The discovery of generalized causal models with mixed variables using MML criterion

One major difficulty frustrating the application of linear causal models is that they are not easily adapted to cope with discrete data. This is unfortunate since most real problems involve both continuous and discrete variables. In this paper, we consider a class of graphical models which allow both continuous and discrete variables, and propose the parameter estimation method and a structure discovery algorithm based on Minimum Message Length and parameter estimation. Experimental results are given to demonstrate the potential for the application of this method.

Modelling grouped survival times in toxicological studies using generalized additive models

A method for combining a proportional-hazards survival time model with a bioassay model where the log-hazard function is modelled as a linear or smoothing spline function of log-concentration combined with a smoothing spline function of time is described. The combined model is fitted to mortality numbers, resulting from survival times that are grouped due to a common set of observation times, using Generalized Additive Models (GAMs). The GAM fits mortalities as conditional binomials using an approximation to the log of the integral of the hazard function and is implemented using freely-available, general software for fitting GAMs. Extensions of the GAM are described to allow random effects to be fitted and to allow for time-varying concentrations by replacing time with a calibrated cumulative exposure variable with calibration parameter estimated using profile likelihood. The models are demonstrated using data from a studies of a marine and a, previously published, freshwater taxa. The marine study involved two replicate bioassays of the effect of zinc exposure on survival of an Antarctic amphipod, Orchomenella pinguides. The other example modelled survival of the daphnid, Daphnia magna, exposed to potassium dichromate and was fitted by both the GAM and the process-based DEBtox model. The GAM fitted with a cubic regression spline in time gave a 61 % improvement in fit to the daphnid data compared to DEBtox due to a non-monotonic hazard function. A simulation study using each of these hazard functions as operating models demonstrated that the GAM is overall more accurate in recovering lethal concentration values across the range of forms of the underlying hazard function compared to DEBtox and standard multiple endpoint probit analyses.

Association between obesity and diabetes using nested linear models

A data source with 7260 cases from the prospective observational study was used. Cases included both sexes and age ranging between 20yr and 7Oyr. Chi-squared tests and linear models were used to examine the effects of age (linear and quadratic), sex, smoking habit (currently a smoker or non-smoker), presence or absence of type 2 diabetes on obesity as measured by body mass index (BMI) and waist to hip ratio (WHR). Both these measures of
obesity were significantly affected by age, sex and the presence of diabetes. Cases with diabetes were significantly more obese (10%) between the ages 40yr to 59yr compared to patients without diabetes. These results accentuate the need to further examine the association between age, diabetes and obesity and other risk factors with cardiovascular diseases.

Detecting contaminated birthdates using generalized additive models.

Erroneous patient birthdates are common in health databases. Detection of these errors usually involves manual verification, which can be resource intensive and impractical. By identifying a frequent manifestation of birthdate errors, this paper presents a principled and statistically driven procedure to identify erroneous patient birthdates.

How socio-economic status contributes to participation in leisure-time physical activity

The aim of this cross-sectional study was to identify individual, social, and environmental contributors (mediators) to individual- and area-level differences in leisure-time physical activity across socio-economic groups. A two-stage stratified sampling design was used to recruit 20–65 year old adults (N = 2194) living in 154 census collection districts of Adelaide, Australia (overall response rate: 12%). Participants completed two surveys six months apart (response rate on the second survey: 83%). Individual-level socio-economic status (SES) was assessed using self-report measures on educational attainment, household income, and household size. Area-level SES was assessed using census data on median household income and household size for each selected census district. Bootstrap generalized linear models were used to examine associations between SES, potential mediators, and leisure-time physical activity. The product-of-coefficient test was used to estimate mediating effects. All SES measures were independently associated with potential individual and social mediators of the SES-activity relationships. Individual- and area-level income was also associated with perceived neighborhood attributes. Self-efficacy and social support for physical activity explained virtually all of the differences in physical activity across educational attainment groups. Physical barriers to walking and access to public open space contributed in part to the explanation of differences in recreational walking across income groups. Yet, self-efficacy and social support were the key mediators of the observed relationships between individual- and area-level income and physical activity. This study suggests that in order to increase physical activity participation in the more disadvantaged segments of the population, comprehensive, multilevel interventions targeting activity-related attitudes and skills as well as social and physical environments are needed.

