PubMed Health⌕ Search

Biomedical subjects

L L Kupper

Publications and source records attributed to L L Kupper.

At least 19 recordsLinked to original sources
PubMed1991

Confidence intervals for post-test probability.

Confidence intervals are a natural way to describe the uncertainty of post-test probability in diagnostic tests. We consider confidence intervals for two different scenarios. At a site, for example, hospital emergency room or student health centre, with measured values of disease prevalence, sensitivity and specificity available, the confidence interval is similar to results in the literature, but at a site where measured values of these indices are unavailable, we develop a method, using the values of disease prevalence, sensitivity and specificity from other sites, to obtain a confidence interval for post-test probability. We use the diagnosis of strep throat to illustrate the results. We also obtain confidence intervals from simulations to compare with the results of both scenarios.

Adult↗
PubMed1991

Effects of exposure misclassification on regression analyses of epidemiologic follow-up study data.

In epidemiologic studies, subjects are often misclassified as to their level of exposure. Ignoring this misclassification error in the analysis introduces bias in the estimates of certain parameters and invalidates many hypothesis tests. For situations in which there is misclassification of exposure in a follow-up study with categorical data, we have developed a model that permits consideration of any number of exposure categories and any number of multiple-category covariates. When used with logistic and Poisson regression procedures, this model helps assess the potential for bias when misclassification is ignored. When reliable ancillary information is available, the model can be used to correct for misclassification bias in the estimates produced by these regression procedures.

Biometry↗
PubMed1990

Sample size determination for pair-matched case-control studies where the goal is interval estimation of the odds ratio.

Samples sizes are calculated for case-control studies where 1:1 matching has been employed, and where the goal is the interval estimation of the odds ratio. The optimal sample size is defined to be the smallest value for which a 100(1 - alpha)% confidence interval for the log odds ratio will not exceed a specified width 2 delta with specified probability (1 - gamma). This approach is similar in spirit to the power-based approach for sample size determination when significance testing is the goal. Tables of sample sizes are presented for various choices of parameters. We also find considerable disagreement with a published method based on expected numbers of discordant pairs.

Case-Control Studies↗
PubMed1990

Sample size requirements for interval estimation of the odds ratio.

Sample sizes are calculated for unmatched case-control (or cohort) studies where the goal is interval estimation of the odds ratio. The procedure used gives the smallest sample size for which a 100(1-alpha)% confidence interval for the log odds ratio will not exceed a specified width with specified probability (1-gamma). Tables of sample sizes for various choices of parameter values are presented. Considerable disagreement is found with a published method which has as its basis expected cell counts.

Case-Control Studies↗
PubMed1990

An exploratory analysis of the occupational correlates of large pigmented nevi at Lawrence Livermore National Laboratory.

To examine the association between exposure to occupational factors and the presence on the body of large pigmented nevi greater than 5 mm, data were collected from 110 employees of Lawrence Livermore National Laboratory employed between 1969 and 1980. In this exploratory prevalent case-control study analysis, 38 employees reported having at least one large pigmented nevus and 72 reported none. Occupational factors associated with the presence of large pigmented nevi (adjusted for age and the number of sunburns per year under age 21) were being an engineer (prevalence odds ratio [POR] = 3.20, 95% confidence interval [CI] = 1.0,10.6) or an electrical engineer (POR = 2.56, 95% CI = 0.3,20.3), being hired at Lawrence Livermore National Laboratory before 1962 (POR = 3.07, 95% CI = 1.2,7.7), and having one's skin exposed to rare earth metals (POR = 3.78, 95% CI = 0.9,15.1).

Adult↗
PubMed1990

Continued fraction representation for expected cell counts of a 2 x 2 table: a rapid and exact method for conditional maximum likelihood estimation.

The expected cell count for a 2 x 2 contingency table, governed by the noncentral (extended) hypergeometric distribution, is expressed as a terminating continued fraction. The coefficients in the continued fraction are better behaved than the multinomial coefficients required for the usual moment calculation. The expected cell count must be calculated repeatedly in a conditional maximum likelihood analysis of K2 x 2 contingency tables. Since the continued fraction can be easily evaluated, a rapid and numerically stable computational algorithm results. Once this first moment is known, higher moments can be obtained as shown by Harkness (1965, Annals of Mathematical Statistics 36, 938-945). A BASIC program to implement the continued fraction algorithm is given in an appendix.

Algorithms↗
PubMed1990

Effective testing of gene-disease associations.

We propose a method for testing any hypothesized association between a candidate allele, for which there is a specific laboratory test, and a common chronic disease. Families in which this allele is segregating are identified through index individuals who are homozygous or heterozygous for the allele. The sample consists of the subset of identified families who also have at least one member with the common disease of interest. For each independent family in this subset, select one person with the disease and determine if he or she is heterozygous for the allele. The observed proportion of heterozygotes in this sample is compared to the proportion expected on the basis of each diseased relative's null probability of being heterozygous for the allele; this null probability depends only on the relative's relationship to the index individual and the population allele frequency. We provide these null probabilities, develop appropriate inference procedures, discuss sample size requirements, and compare this method to a standard case-control design. Results using this method are unlikely to be influenced by confounders, systematic bias, or genetic heterogeneity.

