Dynamic analysis of complex panel count data

ObjectivesPrevious studies suggested protective associations of vegetables and fruits (VF) intake with cognitive function, but evidence on specific types of VF was insufficient. MethodsThe current study included 4066 participants from 1997 to 2006 in the China Health and Nutrition Survey (CHNS) and 6170 participants from 2013 to 2020 in the Health and Retirement Study (HRS). Dietary intake (using 3-day 24-h dietary recalls in CHNS and food frequency questionnaire in HRS) and cognitive function (using the Telephone Interview for Cognitive Status-Modified, TICS-m) were measured. Linear mixed-effects models were used to estimate the beta coefficients (β) and the 95% confidence intervals (CI) to evaluate the association of VF with cognitive function (z-score) and its decline. ResultsHighest intake of total VF was associated with better cognitive function and slower cognitive decline. Differences in cognitive function z-score between the highest and lowest tertiles of VF consumption were 0.039 (95% CI: 0.002, 0.076) for CHNS and 0.063 (95% CI: 0.026, 0.100) for HRS. The corresponding differences in annual cognitive decline were 0.011 (95% CI: 0.002, 0.021) and 0.012 (95% CI: 0.003, 0.020) units respectively. Vegetables and fruits showed independent associations with cognitive function and its decline. In specific VF subgroups, when comparing the highest to the lowest tertile intake, cruciferous vegetables (β = 0.058, 95% CI: 0.017, 0.100 in CHNS and β = 0.067, 95% CI: 0.032, 0.101 in HRS) and green leafy vegetables (β = 0.036, 95% CI: −0.001, 0.073 in CHNS and β = 0.082, 95% CI: 0.046, 0.117 in HRS) was associated with better cognitive function in both cohorts. Similarly, higher intake of dark-colored vegetables (β = 0.019, 95% CI: 0.008, 0.030 for red/yellow vegetables in CHNS and β = 0.004, 95% CI: 0.001, 0.007 for green leafy vegetables in HRS) were associated with slower cognitive decline in subsequent years. Moreover, rigorous sensitivity analyses confirmed the stability of the results. ConclusionsOur findings support the potential beneficial roles of VF, especially cruciferous vegetables, green leafy vegetables, and red/yellow vegetables, in maintaining cognitive function and slowing cognitive decline in middle-aged and older adults.

The aim of the present study was to investigate the effect of body mass index (BMI) distribution, ethnicity and age at menarche on the consistency in the prevalence of underweight and overweight defined by the Centers for Disease Control and Prevention (CDC) and the International Obesity Task Force (IOTF) cut-off points in 2-17-year-old Chinese and US children. Data from the China Health and Nutrition Survey (CHNS) 1991 (n= 3895), CHNS 2004 (n= 1902), and the National Health and Nutrition Examination Survey (NHANES) 2003-2004 (n= 3344) were used to define the prevalence of underweight and overweight using the CDC and IOTF BMI cut-off points. An upward shift in the BMI distribution was observed from the CHNS 1991 to CHNS 2004, and between the CHNS and NHANES 2004. The prevalence of underweight defined by the IOTF cut-off points was higher than that obtained using the CDC cut-off points; absolute differences between IOTF and CDC classifiers were 7%, 5%, and 2% in boys, and 12%, 11%, and 4% in girls in the CHNS 1991, CHNS 2004, and NHANES 2004, respectively. There was a greater consistency in the prevalence of overweight. The consistency of the two classification systems increased with the higher BMI distribution and was not affected significantly by ethnicity or age at menarche. The BMI distribution of a population can affect the classification of childhood underweight and overweight differentially as defined by the CDC and IOTF cut-off points, and thus should be considered in the interpretation of results in clinical and population settings.

Panel count data often occur in a long-term recurrent event study, where the exact occurrence time of the recurrent events is unknown, but only the occurrence count between any two adjacent observation time points is recorded. Most traditional methods only handle panel count data for a single type of event. In this paper, we propose a Bayesian semiparameteric approach to analyze panel count data for multiple types of events. For each type of recurrent event, the proportional mean model is adopted to model the mean count of the event, where its baseline mean function is approximated by monotone I-splines. The correlation between multiple types of events is modeled by common frailty terms and scale parameters. Unlike many frequentist estimating equation methods, our approach is based on the observed likelihood and makes no assumption on the relationship between the recurrent process and the observation process. Under the Poisson counting process assumption, we develop an efficient Gibbs sampler based on novel data augmentation for the Markov chain Monte Carlo sampling. Simulation studies show good estimation performance of the baseline mean functions and the regression coefficients; meanwhile, the importance of including the scale parameter to flexibly accommodate the correlation between events is also demonstrated. Finally, a skin cancer data example is fully analyzed to illustrate the proposed methods.

