colorectal adenomas, but they are consistent with the hypothesis that foods high in fat and sugar and low in fiber and correlated micronutrients increase the risk of adenomas.

Adenomatous Polyps↗
PubMed1999

Multilevel modeling and model averaging.

Multilevel modeling, also known as hierarchical regression, generalizes ordinary regression modeling to allow explicit and flexible compromises between simple and complex models. This article provides an elementary introduction to multilevel modeling as a model-averaging technique. Model averaging provides an alternative to model selection, and it emphasizes the role of prior information in finding good models.

Abortion, Spontaneous↗
PubMed1999

Causal diagrams for epidemiologic research.

Causal diagrams have a long history of informal use and, more recently, have undergone formal development for applications in expert systems and robotics. We provide an introduction to these developments and their use in epidemiologic research. Causal diagrams can provide a starting point for identifying variables that must be measured and controlled to obtain unconfounded effect estimates. They also provide a method for critical evaluation of traditional epidemiologic criteria for confounding. In particular, they reveal certain heretofore unnoticed shortcomings of those criteria when used in considering multiple potential confounders. We show how to modify the traditional criteria to correct those shortcomings.

Epidemiologic Methods↗