Comparison of Different Count Models for Investigation of Some Environmental Factors Affecting Stillbirth in Holsteins

Background: The objective of this study is comparing different count data models for stillbirth data. In modeling this type of data, Poisson regression or alternative models can be preferred. Methods: The poisson, negative binomial, zero-inflated poisson, zero-inflated negative binomial, poisson-logit hurdle and negative binomial-logit hurdle regressions were compared and used to examine the effects of the gender, parity and herd-year-season independent variables on stillbirth. Furthermore, the Log-Likelihood statistics, Akaike Information Criteria, Bayesian Information Criteria and rootogram graphs were used as comparison criteria for performance of the models. According to these criteria, Negative Binomial-Logit Hurdle Regression model was chosen as the best model. Result: The parameter estimates obtained by Negative Binomial-Logit Hurdle Regression model in relation to the effects of the gender, parity and herd-year-season independent variables on stillbirth were found to be significant (p less than 0.01). It was found that while stillbirth incidence was higher in males than females, it was found to decrease as the parity increased. As a result, the Negative Binomial Logit Hurdle model was found the best model for stillbirth count data with overdispersion.

Comparison of Zero Inflated Poisson (ZIP) Regression, Zero Inflated Negative Binomial Regression (ZINB) and Binomial Negative Hurdle Regression (HNB) to Model Daily Cigarette Consumption Data for Adult Population in Indonesia

Smoking is a habit that is not good for health. Smoking habits are generally practiced by adults but it is possible for teenagers to do so.The Report of Southeast Asia Tobacco Control Alliance (SEATCA) entitled The Tobacco Control Atlas, ASEAN Region shows that Indonesia is the country with the highest number of smokers in ASEAN, namely 65.19 million people. This figure is equivalent to 34 percent of the total population of Indonesia in 2016. Based on these data, the authors are interested in modeling the daily cigarette consumption data for adults in Indonesia obtained from the 2015 Indonesia Family Life Survey. The variables used include the variable amount of cigarette consumption, education, level of welfare and income per month. The author wants to compare the best model that can be used to model the daily cigarette consumption of adults in Indonesia. The models being compared are Zero Inflated Poisson Regression (ZIP), Zero Inflated Negative Binomial Regression (ZINB) and Binomial Negative Hurdle Regression (HNB). The comparison results of the three models obtained that the best model is the Zero Inflated Negative Binomial (ZINB) Regression model because it has the smallest Akaike's Information Criterion (AIC) value.

Distributed Learning from Multi-Site Observational Health Data for Zero-Inflated Count Outcomes

BackgroundMulti-site studies facilitate the study of rare outcomes or exposures through integrating patient information from several distinct care sites. Due to patient privacy concerns, sharing of patient-level information among collaborating sites is often prohibited, suggesting a need for privacy-preserving data analysis methods. Several such methods exist, but have been shown to sometimes result in biased estimation or require extensive communication among sites.ObjectiveWe present a communication-efficient, privacy-preserving method for performing distributed regression on Electronic Health Records (EHR) data across multiple sites for zero-inflated count outcomes. Our approach is motivated by two real-world data problems: modeling frequency of serious adverse events and examining risk factors associated with pediatric avoidable hospitalization.MethodsWe use hurdle regression, a two-part (logistic-Poisson) regression model, to characterize the effects of risk factors on zero-inflated count outcomes. Further, we develop a one-shot algorithm for performing hurdle regression (ODAH) across multiple sites, using individual patient data at one site and aggregated data from all other sites to approximate the complete data log likelihood. We evaluate ODAH through extensive simulations and an application to EHR data from the Children’s Hospital of Philadelphia (CHOP) and the OneFlorida Clinical Research Consortium. We compare ODAH estimates to those from meta-analysis and pooled analysis (the gold standard in which all patient data are pooled together).ResultsIn simulations, ODAH estimates exhibited bias relative to the gold standard of less than 0.1% across several settings. In contrast, meta-analysis estimated exhibited relative bias up to 12.7%, largely dependent on the event rate. When applying ODAH to CHOP data, relative biases for estimates in both components of the hurdle model were less than 5.1%, while meta-analysis estimates exhibited relative bias as high as 63.6%. When analyzing OneFlorida data, ODAH relative biases were less than 10% for eight of the ten log relative risks estimated, while meta-analysis estimates again showed substantially greater bias.ConclusionsOur simulations and real-world applications suggest ODAH is a promising method for performing privacy-preserving distributed learning on EHR data when modeling zero-inflated count outcomes.

