Home About us Contact
|
Home About us Contact | ||
|
|
||
Random Effects Models (random + effects_models)
Selected AbstractsMaximum Likelihood Methods for Nonignorable Missing Responses and Covariates in Random Effects ModelsBIOMETRICS, Issue 4 2003Amy L. Stubbendick Summary. This article analyzes quality of life (QOL) data from an Eastern Cooperative Oncology Group (ECOG) melanoma trial that compared treatment with ganglioside vaccination to treatment with high-dose interferon. The analysis of this data set is challenging due to several difficulties, namely, nonignorable missing longitudinal responses and baseline covariates. Hence, we propose a selection model for estimating parameters in the normal random effects model with nonignorable missing responses and covariates. Parameters are estimated via maximum likelihood using the Gibbs sampler and a Monte Carlo expectation maximization (EM) algorithm. Standard errors are calculated using the bootstrap. The method allows for nonmonotone patterns of missing data in both the response variable and the covariates. We model the missing data mechanism and the missing covariate distribution via a sequence of one-dimensional conditional distributions, allowing the missing covariates to be either categorical or continuous, as well as time-varying. We apply the proposed approach to the ECOG quality-of-life data and conduct a small simulation study evaluating the performance of the maximum likelihood estimates. Our results indicate that a patient treated with the vaccine has a higher QOL score on average at a given time point than a patient treated with high-dose interferon. [source] Meta-analysis of standardized incidence and mortality rates of childhood leukaemia in proximity to nuclear facilitiesEUROPEAN JOURNAL OF CANCER CARE, Issue 4 2007P.J. BAKER phd The meta-analysis combined and statistically analysed studies of childhood leukaemia and nuclear facilities. Focus was on studies that calculated standardized rates for individual facilities. Due to variability between study designs, eight separate analyses were performed stratified by age and zone. One hundred and thirty-six sites were used in at least one analysis. Unadjusted, fixed effects and random effects models were used. Meta-rates greater than one were found in all models at all stratification levels often achieving statistical significance. Caution must be used when interpreting these results. The meta-analysis was able to show an increase in childhood leukaemia near nuclear facilities, but does not support a hypothesis to explain the excess. Each type of model utilized has limitations. Fixed effects models give greater weight to larger studies; however, population density may be a risk factor. Random effects models give greater weight to smaller studies that may be more likely to be affected by publication bias. A limitation of the overall study design is that standardized rates must be available for individual sites which led to exclusion of studies that only calculated rates for multiple sites and those that presented other statistical methods. Further, dose-response studies do not support excess rates found near nuclear facilities. However, it cannot be ignored that the majority of studies have found elevated rates, although not usually statistically significant. [source] An 8-Year Prospective Study of the Relationship Between Cognitive Performance and Falling in Very Old AdultsJOURNAL OF AMERICAN GERIATRICS SOCIETY, Issue 8 2006Kaarin J. Anstey PhD OBJECTIVES: To determine whether cognitive performance, as distinct from cognitive impairment, predicts falling during an 8-year follow-up in a community-based sample of very old adults and to evaluate how cognitive change is associated with falling. DESIGN: Prospective cohort study including three waves of data collected in 1992, 1994, and 2000. SETTING: Population based, with the baseline sample drawn from the electoral roll. PARTICIPANTS: Inclusion criteria were completion of at least three cognitive tests at baseline and completion of the falls questionnaire at Wave 6 (N=539). MEASUREMENTS: Assessments of health and medical conditions, visual acuity, cognitive function, functional reach, semitandem stand, and grip strength were conducted in 1992 (baseline), 1994, and 2000. Self-report information on falls in the previous 12 months was obtained on each of these occasions. Marginal models using generalized estimating equations were used to assess the association between baseline cognitive performance and falling over 8 years, adjusting for sociodemographic, health, and sensorimotor variables. Random effects models were used to assess the relationship between change in cognitive performance and change in fall rate and fall risk over 8 years. RESULTS: Mini-Mental State Examination and verbal reasoning at baseline predicted rate of falling over an 8-year period. Within individuals, declines in verbal ability, processing speed, and immediate memory were associated with increases in rates of falling and fall risk. CONCLUSION: Cognitive performance is associated with falling over 8 years in very old adults and should be assessed in clinical practice when evaluating short- and long-term fall risk. [source] Assessing the impacts of grazing levels on bird density in woodland habitat: a Bayesian approach using expert opinionENVIRONMETRICS, Issue 7 2005Petra M. Kuhnert Abstract Many studies on birds focus on the collection of data through an experimental design, suitable for investigation in a classical analysis of variance (ANOVA) framework. Although many findings are