SEARCH · PubMed Health
Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
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.
Correlated observations (within centers, families, subjects,.) are common in epidemiology. Even when one is only interested in the modeling of means according to risk factors, it is also necessary to model the variance-covariance matrix of the observations in order to make correct inferences on the parameters of interest. All the more so when the aim of the survey is the measurement of these correlations or of the variance of the random effects from which they are assumed to originate. We discuss, within the framework of the linear and of the logistic models, the implications of two choices for the modeling of covariances. The mixed model shows the unobserved elements responsible for the similarity between certain observations. In a longitudinal survey, for instance, one can use a random effect, specific to each subject, expressing how much a subject's trajectory is translated as compared to what is expected according to its characteristics (age, sex,.). The marginal approach leads to modeling separately the means and the covariance matrix of the observations. The distinction between these two approaches is important for non linear models, in particular the logistic one. We insist on the interconnection between a mixed model formulation and a marginal one, as well as on the implication of the choice in terms of the parameters' interpretation.
Unit risks used for quantitative cancer risk assessment are defined for constant lifetime exposures. The condition of temporal stability, however, usually is not fulfilled in environmental health applications. In practice, cancer risks for time-dependent exposures are often estimated by calculating lifetime average exposure, assuming a mean life expectancy of 70 years. In the present paper we discuss the question whether this is an appropriate procedure considering various variants of multi stage and epidemiological relative risk models. For this purpose, lifetime risks for time dependent exposures as calculated according to the respective model assumptions, were compared with lifetime risks estimated by the lifetime average exposure approach. As typical exposure histories in environmental health applications we studied exposures either limited to the first 5 years of life (children scenario) or limited to duration of employment (30th to 65th year of age; occupational scenario). The consideration of multistage models (Armitage-Doll- and Moolgavkar-Venzon-Knudson model) in general would not induce serious bias in risk estimation when exposures are limited to middle ages (occupational scenario). On the other hand, when exposures occur only in very young ages or only in very old ages the risk estimated by using lifetime average exposure is not comparable with the predictions of multistage models. Whereas the degree of possible underestimation is bounded by factors well below 10, the amount of possible overestimation is unbounded and may become arbitrarily high, when exposures concentrate in extreme ages. In a second part of the study we investigated different relative risk models, taking lung cancer as an example. The models differed with respect to assumptions on latent periods and moderating effects of age at exposure and age at risk. The simulations showed that the unit risk concept is appropriate for the occupational scenario. For the children scenario results strongly depend on the assumptions made. Whereas the degree of possible underestimation is acceptable, in some models the degree of possible overestimation may become arbitrarily high. Both parts of the study showed that bias induced by using lifetime average exposure is acceptable when exposures are limited to middle ages. On the other hand, the unit risk concept should not uncritically be applied to exposures limited to early childhood (e.g., in kindergartens or due to mouthing activities). Depending on the assumptions made, lifetime risk may either be moderately underestimated or grossly overestimated. Without additional knowledge on mechanisms or latency period risk estimations are of questionable value. With respect to exposures in childhood regulation should concentrate on initiating substances or substances known to have long latent periods, respectively. With respect to cancers which occur relatively frequent already in childhood specific considerations are recommended.
To increase the accuracy of calculation of tolerance doses of the likelihood of radiation-induced complications in normal organs and tissues by using mathematical models, the author has developed a method for local mathematical model parameter adjustment (LMMPA) which included analysis of the structure of a mathematical model, systematization of clinical information and its goal-oriented use to determine the parameters of a model. The necessity of developing the LMMPA method stemmed from the complexity of the body exposed to radiation a system and from the quest for taking into account the impact of the system on the values of mathematical models. LMMPA may be regarded as the extension of determination of the parameters of mathematical models or as the interactive determination of their parameters that describing radiation exposures of complex biological systems. Different aspects of using of LMMPA are indicated how to apply it to the determination of tolerance doses for connective tissue, lung tissue, and the brain by using the Ellis and LQ models. The extended LMMPA is shown how to determine the parameters of a mathematical models for calculation of the likelihood of radiation-induced complications in the lung tissue.
