Simulation on the Zero Inflated Negative Binomial (ZINB) to Model Overdispersed, Poisson Distributed Data

This study aims at evaluating the performance of Zero Inflated Negative Binomial (ZINB) regression analysis using the Maximum Likelihood Estimation (MLE) approach through simulation study. The research data used are secondary data and simulations. Secondary data was obtained from the Ministry of Health of the Republic of Indonesia in 2023 regarding cases of under-five deaths due to pneumonia with a total of 38 samples. The simulation study is conducted to analyze the performance of ZINB regression based on various sample sizes and proportions of zero values. The results show that the ZINB regression model with the MLE approach produces parameter estimates that tend to be more sensitive to sample size, with improved performance at large sample sizes. The ZINB regression model gives results for a large proportion of zeros and satisfies the excess zero condition. Data with a large proportion of zeros reflects high variability as well as the presence of excess zeros, so the ZINB regression model can provide more stable and precise parameter estimates than those with a lower proportion of zeros. Therefore, the ZINB regression model is effective for data with a high proportion of zeros as it fits the characteristics of the data distribution.

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.

Falls among older adults are sudden events that can lead to serious injuries. Particularly for older adults living alone, the lack of social relationships increases the risk of falls. This study aimed to identify the risk factors for falls among older adults living alone, focusing on general, health-related, and social network characteristics. This cross-sectional study utilised data from the 2020 National Survey of Older Koreans, including 3,109 community-dwelling older adults aged 65 years and older who lived alone. A Zero-inflated Negative Binomial Regression analysis was conducted to identify the risk factors for the number of falls using a count model, whereas factors associated with the likelihood of experiencing zero falls were analysed using a logit model. Among the participants, 268 (8.6%) had experienced falls in the past year. In the logit model, participants without family but with acquaintances, as well as those without both family and acquaintances, had a higher risk of falls than those with both family and acquaintances in their social network. According to the count model, the risk factors for the number of falls included low education level, being underweight (body mass index < 18.5), poor perceived health status, and higher depression symptoms. This study found that social network characteristics played a role in protecting older adults living alone from experiencing falls, whereas general and health-related characteristics such as low education level, poor health, and depressive symptoms were associated with a greater number of falls. Healthcare providers should consider not only general and health-related factors but also social network characteristics when assessing the fall risk in this population.

Predictors of Likelihood and Intensity of Past-Year Mental Health Service Use in an Active Canadian Military Sample

BackgroundRespiratory infections continue to be a public health threat, particularly to young children in developing countries. Understanding the geographic patterns of diseases and the role of potential risk factors can help improve future mitigation efforts. Toward this goal, this paper applies a spatial scan statistic combined with a zero-inflated negative-binomial regression to re-examine the impacts of a community-based treatment program on the geographic patterns of acute lower respiratory infection (ALRI) mortality in an area of rural Bangladesh. Exposure to arsenic-contaminated drinking water is also a serious threat to the health of children in this area, and the variation in exposure to arsenic must be considered when evaluating the health interventions.MethodsALRI mortality data were obtained for children under 2 years old from 1989 to 1996 in the Matlab Health and Demographic Surveillance System. This study period covers the years immediately following the implementation of an ALRI control program. A zero-inflated negative binomial (ZINB) regression model was first used to simultaneously estimate mortality rates and the likelihood of no deaths in groups of related households while controlling for socioeconomic status, potential arsenic exposure, and access to care. Next a spatial scan statistic was used to assess the location and magnitude of clusters of ALRI mortality. The ZINB model was used to adjust the scan statistic for multiple social and environmental risk factors.ResultsThe results of the ZINB models and spatial scan statistic suggest that the ALRI control program was successful in reducing child mortality in the study area. Exposure to arsenic-contaminated drinking water was not associated with increased mortality. Higher socioeconomic status also significantly reduced mortality rates, even among households who were in the treatment program area.ConclusionCommunity-based ALRI interventions can be effective at reducing child mortality, though socioeconomic factors may continue to influence mortality patterns. The combination of spatial and non-spatial methods used in this paper has not been applied previously in the literature, and this study demonstrates the importance of such approaches for evaluating and improving public health intervention programs.

