Poisson Mixed Models Research Articles

Impact of IgG Monitoring and IVIG Supplementation on the Frequency of Febrile Illnesses in Pediatric Acute Lymphoblastic Leukemia Patients Undergoing Maintenance Chemotherapy.

Monitoring serum immunoglobulin G (IgG) levels in pediatric oncology patients and treating subtherapeutic levels with intravenous immunoglobulin (IVIG) may prevent infections; however, evidence is limited. This retrospective study assessed pediatric acute lymphoblastic leukemia patients diagnosed 2006 to 2011 to evaluate if monitoring/supplementing IgG would reduce febrile illnesses during maintenance chemotherapy. A subject was categorized as "ever IgG monitored" if they had ≥1 IgG levels checked and their risk days were stratified into not IgG monitored days and IgG monitored days. IgG monitored days were further stratified into IgG monitored with IVIG supplementation, monitored with no IVIG supplementation (IgG level >500 mg/dL) and monitored with no IVIG supplementation days (IgG level <500 mg/dL). Generalized linear mixed effects poisson models were used to compare events (febrile episode, positive blood culture, and febrile upper respiratory infection rates among these groups. In 136 patients, the febrile episode rate was higher in the ever IgG monitored cohort than the never monitored cohort (5.26 vs. 3.78 episodes/1000 d). Among monitored patients, IVIG monitoring and supplementation did not significantly impact the febrile episode, febrile upper respiratory infection, or the positive blood culture rates. These data suggest that monitoring/supplementing low IgG is not indicated for infection prophylaxis in acute lymphoblastic leukemia patients during maintenance chemotherapy.

Short-Term Disability Fluctuations in Late Life.

Late-life disability is highly dynamic but within-person short-term fluctuations have not been assessed previously. We analyze how substantial such late-life disability fluctuations are and whether they are associated with time-to-death, long-term disability trajectories, frailty, and sociodemographics. Monthly survey data (Precipitating Events Project Study) on activities of daily living/instrumental activities of daily living (ADL/IADL) disability (0-9) in the last years of life from 642 deceased respondents providing 56,308 observations were analyzed with a two-step approach. Observation-level residuals extracted from a Poisson mixed regression model (first step), which depict vertical short-term fluctuations from individual long-term trajectories, were analyzed with a linear mixed regression model (second step). Short-term disability fluctuations amounted to about one ADL/IADL limitation, increased in the last 4 years of life, and were closely associated with disability increases. Associations with frailty or sociodemographics characteristics were absent except for living alone. Short-term disability fluctuations in late life were substantial, were linked to mortality-related processes, and represent a concomitant feature of disability increases in late life.

Dependence modeling for multi-type recurrent events via copulas.

When several types of recurrent events may arise, interest often lies in marginal modeling and studying the nature of the dependence structure. In this paper, we propose a multivariate mixed-Poisson model with the dependence between events accommodated by type-specific random effects which are associated through use of a Gaussian copula. Such models retain marginal features with a simple interpretation, reflect the heterogeneity in risk for each type of event, and provide insight into the dependence between the different types of events. Semiparametric inference is proposed based on composite likelihood to avoid high dimensional integration. An application to a study of nutritional supplements in malnourished children is given in which the goal is to evaluate the reduction in the rate of several different kinds of infection.

Fate of Mycobacterium avium subsp. paratuberculosis and changes in bacterial diversity populations in dairy slurry after chemical treatments.

A major drawback of using dairy slurry as fertilizer is that it may contains pathogens such as Mycobacterium avium subsp. paratuberculosis (MAP), and it could represent a risk to animal and public health. Thus, the aim of this study was to evaluate the fate of MAP and bacterial communities in dairy slurry after chemical treatments. Cattle slurry, naturally contaminated with MAP, was collected from a dairy herd and divided into 32 glass bottles which were assigned to eight different treatments (control, 3·0% CaO, 0·5% NaOH; 0·087%, 0·11% and 0·14% H2 SO4 ; and 1·0 and 2·5% KMnO4 ). Treated dairy slurry samples were evaluated at 0, 1, 3, 7, 15, 30 and 60-days following treatment application for viable MAP and dairy slurry pH, and in addition temperature in this material was monitored continuously. Bacterial counts were estimated at each sampling time. A Bayesian zero-inflated Poisson mixed model was fitted to assess the effect of each treatment on the count of MAP cells. Model results indicated that only the 3·0% CaO treatment had a statistically important negative effect on MAP counts during the study period. For most treatments, MAP was undetectable immediately after chemical treatment but re-appeared over time, in some replicates at low concentrations. However, in those cases MAP counts were not statistically different than the control treatment. Regarding the fate of the other bacterial populations, the Firmicutes phylum was the dominant population in the un-treated slurry while Clostridia class members were among the most prevalent bacteria after the application of most chemical treatments. Only 3% CaO treatment had a statistically important negative effect on MAP viability in cattle slurry. This study provides evidence of MAP partial control in dairy slurry. This information should be considered as a best management practice to reduce MAP and other pathogens for slurry management on dairy farms.