A computer model of the olfactory bulb was constructed. The paper describes: 1) the general architecture of a model neuron network that reflects the neurophysiological experimental and theoretical data on the structural and functional organization of the peripheral part of the olfactory system, the olfactory bulb with inputs from olfactory receptor neurons; 2) the organization of each of three levels of the model: receptors, olfactory glomeruli, and basic neurons; and 3) a scenario of the computer model work. In some aspects, in particular, in the principle of information presentation, the treatment of the role of basic neurons (mitral and tufted cells), and their interrelations in modules, the model favorably differs from the available olfactory bulb models. The model is basic and provides further refinement of the architecture, an increase in the number of modules, and the modeling of the learning process.
The Apc(Min/+) mouse model and the azoxymethane (AOM) rat model are the main animal models used to study the effect of dietary agents on colorectal cancer. We reviewed recently the potency of chemopreventive agents in the AOM rat model (D. E. Corpet and S. Tache, Nutr. Cancer, 43: 1-21, 2002). Here we add the results of a systematic review of the effect of dietary and chemopreventive agents on the tumor yield in Min mice. The review is based on the results of 179 studies from 71 articles and is displayed also on the internet http://corpet.net/min.(2) We compared the efficacy of agents in the Min mouse model and the AOM rat model, and found that they were correlated (r = 0.66; P < 0.001), although some agents that afford strong protection in the AOM rat and the Min mouse small bowel increase the tumor yield in the large bowel of mutant mice. The agents included piroxicam, sulindac, celecoxib, difluoromethylornithine, and polyethylene glycol. The reason for this discrepancy is not known. We also compare the results of rodent studies with those of clinical intervention studies of polyp recurrence. We found that the effect of most of the agents tested was consistent across the animal and clinical models. Our point is thus: rodent models can provide guidance in the selection of prevention approaches to human colon cancer, in particular they suggest that polyethylene glycol, hesperidin, protease inhibitor, sphingomyelin, physical exercise, epidermal growth factor receptor kinase inhibitor, (+)-catechin, resveratrol, fish oil, curcumin, caffeate, and thiosulfonate are likely important preventive agents.
OBJECTIVES: Not only architects but also information managers need models and modeling tools for their subject of work. Especially for supporting strategic information management in hospitals, the meta model 3LGM2 is presented as an ontological basis for modeling the comprehensive information system of a hospital (HIS). METHODS: In a case study, requirements for modeling HIS have been deduced. Accordingly 3LGM2 has been designed to describe HIS by concepts on three layers. The domain layer consists of enterprise functions and entity types, the logical tool layer focuses on application components and the physical tool layer describes physical data processing components. In contrast to other approaches a lot of inter-layer-relationships exist. 3LGM2 is defined using the Unified Modeling Language (UML). RESULTS: Models of HIS can be created which comprise not only technical and semantic aspects but also computer-based and paper-based information processing. A software tool supporting the creation of 3LGM2 compliant models in a graphical way has been developed. The tool supports in detecting those shortcomings at the logical or the physical tool layers which make it impossible to satisfy the information needs at the domain layer. 3LGM2 can also be used as an ontology for describing HIS in natural language. CONCLUSIONS: Strategic information management even in large hospitals should be and can be supported by dedicated methods and tools. Although there have been good experiences with 3LGM2 concerning digital document archiving at the Leipzig University Hospital, which are presented in part 2, the benefit of the proposed method and tool has to be further evaluated.
