Poisson Generalized Linear Model Research Articles

Experiencing violence and other predictors of within-person same-day use of multiple substances in youth: a longitudinal study in emergency settings

ABSTRACT Background: Although experiencing violence is a risk factor for substance use among youth, its association with same-day use of multiple substances (a form of polysubstance use) and mitigating factors is less well understood. Objectives: To identify whether prosocial factors modified the effect of experiencing violence on the frequency of same-day use, and examine gender-specific risk/protective factors for same-day use. Methods: We analyzed longitudinal data from a cohort of youth who use drugs aged 14–24 (n = 599; 58% male) presenting to an urban emergency department between 2009–2011 and assessed biannually for two years. Using Poisson-generalized linear models with person-level fixed effects, we estimated within-person associations between self-reported experiencing violence and same-day use and analyzed gender and peer/parent support as effect modifiers. We adjusted for negative peer influence, parental drug and alcohol use, family conflict, anxiety and depression, and age. Results: Overall, positive parental support corresponded to lower rates of same-day use (rate ratio [RR]:0.93, 95% CI:0.87–0.99) and experiencing violence was associated with higher rates of same-day use (RR:1.25, 95% CI:1.10–1.41). Violence exposure was a risk factor among males (RR:1.42, 95% CI:1.21–1.66), while negative peer influences and parental substance use were risk factors among females (RR:1.63, 95% CI:1.36–1.97 and RR:1.58, 95% CI:1.35–1.83, respectively). Positive peer support reduced the association between violence exposure and same-day use among males (RR:0.69, 95% CI:0.57–0.84, p < .05). Conclusions: Tailored interventions may address gender differences in coping with experiencing violence – including interventions that promote parental support among males and reduce influence from parental substance use among females.

A flexible quasi-likelihood model for microbiome abundance count data.

In this article, we present a flexible model for microbiome count data. We consider a quasi-likelihood framework, in which we do not make any assumptions on the distribution of the microbiome count except that its variance is an unknown but smooth function of the mean. By comparing our model to the negative binomial generalized linear model (GLM) and Poisson GLM in simulation studies, we show that our flexible quasi-likelihood method yields valid inferential results. Using a real microbiome study, we demonstrate the utility of our method by examining the relationship between adenomas and microbiota. We also provide an R package "fql" for the application of our method.

Retinal ganglion cells encode differently in the myopic mouse retina?

The etiology of myopia remains unclear. This study investigated whether retinal ganglion cells (RGCs) in the myopic retina encode visual information differently from the normal retina and to determine the role of Connexin (Cx) 36 in this process. Generalized linear models (GLMs), which can capture stimulus-dependent changes in real neurons with spike timing precision and reliability, were used to predict RGCs responses to focused and defocused images in the retinas of wild-type (normal) and Lens-Induced Myopia (LIM) mice. As the predominant subunit of gap junctions in the mouse retina and a plausible modulator in myopia development, Cx36 knockout (KO) mice were used as a control for an intact retinal circuit. The kinetics of excitatory postsynaptic currents (EPSCs) of a single αRGC could reflect projection of both focused and defocused images in the retinas of normal and LIM, but not in the Cx36 knockout mice. Poisson GLMs revealed that RGC encoding of visual stimuli in the LIM retina was similar to that of the normal retina. In the LIM retinas, the linear-Gaussian GLM model with offset was a better fit for predicting the spike count under a focused image than the defocused image. Akaike information criterion (AIC) indicated that nonparametric GLM (np-GLM) model predicted focused/defocused images better in both LIM and normal retinas. However, the spike counts in 33% of αRGCs in LIM retinas were better fitted by exponential GLM (exp-GLM) under defocus, compared to only 13% αRGCs in normal retinas. The differences in encoding performance between LIM and normal retinas indicated the possible amendment and plasticity of the retinal circuit in myopic retinas. The absence of a similar response between Cx36 KO mice and normal/LIM mice might suggest that Cx36, which is associated with myopia development, plays a role in encoding focused and defocused images.