Five-Year Outcomes of Gastric Bypass in Adolescents as Compared with Adults

BackgroundBariatric surgery results in weight loss and health improvements in adults and adolescents. However, whether outcomes differ according to the age of the patient at the time of surgery is unclear.MethodsWe evaluated the health effects of Roux-en-Y gastric bypass in a cohort of adolescents (161 patients enrolled from 2006 through 2012) and a cohort of adults (396 patients enrolled from 2006 through 2009). The two cohorts were participants in two related but independent studies. Linear mixed and Poisson mixed models were used to compare outcomes with regard to weight and coexisting conditions between the cohorts 5 years after surgery. The rates of death and subsequent abdominal operations and selected micronutrient levels (up to 2 years after surgery) were also compared between the cohorts.ResultsThere was no significant difference in percent weight change between adolescents (−26%; 95% confidence interval [CI], −29 to −23) and adults (−29%; 95% CI, −31 to −27) 5 years after surgery (P=0.08). After surgery, adolescents were significantly more likely than adults to have remission of type 2 diabetes (86% vs. 53%; risk ratio, 1.27; 95% CI, 1.03 to 1.57) and of hypertension (68% vs. 41%; risk ratio, 1.51; 95% CI, 1.21 to 1.88). Three adolescents (1.9%) and seven adults (1.8%) died in the 5 years after surgery. The rate of abdominal reoperations was significantly higher among adolescents than among adults (19 vs. 10 reoperations per 500 person-years, P=0.003). More adolescents than adults had low ferritin levels (72 of 132 patients [48%] vs. 54 of 179 patients [29%], P=0.004).ConclusionsAdolescents and adults who underwent gastric bypass had marked weight loss that was similar in magnitude 5 years after surgery. Adolescents had remission of diabetes and hypertension more often than adults. (Funded by the National Institute of Diabetes and Digestive and Kidney Diseases; ClinicalTrials.gov number, NCT00474318.)

Alternative count Bayesian models for genetic evaluation of litter traits in pigs

We compared Hurdle Poisson (Hupoisson), Zero Inflated Poisson (Zipoisson), Poisson and Gaussian Bayesian mixed models for genetic evaluation of litter traits in pigs (number of piglets born alive – NBA; number of piglets weaned – NW; and mortality during the suckling period – MORT). The models’ goodness-of-fit was based on Deviance Information Criteria (DIC), Model Posterior Probabilities (MPP) and Mean Square Error (MSE). The Hupoisson model provided the best fit to MORT, even showing higher heritability estimate (0.13, with credibility interval from 0.0745 to 0.2809). The Hupoisson model proposed in the present study was suitable for genetic evaluation of mortality during the suckling period, and can be indicated as a novel approach to treat count data with excess of zeros under a genetic and breeding viewpoint. The Gaussian model outperformed the other models for NBA and NW.

A Diallel of the Mouse Collaborative Cross Founders Reveals Strong Strain-Specific Maternal Effects on Litter Size.

Reproductive success in the eight founder strains of the Collaborative Cross (CC) was measured using a diallel-mating scheme. Over a 48-month period we generated 4,448 litters, and provided 24,782 weaned pups for use in 16 different published experiments. We identified factors that affect the average litter size in a cross by estimating the overall contribution of parent-of-origin, heterosis, inbred, and epistatic effects using a Bayesian zero-truncated overdispersed Poisson mixed model. The phenotypic variance of litter size has a substantial contribution (82%) from unexplained and environmental sources, but no detectable effect of seasonality. Most of the explained variance was due to additive effects (9.2%) and parental sex (maternal vs. paternal strain; 5.8%), with epistasis accounting for 3.4%. Within the parental effects, the effect of the dam’s strain explained more than the sire’s strain (13.2% vs. 1.8%), and the dam’s strain effects account for 74.2% of total variation explained. Dams from strains C57BL/6J and NOD/ShiLtJ increased the expected litter size by a mean of 1.66 and 1.79 pups, whereas dams from strains WSB/EiJ, PWK/PhJ, and CAST/EiJ reduced expected litter size by a mean of 1.51, 0.81, and 0.90 pups. Finally, there was no strong evidence for strain-specific effects on sex ratio distortion. Overall, these results demonstrate that strains vary substantially in their reproductive ability depending on their genetic background, and that litter size is largely determined by dam’s strain rather than sire’s strain effects, as expected. This analysis adds to our understanding of factors that influence litter size in mammals, and also helps to explain breeding successes and failures in the extinct lines and surviving CC strains.

