Search

The conventional normality assumption for the random effects distribution in logistic mixed models can be too restrictive in some applications. In our data example of a longitudinal study modelling employment participation of Australian women, the random effects exhibit non-normality due to a potential mover–stayer scenario. In such a scenario, the women observed to remain in the same initial response state over the study period may consist of two subgroups: latent stayers—those with extremely small probability of transitioning response states—and latent movers, those with a probability of transitioning response states. The similarities between estimating the random effects using non-parametric approaches and mover–stayer models have previously been highlighted. We explore non-parametric approaches to model univariate and bivariate random effects in a potential mover–stayer scenario. As there are limited approaches available to fit the non-parametric maximum likelihood estimate for bivariate random effects in logistic mixed models, we implement the Vertex Exchange Method (VEM) to estimate the random effects in logistic mixed models. The approximation of the non-parametric maximum likelihood estimate derived by the VEM algorithm induces more flexibility of the random effects, identifying regions corresponding to potential latent stayers in the non-employment category in our data example. [ABSTRACT FROM AUTHOR]

Abstract Background Linear regression models are used to quantitatively predict drug resistance, the phenotype, from the HIV-1 viral genotype. As new antiretroviral drugs become available, new resistance pathways emerge and the number of resistance associated mutations continues to increase. To accurately identify which drug options are left, the main goal of the modeling has been to maximize predictivity and not interpretability. However, we originally selected linear regression as the preferred method for its transparency as opposed to other techniques such as neural networks. Here, we apply a method to lower the complexity of these phenotype prediction models using a 3-fold cross-validated selection of mutations. Results Compared to standard stepwise regression we were able to reduce the number of mutations in the reverse transcriptase (RT) inhibitor models as well as the number of interaction terms accounting for synergistic and antagonistic effects. This reduction in complexity was most significant for the non-nucleoside reverse transcriptase inhibitor (NNRTI) models, while maintaining prediction accuracy and retaining virtually all known resistance associated mutations as first order terms in the models. Furthermore, for etravirine (ETR) a better performance was seen on two years of unseen data. By analyzing the phenotype prediction models we identified a list of forty novel NNRTI mutations, putatively associated with resistance. The resistance association of novel variants at known NNRTI resistance positions: 100, 101, 181, 190, 221 and of mutations at positions not previously linked with NNRTI resistance: 102, 139, 219, 241, 376 and 382 was confirmed by phenotyping site-directed mutants. Conclusions We successfully identified and validated novel NNRTI resistance associated mutations by developing parsimonious resistance prediction models in which repeated cross-validation within the stepwise regression was applied. Our model selection technique is computationally feasible for large data sets and provides an approach to the continued identification of resistance-causing mutations.

Abstract Background During a four-year action research project (2003-2007), a program targeting all type 2 diabetes patients was implemented in a well-defined geographical region in Belgium. The implementation of the program resulted in an increase of the overall Assessment of Chronic Illness Care (ACIC) score from 1.45 in 2003 to 5.5 in 2007. The aim of the follow-up study in 2008 was to assess the effect of the implementation of Chronic Care Model (CCM) elements on the quality of diabetes care in a country where the efforts to adapt primary care to a more chronic care oriented system are still at a starting point. Methods A quasi-experimental study design involving a control region with comparable geographical and socio-economic characteristics and health care facilities was used to evaluate the effect of the intervention in the region. In collaboration with the InterMutualistic Agency (IMA) and the laboratories from both regions a research database was set up. Study cohorts in both regions were defined by using administrative data from the Sickness Funds and selected from the research database. A set of nine quality indicators was defined based on current scientific evidence. Data were analysed by an institution experienced in longitudinal data analysis. Results In total 4,174 type 2 diabetes patients were selected from the research database; 2,425 patients (52.9% women) with a mean age of 67.5 from the intervention region and 1,749 patients (55.7% women) with a mean age of 67.4 from the control region. At the end of the intervention period, improvements were observed in five of the nine defined quality indicators in the intervention region, three of which (HbA1c assessment, statin therapy, cholesterol target) improved significantly more than in the control region. Mean HbA1c improved significantly in the intervention region (7.55 to 7.06%), but this evolution did not differ significantly (p = 0.4207) from the one in the control region (7.44 to 6.90%). The improvement in lipid control was significantly higher (p = 0.0021) in the intervention region (total cholesterol 199.07 to 173 mg/dl) than in the control region (199.44 to 180.60 mg/dl). The systematic assessment of long-term diabetes complications remained insufficient. In 2006 only 26% of the patients had their urine tested for micro-albuminuria and only 36% had consulted an ophthalmologist. Conclusion Although the overall ACIC score increased from 1.45 to 5.5, the improvement in the quality of diabetes care was moderate. Further improvements are needed in the CCM components delivery system design and clinical information systems. The regional networks, as they are financed now by the National Institute for Health and Disability Insurance (NIHDI), are an opportunity to explore how this can be achieved in consultation with the GPs. But it is clear that, simultaneously, action is needed on the health system level to realize the installation of an accurate quality monitoring system and the necessary preconditions for chronic care delivery in primary care (patient registration, staff support, IT support). Trial Registration Trial registration number: ClinicalTrials.gov Identifier: NCT00824499

