Working covariance model selection for generalized estimating equations

We investigate methods for data-based selection of working covariance models in the analysis of correlated data with generalized estimating equations. We study two selection criteria: Gaussian pseudolikelihood and a geodesic distance based on discrepancy between model-sensitive and model-robust regression parameter covariance estimators. The Gaussian pseudolikelihood is found in simulation to be reasonably sensitive for several response distributions and noncanonical mean-variance relations for longitudinal data. Application is also made to a clinical dataset. Assessment of adequacy of both correlation and variance models for longitudinal data should be routine in applications, and we describe open-source software supporting this practice.

Discussion of “Generalized estimating equations: Notes on the choice of the working correlation matrix”

Objective To discuss generalized estimating equations as an extension of generalized linear models by commenting on the paper of Ziegler and Vens "Generalized Estimating Equations. Notes on the Choice of the Working Correlation Matrix". Methods Inviting an international group of experts to comment on this paper. Results Several perspectives have been taken by the discussants. Econometricians have established parallels to the generalized method of moments (GMM). Statisticians discussed model assumptions and the aspect of missing data Applied statisticians; commented on practical aspects in data analysis. Conclusions In general, careful modeling correlation is encouraged when considering estimation efficiency and other implications, and a comparison of choosing instruments in GMM and generalized estimating equations, (GEE) would be worthwhile. Some theoretical drawbacks of GEE need to be further addressed and require careful analysis of data This particularly applies to the situation when data are missing at random.

Working-correlation-structure identification in generalized estimating equations

Selecting an appropriate working correlation structure is pertinent to clustered data analysis using generalized estimating equations (GEE) because an inappropriate choice will lead to inefficient parameter estimation. We investigate the well-known criterion of QIC for selecting a working correlation Structure. and have found that performance of the QIC is deteriorated by a term that is theoretically independent of the correlation structures but has to be estimated with an error. This leads LIS to propose a correlation information criterion (CIC) that substantially improves the QIC performance. Extensive simulation studies indicate that the CIC has remarkable improvement in selecting the correct correlation structures. We also illustrate our findings using a data set from the Madras Longitudinal Schizophrenia Study.

A generalized estimating equations approach for analysis of the impact of new technology on a trawl fishery

The article describes a generalized estimating equations approach that was used to investigate the impact of technology on vessel performance in a trawl fishery during 1988-96, while accounting for spatial and temporal correlations in the catch-effort data. Robust estimation of parameters in the presence of several levels of clustering depended more on the choice of cluster definition than on the choice of correlation structure within the cluster. Models with smaller cluster sizes produced stable results, while models with larger cluster sizes, that may have had complex within-cluster correlation structures and that had within-cluster covariates, produced estimates sensitive to the correlation structure. The preferred model arising from this dataset assumed that catches from a vessel were correlated in the same years and the same areas, but independent in different years and areas. The model that assumed catches from a vessel were correlated in all years and areas, equivalent to a random effects term for vessel, produced spurious results. This was an unexpected finding that highlighted the need to adopt a systematic strategy for modelling. The article proposes a modelling strategy of selecting the best cluster definition first, and the working correlation structure (within clusters) second. The article discusses the selection and interpretation of the model in the light of background knowledge of the data and utility of the model, and the potential for this modelling approach to apply in similar statistical situations.

Unbiased estimating equations from working correlation models for irregularly timed repeated measures

The method of generalized estimating equation-, (GEEs) has been criticized recently for a failure to protect against misspecification of working correlation models, which in some cases leads to loss of efficiency or infeasibility of solutions. However, the feasibility and efficiency of GEE methods can be enhanced considerably by using flexible families of working correlation models. We propose two ways of constructing unbiased estimating equations from general correlation models for irregularly timed repeated measures to supplement and enhance GEE. The supplementary estimating equations are obtained by differentiation of the Cholesky decomposition of the working correlation, or as score equations for decoupled Gaussian pseudolikelihood. The estimating equations are solved with computational effort equivalent to that required for a first-order GEE. Full details and analytic expressions are developed for a generalized Markovian model that was evaluated through simulation. Large-sample ".sandwich" standard errors for working correlation parameter estimates are derived and shown to have good performance. The proposed estimating functions are further illustrated in an analysis of repeated measures of pulmonary function in children.