Excess weight can increase absenteeism of workers and can have a negative influence on their productivity. Current evidence on this association is mostly based on cross-sectional data and there is little evidence concerning the longitudinal relationship between obesity, and disability with workplace absenteeism. Further, gender differences in this association have often ignored in the existing literature. This study aims to examine gender differences in the longitudinal association between obesity, and disability with absenteeism in the workplace. Data from thirteen waves (2006 to 2018) of the Household, Income and Labour Dynamics in Australia (HILDA) survey were pooled, resulting in 117,769 observations for 19,851 adult employees. The Zero-Inflated Negative Binomial (ZINB) regression model was deployed to investigate the links between obesity, and disability with workplace absenteeism for the total sample and stratified by gender. The findings showed that overweight (Incidence Rate Ratio [IRR]: 1.23, 95% confidence interval [CI]: 1.02-1.47), obesity (IRR: 1.35, 95% CI: 1.12-1.64) and disability (IRR: 2.83, 95% CI: 2.36-3.38) were associated with prolonged workplace absenteeism irrespective of gender. This study found that the multiplicative interaction between weight status and gender is significantly associated with absenteeism. The results reveal that the rate of absenteeism was 2.79 times (IRR: 2.79, 95% CI: 1.96-3.97) and 1.73 times (IRR: 1.73, 95% CI: 1.20-2.48) higher among overweight and obese women than male counterparts, respectively. Moreover, this study found that the weight status of male workers is not associated with absenteeism. However, disability (IRR: 3.14, 95% CI: 2.43-4.05) is positively associated with longer days of absence among male workers. Finally, the study results showed that the rate of absenteeism is 1.82 (IRR: 1.82, 95% CI: 1.36-2.44), 1.61 (IRR: 1.61, 95% CI: 1.21-2.13), and 2.63 (IRR: 2.63, 95% CI: 1.99-3.48) times higher among overweight, obese, and female workers with a disability, respectively, compared with their lower weight counterparts. Workplace absenteeism is significantly associated with overweight and obesity among Australian workers. An active workplace health promotion program is very important for weight management of overweight and obese workers and thus to reduce workplace absenteeism. For example, employers may provide incentives for maintaining recommended body weights, encourage exercise, and promote healthy diets amongst their workers.

Background: The desire for universal health coverage in developing countries has brought attention to communitybased health insurance (CBHI) schemes in developing countries. The government of Uganda is currently debating policy for the national health insurance programme, targeting the integration of existing CBHI schemes into a larger national risk pool. However, while enrolment has been largely studied in other countries, it remains a generally under-covered issue from a Ugandan perspective. Using a large CBHI scheme, this study, therefore, aims at shedding more light on the determinants of households’ decisions to enrol and renew membership in these schemes. Methods: We collected household data from 464 households in 14 villages served by a large CBHI scheme in southwestern Uganda. We then estimated logistic and zero-inflated negative binomial (ZINB) regressions to understand the determinants of enrolment and renewing membership in CBHI, respectively. Results: Results revealed that household’s socioeconomic status, husband’s employment in rural casual work (odds ratio [OR]: 2.581, CI: 1.104-6.032) and knowledge of health insurance premiums (OR: 17.072, CI: 7.027-41.477) were significant predictors of enrolment. Social capital and connectivity, assessed by the number of voluntary groups a household belonged to, was also positively associated with CBHI participation (OR: 5.664, CI: 2.927-10.963). More positive perceptions on insurance (OR: 2.991, CI: 1.273-7.029), access to information were also associated with enrolment and renewing among others. Burial group size and number of burial groups in a village, were all significantly associated with increased the likelihood of renewing CBHI. Conclusion: While socioeconomic factors remain important predictors of participation in insurance, mechanisms to promote inclusion should be devised. Improving the participation of communities can enhance trust in insurance and eventual coverage. Moreover, for households already insured, access to correct information and strengthening their social network information pathways enhances their chances of renewing.

Modeling stable and dynamic vulnerabilities in suicide risk: A mechanistic test of fluid vulnerability theory in military personnel with suicidal ideation.