The association between obesity and outcomes in acute pancreatitis: an individual patient data meta-analysis.

There are data to suggest that obesity is associated with local and systemic complications as well as mortality in acute pancreatitis (AP). Cohort studies to date, however, have shown conflicting results from mostly unadjusted analyses. Therefore, we performed an individual patient data meta-analysis with the primary aim to investigate the association between obesity and mortality in AP. Our secondary aim was to investigate the association between obesity and necrosis, organ failure, multiple organ failure, and invasive intervention. We systematically searched four electronic databases for prospective studies on obesity and outcomes in AP. Researchers of eligible studies were invited to share individual patient data using a standardized data collection form. All end points were investigated with a one-stage mixed effects Poisson model with random intercepts and forced entry of relevant confounders. We included five databases with 1302 patients, of whom 418 (32%) were obese. In total, 466 (36%) patients had necrosis, 328 (25%) had organ failure, 188 (14%) had multiple organ failure, 210 (16%) had an intervention, and 84 (7%) patients died. We found no significant association between obesity and mortality [relative risk (RR) 1.40, 95% confidence interval (CI): 0.89-2.20], necrosis (RR: 1.08, 95% CI: 0.90-1.31) or invasive intervention (RR: 1.10, 95% CI: 0.83-1.47) after adjustment for confounders. However, obesity was independently associated with the development of organ failure (RR: 1.38, 95% CI: 1.11-1.73) and multiple organ failure (RR: 1.81, 95% CI: 1.35-2.42). Obesity is independently associated with the development of organ failure and multiple organ failure in AP. However, there is no association between obesity and mortality, necrosis, and an intervention.

Modelling of the Abundance of Malaria Mosquitoes Using Poisson Mixed Model

Anopheles funestus and Anopheles gambiae are malaria vector mosquitoes. Knowing their resting behavior is important for implementing control methods. The aim of this study was to investigate the resting behaviour of the two malaria mosquitoes. The study was conducted in Kilombero River Valley and mosquitoes were collected using indoor and outdoor traps from 2012-2015. Poisson mixed models were used to quantify the impact of environment variables on resting behaviour. A log ratio rate between the type of trap and its interaction with environmental variables was used to determine if there was a change over time in the resting behaviour. A total of 4696 mosquitoes were resting indoors of which 57% were A. funestus and 43% were A. gambiae. Similarly, a total of 12,028 mosquitoes were resting outdoor of which 13% were A. funestus and 87% were A. gambiae. Temperature was significant and affected the resting behaviour of A. funestus. Humidity, saturation deficit and temperature were significant variables influencing the resting behaviour of A. gambiae. A. funestus was resting indoor while A. gambiae was resting outdoor over time generally. The findings of this study on the effects of environmental variables and the variations in the resting behaviour of A. gambiae and A. funestus could be used as a guide to implementing appropriate intervention measures such as indoor residential spraying (IRS), insecticide treated nets (ITNs) and mosquito repellents.

Poisson mixed models for predicting number of fires

Wildfires are considered one of the main causes of forest destruction. In recent years, the number of forest fires and burned area in Mediterranean regions have increased. This problem particularly affects Galicia (north-west of Spain). Conventional modelling of the number of forest fires in small areas may have a high error. For this reason, four area-level Poisson mixed models with time effects are proposed. The first two models contain independent time effects, whereas the random effects of the other models are distributed according to an autoregressive process AR(1). A parametric bootstrap algorithm is given to measure the accuracy of the plug-in predictor of fire number under the temporal models. A significant prediction improvement is observed when using Poisson regression models with random time effects. Analysis of historical data finds significant meteorological and socioeconomic variables explaining the number of forest fires by area and reveals the presence of a temporal correlation structure captured by the area-level Poisson mixed model with AR(1) time effects.

The Effect of Hormonal Contraception and Menstrual Cycle Timing on Genital Herpes Simplex Virus-2 Shedding and Lesions.

