Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “linear mixed model”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9Linked to original sources
PubMed2005

Bayesian ranking of sites for engineering safety improvements: decision parameter, treatability concept, statistical criterion, and spatial dependence.

In recent years, there has been a renewed interest in applying statistical ranking criteria to identify sites on a road network, which potentially present high traffic crash risks or are over-represented in certain type of crashes, for further engineering evaluation and safety improvement. This requires that good estimates of ranks of crash risks be obtained at individual intersections or road segments, or some analysis zones. The nature of this site ranking problem in roadway safety is related to two well-established statistical problems known as the small area (or domain) estimation problem and the disease mapping problem. The former arises in the context of providing estimates using sample survey data for a small geographical area or a small socio-demographic group in a large area, while the latter stems from estimating rare disease incidences for typically small geographical areas. The statistical problem is such that direct estimates of certain parameters associated with a site (or a group of sites) with adequate precision cannot be produced, due to a small available sample size, the rareness of the event of interest, and/or a small exposed population or sub-population in question. Model based approaches have offered several advantages to these estimation problems, including increased precision by "borrowing strengths" across the various sites based on available auxiliary variables, including their relative locations in space. Within the model based approach, generalized linear mixed models (GLMM) have played key roles in addressing these problems for many years. The objective of the study, on which this paper is based, was to explore some of the issues raised in recent roadway safety studies regarding ranking methodologies in light of the recent statistical development in space-time GLMM. First, general ranking approaches are reviewed, which include naïve or raw crash-risk ranking, scan based ranking, and model based ranking. Through simulations, the limitation of using the naïve approach in ranking is illustrated. Second, following the model based approach, the choice of decision parameters and consideration of treatability are discussed. Third, several statistical ranking criteria that have been used in biomedical, health, and other scientific studies are presented from a Bayesian perspective. Their applications in roadway safety are then demonstrated using two data sets: one for individual urban intersections and one for rural two-lane roads at the county level. As part of the demonstration, it is shown how multivariate spatial GLMM can be used to model traffic crashes of several injury severity types simultaneously and how the model can be used within a Bayesian framework to rank sites by crash cost per vehicle-mile traveled (instead of by crash frequency rate). Finally, the significant impact of spatial effects on the overall model goodness-of-fit and site ranking performances are discussed for the two data sets examined. The paper is concluded with a discussion on possible directions in which the study can be extended.

Accidents, Traffic↗
PubMed2006

Incidence and patterns of polydrug use and craving for ecstasy in regular ecstasy users: an ecological momentary assessment study.

BACKGROUND: Previous studies employing retrospective assessments methods found that regular ecstasy users frequently use alcohol, marihuana and other drugs in combination with ecstasy. METHODS: Twenty-two participants (13 males, 9 females) wore a wrist actigraph/data recorder to record real-time drug use and ecstasy craving for 6 weeks. Rates of alcohol and drug use on ecstasy use versus non-use nights, and before, during, and after ecstasy use were analyzed with generalized estimation equations (GEE). Craving was modeled with GEE and linear mixed models. RESULTS: Approximately 70% of ecstasy uses occurred on Friday or Saturday nights. No drug was significantly more likely to be used on ecstasy use nights than comparison Friday and Saturday nights. On nights ecstasy was used, in general across all drugs assessed, use was more likely before and during than after ecstasy intoxication, while alcohol use was also more likely before than during ecstasy intoxication. Though low overall, craving for ecstasy increased over 24 h before use and was higher on Friday nights of weeks ecstasy was used on weekends than weeks it was not used. CONCLUSIONS: Use of ecstasy on a particular night may not be associated with any greater likelihood of using any other intoxicating drug, and use of other drugs on nights involving ecstasy use may simply reflect a "natural history" of drug-use nights that begins with alcohol, progresses to more intoxicating drugs, and ends with little drug use. Confirmation of these findings awaits further advances in the application of ecological momentary assessment methodologies.

Adult↗
PubMed2003

Sleep-wake rhythm in an irregular shift system.