Modeling red-light running behavior using high-resolution event-based data: a finite mixture modeling approach

To effectively reduce the number of red-light violations and crashes, it is crucial to explore RLR behavior at local intersections, understand the contributing factors, and identify the riskiest intersections by estimating RLR frequency. In this study, a finite mixture modeling method was utilized to understand the contributing factors to RLR behavior and estimate this violating behavior. To develop the RLR estimation models, performance metrics and signal phasing data were collected from the Automated Traffic Signal Performance Measures (ATSPMs) system in two jurisdictions in Arizona: Pima County and the Town of Marana. The results from calibrated models showed that an increase in traffic flow, intersection delay, number of approach lanes, and split failure is associated with an increase in the likelihood of observing red-light violations. In addition, it was found that an increase in cycle length is associated with a decrease in the likelihood of observing the red-light violation. The results of comparing the proposed RLR estimation method with several conventional methods, the Poisson Generalized Linear Model (PGLM), Zero-inflated Poisson Regression Model (ZIPM), and Zero-inflated Negative Binomial Regression Model (ZINB) showed the proposed method outperforms all the models in terms of both model fit and accuracy. The application of the proposed method could be used to analyze the intersections with the highest number of red-light violations. Furthermore, the presented transferability results can be advantageous to transportation agencies within Arizona and urban areas with similar characteristics by providing insight into which model specifications may provide the best RLR estimation accuracy.

Application of Poisson and negative binomials models to estimate the frequency of insurance claims

Generalized Linear Models (GLMs) are a modeling approach that allows the modeling of nonlinear behaviors and non-Gaussian distributions of residues. This approach is very useful for general insurance analysis, where the frequency of claims and the amount of claims distributions are usually non-Gaussian. In this article, the application of Poisson and Negative Binomial models to estimate the frequency of claims of auto insurance is discussed. The accuracy of the models was compared to choose the best model for determining pure insurance premiums using R software. The data used are a secondary dataset which is the motor vehicle insurance dataset from Sweden named dataOhlsson and the motor vehicle dataset from Australia named ausprivauto0405. The results of the exploration of the GLMs model are that Poisson's GLM and Negative Binomial models both are suitable models for estimating the number of claims for the dataOhlsson dataset. Both models have relatively similar parameter estimates, as well as the AIC and BIC values for the dataOhlsson dataset, however, both models are not suitable for estimating the number of claims for the ausprivauto0405 dataset. More investigation using different models is needed to ensure which model is more appropriate for estimating the frequency of insurance claims.

Long-term effects of air pollutants on deaths in a semi-urban city in Northwestern Turkey: a time series analysis

ABSTRACT To examine the associations between daily variations of coarse Particulate Matter(PM10) and/or sulfur dioxide(SO2) and mortality. The Poisson Generalized Linear Model(GLM) was employed to analyze the relationship between ambient air pollutants such as PM10 and SO2 and mortality. For each 10 μg/m3 increase in PM10, the overall mortality risk was found to be 1.022-fold high on the previous-eighth-day(lag 7) (RR, 95%CI:1.002–1.042) in the unadjusted model; 1.031-fold high in men (RR, 95%CI:1.005–1.058); 1.024-fold high in those aged 65 and over (RR, 95%CI:1.001–1.048). Also, the risk of death in men was 1.028-fold high in the model adjusted on the previous- eighth-day(lag 7) (RR, 95%CI:1.002–1.055). Mortality risk was found to be 1.088-fold high in 10 μg/m3 increase in SO2 under 65 years in males in the previous-third-day(lag 2) in the unadjusted model, and the risk of death was found to be 1.086-fold (RR, 95%CI:1.007–1.164) high in males in the adjusted model.