Estimating equations for removal data analysis

We consider the problem of estimating a population size from successive catches taken during a removal experiment and propose two estimating functions approaches, the traditional quasi-likelihood (TQL) approach for dependent observations and the conditional quasi-likelihood (CQL) approach using the conditional mean and conditional variance of the catch given previous catches. Asymptotic covariance of the estimates and the relationship between the two methods are derived. Simulation results and application to the catch data from smallmouth bass show that the proposed estimating functions perform better than other existing methods, especially in the presence of overdispersion.

Robust estimating functions and bias correction for longitudinal data analysis

Robust methods are useful in making reliable statistical inferences when there are small deviations from the model assumptions. The widely used method of the generalized estimating equations can be "robustified" by replacing the standardized residuals with the M-residuals. If the Pearson residuals are assumed to be unbiased from zero, parameter estimators from the robust approach are asymptotically biased when error distributions are not symmetric. We propose a distribution-free method for correcting this bias. Our extensive numerical studies show that the proposed method can reduce the bias substantially. Examples are given for illustration.

Iterative estimating equations: Linear convergence and asymptotic properties

We propose an iterative estimating equations procedure for analysis of longitudinal data. We show that, under very mild conditions, the probability that the procedure converges at an exponential rate tends to one as the sample size increases to infinity. Furthermore, we show that the limiting estimator is consistent and asymptotically efficient, as expected. The method applies to semiparametric regression models with unspecified covariances among the observations. In the special case of linear models, the procedure reduces to iterative reweighted least squares. Finite sample performance of the procedure is studied by simulations, and compared with other methods. A numerical example from a medical study is considered to illustrate the application of the method.

Analysing commercial catch and effort data from a penaeid trawl fishery: A comparison of linear models, mixed models, and generalised estimating equations approaches

Statistical methods are often used to analyse commercial catch and effort data to provide standardised fishing effort and/or a relative index of fish abundance for input into stock assessment models. Achieving reliable results has proved difficult in Australia's Northern Prawn Fishery (NPF), due to a combination of such factors as the biological characteristics of the animals, some aspects of the fleet dynamics, and the changes in fishing technology. For this set of data, we compared four modelling approaches (linear models, mixed models, generalised estimating equations, and generalised linear models) with respect to the outcomes of the standardised fishing effort or the relative index of abundance. We also varied the number and form of vessel covariates in the models. Within a subset of data from this fishery, modelling correlation structures did not alter the conclusions from simpler statistical models. The random-effects models also yielded similar results. This is because the estimators are all consistent even if the correlation structure is mis-specified, and the data set is very large. However, the standard errors from different models differed, suggesting that different methods have different statistical efficiency. We suggest that there is value in modelling the variance function and the correlation structure, to make valid and efficient statistical inferences and gain insight into the data. We found that fishing power was separable from the indices of prawn abundance only when we offset the impact of vessel characteristics at assumed values from external sources. This may be due to the large degree of confounding within the data, and the extreme temporal changes in certain aspects of individual vessels, the fleet and the fleet dynamics.

Working correlation structure misspecification, estimation and covariate design: Implications for generalised estimating equations performance

The method of generalised estimating equations for regression modelling of clustered outcomes allows for specification of a working matrix that is intended to approximate the true correlation matrix of the observations. We investigate the asymptotic relative efficiency of the generalised estimating equation for the mean parameters when the correlation parameters are estimated by various methods. The asymptotic relative efficiency depends on three-features of the analysis, namely (i) the discrepancy between the working correlation structure and the unobservable true correlation structure, (ii) the method by which the correlation parameters are estimated and (iii) the 'design', by which we refer to both the structures of the predictor matrices within clusters and distribution of cluster sizes. Analytical and numerical studies of realistic data-analysis scenarios show that choice of working covariance model has a substantial impact on regression estimator efficiency. Protection against avoidable loss of efficiency associated with covariance misspecification is obtained when a 'Gaussian estimation' pseudolikelihood procedure is used with an AR(1) structure.

Estimating equations with nonignorably missing response data

Troxel, Lipsitz, and Brennan (1997, Biometrics 53, 857-869) considered parameter estimation from survey data with nonignorable nonresponse and proposed weighted estimating equations to remove the biases in the complete-case analysis that ignores missing observations. This paper suggests two alternative modifications for unbiased estimation of regression parameters when a binary outcome is potentially observed at successive time points. The weighting approach of Robins, Rotnitzky, and Zhao (1995, Journal of the American Statistical Association 90, 106-121) is also modified to obtain unbiased estimating functions. The suggested estimating functions are unbiased only when the missingness probability is correctly specified, and misspecification of the missingness model will result in biases in the estimates. Simulation studies are carried out to assess the performance of different methods when the covariate is binary or normal. For the simulation models used, the relative efficiency of the two new methods to the weighting methods is about 3.0 for the slope parameter and about 2.0 for the intercept parameter when the covariate is continuous and the missingness probability is correctly specified. All methods produce substantial biases in the estimates when the missingness model is misspecified or underspecified. Analysis of data from a medical survey illustrates the use and possible differences of these estimating functions.