Sleep in shift work has been studied extensively in regular shift systems but to a lesser degree in irregular shifts. Our main aim was to examine the sleep-wake rhythm in shift combinations ending with the night or the morning shift in two irregular shift systems. Three weeks' sleep/work shift diary data, collected from 126 randomly selected train drivers and 104 traffic controllers, were used in statistical analyses including a linear mixed model and a generalized linear model for repeated measurements. The results showed that the sleep-wake rhythm was significantly affected by the shift combinations. The main sleep period before the first night shift shortened by about 2 h when the morning shift immediately preceded the night shift as compared with the combination containing at least 36 h of free time before the night shift (reference combination). The main sleep period before the night shift was most curtailed between two night shifts, on average by 2.9 and 3.5 h among the drivers and the controllers, respectively, as compared with the reference combination. Afternoon napping increased when the morning or the day shift immediately preceded the night shift, the odds being 4.35-4.84 in comparison with the reference combination. The main sleep period before the morning shift became 0.5 h shorter when the evening shift preceded the morning shift in comparison with the sleep period after a free day. The risk for dozing off during the shift was associated only with the shift length, increasing by 17 and 35% for each working hour in the morning and the night shift, respectively. The results demonstrate advantageous and disadvantageous shift combinations in relation to sleep and make it possible to improve the ergonomy of irregular shift systems.

Adult↗
PubMed2004

Influence of previous open synovectomy on the outcome of Souter-Strathclyde total elbow prosthesis.

OBJECTIVES: Open synovectomy of the elbow joint is often performed in early stages of rheumatoid arthritis. Because of poor long-term results after synovectomy, insertion of a total elbow prosthesis is commonly used as a secondary procedure. The aim of this study is to evaluate the influence of previous synovectomy on the outcome after placement of a total elbow prosthesis. METHODS: We inserted 204 primary Souter-Strathclyde total elbow prostheses for rheumatoid arthritis. Two groups could be distinguished: group A with previous synovectomy 3.9 yr (mean) before the elbow replacement (n = 33) and group B without previous synovectomy (n = 171). The mean follow-up was 5.8 yr for group A and 6.3 yr for group B. All patients were assessed clinically and radiologically before the operation, 1 and 2 years later and then at regular intervals. The effect of previous synovectomy was analysed via a Cox model and a generalized linear mixed model for binomial data with multivariate normal random effects. RESULTS: No statistically significant effect of previous synovectomy on pain, function or complaints of the ulnar nerve could be found post-operatively. The post-operative flexion was significantly higher in group B than in group A. The complication-rates were similar for both groups. The overall survival rate for respectively group A and B with revision as endpoint was 66.9% (s.e. 13.4) versus 79.6 (s.e. 4.3) after 10 yr. CONCLUSIONS: Previous synovectomy does not diminish the outcome after total elbow prosthesis in this series and could therefore be considered in early, painful stages of rheumatoid destruction of the elbow joint.

Arthritis, Rheumatoid↗
PubMed2006

Subjective sleepiness and accident risk avoiding the ecological fallacy.

The present study of sleepiness and accident risk in a HI-FI car simulator aimed to provide subject-level relative risks (RR) with 95% confidence intervals (CI) for different levels of subjective sleepiness measured with the Karolinska Sleepiness Scale (KSS), 1 = very alert, 9 = very sleepy, fighting sleep, an effort to staying awake. Five male and five female shift workers, mean age 37 years, participated with a 2-h drive (08:00-10:00 hours) in a dynamic high-fidelity moving base driving simulator, after a night of work and after a night of sleep. Subjective sleepiness was measured with KSS every 5 min and events of incidents (two wheels outside the right lane), accidents (two wheels off the road or four wheels in opposite lane) and crashes (four wheels off the road) were recorded. The probability of an accident was modelled with a Generalized Linear Mixed Model approach to estimate subject-specific effects, rather than group average effects, to avoid the ecological fallacy. The results showed that sleepiness was strongly related to accident risk. An average subject was estimated at 28.2 times (95% CI RR = 10.7-74.1) increased risk at KSS = 8 and at 185 times (95% CI RR = 42-316) at KSS = 9 compared with KSS = 5. There were large individual differences in event propensity that complicates the prediction of absolute accident risk for individual subjects.