The impact of hazes on schizophrenia admissions and the synergistic effect with the combined atmospheric oxidation capacity in Hefei, China

ObjectivesCurrently, most epidemiological studies on haze focus on respiratory diseases, cardiovascular diseases, etc. However, the relationship between haze and mental health has not been adequately explored. The purpose of this study was to investigate the influence of hazes on schizophrenia admissions and to further explore the potential interaction effect with the combined atmospheric oxidative indices (Ox and Oxwt). MethodsWe collected 5328 cases during the cold season from 2013 to 2015 in Hefei, China. By integrating the Poisson Generalized Linear Models with the Distributed Lag Non-linear Models, the association between haze and schizophrenia admissions was evaluated. The interaction between hazes and two combined oxidation indexes was tested by stratifying hazes and Ox, and Oxwt. ResultsHaze was found to be significantly linked to an increased risk of hospitalization for schizophrenia, and a 9-day lag effect on schizophrenia (lag 3-lag 11), with the largest effect on lag 6 (RR = 1.080, 95% confidence interval (CI): 1.046–1.116). Males, females, and <40 y (people under 40 years old) were sensitive to hazes. Furthermore, in the stratified analysis, we found synergies between two combined oxidation indexes and hazes. The interaction relative risk (IRR) and relative excess risk due to interaction (RERI) between Ox and hazes were 1.170 (95% CI: 1.071–1.277) and 0.149 (95% CI: 0.045–0.253), respectively. For Oxwt, the IRR and RERI were 1.179 (95% CI: 1.087–1.281) and 0.159 (95% CI: 0.056–0.263), respectively. It is noteworthy that this synergistic effect was significant in males and <40 y when examining the various subgroups in the interaction analysis. ConclusionsOur findings suggest that exposure to haze significantly increases the risk of hospitalization for schizophrenia. More significant public health benefits can be obtained by prioritizing haze periods with high combined atmospheric oxidation capacity.

Response versus gradient boosting trees, GLMs and neural networks under Tweedie loss and log-link

Thanks to its outstanding performances, boosting has rapidly gained wide acceptance among actuaries. To speed up calculations, boosting is often applied to gradients of the loss function, not to responses (hence the name gradient boosting). When the model is trained by minimizing Poisson deviance, this amounts to apply the least-squares principle to raw residuals. This exposes gradient boosting to the same problems that lead to replace least-squares with Poisson Generalized Linear Models (GLM) to analyze low counts (typically, the number of reported claims at policy level in personal lines). This paper shows that boosting can be conducted directly on the response under Tweedie loss function and log-link, by adapting the weights at each step. Numerical illustrations demonstrate similar or better performances compared to gradient boosting when trees are used as weak learners, with a higher level of transparency since responses are used instead of gradients.

The influence of modes of climate variability on the sub-seasonal temporal clustering of extreme precipitation

SummaryTemporally clustered precipitation extremes can have catastrophic impacts. Therefore, understanding their drivers is paramount for risk assessment in current and future climates. Here, we model for each season 3-week extreme precipitation event counts with Poisson Generalized Linear Models and nine major modes of climate variability as covariates. Model goodness-of-fit is highest in the tropics, particularly over the equatorial Pacific, the Maritime Continent, and East Africa, where ENSO, the Indian Ocean Dipole (IOD) and the MJO are the major drivers of sub-seasonal temporal clustering of extreme precipitation. The IOD and MJO also matter over Southwest Asia during boreal fall and winter. In the Northern Hemisphere, the North Atlantic Oscillation impacts clustering west of the Iberian Peninsula and over Scandinavia and Greenland, and the Pacific North American pattern matters over the central/northern Pacific Ocean. Finally, our models show very little skill in the Southern Hemisphere, where temporal clustering is also less frequent.

Weight of risk factors for mortality and short-term mortality displacement during the COVID-19 pandemic

SummaryBackgroundWe conducted a population-based cohort study to estimate mortality before, during and after the COVID-19 peak and to compare mortality in 2020 with rates reported in previous years, with a view to helping decision makers to apply containment measures for high-risk groups.MethodsAll deaths were collected between 2015 and 2020 from municipal registry database. In 2020, weeks 1-26 were stratified in three periods: before, during and after the COVID mortality peak. The Poisson Generalized Linear regression Model showed the “harvesting effect”. Three logistic regressions for 8 dependent variables (age and comorbidities) and a t-test of differences described all-cause mortality risk factors in 2019 and 2020 and differences between COVID and non-COVID patients.ResultsA total of 47,876 deaths were collected. All-cause deaths increased by 38.5% during the COVID peak and decreased by 18% during the post-peak period in comparison with the average registered during the control period (2015-19), with significant mortality displacement in 2020. Except for chronic renal injuries in subjects aged 45-64 years, diabetes and chronic cardiovascular diseases in those aged 65-84 years, and neuropathies in those aged > 84 years, the weight of comorbidities in deaths was similar or lower in COVID subjects than in non-COVID subjects.DiscussionsSurprisingly, the weight of comorbidities in death, compared to weight in non-COVID subjects allows you to highlight some surprising results such as COPD, IBD and Cancer. The excess mortality that we observed in the entire period were modest in comparison with initial estimates during the peak, owing to the mild influenza season and the harvesting effect starting from the second half of May.