The effect of female sex hormones on herpes simplex virus (HSV)-2 shedding and lesion frequency is poorly understood. Previous studies suggest that hormonal contraception may increase the frequency of HSV-2 shedding. We studied HSV-2 seropositive women who performed daily genital swabbing for HSV DNA and completed diaries for genital lesions and menses. We used Poisson mixed effects models to determine if HSV detection varied throughout the menstrual cycle, or in response to hormonal contraception. We used the Wilcoxon signed-rank test and rank-sum test to determine if lesion frequency differed by cycle phase or hormonal contraceptive use. In 189 women aged 19 to 46 years who collected swabs on 10,715 days and were not using hormonal contraception, HSV-2 DNA was detected on 20.9% of days in the follicular phase and 17.8% of days in the luteal phase (rate ratio, 1.19; 95% confidence interval, 1.03-1.37, P = 0.02). Genital lesions did not differ in the follicular versus luteal phase (12.8% vs. 10.7%, P = 0.07). In analyses of hormonal contraception, including 244 women, HSV-2 DNA was detected on 19.0% of days for women not using hormonal contraception and 18.3% of days for those using hormonal contraception (P = 0.50). Lesions were present on 11.1% of days for women not using hormonal contraception, and 8.7% of days for those using hormonal contraception (P = 0.66). In women with genital HSV-2 infection who are not using hormonal contraception, the follicular phase of the cycle may be associated with a higher frequency of HSV-2 shedding compared to the luteal phase. Lesion frequency is similar during the 2 menstrual phases. Hormonal contraception use was not observed to affect genital HSV-2 DNA detection or lesions.

Reduction in Comorbid Conditions Over 5 Years Following Bariatric Surgery in Medicaid and Commercially Insured Patients

This study sought to determine changes in the prevalence of comorbid disease following bariatric surgery in Medicaid patients compared with commercially insured patients. Data were obtained from the Longitudinal Assessment of Bariatric Surgery, an observational cohort study of adults undergoing bariatric surgery at one of six geographically diverse centers in the United States. A total of 1,201 patients who underwent Roux-en-Y gastric bypass with 5 years of follow-up were identified. Poisson mixed models were used to estimate relative risks (RRs) and compare changes in common comorbidities between insurance groups within 0-1 and 1-5 years post surgery. Propensity scores were used to achieve balance in the baseline comorbidity burden between Medicaid and commercial patients. In the first year, risk of all six comorbidities decreased substantially over time in both groups, ranging from a 32% to a 69% decrease from baseline. After 1 year post surgery, the risk of disease was stable in both groups (RRs ranged from 1.0 to 1.1). After propensity score weighting, the RRs in the first year were more similar in magnitude, while the RRs in the 1- to 5-year period were unchanged. These results suggest that Medicaid patients experience a medium-term reduction in comorbid disease after bariatric surgery.

Sensitivity analysis of longitudinal count responses: a local influence approach and application to medical data

ABSTRACTLongitudinal count responses are often analyzed with a Poisson mixed model. However, under overdispersion, these responses are better described by a negative binomial mixed model. Estimators of the corresponding parameters are usually obtained by the maximum likelihood method. To investigate the stability of these maximum likelihood estimators, we propose a methodology of sensitivity analysis using local influence. As count responses are discrete, we are unable to perturb them with the standard scheme used in local influence. Then, we consider an appropriate perturbation for the means of these responses. The proposed methodology is useful in different applications, but particularly when medical data are analyzed, because the removal of influential cases can change the statistical results and then the medical decision. We study the performance of the methodology by using Monte Carlo simulation and applied it to real medical data related to epilepsy and headache. All of these numerical studies show the good performance and potential of the proposed methodology.

Transportation noise exposure, noise annoyance and respiratory health in adults: A repeated-measures study