Abstract Background Type 2 diabetes mellitus is a complex, progressive disease which requires a variety of quality improvement strategies. Limited information is available on the feasibility and effectiveness of interdisciplinary diabetes care teams (IDCT) operating on the interface between primary and specialty care. A first study hypothesis was that the implementation of an IDCT is feasible in a health care setting with limited tradition in shared care. A second hypothesis was that patients who make use of an IDCT would have significantly better outcomes compared to non-users of the IDCT after an 18-month intervention period. A third hypothesis was that patients who used the IDCT in an Advanced quality Improvement Program (AQIP) would have significantly better outcomes compared to users of a Usual Quality Improvement Program (UQIP). Methods This investigation comprised a two-arm cluster randomized trial conducted in a primary care setting in Belgium. Primary care physicians (PCPs, n = 120) and their patients with type 2 diabetes mellitus (n = 2495) were included and subjects were randomly assigned to the intervention arms. The IDCT acted as a cornerstone to both the intervention arms, but the number, type and intensity of IDCT related interventions varied depending upon the intervention arm. Results Final registration included 67 PCPs and 1577 patients in the AQIP and 53 PCPs and 918 patients in the UQIP. 84% of the PCPs made use of the IDCT. The expected participation rate in patients (30%) was not attained, with 12,5% of the patients using the IDCT. When comparing users and non-users of the IDCT (irrespective of the intervention arm) and after 18 months of intervention the use of the IDCT was significantly associated with improvements in HbA1c, LDL-cholesterol, an increase in statins and anti-platelet therapy as well as the number of targets that were reached. When comparing users of the IDCT in the two intervention arms no significant differences were noted, except for anti-platelet therapy. Conclusion IDCT's operating on the interface between primary and specialty care are associated with improved outcomes of care. More research is required on what team and program characteristics contribute to improvements in diabetes care. Trial registration NTR 1369.

Abstract Background In the working population, back disorders are an important reason for sick leave and permanent work inability. In the context of fitting the job to the worker, one of the primary tasks of the occupational health physician is to evaluate the balance between work-related and individual variables. Since this evaluation of work capacity often consists of a physical examination of the back, the objective of this study was to investigate whether a physical examination of the low back, which is routinely performed in occupational medicine, predicts the development of low back pain (LBP). Methods This study is part of the Belgian Low Back Cohort (BelCoBack) Study, a prospective study to identify risk factors for the development of low back disorders in occupational settings. The study population for this paper were 692 young healthcare or distribution workers (mean age of 26 years) with no or limited back antecedents in the year before inclusion. At baseline, these workers underwent a standardised physical examination of the low back. One year later, they completed a questionnaire on the occurrence of LBP and some of its characteristics. To study the respective role of predictors at baseline on the occurrence of LBP, we opted for Cox regression with a constant risk period. Analyses were performed separately for workers without any back antecedents in the year before inclusion ('asymptomatic' workers) and for workers with limited back antecedents in the year before inclusion ('mildly symptomatic' workers). Results In the group of 'asymptomatic' workers, obese workers showed a more than twofold-increased risk on the development of LBP as compared to non-obese colleagues (RR 2.57, 95%CI: 1.09 – 6.09). In the group of 'mildly symptomatic' workers, the self-reports of pain before the examination turned out to be most predictive (RR 3.89, 95%CI: 1.20 – 12.64). Conclusion This study showed that, in a population of young workers wh no or limited antecedents of LBP at baseline, physical examinations, as routinely assessed in occupational medicine, are not useful to predict workers at risk for the development of back disorders one year later.