Accidents, Traffic↗
PubMed1999

Genetic analysis of fertility in dairy cattle using negative binomial mixed models.

Two negative binomial mixed models with different dispersion specifications were compared for analysis of dairy reproduction count data. The first model was developed previously and had heterogeneous overdispersion in an associated logarithmic scale, assigning greater uncertainty to observations with smaller conditional expectations. The second model postulated homogeneous overdispersion across all data. A simulation study was used to compare marginal modal estimates of additive genetic variance, based on these two negative binomial models, with analogous estimates computed by an overdispersed Poisson mixed model. Estimators from the second negative binomial and overdispersed Poisson models had better frequentist properties than did those from the first negative binomial model. Nevertheless, application to a data set of number of artificial inseminations until conception in Holstein heifers suggested a slightly better fit of the first negative binomial model. A marginal likelihood ratio test indicated that the additive genetic variance was significant. Cross-validation analyses suggested that the two negative binomial mixed models had slightly better predictive ability than a linear mixed model.

Animals↗
PubMed1995

A new method to explore the distribution of interindividual random effects in non-linear mixed effects models.

This article presents a new approach for exploring the distribution of interindividual random effects in nonlinear mixed effect models. The approach introduces a spline function, which transforms an assumed normally distributed interindividual random effect to an arbitrary distribution approximating that of the data. The performance of this tool is illustrated using simulated pharmacokinetic data with non-normally distributed random effects. Results of the analyses of two real kinetic data sets are also presented.

Biometry↗
PubMed2000

A linear mixed-effects model for multivariate censored data.

We apply a linear mixed-effects model to multivariate failure time data. Computation of the regression parameters involves the Buckley-James method in an iterated Monte Carlo expectation-maximization algorithm, wherein the Monte Carlo E-step is implemented using the Metropolis-Hastings algorithm. From simulation studies, this approach compares favorably with the marginal independence approach, especially when there is a strong within-cluster correlation.

Algorithms↗
PubMed1993

Technical note: computing tests of fixed effects in a restricted class of mixed models.

Inferences about fixed effects in mixed linear models are important in a variety of animal science studies. The statistical theory for making such inferences is well known, and if the variance components are known up to a proportionality constant, then optimal exact tests can be performed. Computing the test statistics, however, can still be problematic when the random effects have many levels. In practice, approximate tests that are easily computed but less efficient are usually employed. This article describes reduction in error sum of squares procedures for performing the exact test and for computing associated confidence intervals. By taking advantage of iterative algorithms for solving Henderson's mixed-model equations, the tests can be performed without inverting the covariance matrix or computing a generalized inverse of the mixed-model coefficient matrix. The procedures are illustrated on an animal model that has three random effects, two with 1,372 levels and one with 450 levels.

Algorithms↗
PubMed2003

Comparing linear and nonlinear mixed model approaches to cosinor analysis.

The cosinor model, used for variables governed by circadian and other biological rhythms, is a nonlinear model in the amplitude and acrophase parameters that has a linear representation upon transformation. With linear cosinor analysis, amplitude and acrophase for each harmonic can be computed as nonlinear functions of the estimated linear regression coefficients. Here a flexible mixed model approach to cosinor analysis is considered, where the fixed effect parameters may enter nonlinearly as acrophase and amplitude for each harmonic or linearly after transformation to regression coefficients. In addition, the random effects may enter nonlinearly as subject-specific deviations from the acrophases and amplitudes or linearly as subject-specific deviations from the regression coefficients. It is also possible for the fixed effects to enter nonlinearly while the random effects enter linearly. Additionally, we evaluate whether including higher order linear harmonic terms as random effects, that is, Rao-Khatri 'covariance adjustment', improves precision. Applying the delta method to nonlinear functions of the parameters from linear mixed cosinor models to obtain approximate variances produces results that are often identical to results from nonlinear mixed models. Consequently, traditional linear cosinor analysis can often be used to estimate and compare the nonlinear parameters of interest, that is, amplitudes and acrophases, via the delta method. This is advantageous since the nonlinear mixed model may have convergence difficulties for more complex models. However, for some multiple-group analyses, the linear cosinor transformation should not be used and we clarify when the two methods are equivalent and when they differ.