Panel count data occur in many clinical and observational studies and in some situations the observation process is informative. In this article, we propose a new joint model for the analysis of panel count data with time-dependent covariates and possibly in the presence of informative observation process via two latent variables. For the inference on the proposed model, a class of estimating equations is developed and the resulting estimators are shown to be consistent and asymptotically normal. In addition, a lack-of-fit test is provided for assessing the adequacy of the model. The finite-sample behavior of the proposed methods is examined through Monte Carlo simulation studies which suggest that the proposed approach works well for practical situations. Also an illustrative example is provided.

ABSTRACTPanel count data frequently arise in various applications such as medical research, social sciences and so on. In this paper, a partially varying coefficient model of the panel count data with informative observation times is developed to accommodate the nonlinear interact effects between covariates. For statistical inference of the unknown parameters, quantile regression approaches are proposed, in which the baseline function and the varying coefficients are approximated by B-spline functions. Moreover, asymptotic properties for the estimators are established. Some numerical studies are performed to confirm and evaluate the finite-sample behaviours of the proposed approaches. Finally, the proposed model is applied to the bladder cancer tumour data as an application.

The Poisson regression model is widely used for count data. This model assumes equidispersion. In practice, equidispersion is seldom reflected in data. However, in real-life data, the variance usually exceeds the mean. This situation is known as overdispersion. Negative binomial distribution and other Poisson mix models are often used to model overdispersion count data. Another extension of the negative binomial distribution in another model for count data is the univariate generalized Waring. In addition, the model developed by Famoye can be used in the analysis of count data. When the count data contains a large number of zeros, it is necessary to use zero-inflated models. In this study, different generalized regression models are emphasized for the analysis of excessive zeros count data. For this purpose, a real data set was analysed with the generalized Poisson model, generalized negative binomial model, generalized negative binomial Famoye, generalized Waring model, and the foregoing zero-inflated models. Log-likelihood, Akaike information criterion, Bayes information criterion, Vuong statistics were used for model comparisons.

PurposeDiet plays a fundamental role in promoting resilience against stress-related disorders. We aimed to examine the overall and sex-specific association between food groups and perceived stress in adults.MethodsWe analyzed the prospective data of 7,434 adults who completed both the 2011 and 2015 surveys of the China Health and Nutrition Survey (CHNS). The Diet Quality Questionnaire (DQQ) was used to code all the food items of 2011 dietary intake into 29 food groups, and perceived stress in 2015 was measured using a 14-item perceived stress scale (PSS-14). Univariate analysis and logistic regression models were used to examine the relationship between food groups and perceived stress.ResultsPeople who perceived a higher level of stress (PSS-14 total score > 25) made up 41.5% and 45.1% of the male and female groups, respectively (χ2 = 9.605, p = 0.002). Individuals with increased intake of food groups such as legumes, other vegetables, other fruits, yogurt, poultry, fish & seafood, fluid milk, and fruit juice were less likely to experience a higher level of psychological stress (OR range: 0.544–0.892, p < 0.05). Additionally, we found sex-specific associations between food groups and perceived stress. The difference in the proportion of food groups, such as fluid milk and fish & seafood, between the two stress groups in men was statistically significant (p < 0.025). In the female group, the distribution of eight food groups, like legumes and nuts & seeds, between the two stress groups was statistically significant (p < 0.025).ConclusionThis study indicated that food groups were differentially associated with perceived stress.

The relationship between leisure sedentary behaviors and mental health remains inconclusive. This study aims to provide evidence on leisure sedentary behaviors on mental health using longitudinal data, along with its moderating and mediating roles. We utilized data from four waves (2006-2015) of the China Health and Nutrition Survey (CHNS), focusing on adults who completed their education, with a total sample of 23,693 observations. Leisure sedentary time, the independent variable, was measured based on self-reported data, while mental health issues were the dependent variables based on the Simplified Symptom Self-Rating Scale (SCL) and the Perceived Stress Scale (PSS-14). The primary analysis employed ordinary least squares (OLS) regression. Potential endogeneity was addressed by instrumental variable (IV) method via two-stage least squares (2SLS) regression and a continuous difference-in-differences (DID) design. Our findings indicate that high levels of sedentary time are associated adversely with mental health issues. Moderating roles show that improving health literacy and reducing sedentary time were effective in alleviating this adverse impact. Among older adults, social engagement and support from female caregivers showed greater potential to reduce the detrimental mental health impact of leisure sedentary time. Furthermore, obesity serve as its mediating role. This study highlights the potential adverse impact of increased leisure sedentary time on mental health among Chinese population. These results provide a foundation for public health initiatives aimed at addressing the rising prevalence of sedentary behavior and its association with mental health issues.