Estimating equations for parameters in stochastic growth models from tag-recapture data

James (1991, Biometrics 47, 1519-1530) constructed unbiased estimating functions for estimating the two parameters in the von Bertalanffy growth curve from tag-recapture data. This paper provides unbiased estimating functions for a class of growth models that incorporate stochastic components and explanatory variables. a simulation study using seasonal growth models indicates that the proposed method works well while the least-squares methods that are commonly used in the literature may produce substantially biased estimates. The proposed model and method are also applied to real data from tagged rack lobsters to assess the possible seasonal effect on growth.

Preliminary spatiotemporal analysis of the association between socio-environmental factors and suicide

Background: The seasonality of suicide has long been recognised. However, little is known about the relative importance of socio-environmental factors in the occurrence of suicide in different geographical areas. This study examined the association of climate, socioeconomic and demographic factors with suicide in Queensland, Australia, using a spatiotemporal approach. Methods: Seasonal data on suicide, demographic variables and socioeconomic indexes for areas in each Local Government Area (LGA) between 1999 and 2003 were acquired from the Australian Bureau of Statistics. Climate data were supplied by the Australian Bureau of Meteorology. A multivariable generalized estimating equation model was used to examine the impact of socio-environmental factors on suicide. Results: The preliminary data analyses show that far north Queensland had the highest suicide incidence (e.g., Cook and Mornington Shires), while the south-western areas had the lowest incidence (e.g., Barcoo and Bauhinia Shires) in all the seasons. Maximum temperature, unemployment rate, the proportion of Indigenous population and the proportion of population with low individual income were statistically significantly and positively associated with suicide. There were weaker but not significant associations for other variables. Conclusions: Maximum temperature, the proportion of Indigenous population and unemployment rate appeared to be major determinants of suicide at a LGA level in Queensland.

Temperature variation and emergency hospital admissions for stroke in Brisbane, Australia, 1996-2005

Stroke is a leading cause of disability and death. This study evaluated the association between temperature variation and emergency admissions for stroke in Brisbane, Australia. Daily emergency admissions for stroke, meteorologic and air pollution data were obtained for the period of January 1996 to December 2005. The relative risk of emergency admissions for stroke was estimated with a generalized estimating equations (GEE) model. For primary intracerebral hemorrhage (PIH) emergency admissions, the average daily PIH for the group aged < 65 increased by 15% (95% Confidence Interval (CI): 5, 26%) and 12% (95% CI: 2, 22%) for a 1°C increase in daily maximum temperature and minimum temperature in summer, respectively, after controlling for potential confounding effects of humidity and air pollutants. For ischemic stroke (IS) emergency admissions, the average daily IS for the group aged ≥ 65 decreased by 3% (95% CI: -6, 0%) for a 1°C increase in daily maximum temperature in winter after adjustment for confounding factors. Temperature variation was significantly associated with emergency admissions for stroke, and its impact varied with different type of stroke. Health authorities should pay greater attention to possible increasing emergency care for strokes when temperature changes, in both summer and winter.