Are perceptions of the local environment related to neighbourhood satisfaction and mental health in adults?

Background: Perceptions of environmental attributes can influence satisfaction with where people live and mental health status. We examined the association between perceived environmental characteristics, neighbourhood satisfaction, and self-rated mental health.

Methods: We report cross-sectional data from the Physical Activity in Localities and Community Environments (PLACE) study in Australia (n = 2194). Self-report data included socio-demographics, perceived attributes of the environment, neighbourhood satisfaction (NS) and mental health status. Neighbourhood SES was obtained through census data. Factor analysis was used to identify dimensions of NS. Generalized linear models were used to examine associations between NS and perceived environment characteristics and whether aspects of NS were independently associated with mental health.

Results: NS factors identified were safety and walkability, access to destinations, social network, travel network, and traffic and noise. Perceived environmental characteristics of aesthetics and greenery, land use mix – diversity, street connectivity, traffic safety, infrastructure for walking, access to services and barriers to walking were found to be positively associated with these factors. Traffic load and crime were negatively associated. Three NS factors (safety and walkability, social network, and traffic and noise) were independent predictors of mental health.

Conclusions: Neighbourhood satisfaction may mediate the association between perceived environmental characteristics and measures of mental health in adults.

Maternal plasma ascorbic acid (vitamin C) and risk of gestational diabetes mellitus

Background : Antioxidants, particularly vitamin C (ascorbic acid), have the capacity to influence glucose tolerance. Modification of diet could reduce the likelihood of developing gestational diabetes mellitus.

Methods : In a prospective cohort study of pregnant women, we studied the association of maternal plasma ascorbic acid concentrations, measured at an average of 13 weeks' gestation, with subsequent risk of gestational diabetes. Maternal plasma ascorbic acid concentrations were determined using automated enzymatic procedures. Dietary vitamin C intake during the periconceptional period and early pregnancy was ascertained using a semiquantitative food frequency questionnaire. We fitted generalized linear models to derive estimates of relative risks and 95% confidence intervals (CIs).

Results : Approximately 4% (n = 33) of 755 women who completed pregnancy developed gestational diabetes mellitus. Plasma ascorbic acid concentrations were inversely associated with the risk of gestational diabetes (P for trend = 0.023). After adjusting for maternal age, race, prepregnancy adiposity, parity, family history of type 2 diabetes, and household income, women with plasma ascorbic acid <55.9 micromol/L (lowest quartile) experienced a 3.1-fold increased risk of gestational diabetes (95% CI = 1.0 - 9.7) compared with women whose concentrations were > or = 74.6 micromol/L (upper quartile). Women who consumed <70 mg vitamin C daily experienced a 1.8-fold increased risk of gestational diabetes compared with women who consumed higher amounts (95% CI = 0.8 - 4.4).

Conclusions : If confirmed, our results raise the possibility that current efforts to encourage populations to consume diets rich in antioxidants, including vitamin C, could reduce the occurrence of gestational diabetes mellitus.

Stable isotopes reveal inter-annual and inter-individual variation in the diet of female Australian fur seals

Understanding the temporal and spatial variation of foraging habits of apex predators is central to understanding their role in marine ecosystems and how their populations may respond to environmental variability. In the present study, stable isotope analysis (C and N) of blood was used to investigate inter-individual and inter-annual differences in the diet of adult female Australian fur seals Arctocephalus pusillus doriferus. Positive correlations were observed between red cell and plasma values for δ¹³C and δ¹⁵N (r² = 0.47 and r² = 0.66, respectively, p < 0.001 in both cases), suggesting relatively consistent individual prey choices over 3 or 4 foraging trips. Mean δ¹⁵N values (12.8 to 17.5%) confirm the species occupies the highest marine trophic niche in the region. A significant decrease in plasma δ¹⁵N values, corresponding to two-thirds of a trophic level (ca. 2%), was observed between the 1998 to 2000 and 2003 to 2005 sampling periods. This was associated with a significant decrease in adult female body condition and is consistent with a decline, previously documented by faecal analysis, of the proportion of red cod Pseudophysis bachus, barracouta Thyrsites atun and Gould's squid Nototodarus gouldi in the diet and an increase in redbait Emmelichthys nitidus. While substantial variation in δ¹⁵N was observed within each age cohort, a significant decrease was observed with age, suggesting individual specialisation for particular prey types is evident early in adulthood, but that its composition changes as females age. In addition, generalized linear models indicated body mass had a negative influence on δ¹⁵N, which may reflect larger total body oxygen stores, facilitating individuals hunting cryptic prey of lower trophic level (e.g. octopus) on the sea floor.