Transportation noise leads to sleep disturbance and to psychological and physiological sustained stress reactions, which could impact respiratory health. However, epidemiologic evidence on associations of objective transportation noise exposure and also perceived noise annoyance with respiratory morbidity is limited. We investigated independent associations of transportation noise exposure and noise annoyance with prevalent respiratory symptoms and incident asthma in adults.Using 17,138 observations (from 7049 participants) from three SAPALDIA (Swiss Cohort Study on Lung and Heart Diseases in Adults) surveys, we assessed associations of transportation noise exposure and noise annoyance with prevalent respiratory symptoms, and with incident asthma (in 10,657 nested observations from 6377 participants). Annual day-evening-night transportation noise comprising road, railway and aircraft Lden (Transportation Lden) was calculated for the most exposed façade of participants' residence using Swiss noise models. Transportation noise annoyance was assessed using an 11-point scale, and participants reported respiratory symptoms and doctor-diagnosed asthma at each survey. We estimated associations with transportation Lden (as well as source-specific Lden) and noise annoyance, independent of air pollution and other potential confounders, using mutually-adjusted mixed logistic and Poisson models and applying random intercepts at the level of the participants.Prevalent respiratory symptoms ranged from 5% (nocturnal dyspnoea) to 23% (regular cough/phlegm). Transportation noise annoyance, but not Lden, was independently associated with respiratory symptoms and current asthma in all participants, with odds ratios (OR) and 95% confidence intervals (CI) ranging between 1.03 (95%CI: 1.01, 1.06) and 1.07 (95% CI: 1.04, 1.11) per 1-point difference in noise annoyance. Both noise annoyance and Lden showed independent associations with asthma symptoms among asthmatics, especially in those reporting adult-onset asthma [ORLden: 1.90 (95% CI: 1.25, 2.89) per 10 dB; p-value of interaction (adult-onset vs. childhood-onset): 0.03; ORnoise annoyance: 1.06 (95%CI: 0.97, 1.16) per 1-point difference; p-value of interaction: 0.06]. No associations were found with incident asthma.Transportation noise level and annoyance contributed to symptom exacerbation in adult asthma. This suggests both psychological and physiological noise reactions on the respiratory system, and could be relevant for asthma care. More studies are needed to better understand the effects of objective and perceived noise in asthma aetiology and overall respiratory health.

Increased Impulsivity in Urban-Dwelling Teens: Role of Ambient Air Temperature and Lack of Neighborhood Greenspace

Ambient temperature and lack of residential greenspace have been associated with increased externalizing behavior, but the neuropsychological processes are unclear. We conducted a longitudinal analysis to examine the association between temperature, greenspace, and impulsivity in urban-dwelling adolescents. Participants (N&amp;#61;982) of the Risk Factors for Antisocial Behavior Study, a multi-ethnic cohort of Southern California twins/triplets examined in 2000-2014 (aged 9-21 years), each with repeat measures of impulsivity assessed by the Go/No Go test. Ambient temperature near the residence (6.8&amp;#177;3.0 km) were obtained from local Meteorological Information System and Normalized Difference Vegetation Index was used as a proxy for neighborhood greenspace (in 1000-m buffer). A monthly exposure time-series was created and aggregated in short-term (1-, 3-, and 6-month) and long-term (1-, 2-, and 3-year) averages prior to each behavioral assessment. Generalized Poisson mixed models were used to estimate the exposure effects on number of false alarms, adjusting for within-family/-individual correlations and confounders. In separate adjusted analyses, we found average temperature from 1-month to 2-years significantly (all p&amp;lt;0.05) increased impulsivity, while long-term greenspace decreased impulsivity. In models mutually adjusting for temperature and greenspace, the neurotrophic benefits of greenspace disappeared, while the adverse effects of 1- and 3-month temperature remained. These associations could not be explained by sociodemographics, neighborhood quality and socioeconomic characteristics, or season. This is the first epidemiologic evidence supporting the adverse temperature effects on impulsivity, which may contribute to the behavioral effect of diminished greenspace. Future studies need to examine whether increased impulsivity mediates the adverse behavioral effect and how neural networks underlying such neuropsychological processes are perturbed by ambient temperature.

Zero-Inflated Models for RNA-Seq Count Data

One of the main objectives of many biological studies is to explore differential gene expression profiles between samples. Genes are referred to as differentially expressed (DE) if the read counts change across treatments or conditions systematically. Poisson and negative binomial (NB) regressions are widely used methods for non-over-dispersed (NOD) and over-dispersed (OD) count data respectively. However, in the presence of excessive number of zeros, these methods need adjustments. In this paper, we consider a zero-inflated Poisson mixed effects model (ZIPMM) and zero-inflated negative binomial mixed effects model (ZINBMM) to address excessive zero counts in the NOD and OD RNA-seq data respectively in the presence of random effects. We apply these methods to both simulated and real RNA-seq datasets. The ZIPMM and ZINBMM perform better on both simulated and real datasets.