Socio-environmental factors and suicide in Queensland, Australia

Suicide has drawn much attention from both the scientific community and the public. Examining the impact of socio-environmental factors on suicide is essential in developing suicide prevention strategies and interventions, because it will provide health authorities with important information for their decision-making. However, previous studies did not examine the impact of socio-environmental factors on suicide using a spatial analysis approach. The purpose of this study was to identify the patterns of suicide and to examine how socio-environmental factors impact on suicide over time and space at the Local Governmental Area (LGA) level in Queensland. The suicide data between 1999 and 2003 were collected from the Australian Bureau of Statistics (ABS). Socio-environmental variables at the LGA level included climate (rainfall, maximum and minimum temperature), Socioeconomic Indexes for Areas (SEIFA) and demographic variables (proportion of Indigenous population, unemployment rate, proportion of population with low income and low education level). Climate data were obtained from Australian Bureau of Meteorology. SEIFA and demographic variables were acquired from ABS. A series of statistical and geographical information system (GIS) approaches were applied in the analysis. This study included two stages. The first stage used average annual data to view the spatial pattern of suicide and to examine the association between socio-environmental factors and suicide over space. The second stage examined the spatiotemporal pattern of suicide and assessed the socio-environmental determinants of suicide, using more detailed seasonal data. In this research, 2,445 suicide cases were included, with 1,957 males (80.0%) and 488 females (20.0%). In the first stage, we examined the spatial pattern and the determinants of suicide using 5-year aggregated data. Spearman correlations were used to assess associations between variables. Then a Poisson regression model was applied in the multivariable analysis, as the occurrence of suicide is a small probability event and this model fitted the data quite well. Suicide mortality varied across LGAs and was associated with a range of socio-environmental factors. The multivariable analysis showed that maximum temperature was significantly and positively associated with male suicide (relative risk [RR] = 1.03, 95% CI: 1.00 to 1.07). Higher proportion of Indigenous population was accompanied with more suicide in male population (male: RR = 1.02, 95% CI: 1.01 to 1.03). There was a positive association between unemployment rate and suicide in both genders (male: RR = 1.04, 95% CI: 1.02 to 1.06; female: RR = 1.07, 95% CI: 1.00 to 1.16). No significant association was observed for rainfall, minimum temperature, SEIFA, proportion of population with low individual income and low educational attainment. In the second stage of this study, we undertook a preliminary spatiotemporal analysis of suicide using seasonal data. Firstly, we assessed the interrelations between variables. Secondly, a generalised estimating equations (GEE) model was used to examine the socio-environmental impact on suicide over time and space, as this model is well suited to analyze repeated longitudinal data (e.g., seasonal suicide mortality in a certain LGA) and it fitted the data better than other models (e.g., Poisson model). The suicide pattern varied with season and LGA. The north of Queensland had the highest suicide mortality rate in all the seasons, while there was no suicide case occurred in the southwest. Northwest had consistently higher suicide mortality in spring, autumn and winter. In other areas, suicide mortality varied between seasons. This analysis showed that maximum temperature was positively associated with suicide among male population (RR = 1.24, 95% CI: 1.04 to 1.47) and total population (RR = 1.15, 95% CI: 1.00 to 1.32). Higher proportion of Indigenous population was accompanied with more suicide among total population (RR = 1.16, 95% CI: 1.13 to 1.19) and by gender (male: RR = 1.07, 95% CI: 1.01 to 1.13; female: RR = 1.23, 95% CI: 1.03 to 1.48). Unemployment rate was positively associated with total (RR = 1.40, 95% CI: 1.24 to 1.59) and female (RR=1.09, 95% CI: 1.01 to 1.18) suicide. There was also a positive association between proportion of population with low individual income and suicide in total (RR = 1.28, 95% CI: 1.10 to 1.48) and male (RR = 1.45, 95% CI: 1.23 to 1.72) population. Rainfall was only positively associated with suicide in total population (RR = 1.11, 95% CI: 1.04 to 1.19). There was no significant association for rainfall, minimum temperature, SEIFA, proportion of population with low educational attainment. The second stage is the extension of the first stage. Different spatial scales of dataset were used between the two stages (i.e., mean yearly data in the first stage, and seasonal data in the second stage), but the results are generally consistent with each other. Compared with other studies, this research explored the variety of the impact of a wide range of socio-environmental factors on suicide in different geographical units. Maximum temperature, proportion of Indigenous population, unemployment rate and proportion of population with low individual income were among the major determinants of suicide in Queensland. However, the influence from other factors (e.g. socio-culture background, alcohol and drug use) influencing suicide cannot be ignored. An in-depth understanding of these factors is vital in planning and implementing suicide prevention strategies. Five recommendations for future research are derived from this study: (1) It is vital to acquire detailed personal information on each suicide case and relevant information among the population in assessing the key socio-environmental determinants of suicide; (2) Bayesian model could be applied to compare mortality rates and their socio-environmental determinants across LGAs in future research; (3) In the LGAs with warm weather, high proportion of Indigenous population and/or unemployment rate, concerted efforts need to be made to control and prevent suicide and other mental health problems; (4) The current surveillance, forecasting and early warning system needs to be strengthened, to trace the climate and socioeconomic change over time and space and its impact on population health; (5) It is necessary to evaluate and improve the facilities of mental health care, psychological consultation, suicide prevention and control programs; especially in the areas with low socio-economic status, high unemployment rate, extreme weather events and natural disasters.