Additive Mixed Effects Models Research Articles

The Associations between Types of Ambient PM2.5 and Under-Five and Maternal Mortality in Africa.

Exploring the effects of different types of PM2.5 is necessary to reduce associated deaths, especially in low- and middle-income countries (LMICs). Hence we determined types of ambient PM2.5 before exploring their effects on under-five and maternal mortality in Africa. The spectral derivate of aerosol optical depth (AOD) from Moderate Resolution Imaging Spectroradiometer (MODIS) products from 2000 to 2015 were employed to determine the aerosol types before using Generalized Linear and Additive Mixed-Effect models with Poisson link function to explore the associations and penalized spline for dose-response relationships. Four types of PM2.5 were identified in terms of mineral dust, anthropogenic pollutant, biomass burning and mixture aerosols. The results demonstrate that biomass PM2.5 increased the rate of under-five mortality in Western and Central Africa, each by 2%, and maternal mortality in Central Africa by 19%. Anthropogenic PM2.5 increased under-five and maternal deaths in Northern Africa by 5% and 10%, respectively, and maternal deaths by 4% in Eastern Africa. Dust PM2.5 increased under-five deaths in Northern, Western, and Central Africa by 3%, 1%, and 10%, respectively. Mixture PM2.5 only increased under-five deaths and maternal deaths in Western (incidence rate ratio = 1.01, p < 0.10) and Eastern Africa (incidence rate ratio = 1.06, p < 0.01), respectively. The findings indicate the types of ambient PM2.5 are significantly associated with under-five and maternal mortality in Africa where the exposure level usually exceeds the World Health Organization’s (WHO) standards. Appropriate policy actions on protective and control measures are therefore suggested and should be developed and implemented accordingly.

Optimal block designs with non additive mixed effects interference

ABSTRACTThis paper deals with optimality aspects of block designs balanced for interference effects from neighboring units on both sides under a general non additive model along with random block effects. Here, a class of complete, circular block designs strongly balanced for interference effects has been shown to be universally optimal for the estimation of direct effects and interference effects (left and right) of treatments under a non additive mixed effects model.

Estimation of measurement error in plasma HIV-1 RNA assays near their limit of quantification.

BackgroundPlasma HIV-1 RNA levels (pVLs), routinely used for clinical management, are influenced by measurement error (ME) due to physiologic and assay variation.ObjectiveTo assess the ME of the COBAS HIV-1 Ampliprep AMPLICOR MONITOR ultrasensitive assay version 1.5 and the COBAS Ampliprep Taqman HIV-1 assay versions 1.0 and 2.0 close to their lower limit of detection. Secondly to examine whether there was any evidence that pVL measurements closest to the lower limit of quantification, where clinical decisions are made, were susceptible to a higher degree of random noise than the remaining range.MethodsWe analysed longitudinal pVL of treatment-naïve patients from British Columbia, Canada, during their first six months on treatment, for time periods when each assay was uniquely available: Period 1 (Amplicor): 08/03/2000–01/02/2008; Period 2 (Taqman v1.0): 07/01/2010–07/03/2012; Period 3 (Taqman v2.0): 08/03/2012–30/06/2014. ME was estimated via generalized additive mixed effects models, adjusting for several clinical and demographic variables and follow-up time.ResultsThe ME associated with each assay was approximately 0.5 log10 copies/mL. The number of pVL measurements, at a given pVL value, was not randomly distributed; values ≤250 copies/mL were strongly systematically overrepresented in all assays, with the prevalence decreasing monotonically as the pVL increased. Model residuals for pVL ≤250 copies/mL were approximately three times higher than that for the higher range, and pVL measurements in this range could not be modelled effectively due to considerable random noise of the data.ConclusionsAlthough the ME was stable across assays, there is substantial increase in random noise in measuring pVL close to the lower level of detection. These findings have important clinical significance, especially in the range where key clinical decisions are made. Thus, pVL values ≤250 copies/mL should not be taken as the “truth” and repeat pVL measurement is encouraged to confirm viral suppression.