confirmed by one or more experts, expert information is rarely used in conjunction with the survey data to enhance the explanatory and predictive power of the model. We explore this neglected aspect of ecological modelling through a study on Australian woodland birds, focusing on the potential impact of different intensities of commercial cattle grazing on bird density in woodland habitat. We examine a number of Bayesian hierarchical random effects models, which cater for overdispersion and a high frequency of zeros in the data using WinBUGS and explore the variation between and within different grazing regimes and species. The impact and value of expert information is investigated through the inclusion of priors that reflect the experience of 20 experts in the field of bird responses to disturbance. Results indicate that expert information moderates the survey data, especially in situations where there are little or no data. When experts agreed, credible intervals for predictions were tightened considerably. When experts failed to agree, results were similar to those evaluated in the absence of expert information. Overall, we found that without expert opinion our knowledge was quite weak. The fact that the survey data is quite consistent, in general, with expert opinion shows that we do know something about birds and grazing and we could learn a lot faster if we used this approach more in ecology, where data are scarce. Copyright © 2005 John Wiley & Sons, Ltd. [source] Meta-analysis of standardized incidence and mortality rates of childhood leukaemia in proximity to nuclear facilitiesEUROPEAN JOURNAL OF CANCER CARE, Issue 4 2007P.J. BAKER phd The meta-analysis combined and statistically analysed studies of childhood leukaemia and nuclear facilities. Focus was on studies that calculated standardized rates for individual facilities. Due to variability between study designs, eight separate analyses were performed stratified by age and zone. One hundred and thirty-six sites were used in at least one analysis. Unadjusted, fixed effects and random effects models were used. Meta-rates greater than one were found in all models at all stratification levels often achieving statistical significance. Caution must be used when interpreting these results. The meta-analysis was able to show an increase in childhood leukaemia near nuclear facilities, but does not support a hypothesis to explain the excess. Each type of model utilized has limitations. Fixed effects models give greater weight to larger studies; however, population density may be a risk factor. Random effects models give greater weight to smaller studies that may be more likely to be affected by publication bias. A limitation of the overall study design is that standardized rates must be available for individual sites which led to exclusion of studies that only calculated rates for multiple sites and those that presented other statistical methods. Further, dose-response studies do not support excess rates found near nuclear facilities. However, it cannot be ignored that the majority of studies have found elevated rates, although not usually statistically significant. [source] Ranking State Environmental Outputs: Evidence from Panel DataGROWTH AND CHANGE, Issue 1 2000John A. List This paper uses state-level pollution data from 1986,1997 to construct two indices that rank U.S. states according to environmental outputs. A major finding is that marginal performers in other indices, such as Wyoming, garner top spots in these ranking systems. The paper also presents findings from fixed and random effects models of panel data that imply state income levels are positively associated with environmental outputs after a threshold level of income is obtained. [source] Distinguishing between heterogeneity and inefficiency: stochastic frontier analysis of the World Health Organization's panel data on national health care systemsHEALTH ECONOMICS, Issue 10 2004William Greene Abstract The most commonly used approaches to parametric (stochastic frontier) analysis of efficiency in panel data, notably the fixed and random effects models, fail to distinguish between cross individual heterogeneity and inefficiency. This blending of effects is particularly problematic in the World Health Organization's (WHO) panel data set on health care delivery, which is a 191 country, 5-year panel. The wide variation in cultural and economic characteristics of the worldwide sample produces a large amount of unmeasured heterogeneity in the data. This study examines several alternative approaches to stochastic frontier analysis with panel data, and applies some of them to the WHO data. A more general, flexible model and several measured indicators of cross country heterogeneity are added to the analysis done by previous researchers. Results suggest that there is considerable heterogeneity that has masqueraded as inefficiency in other studies using the same data. Copyright © 2004 John Wiley & Sons, Ltd. [source] A modelling strategy for the analysis of clinical trials with partly missing longitudinal dataINTERNATIONAL JOURNAL OF METHODS IN PSYCHIATRIC RESEARCH, Issue 3 2003Ian R. White Abstract Standard statistical analyses of randomized controlled trials with partially missing outcome data often exclude valuable information from individuals with incomplete follow-up. This may lead to biased estimates of the intervention effect and loss of precision. We consider a randomized trial with a repeatedly measured outcome, in which the value of the outcome on the final occasion is of primary interest. We propose