The association between air pollution parameters and emergency department visits and hospitalizations due to cardiovascular and respiratory diseases: a time-series analysis.

In this study, the association between cardiovascular and respiratory Emergency Department (ED) visits and hospitalizations and Sulfur Dioxide (SO2) and particulate matter with an aerodynamic diameter ≤ 10µm (PM10) was determined. The records of a total of 632,223 people diagnosed with cardiovascular diseases (ICD-10 code I00-I99) and respiratory diseases (ICD-10 code J00-J99) at and over the age of 18 between 2012 and 2018 in three hospitals in Karaman in Turkey were examined in the present study. The daily 24-h averages for SO2 and PM10 concentrations were acquired in National Air Quality Monitoring. A time-series analysis with Poisson Generalized Linear Model was used. Among the air pollution parameters, the mean ± standard deviation of SO2 and PM10 were 16.2 ± 22.1 and 75.8 ± 48.2μg/m3, respectively. An increase of 10μg/m3 in SO2 was associated with a 3% [Relative Risk (RR), 95% Confidence Interval (CI): 2-4)] increase in cardiovascular ED visit and 2% (RR, 95% CI 2-3) increase in respiratory ED visit, and 1% (RR, 95% CI 1-3) increase in respiratory hospitalizations. An increase of 10μg/m3 in PM10 was associated with a 1% (RR, 95% CI 1-2) increase in cardiovascular hospitalizations. Increases in SO2 are a risk factor for cardiovascular and respiratory ED visits, and respiratory hospitalizations. Increases in PM10 are a risk factor for cardiovascular hospitalizations.

Drug repurposing of dextromethorphan as a cellular target for the management of influenza.

Influenza viruses are responsible for seasonal epidemics and sporadic pandemics of varying severity in humans, and additional treatment options are needed. High-throughput siRNA screens and a pre-clinical research model demonstrated that dextromethorphan (DM) has anti-viral activity as a cellular target for treatment of influenza. This study examined DM usage and hospitalization rates among patients with laboratory-confirmed influenza in a national cohort of United States veterans. We aimed to evaluate the potential drug repurposing of DM as a cellular target for the management of influenza utilizing a large, national claims and electronic health record database. This retrospective drug-disease association cohort study was conducted using data from the Veterans Affairs Informatics and Computing Infrastructure (VINCI). We used a cohort with laboratory-confirmed diagnosis of influenza and international classification of disease (ICD)-9/10 diagnosis codes of fever, cough, influenza, or acute upper respiratory infection in an outpatient setting. The study outcome is inpatient hospitalization (all-cause and respiratory) within 30days of influenza diagnosis. We estimated the relative risk for all-cause and respiratory hospitalizations using Poisson generalized linear model (GLM) and a greedy nearest neighbor propensity score 1:1 matched sub-analysis for both hospitalization models. A total of 18,677 patients met the inclusion and exclusion criteria and were evaluated in our study. The cohorts consisted of 2801 patients dispensed DM and 15,876 untreated patients (no DM). The Poisson GLM adjusted for covariates demonstrated a relative risk reduction of 34% for all-cause hospitalizations (Relative Risk (RR) 0.66, 95% Confidence Interval (CI) 0.525-0.832) and 40% for respiratory hospitalizations (RR 0.597, 95% CI 0.423-0.843) in patients with influenza treated with DM. Influenza viruses continue to emerge and cause infection (including pandemics) in humans, so there remains a critical need to advance the understanding of influenza treatment. Our results demonstrated reduced hospitalization rates for influenza patients treated with DM. Further research on cellular targets and/or DM is warranted for the treatment of influenza.