Statistical approaches to testing the relationships of the built environment with resident-level physical activity behavior and health outcomes in cross-sectional studies with cluster sampling

To achieve valid conclusions, studies exploring associations of the built environment with residents' physical activity and health-related outcomes need to employ statistical approaches accounting for clustered data. This article discusses the following main statistical approaches: analysis of covariance, regression models with robust standard errors, generalized estimating equations, and multilevel generalized linear models. The choice of a statistical method depends on the characteristics of the study and research questions. While the first three approaches are employed to account for clustering in the data, multilevel models can also help unravel more substantive issues within a social ecological theoretical framework of health behavior.

Associations of leisure-time physical activity with quality of life in a large, population-based sample of colorectal cancer survivors

Objective : To examine the associations between physical activity and quality of life for colorectal cancer survivors; and to describe the associations of medical and sociodemographic attributes with overall quality of life, and their moderating effects on the relationships between physical activity and quality of life.

Methods : Telephone interviews were conducted with 1,996 colorectal cancer survivors recruited through the Queensland Cancer Registry. Data were collected on current quality of life; leisure-time physical activity pre- and post-diagnosis; cancer treatment and side-effects; and general sociodemographic attributes. Hierarchical generalized linear models identified variables significantly associated with quality of life.

Results : After controlling for sociodemographic variables, disease-specific variables, treatment side-effects, and pre-diagnosis leisure-time physical activity, there were significant differences in quality of life scores by post-diagnosis physical activity category. Compared to participants who were inactive after their diagnosis, those who were sufficiently active had a 17.0% higher total quality of life score. Physical activity also had a significant independent positive association with the physical well-being, functional well-being, and additional concerns subscales of the FACT-C.

Conclusions : Our findings demonstrate that quite modest changes in leisure-time physical activity are associated with quality of life. Colorectal cancer survivors may benefit from a more active lifestyle.

Predicting avian distributions to evaluate spatiotemporal overlap with locust control operations in eastern Australia.

Locusts and grasshoppers cause considerable economic damage to agriculture worldwide. The Australian Plague Locust Commission uses multiple pesticides to control locusts in eastern Australia. Avian exposure to agricultural pesticides is of conservation concern, especially in the case of rare and threatened species. The aim of this study was to evaluate the probability of pesticide exposure of native avian species during operational locust control based on knowledge of species occurrence in areas and times of application. Using presence-absence data provided by the Birds Australia Atlas for 1998 to 2002, we developed a series of generalized linear models to predict avian occurrences on a monthly basis in 0.5 degrees grid cells for 280 species over 2 million km2 in eastern Australia. We constructed species-specific models relating occupancy patterns to survey date and location, rainfall, and derived habitat preference. Model complexity depended on the number of observations available. Model output was the probability of occurrence for each species at times and locations of past locust control operations within the 5-year study period. Given the high spatiotemporal variability of locust control events, the variability in predicted bird species presence was high, with 108 of the total 280 species being included at least once in the top 20 predicted species for individual space-time events. The models were evaluated using field surveys collected between 2000 and 2005, at sites with and without locust outbreaks. Model strength varied among species. Some species were under- or over-predicted as times and locations of interest typically did not correspond to those in the prediction data set and certain species were likely attracted to locusts as a food source. Field surveys demonstrated the utility of the spatially explicit species lists derived from the models but also identified the presence of a number of previously unanticipated species. These results also emphasize the need for special consideration of rare and threatened species that are poorly predicted by presence-absence models. This modeling exercise was a useful a priori approach in species risk assessments to identify species present at times and locations of locust control applications, and to discover gaps in our knowledge and need for further focused data collection.