Estimating plant abundances from crown cover and forest structure data reveals size‐dependent patterns of rarity in subtropical Australia

AbstractAimsAccurate field quantification of stem counts is time consuming and laborious, especially in ecosystems containing many small individuals where counts are immense. There are many large, long‐term data sets worldwide that include various measures of woody plant abundance, such as crown cover, that are faster and easier to measure than true counts of individuals. Quantifying the number of individuals allows in‐depth research into various topics including functional ecology, carbon accounting, conservation and metabolic theory. Here we test the accuracy of predicting the abundance of individual woody plants from cover and forest structure measures, and explore the implications of refined abundance measures for understanding the rarity of woody plant species across southeast Queensland.LocationSoutheast Queensland (SEQ), Australia.MethodsTo supplement an existing data set of 1251 sites with measured crown covers for woody plants in various strata, we sampled stem counts in an additional 30 sites, forming a calibration data set. We ensured that these 30 sites covered the environmental variation in the study area. Generalized additive mixed effects models were used to predict stem counts for all woody plants in the original data set from their crown cover and forest structure values and a site‐level variable accounting for water balance. We used these refined abundance data to investigate relationships between local and regional abundance for different size classes.ResultsStem densities increased with crown cover in a non‐linear fashion; slowing once crown cover reached ca. 5% for shrubs and ca. 50% for trees, depending on their maximum potential height. Using the predicted stem count data, we found that shrubs were more locally abundant than trees, but less regionally abundant. Larger trees (&gt;20 m maximum potential height) showed a negative relationship between local and regional abundance; there was no significant relationship between other size classes.ConclusionsOur study demonstrates how crown cover can be converted to stem counts using a calibration data set and flexible mixed‐effects models. This enables further ecological analysis at a fraction of the investment required to re‐survey. Analyses of abundances revealed that tree species are less locally abundant than smaller woody species, likely due to space filling constraints, but more regionally abundant.

Congenital Anomalies of the Male Genitalia and Risk of Autism Spectrum Disorders

Introduction Prenatal exposure to ortho-phthalates has been suggested as a possible risk factor for Autism Spectrum Disorders (ASD) through their anti-androgenic activity. In rats, prenatal exposure to phthalates causes hypospadias (abnormal positioning of the urethral opening) and cryptorchidism (undescended testis), two common male reproductive tract disorders. An ecological analysis of human newborn infants showed a strong correlation between these two disorders and ASD, but this association has not been extensively studied on an individual level. Methods We analyzed all male singleton live births (n=147,171) from January 1, 1999 through December 31, 2008 occurring in a large Israeli healthcare organization. Boys with cryptorchidism (n=3,649) or hypospadias (n=2,342) were identified via ICD-9 codes. ASD cases (n=1,329) were identified by ICD-9 code and verified by review of medical records. Analyses were conducted using generalized additive mixed effect models with random component accounting for family clusters. Results In multivariable-adjusted analyses, the odds ratio (OR) for ASD among boys with hypospadias or cryptorchidism was 1.31 (95% confidence interval (CI): 1.01-1.68). For cryptorchidism only the OR was 1.44 (95% CI: 1.06-1.96) and for hypospadias only was 1.07 (95% CI: 0.70-1.64). Restricting to boys who underwent surgical correction strengthened the results for cryptorchidism (OR: 1.77; 95% CI: 1.16-2.72), but not hypospadias. Conclusions Boys with cryptorchidism were at higher risk of ASD, but we did not observe an increased risk among boys with hypospadias. Differences in ASD risk between the disorders may reflect variations in critical windows of exposure to a hypoandrogenic in-utero environment, or effects of other singling systems involved in the etiology of the disorders. Understanding the links between cryptorchidism and ASD could shed light on mechanisms of ASD pathogenesis.