Caregiver Social Status and Health-Related Quality of Life in Neurologically Impaired Children on Home Enteral Nutrition.

We aimed to investigate the association between caregiver social status and health-related quality of life (HRQoL) in children with neurological impairment (NI) on home enteral nutrition (HEN). This was an ancillary study of a multicenter, cross-sectional study which explored HRQoL in 75 NI children on HEN. All the caregivers from the original cohort were contacted, and data on education level, occupation and marital status were collected. The association between social status and HRQoL was investigated using a multiple Poisson Generalized Linear Model. In total, 93 caregivers were included, responsible for the care of 71 children. The caregivers of four children of the original cohort did not answer the questionnaire. Mothers with high-level education presented lower HRQoL in comparison to mothers with low-level (β: −5.97; 95% CI −11.51, −0.10; p = 0.027) or medium-level education (β: 4.85; 95% CI −9.87, 0.53; p = 0.044). The analysis of the subgroup of cases in which the main caregiver was represented by both parents gave similar findings, with education level of the father being negatively correlated with HRQoL. Our data showed that higher education level may negatively affect quality of life of caregivers of NI children. This could be helpful in identifying at-risk families and addressing supportive efforts.

Distribution patterns of epiphytic reed-associated macroinvertebrate communities across European shallow lakes

So far, research on plant-associated macroinvertebrates, even if conducted on a large number of water bodies, has mostly focused on a relatively small area, permitting limited conclusions to be drawn regarding potentially broader geographic effects, including climate. Some recent studies have shown that the composition of epiphytic communities may differ considerably among climatic zones. To assess this phenomenon, we studied macroinvertebrates associated with the common reed Phragmites australis (Cav.) Trin. ex Steud in 46 shallow lakes using a common protocol. The lakes, located in nine countries, covered almost the entire European latitudinal range (from <48°N to 61°N) and captured much of the variability in lake size and nutrient content in the region. A Poisson Generalized Linear Mixed Model (GLMM) showed the number of macroinvertebrate epiphytic taxa to be negatively associated with water conductivity and positively associated with medium ice cover duration (approximately 1 month). A Gamma GLMM showed a positive effect of chlorophyll a on the density of macroinvertebrates, and a significantly greater density in lakes located at the lowest and highest latitudes. Individual taxa responded differently to lake environmental conditions across climate zones. Chironomidae dominated in all climate zones, but their contribution to total density decreased with increasing latitude, with progressively greater proportions of Naidinae, Asellidae, Ephemeroptera and Trichoptera. Our study demonstrates that epiphytic macroinvertebrate fauna, even when analyzed at low taxonomic resolution, exhibits clear differences in diversity, relative abundance of individual taxa and total density, shaped both by geographic and anthropogenic variables. The results were discussed in the context of climate change. To our best knowledge this is the first study to examine epiphytic fauna carried out on a European scale.

