Biomedical subjects

J B Greenhouse

Publications and source records attributed to J B Greenhouse.

At least 19 recordsLinked to original sources

PubMed1996

Applications of a mixture survival model with covariates to the analysis of a depression prevention trial.

This paper presents a case study of model selection for survival analysis data. We use an approximate Bayesian method for model selection based on assessing the posterior probability of competing models given the data. We introduce the Schwarz criteria, an approximation to the logarithm of the Bayes factor, to provide an indication of evidence in favour of one model compared to another. Specifically, in the context of a depression prevention clinical trial we evaluate the efficacy of treatment in preventing or delaying the time to recurrence of depression, and evaluate how differences in the survival distributions between the two treatment groups depend on explanatory variables of interest. This investigation is based on a mixture survival model that explicitly incorporates the possibility of a surviving fraction.

Antidepressive Agents, Tricyclic↗

PubMed1996

Bone lead levels and delinquent behavior.

OBJECTIVE: To evaluate the association between body lead burden and social adjustment. DESIGN: Retrospective cohort study. SETTING: Public school community. PARTICIPANTS: From a population of 850 boys in the first grade at public schools, 503 were selected on the basis of a risk scale for antisocial behavior. All of the 850 boys who scored in the upper 30th percentile of the distribution on a self-reported antisocial behavior scale were matched with an equal number drawn by lot from the lower 70% of the distribution. From this sample, 301 students accepted the invitation to participate. EXPOSURE MEASURE: K x-ray fluorescence spectroscopy of tibia at subjects' age of 12 years. MAIN OUTCOME MEASURES: Child Behavior Checklist (CBCL), teachers' and parents' reports, and subjects' self-report of antisocial behavior and delinquency at 7 and 11 years of age. RESULTS: Subjects, teachers, and parents were blind to the bone lead measurements. At 7 years of age, borderline associations between teachers' aggression, delinquency, and externalizing scores and lead levels were observed after adjustment for covariates. At 11 years of age, parents reported a significant lead-related association with the following CBCL cluster scores: somatic complaints and delinquent, aggressive, internalizing, and externalizing behavior. Teachers reported significant associations of lead with somatic complaints, anxious/depressed behavior, social problems, attention problems, and delinquent, aggressive, internalizing, and externalizing behavior. High-lead subjects reported higher scores in subjects' self-reports of delinquency at 11 years. High-lead subjects were more likely to obtain worse scores on all items of the CBCL during the 4-year period of observation. High bone lead levels were associated with an increased risk of exceeding the clinical score (T > 70) for attention, aggression, and delinquency. CONCLUSION: Lead exposure is associated with increased risk for antisocial and delinquent behavior, and the effect follows a developmental course.

Analysis of Variance↗

PubMed1995

Robust Bayesian methods for monitoring clinical trials.

Bayesian methods for the analysis of clinical trials data have received increasing attention recently as they offer an approach for dealing with difficult problems that arise in practice. A major criticism of the Bayesian approach, however, has focused on the need to specify a single, often subjective, prior distribution for the parameters of interest. In an attempt to address this criticism, we describe methods for assessing the robustness of the posterior distribution to the specification of the prior. The robust Bayesian approach to data analysis replaces the prior distribution with a class of prior distributions and investigates how the inferences might change as the prior varies over this class. The purpose of this paper is to illustrate the application of robust Bayesian methods to the analysis of clinical trials data. Using two examples of clinical trials taken from the literature, we illustrate how to use these methods to help a data monitoring committee decide whether or not to stop a trial early.

Bayes Theorem↗

PubMed1995

An introduction to logistic regression with an application to the analysis of language recovery following a stroke.

The aim of a statistical model is to present a simplified representation of the underlying structure in a data set by separating systematic features from random variation. Sometimes the purpose of a statistical model is to provide a simple descriptive summary of the data and sometimes it is to use the data for comparative or inferential purposes. In practice, the specification of a statistical model requires a thorough understanding of the substantive area of application, an assessment of the validity of the assumptions of the model, and an evaluation of the fit of the model to the data. In this paper, as an illustration of these aspects of the statistical modeling of data, we consider the specification, application, and interpretation of a logistic regression model for the investigation of relationships between binary response data and a collection of explanatory variables. We illustrate applications of the methodology using data from a prospective study of spontaneous language recovery following a stroke (Holland, Greenhouse, Fromm, & Swindell, 1989).