Research in cancer chemoprevention involves a number of activities, the first and foremost of which is acquisition of detailed knowledge concerning the process of carcinogenesis and identification of points of intervention whereby the process can be reversed or stalled. Parallel to this is the search for ideal chemopreventive agents--natural or synthetic--and screening for their activity and efficacy in vitro and in vivo. For ethical reasons it is not possible to test new agents on humans, so preclinical studies are dependent on results first being obtained with suitable animal models. Since it is not possible for a single model to reflect the diversity and heterogeneity of human cancers, it is necessary to have as many different models as possible, depending on the requirement of the studies on different aspects of cancer biology. Advances in research on carcinogenesis and chemoprevention therefore have to be accompanied by development of appropriate laboratory animal models using a variety of carcinogens that produce tumours at different sites. Animal models have contributed significantly to our understanding of carcinogenesis and ways to intervene in the underlying processes. Many animal carcinogenesis and tumour models have been found to mirror corresponding human cancers with respect to cell of origin, morphogenesis, phenotype markers and genetic alteration. In spite of the fact that interpolation of data from animal studies to humans is difficult for various reasons, animal models are widely used for assessment of new compounds with cancer chemopreventive potential and for preclinical trials. So despite the movements of animal rights activists, animal models will continue to be used for biomedical research for saving human lives. In doing so, care should be taken to treat and handle the animals with minimal discomfort to them and ensuring that alternatives are used whenever possible.
Modeling vocal-fold vibration is extremely significant in realizing the vibration properties of human vocal folds and investigating their physiological and pathological characteristics. A combined model presented is two mass-finite element (T-F) model, which integrates all merits of both the finite element method (FEM) model and the asymmetric two-mass model of vocal folds. The high-speed glottis graph (HGG) can also be synthesized by the model. The result shows that T-F model can simulate the vibration behavior of normal and pathological vocal folds in a more realistic way with competitively computational speed. Therefore, the T-F model is helpful to gaining a thorough understanding of the vibration properties of vocal folds.
Mixed models take the dependency between observations based on the same person into account by introducing one or more random effects. After introducing the mixed model framework, it is explained, by taking the Rasch model as a generic example, how item response models can be conceptualized as generalized linear and nonlinear mixed models. Common estimation methods for generalized linear and nonlinear models are discussed. In a simulation study, the performance of four estimation methods is assessed for the Rasch model under different conditions regarding the number of items and persons, and the degree of interindividual differences. The estimation methods included in the study are: an approximation of the integral over the random effect by means of Gaussian quadrature; direct maximization with a sixth-order Laplace approximation to the integrand; a linearized approximation of the nonlinear model employing PQL2; and finally a Bayesian MCMC method. It is concluded that the estimation methods perform almost equally well, except for a slightly worse recovery of the variance parameter for PQL2 and MCMC.
During the past decade, the rice agroecosystem and its associated mosquitoes have been the subject of an extensive research effort directed toward the development and implementation of integrated pest management (IPM) strategies. The objective of this work was to synthesize the literature and unpublished data on the rice agroecosystem into a comprehensive simulation model of the key elements of the system known to influence the population dynamics of Psorophora columbiae. Subsequent companion papers will present a validation of these models, provide an in-depth analysis of the population dynamics of Ps. columbiae, and evaluate current and proposed IPM strategies for this mosquito. This paper describes the development of 2 models: WaterMod: Because spatial and temporal distributions of surface water and soil moisture play a decisive role in the dynamics of Ps. columbiae, an essentially hydrological simulator was developed. Its purpose is to provide environmental inputs for a second model (PcSim) which simulates the population dynamics of Ps. columbiae. WaterMod utilizes data on weather, agricultural practices, and soil characteristics for a particular region to generate a data set containing daily estimates of soil moisture and depth of water table for 12 representative areas comprising the rice agroecosystem. This model could be used to provide hydrologic inputs for additional simulation models of other riceland mosquito species. PcSim: This model simulates the population dynamics of Ps. columbiae by using the computer to maintain a daily accounting of the absolute number of mosquitoes within each daily age class for each life stage. The model creates estimates of the number of eggs, larvae, pupae, and adults for a representative l-ha area of a rice agroecosystem.