Adoption of ICT-based Information Sources and Market Participation among Smallholder Livestock Farmers in South Africa

The study explored the contribution of ICT-based information sources to market participation among smallholder livestock farmers. Use of ICTs is considered paramount for providing smallholder farmers with required market information, in order to reduce market asymmetries. A Double Hurdle regression was utilized to analyze data collected from 150 smallholder livestock farmers in the study area. The results show that while use of ICT-based market information sources significantly influenced market participation, the effect of using ICT-based information sources on intensity of market participation was not significant. Other variables shown to influence both market participation and the intensity of market participation were age, additional income and membership of farmer cooperatives. This suggests the need to also consider other associated factors in the application of interventions which utilize ICT-based information sources in achieving planned market interventions.

Contextual Influences on the Sentencing of Individuals Convicted of Sexual Crimes

This study examines effects of court and community contextual factors on sentencing outcomes for individuals convicted of sexual crimes using indicators from two perspectives—focal concerns and populist punitiveness. Sourced from the Pennsylvania Commission on Sentencing, the sample includes 9,431 persons convicted of sexual crimes and a precision-matched sample of persons convicted of non-sexual violent crimes for comparison. Based on multilevel hurdle regression models for both incarceration and sentence length decisions, results indicate that individuals convicted of sexual crimes face enhanced sentence severity in judicial districts with smaller courts, increased jail capacity, stronger political competition, and higher religious homogeneity. The results also suggest statistically significant differences between effects for persons convicted of sexual crimes and a matched sample of persons convicted of violent crimes. Overall, results suggest that specific contextual factors have a distinguishable impact on sentencing of individuals convicted of sexual crimes.

World Pneumonia Day 2011–2016: Twitter contents and retweets

Background Twitter is used for World Pneumonia Day (WPD; November 12) communication. We evaluate if themes of #pneumonia tweets were associated with retweet frequency. Methods A total of 28 181 original #pneumonia tweets were retrieved (21 November 2016), from which six subcorpora, 1 mo before and 1 mo after WPD 2011–2016, were extracted (n=6721). Underlying topics were identified via latent Dirichlet allocation and were manually coded into themes. The association of themes with retweet count was assessed via multivariable hurdle regression. Results Compared with personal experience tweets, tweets that both raised awareness and promoted intervention were 2.62 times as likely to be retweeted (adjusted odds ratio [aOR] 2.62 [95% 1.79 to 3.85]) and if retweeted had 37% more retweets (adjusted prevalence ratio [aPR] 1.37 [95% CI 1.06 to 1.78]). Tweets that raised concerns about vaccine price were twice as likely to be retweeted (aOR 2.29 [95% CI 1.36 to 3.84]) and if retweeted, had double the retweet count (aPR 2.05 [95% CI 1.27 to 3.29]) of tweets sharing personal experience. Conclusions The #pneumonia tweets that both raised awareness and promoted interventions and those discussing vaccine price were more likely to engage users than tweets about personal experience. These results help health professionals craft WPD messages that will engage the audience.

Relationship between Neighborhood Characteristics and Demand for Emergency Health Service Vehicles: A Poisson Hurdle Regression Modeling Approach

The main objective of this study is to explore the spatial and temporal variability of demand for emergency health service vehicles, measured at the 1 km-by-1 km grid level in Halifax, Nova Scotia, Canada. This study utilizes and compares a Poisson regression and Poisson hurdle regression model that examine the effects of neighborhood characteristics on emergency health service vehicle demand. It also develops a time-segmented model to investigate the temporal variability of the effects of factors considered in relation to demand. It analyzes the Nova Scotia emergency health service administrative database for the period from January 2012 to December 2012. A comprehensive set of socio-demographic attributes, land use characteristics, and measures of accessibility to services are used to achieve the objectives of this research. Results found that demand for emergency health service vehicles was higher in areas where there was higher population density, more heterogeneous land uses, and a larger proportion of the population aged 40 years and above. The time-segmented model shows demand was highest during morning peak periods for residential areas, afternoon peak for commercial areas, and morning peak and midday for members of the population aged over 75 years. An assessment of the time-segmented model suggests that the generic model is sufficient for predicting how neighborhood characteristics relate to the demand for emergency health service vehicles. The findings of this study will be beneficial for urban planners and health professionals in designing healthy cities and targeting health promotions to reduce the need for emergency services.