A marginalized two-part model with heterogeneous variance for semicontinuous data

The Effect of Left-Behind Women on Fertilizer Use: Evidence from China's Rural Households Engaging in Rural-Urban Migration

Rural-urban migration in developing countries has required left-behind women to assume the role of key decision makers and take increasing responsibility for agricultural production. However, little is known about the effect of rural-urban migration on fertilizer use when left-behind women assume key decision-maker roles. This study highlights the effect of left-behind women on fertilizer use, drawing on nationwide household survey data in China. The results indicate that households with recognized left-behind women heads use less fertilizer than those with recognized men heads, while households with de facto left-behind women heads use more fertilizer. Moreover, the average nexus between households with recognized left-behind women heads whose major agricultural income comes from grain crops and fertilize use is negative but small in size and statistically insignificant. The findings indicate that future policy efforts aimed at decreasing rural environmental degradation should place greater emphasis on left-behind women's empowerment in socioeconomic decision-making within and outside the household, thereby contributing to an environment in which left-behind women farmers can succeed in a sustainable way.

A semi-parametric Bayesian model for semi-continuous longitudinal data

Semi-continuous data present challenges in both model fitting and interpretation. Parametric distributions may be inappropriate for extreme long right tails of the data. Mean effects of covariates, susceptible to extreme values, may fail to capture relevant information for most of the sample. We propose a two-component semi-parametric Bayesian mixture model, with the discrete component captured by a probability mass (typically at zero) and the continuous component of the density modeled by a mixture of B-spline densities that can be flexibly fit to any data distribution. The model includes random effects of subjects to allow for application to longitudinal data. We specify prior distributions on parameters and perform model inference using a Markov chain Monte Carlo (MCMC) Gibbs-sampling algorithm programmed in R. Statistical inference can be made for multiple quantiles of the covariate effects simultaneously providing a comprehensive view. Various MCMC sampling techniques are used to facilitate convergence. We demonstrate the performance and the interpretability of the model via simulations and analyses on the National Consortium on Alcohol and Neurodevelopment in Adolescence study (NCANDA) data on alcohol binge drinking.

Financial Burden of Traumatic Injury Amongst the Privately Insured

OBJECTIVE

We sought to evaluate the overall financial burden associated with traumatic injury amongst patients with private insurance and assess the effect of high deductible plans on out-of-pocket costs (OOPCs).

SUMMARY OF BACKGROUND DATA

Traumatic injury can be a source of unexpected financial burden for households. However, the effect of increasing participation in higher cost-sharing private health insurance plans remains unknown.

METHODS

We conducted a retrospective cohort observational study, using the Clinformatics Data Mart Database, a nationwide single-payer administrative claims database to identify US adults who required emergency department services or hospital admission for single traumatic injury from 2008 to 2018. A 2-part model using a logistic regression and a generalized linear model with gamma distribution and log link was used to evaluate 12-month OOPCs after traumatic injury. Multivariable logistic regression was used to evaluate the likelihood of catastrophic health expenditure (CHE) after injury.

RESULTS

Of 426,945 included patients, 53% were male, 71% were white, and median age was 42 years. Patients faced monthly OOPC of $660 at the time of their injury. High deductible plan enrollment was associated with an increase of $1703 in 12-month OOPC after trauma, compared to those covered by traditional health plans. In addition to high deductible health plan enrollment, worsening injury severity and longer hospital stays were also associated with increased 12-month OOPC after trauma. Non-white minorities paid less 12-month OOPC after trauma compared to non-Hispanic white patients, but also used fewer services. Overall, the incidence of CHE was 5%; however high-deductible health plan enrollees faced a 13% chance of CHE.

CONCLUSIONS

Privately insured trauma patients face substantial OOPCs at the time of their injuries. High-deductible health plans are associated with increased financial vulnerability after trauma.

Fitting marginalized two-part models to semicontinuous survey data arising from complex samples

OBJECTIVE

To accurately model semicontinuous data from complex surveys, we extend marginalized two-part models to a design-based inferential framework and provide guidance on incorporating complex sample designs.

DATA SOURCES

2014 Medical Expenditure Panel Survey (MEPS).

STUDY DESIGN

We describe the use of pseudo-Maximum Likelihood Estimation and Jackknife Repeated Replication for estimating model parameters and sampling variance, respectively. We illustrate our approach using MEPS, modeling total healthcare expenditures in 2014 as a function of respondents' age and family income. We provide SAS and R code for implementing the extension, assessing model-fit indices, and evaluating the need to incorporate complex sampling features.

DATA EXTRACTION METHODS

Data obtained from www.meps.ahrq.gov.

PRINCIPLE FINDINGS

A 100 percentage-point increase in family income as a percent of the federal poverty level was associated with a 5%-6% increase in healthcare spending. People over 65 had an increase of 4-5 times compared to those younger. Accounting for complex sampling in the models led to different parameter estimates and wider confidence intervals than the unweighted models. Ignoring complex sampling could lead to inaccurate finite population inference.

CONCLUSION

Researchers should account for complex sampling features when analyzing semicontinuous data from surveys.