Temporal patterns in habitat use by small cetaceans at an oceanographically dynamic marine renewable energy test site in the Celtic Sea

Shelf-seas are highly dynamic and oceanographically complex environments, which likely influences the spatio-temporal distributions of marine megafauna such as marine mammals. As such, understanding natural patterns in habitat use by these animals is essential when attempting to ascertain and assess the impacts of anthropogenically induced disturbances, such as those associated with marine renewable energy installations (MREIs). This study uses a five year (2009–2013) passive acoustics (C-POD) dataset to examine the use of an oceanographically dynamic marine renewable energy test site by small cetaceans, dolphins (unspecified delphinids) and harbour porpoises Phocoena phocoena, in the southern Celtic Sea. To examine how temporal patterns in habitat use across the site related to oceanographic changes occurring over broad seasonal scales as well as those driven by fine scale (bi-weekly) localised processes (that may be masked by seasonal trends), separate analyses were conducted using (1) all daily animal detection rates spanning the entire five year dataset and (2) daily animal detection rates taken only during the summer months (defined as mid-June to mid-October) of 2010 (when continuous monitoring was carried out at multiple discrete locations across the site). In both instances, generalised additive mixed effects models (GAMMs) were used to link detection rates to a suite of environmental variables representative of the oceanography of the region. We show that increased harbour porpoise detection rates in the late winter/early spring (January–March) are associated with low sea surface temperatures (SST), whilst peaks in dolphin detection rates in the summer (July–September) coincide with increased SSTs and the presence of a tidal-mixing front. Moreover, across the summer months of 2010, dolphin detection rates were found to respond to small scale changes in SST and position in the spring-neap cycle, possibly reflective of a preference for the stratified waters immediately offshore of the front. Together, these findings suggest that habitat use by small cetaceans within shelf-seas is temporally variable, species specific and likely driven by complex bottom-up processes. As such, the effective conservation management of shelf-seas requires that we understand the dynamic complexities of these systems and the species that inhabit them. In particular, we emphasise the need for a good understanding of the natural drivers of habitat use by marine megafauna before the potential impacts of anthropogenically induced disturbances, such as those associated with the construction, maintenance and operation of MREIs, can be assessed.

Additive mixed effect model for recurrent gap time data.

Gap times between recurrent events are often of primary interest in medical and observational studies. The additive hazards model, focusing on risk differences rather than risk ratios, has been widely used in practice. However, the marginal additive hazards model does not take the dependence among gap times into account. In this paper, we propose an additive mixed effect model to analyze gap time data, and the proposed model includes a subject-specific random effect to account for the dependence among the gap times. Estimating equation approaches are developed for parameter estimation, and the asymptotic properties of the resulting estimators are established. In addition, some graphical and numerical procedures are presented for model checking. The finite sample behavior of the proposed methods is evaluated through simulation studies, and an application to a data set from a clinic study on chronic granulomatous disease is provided.

Joint analysis of longitudinal data with additive mixed effect model for informative observation times

Longitudinal data occur in many clinical and observational studies, and in many situations, longitudinal responses are often correlated with observation times. In this article, we propose a new joint model for the analysis of longitudinal data with informative observation times via two random effects. In particular, an additive mixed effect model is used for observation times. Estimating equation approaches are developed for parameter estimation, and asymptotic properties of the resulting estimators are established. In addition, some graphical and numerical procedures are provided for model checking. The finite-sample behavior of the proposed method is evaluated through simulation studies, and an application to a bladder cancer study is illustrated.

Endotoxin and gender modify lung function recovery after occupational organic dust exposure: a 30-year study

ObjectivesThe purpose of this study is to determine the trajectory of lung function change after exposure cessation to occupational organic dust exposure, and to identify factors that modify improvement.MethodsThe Shanghai...

Daily activity budgets reveal a quasi-flightless stage during non-breeding in Hawaiian albatrosses.