Alleles↗
PubMed1989

On assessing interrater agreement for multiple attribute responses.

New methods are developed for assessing the extent of interrater agreement when each unit to be rated is characterized by a (possibly empty) subset of a specified set of distinct nominal attributes. For such multiple attribute response data, a two-rater concordance statistic is derived, and associated statistical inference-making procedures are provided. This concordance statistic is corrected for chance agreement by using an underlying hypergeometric model. Numerical examples are given to illustrate the proposed methodology, and comparisons to other agreement statistics (e.g., kappa) are made.

Biometry↗
PubMed1979

Analysis of dichotomous response data from certain toxicological experiments.

In certain toxicological experiments with laboratory animals, littermate data are frequently encountered. It is generally recognized that one characteristic of this type of data is the "litter effect", i.e., the tendency for animals from the same litter to respond more alike than animals from different litters. In this paper attention is restricted to dichotomous response variables that frequently arise in toxicological studies, such as the occurrence of fetal death or a particular malformation. Various techniques for estimating the underlying probability of response are discussed. A number of generalized models that have recently been proposed to take the litter effect into account are breifly reviewed and compared to the simpler binomial and Poisson models. Various procedures for assessing the significance of treatment-control differences are presented and their relative merits discussed. Finally, future research needs in this area are outlined.

Animals↗
PubMed1978

The use of a correlated binomial model for the analysis of certain toxicological experiments.

In certain toxicological experiments with laboratory animals, the outcome of interest is the occurrence of dead or malformed fetuses in a litter. Previous investigations have shown that the simple one-parameter binomial and Poisson models generally provide poor fits to this type of binary data. In this paper, a type of correlated binomial model is proposed for use in this situation. First, the model is described in detail and is compared to a beta-binomial model proposed by Williams (1975). These two-parameter models are then contrasted for goodness of fit to some real-life data. Finally, numerical examples are given in which likelihood ratio tests based on these models are employed to assess the significance of treatment-control differences.

Abnormalities, Drug-Induced↗
PubMed1978

Alternatives to Rothman's approach for assessing synergism (or antagonism) in cohort studies.

K.J. Rothman has explored in some detail the issue of assessing the potential presence of synergism (or antagonism) in data generated from either a cohort or a case-control study. Arguing that the "natural" scale for quantifying the joint effects of two or more factors acting in combination is the probability scale, he has proposed a procedure based on a ratio-type index for evaluating two-factor interaction in the presence of non-zero background effects. In this paper, the authors review the rationale underlying Rothman's approach for a cohort study. They then present what they maintain is a simpler and more appropriate test procedure (utilizing a linear contrast of the observed risks) for the additive approximation to his basic probabilistic model of "no interaction." A likelihood ratio test based on his original model is also proposed, as well as a closed form approximation to it. Finally, the assessment of interaction in cohort studies involving exposure factors measured at more than two levels is addressed.

Epidemiology↗
PubMed1976

Physician management in primary care.

Minimal explicit consensus criteria in the management of patients with four indicator conditions were established by an ad hoc committee of primary care physicians practicing in different locations. These criteria were then applied to the practices of primary care physicians located in a single community by abstracting medical records and obtaining questionnaire data about patients with the indicator conditions. A standardized management score for each physician was used as the dependent variable in stepwise regression analysis with physician/practice and patient/disease characteristics as the candidate independent variables. For all physicians combined, the mean management scores were high, ranging from .78 to .93 for the four conditions. For two of the conditions, care of the normal infant and pregnant woman, the management scores were better for pediatricians and obstetricians respectively than for family physicians. For the other two conditions, adult onset diabetes and congestive heart failure, there were no differences between the management scores of family physicians and internists. Patient/disease characteristics did not contribute significantly to explaining the variation in the standardized management scores.

Adult↗
PubMed1976

Communication, compliance, and concordance between physicians and patients with prescribed medications.

Forty-six practicing physicians and 357 patients with diabetes mellitus or congestive heart failure were the subjects for this study, which focuses on the impact of medication regimen and doctor-patient communication in affecting patient medication-taking behavior and physician awareness of these behaviors. Four types of medication errors were defined: omissions, commissions, scheduling misconceptions and scheduling non-compliance. The average error rates were 19 per cent, 19 per cent, 17 per cent and 3 per cent, respectively. The combined average error was 58 per cent; scheduline non-compliance on the part of the patient was a minor component. Specific aspects of the medication regimen were associated with increased errors: (1) the more drugs involved between the doctor-patient pair, the greater the errors of omission and commission; and (2) the greater the complexity of the scheduling, the greater the errors of commission and scheduling misconceptions. If the patient did not know the function of all his drugs, errors of commission and scheduling misconception increased. Neither characteristics of patients nor the severity of disease were influential in determining the extent of medication errors. For patients with congestive heart failure, good communication of instructions and information from physician to patient was associated with low levels of all types of errors.

Aged↗