Brain↗

PubMed1993

Some conceptual and statistical issues in analysis of longitudinal psychiatric data. Application to the NIMH treatment of Depression Collaborative Research Program dataset.

Longitudinal studies have a prominent role in psychiatric research; however, statistical methods for analyzing these data are rarely commensurate with the effort involved in their acquisition. Frequently the majority of data are discarded and a simple end-point analysis is performed. In other cases, so called repeated-measures analysis of variance procedures are used with little regard to their restrictive and often unrealistic assumptions and the effect of missing data on the statistical properties of their estimates. We explored the unique features of longitudinal psychiatric data from both statistical and conceptual perspectives. We used a family of statistical models termed random regression models that provide a more realistic approach to analysis of longitudinal psychiatric data. Random regression models provide solutions to commonly observed problems of missing data, serial correlation, time-varying covariates, and irregular measurement occasions, and they accommodate systematic person-specific deviations from the average time trend. Properties of these models were compared with traditional approaches at a conceptual level. The approach was then illustrated in a new analysis of the National Institute of Mental Health Treatment of Depression Collaborative Research Program dataset, which investigated two forms of psychotherapy, pharmacotherapy with clinical management, and a placebo with clinical management control. Results indicated that both person-specific effects and serial correlation play major roles in the longitudinal psychiatric response process. Ignoring either of these effects produces misleading estimates of uncertainty that form the basis of statistical tests of hypotheses.

Analysis of Variance↗

PubMed1992

On some applications of Bayesian methods in cancer clinical trials.

The NCCTG randomized controlled clinical trial for the treatment of advanced colorectal carcinoma is a wonderful case study of the dynamic interplay between scientific learning and statistical inference. Ethical concerns for minimizing the number of patients assigned to an inferior treatment and interest in identifying subsets of patients for whom a treatment is most likely efficacious pose challenging problems for the practice of statistics. In the first part of this paper, I comment on the applications of Bayesian methods to these problems in the NCCTG trial as presented by Freedman and Spieglehalter and Dixon and Simon, respectively. In the second part of this paper, I discuss and illustrate a Bayesian approach to model sensitivity analysis with a particular focus on model specification and criticism. The Bayesian approach provides a formal methodology to assess the sensitivity of inferences to the inputs into an analysis so that it is possible to investigate the consequences of the specification of the model. I apply these methods to the specification and criticism of a class of survival models for the analysis of survival times in the NCCTG trial.

Antineoplastic Combined Chemotherapy Protocols↗

PubMed1992

Bayesian methods for phase I clinical trials.

Phase I clinical trials are conducted to determine the dose-response curve of a new drug with respect to toxic side effects and, in particular, to estimate the maximum tolerated dose (MTD). In this paper we take a Bayesian approach to the problem of making inferences about the MTD. Working with broad classes of priors, we obtain the posterior distribution of the MTD and study its properties. We also address the question of providing updated assessments of the risk of toxicity for new patients entering the study at a specific dose level. These assessments would be useful in deciding issues of study management and ethics. Our analysis pays particular attention to the sensitivity of the inferences and risk assessments to the choice of prior and the choice of model for the dose-response relationship.

Algorithms↗

PubMed1992

Exploratory statistical methods, with applications to psychiatric research.

This article introduces statistical methods for describing and summarizing the results of studies and introduces statistical principles that will guide the psychiatric researcher in the evaluation and interpretation of clinical research. We discuss relatively easy-to-use and informal methods for describing and comparing data. Our aim is to develop methods for investigating relationships among variables, to learn about the effect of one variable upon another. Once we have observed an apparent relationship between variables, an important question to be addressed is whether or not this observed relationship is causal in the sense that a change in one variable causes a changes in the other. We discuss and illustrate principles related to the evaluation of the nature of the association among variables. Throughout the article, principles and methods will be illustrated by examples and case studies based on data sets primarily from the psychiatric research literature.