BackgroundAnimals adjust activity budgets as competing demands for limited time and energy shift across life history phases. For far-ranging migrants and especially pelagic seabirds, activity during breeding and migration are generally well studied but the “overwinter” phase of non-breeding has received less attention. Yet this is a critical time for recovery from breeding, plumage replacement and gaining energy stores for return migration and the next breeding attempt. We aimed to identify patterns in daily activity budgets (i.e. time in flight, floating on the water’s surface and active foraging) and associated spatial distributions during overwinter for the laysan Phoebastria immutabilis and black-footed P. nigripes albatrosses using state-space models and generalized additive mixed-effects models (GAMMs). We applied these models to time-series of positional and immersion-state data from small light- and conductivity-based data loggers.ResultsDuring overwinter, both species exhibited a consistent ‘quasi-flightless’ stage beginning c. 30 days after initiating migration and lasting c. 40 days, characterized by frequent long bouts of floating, very little sustained flight, and infrequent active foraging. Minimal daily movements were made within localized areas during this time; individual laysan albatross concentrated into the northwest corner of the Pacific while black-footed albatross spread widely across the North Pacific Ocean basin. Activity gradually shifted toward increased time in flight and active foraging, less time floating, and greater daily travel distances until colony return c. 155 days after initial departure.ConclusionsOur results demonstrate that these species make parallel adjustments to activity budgets at a daily time-scale within the overwinter phase of non-breeding despite different at-sea distributions and phenologies. The ‘quasi-flightless’ stage likely reflects compromised flight from active wing moult while the subsequent increase in activity may occur as priorities shift toward mass gain for breeding. The novel application of a GAMM-based approach used in this study offers the possibility of identifying population-level patterns in shifting activity budgets over extended periods while allowing for individual-level variation in the timing of events. The information gained can also help to elucidate the whereabouts of areas important at different times across life history phases for far-ranging migrants.

Lexical differences between Tuscan dialects and standard Italian: Accounting for geographic and sociodemographic variation using generalized additive mixed modeling

This study uses a generalized additive mixed-effects regression model to predict lexical differences in Tuscan dialects with respect to standard Italian. We used lexical information for 170 concepts used by 2,060 speakers in 213 locations in Tuscany. In our model, geographical position was found to be an important predictor, with locations more distant from Florence having lexical forms more likely to differ from standard Italian. In addition, the geographical pattern varied significantly for low- versus high-frequency concepts and older versus younger speakers. Younger speakers generally used variants more likely to match the standard language. Several other factors emerged as significant. Male speakers as well as farmers were more likely to use lexical forms different from standard Italian. In contrast, higher-educated speakers used lexical forms more likely to match the standard. The model also indicates that lexical variants used in smaller communities are more likely to differ from standard Italian. The impact of community size, however, varied from concept to concept. For a majority of concepts, lexical variants used in smaller communities are more likely to differ from the standard Italian form. For a minority of concepts, however, lexical variants used in larger communities are more likely to differ from standard Italian. Similarly, the effect of the other community- and speaker-related predictors varied per concept. These results clearly show that the model succeeds in teasing apart different forces influencing the dialect landscape and helps us to shed light on the complex interaction between the standard Italian language and the Tuscan dialectal varieties. In addition, this study illustrates the potential of generalized additive mixed-effects regression modeling applied to dialect data.

Trends in stratospheric ozone profiles using functional mixed models

Abstract. This paper is devoted to the modeling of altitude-dependent patterns of ozone variations over time. Umkehr ozone profiles (quarter of Umkehr layer) from 1978 to 2011 are investigated at two locations: Boulder (USA) and Arosa (Switzerland). The study consists of two statistical stages. First we approximate ozone profiles employing an appropriate basis. To capture primary modes of ozone variations without losing essential information, a functional principal component analysis is performed. It penalizes roughness of the function and smooths excessive variations in the shape of the ozone profiles. As a result, data-driven basis functions (empirical basis functions) are obtained. The coefficients (principal component scores) corresponding to the empirical basis functions represent dominant temporal evolution in the shape of ozone profiles. We use those time series coefficients in the second statistical step to reveal the important sources of the patterns and variations in the profiles. We estimate the effects of covariates – month, year (trend), quasi-biennial oscillation, the solar cycle, the Arctic oscillation, the El Niño/Southern Oscillation cycle and the Eliassen–Palm flux – on the principal component scores of ozone profiles using additive mixed effects models. The effects are represented as smooth functions and the smooth functions are estimated by penalized regression splines. We also impose a heteroscedastic error structure that reflects the observed seasonality in the errors. The more complex error structure enables us to provide more accurate estimates of influences and trends, together with enhanced uncertainty quantification. Also, we are able to capture fine variations in the time evolution of the profiles, such as the semi-annual oscillation. We conclude by showing the trends by altitude over Boulder and Arosa, as well as for total column ozone. There are great variations in the trends across altitudes, which highlights the benefits of modeling ozone profiles.