Introduction: Catheter ablation is used increasingly as a 1st-line (1L) alternative to antiarrhythmic drugs (AAD) including dronedarone in patients with atrial fibrillation (AF). Healthcare resource utilization (HCRU)-related costs with dronedarone vs 1L ablation are unknown. Methods: We conducted a retrospective, observational cohort study using Optum Clinformatics Data Mart® from Jan 2012-Jan 2022 among US adults with AF and no prior rhythm control therapy who received new AAD therapy with dronedarone (index: date of incident dronedarone fill) vs those who received 1L ablation, or a non-dronedarone AAD followed by ablation within 3 mo (index: date of first recorded ablation or AAD). Patients were required to have ≥24 mo of pre-index data, and ≥3 mo follow-up. Patients in the 1L ablation cohort were propensity score matched 2:1 to the dronedarone cohort. Mean payer costs per patient per month (PPPM) during the 24-mo post-index period were calculated for total HCRU, inpatient visits, emergency room (ER) visits, outpatient physicians’ office visits, and all outpatient visits, and compared by zero-inflated negative binomial (ZINB) regression model. Results: Post-matching, the dronedarone (n=1440) and 1L ablation (n=2253) cohorts had similar baseline characteristics (mean age at index: 68.4 vs 67.7 years; male: 57.7% vs 59.3%; mean time from AF diagnosis to index: 80.2 vs 91.6 days; Charlson comorbidity index: 2.3 vs 2.2; CHADS2-VASc score: 3.4 vs 3.2). Mean PPPM costs were lower with dronedarone vs 1L ablation for all-cause total HCRU, inpatient visits, any outpatient visit, ER visit, and outpatient office visit (Table). ZINB analyses showed significant cost differences for all-cause total HCRU and any outpatient visit events. Conclusion: In patients with AF and no prior rhythm control therapy, 1L dronedarone was associated with lower total HCRU and outpatient visit costs during 24-mo follow-up vs 1L ablation; future studies will assess cost-effectiveness.

652 Background: Geographic disparities of phase I-III cancer clinical trials in the US are well described (Gu et al., JMCP 2024), but specific patterns for BC remain unexplored. We evaluated BC trial availability across US counties, interrogating associations with epidemiologic and socioeconomic factors. Methods: We queried ClinicalTrials.gov for adult, interventional BC trials from June 1, 2019 to June 1, 2025, and mapped the US site ZIP codes to counties using US Postal Service data. County-level age-adjusted incidence and mortality, and Social Vulnerability Index (SVI) were obtained from the CDC, NCI, and Agency for Toxic Substances and Disease Registry. The SVI reflects 16 socioeconomic and demographic variables. Trial characteristics were summarized descriptively. To assess geographic factors linked to trial availability, we modeled the number of trials available per county using a zero-inflated negative binomial (ZINB) regression—a model that accounts for the many counties without trials and estimates how county characteristics relate to the trial rate. The model was adjusted for county size, effectively analyzing the rate of trials per household. Results are expressed as incidence rate ratios (IRRs), where IRR &gt; 1 indicates a higher trial rate as the predictor increases. Analyses used R 4.4.3. Results: We identified 436 eligible trials with active sites across 713 US counties; 77.3% of counties had no trials. Trials were mostly sponsored by pharmaceutical companies (48.4%) followed by academia (39%). Most trials focused on drug therapy (80.2%) over surgery (6.4%) or radiation (4.6%); and most were early phase (I/II, 72%). Using the ZINB model, higher BC incidence was associated with slightly higher trial rate per household (IRR 1.033, 95% CI 1.003–1.063). Notably, higher BC mortality was associated with fewer trials (IRR 0.80, 95% CI 0.73–0.88). Compared with high-SVI (most socially vulnerable) counties, less vulnerable counties had higher trial rates: lower-middle SVI (IRR 1.57, 95% CI 1.18–2.08), middle-high SVI (IRR 1.60, 95% CI 1.22–2.11), and low-SVI counties (IRR 2.53, 95% CI 1.89–3.38). Of 7,091 trial sites in the US, 657 (9.3%) were in counties with the highest quintile (Q5) for BC incidence. However, only 225 (3.2%) sites were in Q5 counties for BC mortality. Metastatic and non-metastatic trials were similarly represented across mortality quintiles. While most trials were early phase overall, high mortality counties had proportionally less early phase trial sites (56.4% Q5 vs. 63.8% Q1). Conclusions: Most US counties lacked a BC trial. Trial availability was higher where disease is diagnosed, not where patients most often die from it. Research access is concentrated in counties with high BC incidence and low social vulnerability. Expanding trial sites to high mortality or high-SVI counties may be an actionable step for investigators to improve equitable trial access nationwide.