Clinical Trials as Topic↗

PubMed1991

Methodologic issues in maintenance therapy clinical trials.

In the early 1980s, the National Institute of Mental Health supported a multicenter, randomized, controlled, clinical trial on unipolar and bipolar disorder to evaluate the comparative efficacies of lithium carbonate, imipramine hydrochloride, a lithium-imipramine combination, and placebo in preventing the recurrence of affective disorders. The objective of this report is to present a reanalysis of the relative efficacies of these treatments in patients with unipolar disorder to focus attention on general issues related to the design and conduct of maintenance therapy trials. We show that the earlier conclusions of that study that imipramine and the combination therapy are more effective than lithium and placebo in preventing the recurrence of depression in unipolar patients can be accounted for by alternative explanations that are a consequence of the design of the study. Our findings have important implications for the design, conduct, and interpretation of results of maintenance therapy clinical trials in general.

Adult↗

PubMed1991

A note on randomization and selection bias in maintenance therapy clinical trials.

In this article we demonstrate that even in randomized controlled clinical trials, unobserved confounding variables can bias the outcome of a study. For the case of a two-phase maintenance therapy trial where patients who respond to treatment during the acute phase are then randomized to a maintenance therapy, we show explicitly the role that confounding may play in biasing the interpretation of the results of such a trial. We suggest an alternative design to deal with the problem of a selection effect for treatment responders in the acute phase of the trial by randomizing patients at the outset of the study to both an acute and maintenance therapy.

Bias↗

PubMed1990

Minute-by-minute analysis of REM sleep timing in major depression.

Sleep changes described in depressed patients may represent alterations in the timing of rapid-eye-movement (REM) sleep or sleep onset. We examined these variables in groups of healthy control subjects (n = 47), depressed outpatients (n = 98), and depressed inpatients (n = 41). Outpatient depressives had greater severity of clinical symptoms than inpatients using the Hamilton Rating Scale for Depression. The depressed inpatient group had a later mean sleep onset time than the other groups, and the depressed outpatient group had a wider range of good night times than control subjects. REM timing in each group was examined as a relative frequency distribution of REM sleep (FDRS) for each minute across the night. The FDRSs for the three groups were statistically compared using the parameters from a two-component model, which includes a deterministic sinusoidal function and a time series process for errors. The slope of the linear trend in the FDRS rhythm was smaller (less positive) for both depressed groups than for controls. The ultradian FDRS rhythm occurred at an earlier phase, relative to sleep onset, in the inpatient depressed group compared to the control group. The ultradian FDRS rhythm had a longer period in the outpatient group compared to the control and inpatient groups. When referenced to 24-hr clock time in an exploratory analysis, the depressed groups appeared to have less robust FDRS ultradian rhythms than controls, but they did not appear to have a systematic phase alteration compared to controls. Abnormalities of REM sleep timing in groups of depressed patients may reflect a disturbance of sleep initiation and generation, or difficulty in entrainment of REM, rather than a systematic phase alteration in REM sleep propensity.

Adult↗

PubMed1990

A reexamination of the relationship between growth hormone secretion and slow wave sleep using delta wave analysis.

Sleep onset growth hormone secretion is a reliable and reproducible finding in young adults and children. Secretion typically occurs during the first non-REM period of sleep and, despite some evidence to the contrary, growth hormone secretion has frequently been associated with the first period of slow wave sleep. By measuring delta wave activity (0.5-2 Hz) instead of slow wave sleep and, accounting for the within subject variability, it has not been possible to demonstrate a consistent or statistically significant linear relationship between delta wave activity and sleep-related growth hormone secretion. This suggests the presence of more complex mediating factors and the possibility that sleep onset and growth hormone secretion are two separate processes which are independently stimulated by events associated with sleep onset.

Adolescent↗

PubMed1989

An introduction to survival analysis: statistical methods for analysis of clinical trial data.