Advances in the understanding of human respiratory disease can come from careful clinical studies of the diseases as they occur in man, but such studies are naturally limited in terms of experimental manipulation. In the last 2 decades, an increasingly complex plethora of experimental respiratory disease models has been developed and utilized by investigators, but relatively less attention has been paid to the naturally occurring pulmonary diseases of animals as potential models. This paper is aimed at presenting selected examples of spontaneous pulmonary disease in animals that may serve as exploitable models for human chronic bronchitis, bronchiectasis, emphysema, interstitial lung disease, hypersensitivity pneumonitis, hyaline membrane disease, and bronchial asthma. Chronic bronchitis in dogs is characterized by chronic cough, excessive mucus production, and chronic inflammatory changes in bronchial walls. The disease affects mainly smaller-breed dogs of middle age or older. Equine chronic bronchitis tends to be a small airway disease with marked goblet cell proliferation and excessive mucus production, which may be accompanied by alveolar emphysema. Many animals develop bronchiectasis or bronchiolitis obliterans secondary to chronic suppurative bronchopneumonia, but chronic respiratory disease (CRD) of rats may be the most useful model of bronchiectasis. Models for emphysema must include actual alveolar destruction and ideally should be accompanied by appropriate pathophysiologic decrements. Many animals occasionally develop emphysema, but the disease has not been well documented, except possibly in horses. The interstitial lung diseases of man represent a complicated and poorly understood group of entities and near-entities. The same is true for animals, although interstitial lung disease in animals is much less common than bronchopneumonia. Cattle seem prone to develop interstitial lesions. Proliferative interstitial pneumonia of cattle includes many morphologic similarities to the spectrum of human interstitial pneumonitides. Fibrosing alveolitis of cattle is a morphologic end point that may have its origins in different forms of interstitial injury. Hypersensitivity pneumonitis has been best detailed in cattle and in horses and is clinically, etiologically, immunologically, and morphologically similar to the disease in man. Hyaline membrane disease has been poorly documented in animals, with the possible exception of the neonatal respiratory distress syndromes of foals and piglets. Bronchial asthma is similarly not well established as a spontaneous disease in animals, although experimental models exist. Eosinophilic bronchiolitis of cattle may represent a useful asthma model but has been poorly detailed. In order to make them useful as models, more attention should be paid to detailing the clinical, morphologic, and etiologic aspects of these naturally occurring animal pulmonary diseases.
In contrast to the great number of etiological models for bulimia nervosa, few conceptions determine the therapeutic practice. Two of these models, the ego-psychological model of reduced impulse-control and the behavioural model focussing the influence of "restrained eating" are empirically compared. Both models are not supposed to be of ubiquitous validity, but valid for subgroups of bulimic women. The theoretical assumptions are translated into structural equations and a sample of 127 bulimic women is divided up into 2 subsamples according to criteria which are presumed to define the validity of both etiological models. In agreement with the assumptions the "restrained eating" model shows the better "fit" with the one subsample while the ego-psychological model is more adequate for the other group.
One of the key challenges within medical information sciences is the development of useful models for biological structure and its variability. Many biomedical problems involve the elucidation of structure (for example, from experimental data or from imaging studies), and structural models can often drive the process of inferring precise structure from data. Ideally, model-driven data interpretation combines knowledge about the generic features of a class of biological structures (as contained within a model) with data that provide specific information (often noisy) about a particular instance of the class. In this paper we briefly discuss model-driven determination of biological structure as an example of a structural constraint satisfaction problem. We describe a probabilistic implementation of structural constraint satisfaction, and show that our formulation of a particular organ modeling technology (Radial Contour Models) exhibits promising performance. Our results demonstrate the utility of probabilistic models for the solution of structural constraint satisfaction problems.