Background: It may take a long time to diagnose multiple sclerosis (MS) since the emergence of primary symptoms. This study aimed to use count regression models to compare their fit and to identify factors affecting delay in the diagnosis of MS. Methods: Data were collected from the Nationwide MS Registry of Iran (NMSRI) for Mazandaran Province, Iran, using census sampling until April 2022. The four models of Poisson regression, negative binomial (NB) regression, zero-inflated Poisson (ZIP) regression, and zero-inflated negative binomial (ZINB) regression were used in this study. Results: In this study on 2894 patients, 74.0% were women, and 8.5% had a family history of MS. The mean ± standard deviation (SD) of the patients' age was 34.96 ± 9.41 years, and the mean delay in diagnosis was 12.32 ± 33.26 months, with a median of 0 (Q1-Q3: 0-9). The NB regression model showed the best performance, and factors, including a history of hospitalization and the year of symptom onset, had significant effects on a delayed diagnosis. Besides, the Expanded Disability Status Scale (EDSS) score was significantly different before and after 2017; it was also associated with sex, type of MS, and history of hospitalization. Conclusion: The mean diagnostic delay and the mean age of MS diagnosis are critical in Mazandaran Province. Patients with MS develop the disease at an early age and are diagnosed with a long delay. The time of symptom onset is a significant factor in the diagnosis of MS, and in recent years, there have been improvements in the diagnostic process.

Water theft (WT) jeopardizes equitable water-use rights and potentially threatens the effectiveness and sustainability of irrigation water management. This paper examines whether irrigation-member network size, kinship network size and connections with linking networks affect WT probability and frequency in the local reality characterized by strict water quantity management and weak formal sanctions but conditionally strong informal sanctions. Our empirical analysis applies a zero-inflated negative binomial (ZINB) regression model and survey data involving 739 households in Awati County, Xinjiang, China. We find that the increasing irrigation member network size decreases WT probability but increases WT frequency, the increasing kinship network size increases WT probability but insignificantly affects WT frequency, and the connection with linking networks insignificantly affects WT probability but increases WT frequency. Consequently, these social network variables contribute little to the local execution of irrigation water management. Our findings’ policy implications and academic contributions are discussed in detail.

Background. Women with lower-extremity arterial disease (LEAD) are often underdiagnosed, present themselves with more advanced disease at diagnosis, and fare worse than men. Objective. To investigate to what extent potential gender differences exist in the frequency and reasons for general practitioner (GP) consultation six months prior to the diagnosis of LEAD, as potential indicators of diagnostic delay. Methods. Individuals older than 18 years diagnosed with LEAD, sampled from the Julius General Practitioner’s Network (JGPN), were included and compared with a reference population, matched (1:2.6 ratio) in terms of age, sex, and general practice. We applied a zero-inflated negative binomial (ZINB) regression model. Results. The study population comprised 4044 patients with LEAD (43.5% women) and 10,486 subjects in the reference population (46.3% women). In the LEAD cohort, the number of GP contacts was 2.70 (95% CI: 2.42, 3.02) in women and 2.54 (2.29, 2.82) in men. In the reference cohort, 1.77 (95% CI: 1.62, 1.94) in women and 1.63 (95% CI: 1.50, 1.78) in men. In the LEAD cohort, 21.9% of GP contacts occurred one month prior to diagnosis. In both cohorts and both sexes, the most common cause of consultation during the last month before the index date was cardiovascular problems. Conclusions. Six months preceding the initial diagnosis of LEAD, patients visit the GP more often than a similar population without LEAD, regardless of gender. Reported gender differences in the severity of LEAD at diagnosis do not seem to be explained by a delay in presentation to the GP.

Identification of defect-prone classes in telecommunication software systems using design metrics

Purpose Social media have escalated the frequency and intensity of brands’ online controversial events (OCEs), which differs conceptually from negative events. Despite this, there remains a scarcity of research exploring the nature of OCEs. This paper aims to investigate the impact of positive buzz on consumer engagement during OCEs. Design/methodology/approach Using 47,468 posts from two popular Chinese social media (i.e. Weibo and Zhihu), we employ a zero-inflated negative binomial regression and content analysis to test our hypotheses. Findings The results indicate that positive buzz informativeness and sentiment positively affect consumer engagement in online brand communities, moderated by community type and time-related factors. Expert communities (vs social communities) weaken main effects, while date distance strengthens them. Originality/value This study is the first to propose the nature of brand’s OCEs and explores how positive buzz affects consumer engagement, highlighting the moderating roles of community type and date distance. This paper contributes to literature on user-generated content (UGC), OCEs and dual process theory, offering valuable insights for brands, consumers and community owners.