S3090 Colon Polyps in Cirrhotics: A Community Practice Experience

INTRODUCTION: Colorectal cancer (CRC) is the third most common cancer and leading cause of cancer death in men and women in the US. Premalignant polyps are common, occurring in 25-50% of all people of screening age, and about 5% of the population will develop CRC in their lifetime. In patients with cirrhosis, colonoscopy is performed routinely for CRC screening and for a variety of indications related to GI complications of cirrhosis as well as to assess eligibility for liver transplantation. We aimed to analyze our experience with colonoscopy findings in patients with liver cirrhosis in a community setting. Methods: All colonoscopies performed in patients with a diagnosis of liver cirrhosis at South Denver Gastroenterology, P.C. in between 2006 and 2019 were retrospectively analyzed. First, a Poisson Generalized Linear Mixed Model (GLMM) was used to analyze the effect size of gender and age. Then, the effect of multiple colonoscopies was evaluated in a sub-cohort by including it as an effect in the model along with sex and age. Last, a binary GLMM was used to detect associations linking colonoscopy indications to findings. Results: Our study included 351 outpatient elective colonoscopies performed in patients with liver cirrhosis for the following indications: colon cancer screening (n = 147), prior personal or family history of colon polyps or cancer (n = 109), diarrhea (n = 31), constipation (n = 8), anemia (n = 31), GI bleed (n = 28) and abdominal pain (n = 12). Mean age of the group was 62.3 +/- 9.4 years; 158 patients were female and 193 were male. Adenomatous polyps were found in 159 procedures (49.07%, Table 1) and were 4 times more likely to be in the right colon (76.73%) than the left colon (18.87%). Left-sided adenoma occurrence was 61% lower in women than men (P = 0.042). Neither indication for the procedure (P = 0.08) nor advancing age (P = 0.94) impacted ADR. In patients with multiple colonoscopies (n = 49), there was no increase in number of adenomas found in subsequent exams (P = 0.11) on the contrary, significantly less right sided adenomatous polyps were detected (P = 0.025). Conclusion: Adenoma detection rate in elective colonoscopies performed in a community setting in patients with liver cirrhosis is greater than that in the general population, regardless of indication, gender and age. Our ADR (49%) was greater than that usually reported in the literature (23-30%). Male patients were more likely to have left sided colon adenomas than females, the impact of cirrhosis etiology and hormonal milieu is raised.Figure 1.: Frequency and number of adenomas found in colonoscopies in patients with liver cirrhosis.Table 1.: Frequency of adenomas found in patients with liver cirrhosis according to age

Compound Drought and Heatwaves at a Global Scale: The Role of Natural Climate Variability‐Associated Synoptic Patterns and Land‐Surface Energy Budget Anomalies

AbstractCompound drought and heatwave (CDHW) events have garnered much attention in recent studies. However, thus far, the identification of such events is oversimplified, and their association with natural climate variability is not fully explored. Here, we derive anomalies in the weekly self‐calibrated Palmer Drought Severity Index (sc_PDSI) and daily maximum temperatures to identify CDHW events from 1982 to 2016 over 26 climate regions across the globe. Using a Poisson Generalized Linear Model (GLM), we analyze yearly occurrences of seasonal CDHW events and their association with the warm and cold phases of El Nino Southern Oscillation (ENSO), Pacific Decadal Oscillation (PDO), and North Atlantic Oscillation (NAO). ENSO exhibits robust association with CDHW events over the Southern Hemisphere during the austral summer and fall, while PDO influences their occurrences over the Western North America in the Northern Hemisphere during the boreal summer, which is supported by the composites of anomalies in the atmospheric circulations and surface energy budget. However, NAO association with CDHW events is relatively weak. The CDHW occurrence over other regions is driven by a combination of these large‐scale natural forcing. Our analyses also highlight that the cooccurrence of weekly to submonthly scale anomalies in the observed temperature and precipitation may not be always aligned between the observations and the reanalysis. Therefore, caution must be exercised while explaining such observed anomalies on the basis of reanalysis‐based circulations and surface energy budget. Overall, our analyses provide a new insight towards concurrent extremes and should help foster research efforts in this area.

Interactions between climate factors and air quality index for improved childhood asthma self-management

BackgroundDaily air quality index (AQI) forecast can provide early warning information, and it is not clear whether it is appropriate for childhood asthma hospitalizations (CAHs). Furthermore, little is known about the effects of AQI on CAHs, as well as the interactions between temperature, humidity and AQI. MethodsWe collected 32,238 cases in Hefei from 2013 to 2016 and estimated the association between daily CAHs and AQI by combining the Poisson Generalized Linear Models (PGLMs) with the Distributed Lag Nonlinear Models (DLNMs). The interaction between AQI and temperature was tested by stratifying AQI and temperature, as well as humidity. ResultsAQI was associated with an increased risk of hospitalizations for childhood asthma. The adverse effect first appeared on the 3rd day, with the RR of 1.011 (95%CI: 1.000–1.023) and continued until the 19th day of lag (RR = 1.010, 95%CI: 1.001–1.020). In the subgroup analysis, the male and pre-school children were more sensitive to AQI, and there are seasonal differences in the effects of AQI on CAHs. Besides, in a stratified analysis with an AQI of 150, we found synergies between temperature, humidity and AQI. The interaction relative risk (IRR) and relative excess risk due to interaction (RERI) for the interaction between temperature and AQI were 1.157 (95%CI: 1.029–1.306) and 0.122 (95%CI: 0.022–0.223) respectively. For the humidity, the IRR and RERI were 1.090 (95%CI: 1.056–1.206) and 0.083 (95%CI: 0.083–0.143) respectively. Exploring different subgroups in the interaction analyses, it was worth noting that female and pre-school children were more sensitive to the interaction between AQI and temperature, while school-age children were more sensitive to the interaction between AQI and humidity. ConclusionsThe study found that not only AQI can significantly increase the risk of CAHs, but also that under the context of climate change, temperature and humidity have a synergistic effect on AQI, suggesting that considering only the warning information of air pollution is not enough to strengthen the prevention of childhood asthma hospitalization.