Additive Mixed Effect Model for Clustered Doubly Censored Data

Cai and Zeng (2011) proposed an additive mixed effect model to analyze clustered right-censored data. In this article, we demonstrate that the approach of Cai and Zeng (2011) can be extended to clustered doubly censored data. Furthermore, when both left- and right-censoring variables are always observed, we propose alternative estimators using the approach of Cai and Cheng (2004). A simulation study is conducted to investigate the performance of the proposed estimators.

The impact of co-infections on the haematological profile of East African Short-horn Zebu calves.

The cumulative effect of co-infections between pathogen pairs on the haematological response of East African Short-horn Zebu calves is described. Using a longitudinal study design a stratified clustered random sample of newborn calves were recruited into the Infectious Diseases of East African Livestock (IDEAL) study and monitored at 5-weekly intervals until 51 weeks of age. At each visit samples were collected and analysed to determine the infection status of each calf as well as their haematological response. The haematological parameters investigated included packed cell volume (PCV), white blood cell count (WBC) and platelet count (Plt). The pathogens of interest included tick-borne protozoa and rickettsias, trypanosomes and intestinal parasites. Generalized additive mixed-effect models were used to model the infectious status of pathogens against each haematological parameter, including significant interactions between pathogens. These models were further used to predict the cumulative effect of co-infecting pathogen pairs on each haematological parameter. The most significant decrease in PCV was found with co-infections of trypanosomes and strongyles. Strongyle infections also resulted in a significant decrease in WBC at a high infectious load. Trypanosomes were the major cause of thrombocytopenia. Platelet counts were also affected by interactions between tick-borne pathogens. Interactions between concomitant pathogens were found to complicate the prognosis and clinical presentation of infected calves and should be taken into consideration in any study that investigates disease under field conditions.

Amino Acid Usage Is Asymmetrically Biased in AT- and GC-Rich Microbial Genomes

IntroductionGenomic base composition ranges from less than 25% AT to more than 85% AT in prokaryotes. Since only a small fraction of prokaryotic genomes is not protein coding even a minor change in genomic base composition will induce profound protein changes. We examined how amino acid and codon frequencies were distributed in over 2000 microbial genomes and how these distributions were affected by base compositional changes. In addition, we wanted to know how genome-wide amino acid usage was biased in the different genomes and how changes to base composition and mutations affected this bias. To carry this out, we used a Generalized Additive Mixed-effects Model (GAMM) to explore non-linear associations and strong data dependences in closely related microbes; principal component analysis (PCA) was used to examine genomic amino acid- and codon frequencies, while the concept of relative entropy was used to analyze genomic mutation rates.ResultsWe found that genomic amino acid frequencies carried a stronger phylogenetic signal than codon frequencies, but that this signal was weak compared to that of genomic %AT. Further, in contrast to codon usage bias (CUB), amino acid usage bias (AAUB) was differently distributed in AT- and GC-rich genomes in the sense that AT-rich genomes did not prefer specific amino acids over others to the same extent as GC-rich genomes. AAUB was also associated with relative entropy; genomes with low AAUB contained more random mutations as a consequence of relaxed purifying selection than genomes with higher AAUB.ConclusionGenomic base composition has a substantial effect on both amino acid- and codon frequencies in bacterial genomes. While phylogeny influenced amino acid usage more in GC-rich genomes, AT-content was driving amino acid usage in AT-rich genomes. We found the GAMM model to be an excellent tool to analyze the genomic data used in this study.