In regression analysis, the data sample is often composed of diverse sub-populations such as ethnicities and geographical regions. In multiple application areas, it is important to identify the groups where each covariate has a positive, negative or null impact on the response. If the number of sub-populations is small, a full interaction model may be fit with group-specific covariate effects. However, if the number of groups is very large, for example, hospitals or other clustering units, such a model is not identifiable. Therefore, we propose a prior distribution which combines the information across sub-populations with a similar covariate effect. This Bayesian analysis of differential effects (BADE) classifies the group-specific covariate effects as positive, negative or null. Besides allowing the analysis of differential effects for many sub-populations, the proposed approach improves substantially the identification of important interactions in cases with few groups. This is illustrated via simulations. The procedure is motivated on, and applied to, a large study related to patients' satisfaction with hospitals, where we show that classifying group-specific covariate effects based on methods such as mixed-effects models may be strongly misleading. [ABSTRACT FROM AUTHOR]

We propose a Bayesian latent Ornstein‐Uhlenbeck (OU) model to analyze unbalanced longitudinal data of binary and ordinal variables, which are manifestations of fewer continuous latent variables. We focus on the evolution of such latent variables when they continuously change over time. Existing approaches are limited to data collected at regular time intervals. Our proposal makes use of an OU process for the latent variables to overcome this limitation. We show that assuming real eigenvalues for the drift matrix of the OU process, as is frequently done in practice, can lead to biased estimates and/or misleading inference when the true process is oscillating. In contrast, our proposal allows for both real and complex eigenvalues. We illustrate our proposed model with a motivating dataset, containing patients with amyotrophic lateral sclerosis disease. We were interested in how bulbar, cervical, and lumbar functions evolve over time. [ABSTRACT FROM AUTHOR]

Abstract Background Most quality improvement programs in diabetes care incorporate aspects of clinician education, performance feedback, patient education, care management, and diabetes care teams to support primary care physicians. Few studies have applied all of these dimensions to address clinical inertia. Aim To evaluate interventions to improve adherence to evidence-based guidelines for diabetes and reduce clinical inertia in primary care physicians. Design Two-arm cluster randomized controlled trial. Participants Primary care physicians in Belgium. Interventions Primary care physicians will be randomly allocated to 'Usual' (UQIP) or 'Advanced' (AQIP) Quality Improvement Programs. Physicians in the UQIP will receive interventions addressing the main physician, patient, and office system factors that contribute to clinical inertia. Physicians in the AQIP will receive additional interventions that focus on sustainable behavior changes in patients and providers. Outcomes Primary endpoints are the proportions of patients within targets for three clinical outcomes: 1) glycosylated hemoglobin < 7%; 2) systolic blood pressure differences ≤130 mmHg; and 3) low density lipoprotein/cholesterol < 100 mg/dl. Secondary endpoints are individual improvements in 12 validated parameters: glycosylated hemoglobin, low and high density lipoprotein/cholesterol, total cholesterol, systolic blood pressure, diastolic blood pressure, weight, physical exercise, healthy diet, smoking status, and statin and anti-platelet therapy. Primary and secondary analysis Statistical analyses will be performed using an intent-to-treat approach with a multilevel model. Linear and generalized linear mixed models will be used to account for the clustered nature of the data, i.e., patients clustered withinimary care physicians, and repeated assessments clustered within patients. To compare patient characteristics at baseline and between the intervention arms, the generalized estimating equations (GEE) approach will be used, taking the clustered nature of the data within physicians into account. We will also use the GEE approach to test for differences in evolution of the primary and secondary endpoints for all patients, and for patients in the two interventions arms, accounting for within-patient clustering. Trial Registration number: NTR 1369.