Circadian Rhythm↗
PubMed2004

A quality-adjusted survival (Q-TWiST) model for evaluating treatments for advanced stage cancer.

Quality of life is an important component in the evaluation of therapies, especially in advanced cancer. Methods available for the analysis of longitudinal quality-of-life data include linear mixed models (including growth curve models), generalized linear models, generalized estimating equations, and joint modeling of quality of life and the missingness process. Quality-adjusted survival (Q-TWiST) has also been useful to compare treatments. By weighting the durations of health states according to their quality of life, one arrives at a single end point reflecting the duration of survival and the quality of life. We propose methods for incorporating longitudinal quality-of-life data into quality-adjusted survival. We divide follow-up time into two states, "poor" and "good," based on a cut-off applied to observed quality-of-life scores. Disease progression is handled as a separate state. We then use survival analysis methods to estimate the mean duration of each state as well as mean quality-adjusted time. The analysis is repeated by varying the cut-off to illustrate the range of possible results. Finally a single summary analysis is achieved by averaging (possibly with weights) across the cut-offs used. We illustrate the methodology using data from a cancer clincial trial.

Glioma↗
PubMed1996

Meta-analysis of published data using a linear mixed-effects model.

This paper describes the use of a linear mixed-effects regression model as a framework for the meta-analysis of published data. It generalizes the random-effects models used by DerSimonian and Laird (1986, Controlled Clinical Trials 7, 177-188) and Begg and Pilote (1991, Biometrics 47, 899-906), and describes the use of the model using examples from these papers and the data given by Tori et al.

Biometry↗
PubMed1996

A generalized Michaelis-Menten response surface.

We describe an application of recently developed generalized Michaelis-Menten response surface and non-linear mixed model methodologies to model glucose utilization in foetal sheep. More specifically, we model the response surface of glucose utilization rate in the foetal sheep as a function of glucose and insulin concentrations using a three-dimensional analogue of the Michaelis-Menten pharmacokinetic model. To account for multiple measurements per sheep, we apply the non-linear mixed effects model proposed by Lindstrom and Bates using the EM algorithm computational scheme presented by Hirst et al.

Animals↗
PubMed2004

Modeling antitumor activity by using a non-linear mixed-effects model.

The response of solid tumors to antitumor treatment generally declines markedly with treatment time. Sometimes, a tumor regrows (rebounds) before the end of the treatment period. Studies of the patterns of tumor response to treatment are important, because they may provide useful information for clinical decision-making. We have investigated patterns of tumor response in mouse xenograft tumors by using data from a study conducted at St. Jude Children's Research Hospital. We applied a biexponential non-linear mixed-effects model to an analysis of changes in tumor volume over a given period of treatment. The model gives a good fit to the data, even for small sample sizes. We addressed the relation between the baseline tumor volumes and the decay rates of the first and second stages of the tumor's response to treatment, and we applied sensitive analysis to determine the effect of using different imputed values for missing data. We also proposed a novel approach to a comparison of the antitumor effects of three different treatments, and we used the data from a St. Jude study to demonstrate the potential of this comparison approach in cancer clinical decision-making.

Algorithms↗
PubMed2005

Non-linear mixed-effects models with stochastic differential equations: implementation of an estimation algorithm.