Measuring bone mineral density (BMD) is currently the best modality to diagnose osteoporosis and predict future fractures. The use of risk factors to predict BMD and fracture risk has been considered to be inadequate for precise diagnostic purpose, but it may be helpful as a screening tool to determine who actually needs BMD assessment. Recently, artificial neural network (ANN), a nonlinear computational model, has been used in clinical diagnosis and classification. In the present study, we evaluated the risk factors associated with low BMD in Thai postmenopausal women and assessed the prediction of low BMD using an ANN model compared to a logistic regression model. The subjects consisted of 129 Thai postmenopausal women divided into 2 groups, 100 subjects in the training set and the remaining 29 subjects in the validation set. The subjects were classified as having either low BMD or normal BMD by using BMD value 1 SD lower than the mean value of young adults as the cutoff point. Decreased body weight, decreased hip circumference and increased years since menopause were found to be associated with low BMD at the lumbar spine by logistic regression. For the femoral neck, increased age and decreased urinary calcium were associated with low BMD. The models had a sensitivity of 85.0 per cent, a specificity of 11.1 per cent and an accuracy of 62.0 per cent for the diagnosis of low BMD at the lumbar spine when tested in the validation group. For the femoral neck, the sensitivity, specificity and accuracy were 90.5 per cent, 12.5 per cent, and 69.0 per cent, respectively. Models based on ANN correctly classified 65.5 per cent of the subjects in the validation group according to BMD at the lumbar spine with a sensitivity of 80.0 per cent and a specificity of 33.3 per cent while it correctly classified 58.6 per cent of the subjects at the femoral neck with a sensitivity of 76.2 per cent and a specificity of 12.5 per cent. There was no significant difference in terms of accuracy, sensitivity and specificity in the prediction of low BMD at the lumbar spine or the femoral neck between ANN model and logistic regression model. We concluded that ANN does not perform better than convention statistical methods in the prediction of low BMD. The less than perfect performance of the prediction rules used in the prediction of low BMD may be due to the lack of adequate association between the commonly used risk factors and BMD rather than the nature of the computational models.
The aim of the paper is to describe the development of nursing theory. As an example, a theoretical model of the compliance of young diabetics and related factors is built. The content of the theoretical model is not described but the process of developing the model is presented. In the first phase, a hypothetical model of young diabetics' compliance was developed inductively. The data were collected by interviewing 51 young diabetics aged 13-17 years, by observing the behaviour of 18 of these young people during an adaptation course and by analysing their drawings (N = 17). The data were analysed by using continuous comparative analyses and content analyses. In the second phase, an instrument for testing the model was developed. For this purpose, data were collected with a questionnaire from young diabetics aged 12-17 years (N = 91). The content validity, construct validity and reliability of the instrument turned out to be quite good. In the third phase, the hypothetical model was tested and developed further by using LISREL (linear structural relations) analyses. The data were collected with a questionnaire from young diabetics aged 13-17 years (N = 346). In the fourth phase, the model was expanded using the qualitative data (N = 51). The categories which had been discovered by continuous comparative analyses were quantified. The data were analysed by cross-tabulation, the chi square test and discriminant analyses. A summary of the results has been presented by building a theoretical model of young diabetics' compliance and related factors.
We consider problems involving the comparison of two or more treatments where we have the opportunity to adjust for relevant covariates either conditionally in a regression model or implicitly in repeated measures data, for example, in crossover trials. It is seen that for data arising from non-Normal distributions there is the possibility that models adjusting for covariates and those not adjusting for covariates will be inconsistent, that is, at most one of the models can be valid. Alternatively, even if conditional and unconditional models are valid, parameters in each model may have different interpretations. We note that this presents difficulties for the specification and interpretation of the analysis. It is also clear that model validation is critical. Specific attention is paid to survival data analysed by the Cox proportional hazards model.
PURPOSE: A population pharmacokinetic-pharmacodynamic model accounting for placebo effect was used to relate nicotine concentration and enforced smoking cessation craving score measured by the Tiffany rating scale short form. METHODS: Twenty-four smokers were enrolled in a placebo-controlled, randomized, double-blind, three periods, crossover trial. The study objective was to describe the nicotine-induced changes on craving scores. Two modeling strategies based on a mechanistic (indirect response models with drug-related inhibition on the k(in) synthesis rate and with a drug-related stimulation of the k(out) removal rate were evaluated) and a probabilistic (logistic regression) approach were used. RESULTS: Placebo response model properly fitted the circadian changes on craving scores. The analysis revealed that the indirect response model with inhibition on k(in) was the preferred model for the smoking data whereas the preferred model for the Nicotine Replacement Therapy data was the one with stimulation on k(out). The logistic analysis showed that the nicotine concentration was a significant predictor of reduction in craving during the free-smoking period. CONCLUSIONS: Nicotine dosage regimen can influence the nicotine mechanism of action: an instantaneous delivery at an individually selected time seems to inhibit the onset of craving while constant delivery at a pre-defined time seems to attenuate the craving.