Bi-phasic trends in mercury concentrations in blood of Wisconsin common loons during 1992–2010

We assessed the ecological risk of mercury (Hg) in aquatic systems by monitoring common loon (Gavia immer) population dynamics and blood Hg concentrations. We report temporal trends in blood Hg concentrations based on 334 samples collected from adults recaptured in subsequent years (resampled 2-9 times) and from 421 blood samples of chicks collected at lakes resampled 2-8 times 1992-2010. Temporal trends were identified with generalized additive mixed effects models and mixed effects models to account for the potential lack of independence among observations from the same loon or same lake. Trend analyses indicated that Hg concentrations in the blood of Wisconsin loons declined over the period 1992-2000, and increased during 2002-2010, but not to the level observed in the early 1990s. The best fitting linear mixed effects model included separate trends for the two time periods. The estimated trend in Hg concentration among the adult loon population during 1992-2000 was -2.6% per year, and the estimated trend during 2002-2010 was +1.8% per year; chick blood Hg concentrations decreased -6.5% per year during 1992-2000, but increased 1.8% per year during 2002-2010. This bi-phasic pattern is similar to trends observed for concentrations of methylmercury and SO(4) in lake water of an intensely studied seepage lake (Little Rock Lake, Vilas County) within our study area. A cause-effect relationship between these independent trends is hypothesized.

P1-S1.12 Age-specific incidence of chlamydia, gonorrhoea, and trichomoniasis infection in adolescent women

BackgroundAge-specific incidence rates of sexually transmitted infections (STI) with Chlamydia trachomatis (CT), Neisseria gonorrhoeae (GC), and Trichomonas vaginalis (TV) are not well characterised in adolescents and young adult women. In...

Additive Mixed Effect Model for Clustered Failure Time Data

We propose an additive mixed effect model to analyze clustered failure time data. The proposed model assumes an additive structure and includes a random effect as an additional component. Our model imitates the commonly used mixed effect models in repeated measurement analysis but under the context of hazards regression; our model can also be considered as a parallel development of the gamma-frailty model in additive model structures. We develop estimating equations for parameter estimation and propose a way of assessing the distribution of the latent random effect in the presence of large clusters. We establish the asymptotic properties of the proposed estimator. The small sample performance of our method is demonstrated via a large number of simulation studies. Finally, we apply the proposed model to analyze data from a diabetic study and a treatment trial for congestive heart failure.

Additive Mixed Effect Model for Clustered Failure Time Data

Additive Mixed Effect Model for Clustered Failure Time Data

Geographical variation in the behaviour of a central place forager: Antarctic fur seals foraging in contrasting environments

Foragers show adaptive responses to changes within their environment, and such behavioural plasticity can be a significant driving force in speciation. We investigated how lactating Antarctic fur seals, Arctocephalus gazella, adapt their foraging within two contrasting ecosystems. Location and diving data were collected concurrently, between December 2003 and February 2004, from 43 seals at Bird Island, where krill, Euphausia superba, are the main prey, and 39 at Heard Island, where mostly fish are consumed. Seals at Heard Island were shorter and lighter than those at Bird Island and they spent longer at sea, dived more frequently and spent more time in the bottom phase of dives. Generalized additive mixed effects models showed that diving behaviours differed between the islands. Both populations exploited diel vertically migrating prey species but, on average, Heard Island seals dived deeper and exceeded their estimated aerobic dive limits. We propose that the recovery of the Heard Island population may be limited by the relative inaccessibility and scarcity of food, whereas at Bird Island, the presence of abundant krill resources helps sustain extremely high numbers of seals, even with increased intra- and inter-specific competition. Both populations of fur seals appear to be constrained by their physiological limits, in terms of their optimal diving behaviour. However, there does appear to be some flexibility in strategy at the level of trip with animals adjusting their time at sea and foraging effort, in order to maximize the rate of delivery of energy to their pups.