Pseudo-Bayesian D-optimal designs for longitudinal Poisson mixed models with correlated errors

This paper is concerned with the problem of pseudo-Bayesian D-optimal designs for the first-order Poisson mixed model for longitudinal data with time-dependent correlated errors. A standard approximate covariance matrix of the parameter estimation is obtained based on the quasi-likelihood method. Furthermore, to overcome the dependence of pseudo-Bayesian D-optimal designs on the choice of the prior mean, a hierarchical pseudo-Bayesian D-optimal designs based on the hierarchical prior distribution of unknown parameters is proposed. The results show that the optimal number of time points depends on both the interclass autoregressive coefficients and different cost constraints. The relative efficiency of equidistant designs compared with the hierarchical pseudo-Bayesian D-optimal designs is also discussed.

Wildlife warning reflectors do not mitigate wildlife–vehicle collisions on roads

Wildlife–vehicle collisions cause human fatalities and enormous economic and ecological losses on roads worldwide. A variety of mitigation measures have been developed over the past decades to separate traffic and wildlife, warn humans, or prevent wildlife from entering a road while vehicles are passing by, but only few are economical enough to be applied comprehensively. One such measure, wildlife warning reflectors, has been implemented over the past five decades. However, their efficacy is questioned because of contradictory study results and the variety of applied study designs and reflector models. We used a prospective, randomized non-superiority cross-over study design to test our hypothesis of the inefficacy of modern wildlife warning reflectors. We analyzed wildlife–vehicle collisions on 151 testing sites of approximately 2 km in length each. During the 24-month study period, 1984 wildlife–vehicle collisions were recorded. Confirmatory primary and exploratory secondary analyses using a log-link Poisson mixed model with normal nested random intercepts of observation year in road segment, involved species, and variables of the road segment and the surrounding environment showed that reflectors did not lower the number of wildlife–vehicle collisions by a relevant amount. In addition, variables of the road segment and the surrounding environment did not indicate differential effects of wildlife warning reflectors. Based on our results, we conclude that wildlife warning reflectors are not an effective tool for mitigating wildlife–vehicle collisions on roads.

Accumulation of unhealthy behaviors: Marked social inequalities in men and women.

The objective of this study was to compare the accumulation of unhealthy behaviors at the bottom of the social scale in men and women and, secondarily, to compare social and gender-based inequalities.Fifty-two general practitioners from the Paris area volunteered to participate. A sample of 70 patients (stratified by gender) aged 40–74 years was randomly chosen from each physician's patient panel and asked to complete a questionnaire about their social position and health behaviors: tobacco and alcohol use, unhealthy diet, and physical inactivity. Mixed Poisson models were used to describe, with relative risks (RRs) and relative inequality indexes (RIIs), the social inequalities in the accumulation of these four unhealthy behaviors.In 2008–2009, 71% of the 3640 patients returned their questionnaires. Men had an average of 1.59 of the 4 unhealthy behaviors we studied, and women 1.35 (RR = 1.18; 95% CI [1.11–1.25]). The mean number of unhealthy behaviors increased significantly for both genders from the top to the bottom of the social scale. The order of magnitude of RIIs was similar among men and women, ranging from 1.33 (occupational RII among women, 95% CI [1.11–1.60]) through 1.69 (financial RII among women, 95% CI [1.43–1.99]). None of the interaction tests between gender and social position was significant. The social inequalities had significantly wider amplitudes than those between genders for two of the three indicators of social position. The amplitude of social gradients related to unhealthy behaviors was similar between men and women and exceeded the gender inequality between them.

Modeling Count Data from Dependent Clusters with Poisson Mixed Models

&#x0D; &#x0D; &#x0D; &#x0D; Responses collected from dependent clusters are affected by the dependence among clusters and it should be taken into account in modeling such responses. In this study, a new approach was evaluated to incorporate cluster dependence in generalized linear Poisson mixed models for count responses from dependent clusters. Performance of this approach was evaluated by using a simulation process under three different designs and different covariates. The Marginal Generalized Quasi-likelihood (GQL) method was used for estimation of parameters with the cluster dependence. Monte Carlo likelihood (MCL) and Penalized Quasi-likelihood (PQL) estimates also were obtained for the purpose of comparison. Proposed approach was tested with a real data set also.&#x0D; The proposed approach, with the incorporation of cluster dependence, gives better estimates for both fixed effects and variance of random effects with low standard errors with compared to the estimates obtained by ignoring the cluster dependence. Therefore, the proposed approach can be used for modeling count responses from a dependent cluster set up.&#x0D; &#x0D; &#x0D; &#x0D;