We propose a latent linear mixed model to analyze multivariate longitudinal data of multiple ordinal variables, which are manifestations of fewer continuous latent variables. We focus on the latent level where the effects of observed covariates on the latent variables are of interest. We incorporate serial correlation into the variance component rather than assuming independent residuals. We show that misleading inference may be drawn when misspecifying the variance component. Furthermore, we provide a graphical tool depicting latent empirical semi-variograms to detect serial correlation for latent stationary linear mixed models. We apply our proposed model to examine the treatment effect on patients having the amyotrophic lateral sclerosis disease. The result shows that the treatment can slow down progression of latent cervical and lumbar functions. [ABSTRACT FROM AUTHOR]

We propose a Bayesian latent vector autoregressive (LVAR) model to analyze multivariate longitudinal data of binary and ordinal variables (items) as a function of a small number of continuous latent variables. We focus on the evolution of the latent variables while taking into account the correlation structure of the responses. Often local independence is assumed in this context. Local independence implies that, given the latent variables, the responses are assumed mutually independent cross-sectionally and longitudinally. However, in practice conditioning on the latent variables may not remove the dependence of the responses. We address local dependence by further conditioning on item-specific random effects. A simulation study shows that wrongly assuming local independence may give biased estimates for the regression coefficients of the LVAR process as well as the item-specific parameters. Novel features of our proposal include (i) correcting biased estimates of the model parameters, especially the regression coefficients of the LVAR process, obtained when local dependence is ignored and (ii) measuring the magnitude of local dependence. We applied our model on data obtained from a registry on the elderly population in Belgium. The purpose was to examine the values of oral health information on top of general health information. [ABSTRACT FROM AUTHOR]

For count responses, there are situations in biomedical and sociological applications in which extra zeroes occur. Modeling correlated (e.g. repeated measures and clustered) zero-inflated count data includes special challenges because the correlation between measurements for a subject or a cluster needs to be taken into account. Moreover, zero-inflated count data are often faced with over/under dispersion problem. In this paper, we propose a random effect model for repeated measurements or clustered data with over/under dispersed response called random effect zero-inflated exponentiated-exponential geometric regression model. The proposed method was illustrated through real examples. The performance of the model and asymptotical properties of the estimations were investigated using simulation studies. [ABSTRACT FROM AUTHOR]

We consider multiple imputation as a procedure iterating over a set of imputed datasets. Based on an appropriate stopping rule the number of imputed datasets is determined. Simulations and real-data analyses indicate that the sufficient number of imputed datasets may in some cases be substantially larger than the very small numbers that are usually recommended. For an easier use in various applications, the proposed method is implemented in the R package imi. [ABSTRACT FROM AUTHOR]

Linear mixed models (LMMs) are popular to analyze repeated measurements with a Gaussian response. For longitudinal studies, the LMMs consist of a fixed part expressing the effect of covariates on the mean evolution in time and a random part expressing the variation of the individual curves around the mean curve. Selecting the appropriate fixed and random effect parts is an important modeling exercise. In a Bayesian framework, there is little agreement on the appropriate selection criteria. This paper compares the performance of the deviance information criterion (DIC), the pseudo-Bayes factor and the widely applicable information criterion (WAIC) in LMMs, with an extension to LMMs with skew-normal distributions. We focus on the comparison between the conditional criteria (given random effects) versus the marginal criteria (averaged over random effects). In spite of theoretical arguments, there is not much enthusiasm among applied statisticians to make use of the marginal criteria. We show in an extensive simulation study that the three marginal criteria are superior in choosing the appropriate longitudinal model. In addition, the marginal criteria selected most appropriate model for growth curves of Nigerian chicken. A self-written R function can be combined with standard Bayesian software packages to obtain the marginal selection criteria. [ABSTRACT FROM AUTHOR]

Missing data is almost inevitable in correlated-data studies. For non-Gaussian outcomes with moderate to large sequences, direct-likelihood methods can involve complex, hard-to-manipulate likelihoods. Popular alternative approaches, like generalized estimating equations, that are frequently used to circumvent the computational complexity of full likelihood, are less suitable when scientific interest, at least in part, is placed on the association structure; pseudo-likelihood (PL) methods are then a viable alternative. When the missing data are missing at random, Molenberghs et al. (2011, Statistica Sinica, 21,187–206) proposed a suite of corrections to the standard form of PL, taking the form of singly and doubly robust estimators. They provided the basis and exemplified it in insightful yet primarily illustrative examples. We here consider the important case of marginal models for hierarchical binary data, provide an effective implementation and illustrate it using data from an analgesic trial. Our doubly robust estimator is more convenient than the classical doubly robust estimators. The ideas are illustrated using a marginal model for a binary response, more specifically a Bahadur model. [ABSTRACT FROM AUTHOR]