a modelling strategy in which the model is successively extended to include baseline values of the outcome, then intermediate values of the outcome, and finally values of other outcome variables. Likelihood-based estimation of random effects models is used, allowing the incorporation of data from individuals with some missing outcomes. Each estimated intervention effect is free of non-response bias under a different missing-at-random assumption. These assumptions become more plausible as the more complex models are fitted, so we propose using the trend in estimated intervention effects to assess the nature of any non-response bias. The methods are applied to data from a trial comparing intensive case management with standard case management for severely psychotic patients. All models give similar estimates of the intervention effect and we conclude that non-response bias is likely to be small. Copyright © 2003 Whurr Publishers Ltd. [source] Anemia and 9-Year Domain-Specific Cognitive Decline in Community-Dwelling Older Women: The Women's Health and Aging Study IIJOURNAL OF AMERICAN GERIATRICS SOCIETY, Issue 9 2009Jennifer A. Deal MHS OBJECTIVES: To test the hypothesis that anemia (hemoglobin <12 g/dL) is associated with a faster rate of cognitive decline over 9 years in a community-dwelling sample of women aged 70 to 80 at baseline. DESIGN: A population-based, prospective cohort study. SETTING: East Baltimore, Maryland. PARTICIPANTS: Four hundred thirty-six women sampled to be representative of the two-thirds least-disabled women aged 70 to 80 at baseline (1994,1996). MEASUREMENTS: Nine-year trajectories of cognitive decline, analyzed using linear random effects models, in the domains of immediate verbal recall, delayed verbal recall, psychomotor speed, and executive function. RESULTS: At baseline and after adjustment for demographic and disease covariates, women with anemia were slower to complete a test of executive function; the difference in baseline function between women with anemia and those without was ,0.43 standard deviations (SDs) (95% confidence interval (CI)=,0.74 to ,0.13) on the Trail Making Test Part B. During follow-up, anemia was associated with a faster rate of decline in memory. Between baseline and Year 3, the difference in the rates of decline between women with anemia and those without was ,0.18 SDs per year (95% CI=,0.29 to ,0.06) on the Hopkins Verbal Learning Test (HVLT) and ,0.15 SDs per year (95% CI=,0.26 to ,0.04) on the HVLT-Delayed. CONCLUSION: Anemia was associated with poorer baseline performance on a test of executive function and with faster rates of decline on tests of immediate and delayed verbal recall. If this relationship is causal, it is possible that treatment of anemia could prevent or postpone cognitive decline. [source] Low Serum Vitamin D Does Not Predict New Disability or Loss of Muscle Strength in Older WomenJOURNAL OF AMERICAN GERIATRICS SOCIETY, Issue 5 2002René Verreault MD OBJECTIVES: To determine whether serum levels of 25-hydroxyvitamin D (25(OH)D) and parathyroid hormone (PTH) predict accelerated decline in muscular strength or onset of new disability in mobility and upper extremity functioning over a 3-year follow-up. DESIGN: A community-based prospective cohort study. PARTICIPANTS: Six hundred twenty-eight moderately to severely disabled women aged 65 and older living in the community. MEASUREMENTS: Subjects were divided into three groups of baseline 25(OH)D serum levels (deficiency: <25 nmol/L; low normal: 25,52 nmol/L; high normal: ,53 nmol/L) and into tertiles of PTH levels. Objective performance measures (hip flexor, knee extensor, and grip strengths; walking speed; and time for repeated chair stands) and disability in activities involving mobility and upper extremity function were assessed at baseline and every 6 months for 3 years. Decline in performance measures and onset of new disability were compared between 25(OH)D and PTH groups using random effects models and proportional hazards models, respectively, while adjusting for age, race, education, body mass index, baseline performance, and chronic conditions. RESULTS: The annual rate of decline over 3 years in muscular strength, walking speed, and time to perform repeated chair stands was similar across 25(OH)D groups. We observed a nonsignificantly faster decline in proximal muscle strength and walking speed with increasing PTH levels. There was no association for either measure between serum levels and the risk of incident disability in activities relating to mobility and upper extremity function. CONCLUSION: This study does not support the hypothesis that vitamin D deficiency is associated with loss in muscular strength and decline in mobility and upper extremity functioning over time in older women who were moderately to severely disabled at baseline. [source] Therapist effects in randomised controlled trials: what to do about themJOURNAL OF CLINICAL NURSING, Issue 7-8 2010Stephen J Walters Aims and objectives., The aim of this study is to describe and compare three statistical methods to allow for therapist effects in individually randomised controlled trials. Background., In an individually randomised controlled trial where the intervention is delivered by a health professional it seems likely that the effectiveness of the intervention, independent of any treatment effect, could depend on the skill of the health professional delivering it. This leads to a potential clustering of the outcomes for the patients being treated by the same health professional. Design., Retrospective statistical