Pharmacokinetic/pharmacodynamic modelling is most often performed using non-linear mixed-effects models based on ordinary differential equations with uncorrelated intra-individual residuals. More sophisticated residual error models as e.g. stochastic differential equations (SDEs) with measurement noise can in many cases provide a better description of the variations, which could be useful in various aspects of modelling. This general approach enables a decomposition of the intra-individual residual variation epsilon into system noise w and measurement noise e. The present work describes implementation of SDEs in a non-linear mixed-effects model, where parameter estimation was performed by a novel approximation of the likelihood function. This approximation is constructed by combining the First-Order Conditional Estimation (FOCE) method used in non-linear mixed-effects modelling with the Extended Kalman Filter used in models with SDEs. Fundamental issues concerning the proposed model and estimation algorithm are addressed by simulation studies, concluding that system noise can successfully be separated from measurement noise and inter-individual variability.

Algorithms↗
PubMed2006

Short communication: genetic evaluation of the interval from first to last insemination with survival analysis and linear models.

Sire breeding values for the interval between the first and last insemination were predicted using 4 proportional hazards models (survival analyses) and 2 linear mixed models to determine which would result in a more accurate genetic evaluation. A stochastic simulation describing the reproductive cycle of first-parity cows was conducted, in which true breeding values for conception rate were created. The model included the effects of sire and herd. The highest correlations between true breeding values for conception rate and breeding values for the interval between first and last insemination predicted by the survival analysis model and the linear model were 0.803 and 0.744, respectively. The results showed that when pregnancy status was known, survival models were more accurate than linear models to predict breeding values for conception rate when using observations on the interval between first and last insemination.

Animals↗
PubMed2007

Impact of modelling intra-subject variability on tests based on non-linear mixed-effects models in cross-over pharmacokinetic trials with application to the interaction of tenofovir on atazanavir in HIV patients.

We evaluated the impact of modelling intra-subject variability on the likelihood ratio test (LRT) and the Wald test based on non-linear mixed effects models in pharmacokinetic interaction and bioequivalence cross-over trials. These tests were previously found to achieve a good power but an inflated type I error when intra-subject variability was not taken into account. Trials were simulated under H0 and several H1 and analysed with the NLME function. Different configurations of the number of subjects n and of the number of samples per subject J were evaluated for pharmacokinetic interaction and bioequivalence trials. Assuming intra-subject variability in the model dramatically improved the type I error of both interaction tests. For the Wald test, the type I error decreased from 22, 14 and 7.7 per cent for the original (n = 12, J = 10), intermediate (n = 24, J = 5) and sparse (n = 40, J = 3) designs, respectively, down to 7.5, 6.4 and 3.5 per cent when intra-subject variability was modelled. The LRT achieved very similar results. This improvement seemed mostly due to a better estimation of the standard error of the treatment effect. For J = 10, the type I error was found to be closer to 5 per cent when n increased when modelling intra-subject variability. Power was satisfactory for both tests. For bioequivalence trials, the type I error of the Wald test was 6.4, 5.7 and 4.2 per cent for the original, intermediate and sparse designs, respectively, when modelling intra-subject variability. We applied the Wald test to the pharmacokinetic interaction of tenofovir on atazanavir, a novel protease inhibitor. A significant decrease of the area under the curve of atazanavir was found when patients received tenofovir.

Adenine↗
PubMed2000

Nonconjugate Bayesian analysis of variance component models.

We consider the usual normal linear mixed model for variance components from a Bayesian viewpoint. With conjugate priors and balanced data, Gibbs sampling is easy to implement; however, simulating from full conditionals can become difficult for the analysis of unbalanced data with possibly nonconjugate priors, thus leading one to consider alternative Markov chain Monte Carlo schemes. We propose and investigate a method for posterior simulation based on an independence chain. The method is customized to exploit the structure of the variance component model, and it works with arbitrary prior distributions. As a default reference prior, we use a version of Jeffreys' prior based on the integrated (restricted) likelihood. We demonstrate the ease of application and flexibility of this approach in familiar settings involving both balanced and unbalanced data.

Algorithms↗