Background: Even though it is well-known that iron deficiency is the result of chronic kidney disease (CKD), whether iron will affect kidney function and disease in the general population is not clear. We thus conducted a nationwide cross-sectional study using data from the China Health and Nutrition Survey (CHNS) to assess the relationship of iron status with estimated glomerular filtration rate (eGFR) and CKD among general adults.Methods: A total of 8,339 adults from the China Health and Nutrition Survey in the wave of 2009 were included to assess the association between iron status and eGFR/CKD. Serum ferritin (SF), transferrin, soluble transferrin receptor (sTfR), and hemoglobin (Hb) were measured. The relationship of iron status and eGFR was evaluated by using multi-variable linear regression model. The effect of iron status on the odds of CKD was calculated by logistic regression model.Results: For the association between iron status and eGFR, every 100 μg/L increase in SF was correlated with 0.26 ml/min per 1.73 m2 (95% CI: 0.08–0.44) decrease in eGFR, and every 5 mg/L increase in sTfR was associated with a decrease of 6.00 ml/min per 1.73 m2 (95% CI: 3.79–8.21) in eGFR. There were no significant associations between Hb or transferrin with eGFR. For the association between iron status and CKD, every 5 g/L increase in sTfR was associated with an odds ratio of 3.72 (95% CI: 2.16–6.13) for CKD. The concentrations of Hb were associated with the odds of CKD in a U-shaped manner, with the lowest risk in the Hb range of 136–141 g/L. There was a positive correlation between SF concentration and CKD prevalence but not in a dose–response manner. The odds of CKD for participants in the highest tertile increased by 28% (98% CI: 1–63%) compared with those in the lowest tertile.Conclusion: The concentration of SF and sTfR was positively correlated with the odds of CKD, and Hb was associated with the odds of CKD in a U-shaped manner. Further large prospective researches are warranted to confirm these findings.

Dietary acid load (DAL) is closely related to several chronic diseases. However, the link between DAL and chronic kidney disease (CKD) remains scarce and without data from the Chinese populations whose diet is quite different from people in Western countries. This study evaluated DAL by potential renal acid load (PRAL) and net endogenous acid production (NEAP). We clarified the relationship between DAL and CKD by logistic regression analysis based on data from the China Health and Nutrition Survey (CHNS). The final analysis included 7699 individuals, of whom 811 (11.44%) were CKD patients. Although there was no notable link between PRAL and CKD, higher NEAP levels were independently correlated with CKD. As NEAP values rise, so does CKD prevalence. This trend remains highly significant even after adjustments. In subgroup analyses, the relationship between NEAP and CKD was more consistent in the elderly and subjects with a waistline of less than 82 cm and those without diabetes and heart disease. RCS analysis further confirmed the clear linear relationship between the OR of CKD and NEAP score. This study highlighted that higher NEAP was positively correlated with the risk of CKD.

An econometric analysis of three health economic issues in China

As an essential trace element in the body, selenium is associated with the development of many diseases. The purpose of this study was to explore the association between dietary selenium intake and new-onset stroke risk in Chinese adults. Adults aged ≥18 years in the China Health and Nutrition Survey (CHNS) from 2004 to 2015 were enrolled. Participants were divided into five groups according to the quintile of dietary selenium intake: Q1 (≤ 29.80 μg/day), Q2 (29.80-38.53 μg/day), Q3 (38.53-47.23 μg/day), Q4 (47.23-60.38 μg/day), Q 5(>60.38 μg/day). Cox proportional-hazards model was used to explore the effect of dietary selenium on new-onset stroke. Restricted cubic spline (RCS) was used to visualize the dose-response relationship between dietary selenium and the risk of morbidity. A total of 11,532 subjects were included, and 271 (2.35%) of them developed stroke during a mean follow-up of 6.78 person-years. Compared with the lowest selenium intake group, the HR and 95%CI of stroke in the participants with selenium intake of Q2, Q3, Q4 and Q5 were: 0.85 (0.59, 1.21), 0.62 (0.42, 0.92), 0.43 (0.28, 0.68), 0.49 (0.30, 0.82), respectively. There was an L-shaped relationship between dietary selenium and stroke (nonlinear P-value = 0.0420). The HR and 95%CI of developing stroke was 0.75 (0.65, 0.87) in participants with selenium intake ≤ 60 μg/day. The L-shaped negative association between dietary selenium and stroke in Chinese adults which indicated that dietary selenium should be improved to a certain level to prevent stroke.