The randomized controlled clinical trial (RCT) is a prospective study using random assignment of subjects to treatment groups to compare the effect and value of a therapeutic intervention against a control. The RCT is the most definitive clinical research tool for evaluating the efficacy of a new therapy in human subjects. Often the outcome of interest in an RCT is the length of time until an event occurs after treatment or intervention. In this article we introduce statistical methods for evaluating differences in the patterns of time to response between two groups of subjects to determine whether one therapy is better than another. The collection of methods for analyzing such data, known as survival data, is called survival analysis. Using data from a hypothetical clinical trial for the prevention of the recurrence of depression, we illustrate two elementary methods for analyzing survival data. We also discuss generalizations of these methods to incorporate covariates and conclude with a general discussion of clinical trials of psychiatric therapies.

Clinical Trials as Topic↗

PubMed1989

Predictors of language restitution following stroke: a multivariate analysis.

A consecutive sample of 50 language-impaired patients was evaluated prospectively during the first 3 to 4 months following unilateral left- or right-hemisphere stroke. A multiple logistic linear regression model was used to assess the relative importance of eight predictor variables on the likelihood of language recovery. Those found to be significantly associated with language recovery included age (favoring younger patients) and length of hospital stay (favoring shorter stays). Gender (favoring males), type of stroke (favoring hemorrhages), and side of lesion (favoring right) were only moderate correlates of recovery. Neither race nor history of previous stroke was a significant predictor of language recovery. Multivariate statistical analysis was useful in illuminating the joint relationship between clinical and demographic predictor variables and language recovery.

Age Factors↗

PubMed1988

Characteristics of recovery of drawing ability in left and right brain-damaged patients.

Drawings of a "person" were obtained from 21 patients with left brain damage (LBD) and 13 patients with right brain damage (RBD) at hospital discharge and 1 month postdischarge, corresponding to 2-4 and 6-8 weeks postonset, respectively. Early LBD drawings were simplistic and often incoherent. Most were placed in the upper left quadrant of the page. RBD drawings were scattered, fragmented, and elaborately detailed. Many showed evidence of left-sided neglect. The resolution of drawing disability also differed between the two groups: LBD subjects recovered drawing abilities more rapidly and more completely than RBD subjects. These differences in drawing characteristics and in the resolution of drawing disability were discernable for experienced neuropsychologists.

Adult↗

PubMed1988

The EM algorithm for maximum likelihood estimation in the mover-stayer model.

The discrete-time mover-stayer model (Blumen, Kogan, and McCarthy, 1955, The Industrial Mobility of Labor as a Probability Process, Ithaca, New York: Cornell University Press) is a useful model for studying changes over time in heterogeneous populations. Using the EM algorithm, we present an alternative method for obtaining maximum likelihood estimates of the parameters of the mover-stayer model, and consider an extension of the basic model to the problem of incomplete follow-up in panel studies. The models and the methods are illustrated with data from a community-based survey of changes in mental health status over a 1-year period.

Algorithms↗

PubMed1987

Methodology in psychiatric research. Report on the 1986 MacArthur Foundation Network I Methodology Institute.

This report discusses many of the issues raised during a two-day institute that focused on methodological problems encountered in psychiatric research. The topics range from the problems with psychiatric diagnosis and measurement to sampling issues and biases to specific statistical concerns. Attention is given to the need for improved communication and collaboration between psychiatric researchers and biostatisticians.

Biometry↗

PubMed1987

Fitting nonlinear models with ARMA errors to biological rhythm data.

Many behavioural and physiological processes are periodic with a period of approximately 24 hours. For descriptive purposes, linear regression using a simple sinusoid with a fixed 24 hour period is sometimes an adequate tool for analysing data from such processes. Inference based on regression models under the assumption of independent and identically distributed errors, however, can often mislead seriously. In this paper we present a general class of models for fitting biological rhythms, with use of higher-order harmonic terms of one or more unknown fundamentals and ARMA processes for the errors. We describe the procedures for model specification and estimation and give the theoretical justification for these procedures. Analysis of a series of human core body temperature illustrates the methodology.

Biometry↗