Ambient temperature and non-accidental mortality: a time series study.

Under the background of global climate change, the present study aimed to evaluate the effects of daily mean temperature and diurnal temperature range (DTR) on the non-accidental mortality. Poisson generalized linear model (PGLM) combined with distributed lag non-linear model (DLNM) was used to evaluate these effects after adjusting the relative humidity and major air pollutants. All effects were presented as relative risk (RR), with 75th percentiles of daily mean temperature and DTR compare with their lowest RRs corresponding values. Daily mean temperature was associated with the non-accidental mortality with a U-shaped curve, and the non-accidental mortality increased by 1.8% (95% CI: 0.7~3.0%) when the temperature was 24.4°C (20°C as the reference). Additionally, the non-accidental mortality increased by 1.4% (95% CI: 0.1~2.7%) on lag6 day when DTR was 11.3°C (7°C as the reference). The elderly (≥ 65years) were more susceptible to daily mean temperature and DTR, and females were more susceptible to high DTR effect than males. Our study provides evidence that daily mean temperature and DTR are significantly associated with non-accidental mortality and have delayed effects. Both females and elderly people are vulnerable to the potential adverse effects.

Effects of logging on wildlife communities in certified tropical rainforests in East Kalimantan, Indonesia

Logging has been operated throughout Bornean tropical forests that harbor diverse fauna species. We examined the responses of ground-dwelling wildlife to logging intensity in two certified forest management units under the same climate and vegetation in East Kalimantan, Indonesia, in 2012–2016. A total of ten and 13 circular plots with a diameter of 1 km were established respectively, and 10–18 camera points were distributed randomly in each plot to record wildlife for two years. A total of 41 species of medium to large-sized terrestrial wildlife were recorded in the two management units. Canonical Correspondence Analysis (CCA) was performed to examine the response of wildlife community to the logging influences with the following three variables: forest intactness (as a surrogate of tree-species composition), above-ground carbon (as a surrogate of forest structure), and distance to the nearest village to represent hunting pressure. Forest intactness was the strongest factor affecting the assemblage of wildlife species in both management units, suggesting that changes in tree-species composition due to logging were a decisive factor influencing animal assemblage. Poisson Generalized Linear Model (GLM) was employed to examine the response of each species to forest intactness in each management unit. We found that eight out of the 32 tested wildlife species responded to forest intactness consistently between the two units, while the other species responded more individualistically. We suggest that consistent responses of several species, as well as individualistic responses of the other species to the changes of tree-species composition, determine the animal assemblage in logged-over Bornean rainforests. Sustainably managed logged-over forests in Borneo are a habitat for diverse wildlife community and therefore also require attention for biodiversity conservation.

Monitoring multinomial logistic profiles in Phase I using log-linear models

PurposeThe purpose of this paper is to apply Poisson generalized linear model (PGLM) with log link instead of multinomial logistic regression to monitor multinomial logistic profiles in Phase I. Hence, estimating the coefficients becomes easier and more accurate.Design/methodology/approachSimulation technique is used to assess the performance of the proposed algorithm using four different control charts for monitoring.FindingsThe proposed algorithm is faster and more accurate than the previous algorithms. Simulation results also indicate that the likelihood ratio test method is able to detect out-of-control parameters more efficiently.Originality/valueThe PGLM with log link has not been used to monitor multinomial profiles in Phase I.