analysis of outcomes from four example randomised controlled trial datasets with potential clustering by health professional. Methods., Three methods to allow for clustering are described: cluster level analysis; random effects models and marginal models. These models were fitted to continuous outcome data from four example randomised controlled trial datasets with potential clustering by health professional. Results., The cluster level models produced the widest confidence intervals. Little difference was found between the estimates of the regression coefficients for the treatment effect and confidence intervals between the individual patient level models for the datasets. The conclusions reached for each dataset match those published in the original papers. The intracluster correlation coefficient ranged from <0·001,0·04 for the outcomes, which shows only minor levels of clustering within the datasets. Conclusions., The models, which use individual level data are to be preferred. Treatment coefficients from these models have different interpretations. The choice of model should depend on the scientific question being asked. Relevance to clinical practice., We recommend that researchers should be aware of any potential clustering, by health professional, in their randomised controlled trial and use appropriate methods to account for this clustering in the statistical analysis of the data. [source] The Utility of Third International Mathematics and Science Study Scales in Predicting Students' State Examination PerformanceJOURNAL OF EDUCATIONAL MEASUREMENT, Issue 4 2004Nick Sofroniou To examine the predictive utility of three scales provided in the released database of the Third International Mathematics and Science Study (TIMSS) (international plausible values, standardized percent correct score, and national Rasch score), information was obtained on the performance in state examinations in mathematics and science in 1996 (2,969 Grade 8 students) and in 1997 (2,898 Grade 7 students) of students in the Republic of Ireland who had participated in TIMSS in 1995. Performance on TIMSS was related to later performance in the state examinations using normal and nonparametric maximum likelihood (NPML) random effects models. In every case, standardized percent correct scores were found to be the best predictors of later performance, followed by national Rasch scores, and lastly, by international plausible values. The estimates for normal mixing distributions are close to those estimated by the NPML approach, lending support to the validity of estimates. [source] Suicide and Alcohol: Do Outlets Play a Role?ALCOHOLISM, Issue 12 2009Fred W. Johnson Background:, The purpose of this study was to determine whether the number of alcohol outlets in local and adjacent areas, in particular bars, was related over time to completed suicide and suicide attempts. There is evidence both from studies of individuals and time series aggregate studies, mostly at the national level, of substantial alcohol involvement in suicide, but no small-area, longitudinal studies have been carried out. The present study is the first that is both longitudinal and based on a large number of small spatial units, California zip codes, a level of resolution permitting analysis of the relationship between local alcohol access and suicide rates over time. Method:, Longitudinal data were obtained from 581 consistently defined zip code areas over 6 years (1995,2000) using data from the California Index Locations Database, a geographic information system that contains both population and place information with spatial attributes for the entire state. Measures obtained from each zip code included population characteristics (e.g., median age) and place characteristics (e.g., numbers of retail and alcohol outlets) which were related in separate analyses to (i) suicide mortality and (ii) the number of hospitalizations for injuries caused by suicide attempts. The effect of place characteristics in zip code areas adjacent to each of the 581 local zip codes (spatial lags) was also assessed. Analysis methods were random effects models corrected for spatial autocorrelation. Results:, Completed suicide rates were higher in zip code areas with greater local and lagged bar densities; and higher in areas with greater local but not lagged off-premise outlet densities. Whereas completed suicide rates were lower among blacks and Hispanics, completed suicide rates were higher among low income, older whites living in less densely populated areas, that is, rural areas. Rates of suicide attempts were higher in zip code areas with greater local but not lagged bar densities, and higher among low income younger whites living in smaller households and in rural areas. Rates of attempted suicide were also higher among blacks. Completed suicide and suicide attempt rates were lower in zip code areas with greater local restaurant densities; there were no lagged effects for restaurants. Conclusions:, Bar densities in particular appear related to suicide, meaning, because this is an aggregate-level spatial analysis, that suicides, both attempted and completed, occur at greater rates in rural community areas with greater bar densities. Because the suicide rate is highest in rural areas, this study suggests that although the number of completed and attempted suicides is no doubt greater in absolute numbers in urban areas, the suicide rate, both completed and attempted, is greater in rural areas, which draws attention, perhaps much needed, to the problems of rural America. [source] | |||||||||||||