How long is too long in contemporary peer review? Perspectives from authors publishing in conservation biology journals

Delays in peer reviewed publication may have consequences for both assessment of scientific prowess in academics as well as communication of important information to the knowledge receptor community. We present an analysis on the perspectives of authors publishing in conservation biology journals regarding their opinions on the importance of speed in peer-review as well as how to improve review times. Authors were invited to take part in an online questionnaire, of which the data was subjected to both qualitative (open coding, categorizing) and quantitative analyses (generalized linear models). We received 637 responses to a total of 6,547 e-mail invitations sent. Peer-review speed was generally perceived as slow, with authors experiencing a typical turnaround time of 14 weeks while their perceived optimal review time is six weeks. Male and younger respondents seem to have higher expectations of review speed than females and older respondents. Majority of participants attributed lengthy review times to the 'stress' on the peer-review system (i.e., reviewer and editor fatigue), while editor persistence and journal prestige were believed to speed up the review process. Negative consequences of lengthy review times appear to be greater for early career researchers and can also have impact on author morale (e.g. motivation or frustration). Competition among colleagues were also of concern to respondents. Incentivizing peer review was among the top suggested alterations to the system along with training graduate students in peer review, increased editorial persistence, and changes to the norms of peer-review such as opening the peer-review process to the public. It is clear that authors surveyed in this study view the peer-review system as under stress and we encourage scientists and publishers to push the envelope for new peer review models.

Ecology of urban malaria vectors in Niamey, Republic of Niger

BACKGROUND: Urbanization in African cities has major impact on malaria risk. Niamey, the capital of the Republic of Niger, is situated in the West African Sahel zone. The short rainy season and human activities linked with the Niger River influence mosquito abundance. This study aimed at deciphering the factors of distribution of urban malaria vectors in Niamey.

METHODS: The distribution of mosquito aquatic stages was investigated monthly from December 2002 to November 2003, at up to 84 breeding sites, throughout Niamey. An exploratory analysis of association between mosquito abundance and environmental factors was performed by a Principal Component Analysis and confirmed by Kruskall-Wallis non-parametric test. To assess the relative importance of significant factors, models were built for Anopheles and Culicinae. In a second capture session, adult mosquitoes were collected weekly with pyrethrum sprays and CDC light-traps from June 2008 to June 2009 in two differentiated urban areas chosen after the study's first step. Members of the Anopheles gambiae complex were genotyped and Anopheles females were tested for the presence of Plasmodium falciparum circumsporozoite antigens using ELISA.

RESULTS: In 2003, 29 % of 8420 mosquitoes collected as aquatic stages were Anopheles. They were significantly more likely to be found upstream, relatively close to the river and highly productive in ponds. These factors remained significant in regression and generalized linear models. The Culicinae were found significantly more likely close to the river, and in the main temporary affluent stream. In 2009, Anopheles specimens, including Anopheles gambiae s.l. (95 %), but also Anopheles funestus (0.6 %) accounted for 18 % of the adult mosquito fauna, with a large difference between the two sampled zones. Three members of the An. gambiae complex were found: Anopheles arabiensis, Anopheles coluzzii, and An. gambiae. Nineteen (1.3 %) out of 1467 females tested for P. falciparum antigen were found positive.

CONCLUSION: The study provides valuable update knowledge on malaria vector ecology and distribution in Niamey. The identification of spatial and environmental risk factors could pave the way to larval source management strategy and allow malaria vector control to focus on key zones for the benefit of the community.

Predictive distribution models and their application in wildlife conservation

Wildlife managers are often faced with the difficult task of determining the distribution of species, and their preferred habitats, at large spatial scales. This task is even more challenging when the species of concern is in low abundance and/or the terrain is largely inaccessible. Spatially explicit distribution models, derived from multivariate statistical analyses and implemented in a geographic information system (GIS), can be used to predict the distributions of species and their habitats, thus making them a useful conservation tool. We present two such models: one for a dasyurid, the Swamp Antechinus (Antechinus minimus), and the other for a ground-dwelling bird, the Rufous Bristlebird (Dasyornis broadbenti), both of which are rare species occurring in the coastal heathlands of south-western Victoria. Models were generated using generalized linear modelling (GLM) techniques with species presence or absence as the independent variable and a series of landscape variables derived from GIS layers and high-resolution imagery as the predictors. The most parsimonious model, based on the Akaike Information Criterion, for each species then was extrapolated spatially in a GIS. Probability of species presence was used as an index of habitat suitability. Because habitat fragmentation is thought to be one of the major threats to these species, an assessment of the spatial distribution of suitable habitat across the landscape is vital in prescribing management actions to prevent further habitat fragmentation.