We aimed to investigate the association of long-term dietary iron intake with the risk of non-fatal cardiovascular diseases (CVDs), myocardial infarction (MI), and stroke in Chinese populations with predominantly plant-based diets by sex. A total of 17 107 participants (8569 men and 8538 women) aged 18-80 years in the China Health and Nutrition Survey (CHNS) 1989-2015 were included. Dietary intake was assessed repeatedly by three consecutive 24-h dietary recalls. Cox proportional hazard models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). During a median follow-up of 11.1 years, the adjusted HRs (95% CIs) for non-fatal CVDs risk across quintiles of total iron intake in men were 1.00, 0.65 (0.46-0.93), 0.54 (0.37-0.78), 0.66 (0.46-0.94), 0.69 (0.47-1.03), but no significant association in women. Similar associations were found for stroke risk, but not for MI risk. The dose-response curves for the association of total iron and non-heme iron intake with the risk of non-fatal CVDs and stroke followed a reverse J-shape only in men and similar reverse J-shaped association of heme iron intake with non-fatal CVDs and stroke risk in both men and women (P-non-linearity <0.05). Moderate dietary iron intake may protect against non-fatal CVDs and stroke, especially in Chinese men consuming plant-based diets. Both quantity and quality of dietary iron intake should be considered in the prevention of non-fatal CVDs due to differences in dietary patterns among diverse populations.

The prevalence of metabolic syndrome (MetS) is increasing and the relationship between ultra-processed food (UPF) consumption and MetS remains uncertain in Chinese adults. This study aimed to examine the longitudinal association of UPF consumption with the risk of MetS and its components in Chinese adults. Adults aged 18 years and above who participated in at least two waves of the China Health and Nutrition Survey (CHNS) in 2009, 2015, and 2018 were included in this analysis. Dietary intake data were collected by three consecutive 24 h dietary recalls and weighing household foods and condiments. Depending on the purpose and extent of food processing, UPFs were classified using the NOVA food classification system. A multivariate Cox proportional risk model was used to explore the association between UPF consumption (grouped by quartile: quartile 1 (Q1), quartile 2 (Q2), quartile 3 (Q3), and quartile 4 (Q4)) and risk of MetS and its components. A total of 5147 adults were included. During a median (IQR) 6.0 (3.0, 9.0) year follow-up with 31,878 person-years, 1712 MetS cases were identified, with an incidence of 33.26%. After multivariable adjustment, the risk of MetS was increased by 17% in the highest quartile with UPF consumption (HR: 1.17, 95% CI: 1.01–1.35, p trend: 0.047), with the lowest quartile as a reference. For the components of MetS, the risk of central obesity, raised triglycerides (TG), reduced high-density lipoprotein cholesterol (HDL-C), and raised blood pressure (BP) was increased by 33% (HR: 1.33, 95% CI: 1.18–1.51, p trend: <0.001), 26% (HR: 1.26, 95% CI: 1.08–1.48, p trend: 0.003), 25% (HR: 1.25, 95% CI: 1.07–1.46, p trend: 0.007), and 16% (HR: 1.16, 95% CI: 1.03–1.32, p trend: 0.018) in the highest quartile with UPF consumption, respectively. Adults aged 45–59 years and living in urban areas with higher UPF consumption had higher odds of MetS. These results indicate that higher long-term UPF consumption was associated with an increased risk of MetS in Chinese adults. Further studies such as intervention trials are needed to confirm the mechanism of correlation between UPF consumption and health-related outcomes. Nutritional education actions are warranted to promote a balanced diet and improve the overall dietary quality of residents to reduce the risk of MetS effectively.

In this article, we extend the joint frailty models proposed by Zhao and Tong (2011) to panel count data with the time-dependent covariates and informative observation and censoring times. A novel estimating equation approach that does not depend on the distribution of frailty variables and the link function is proposed for estimation of parameters, and the asymptotic properties of the proposed estimators are established. Simulation studies demonstrate that the proposed inference procedure performs well. The analysis of a bladder tumor data is presented to illustrate the method.