Background: Missing data are common in statistical analyses, and imputation methods based on random forests (RF) are becoming popular for handling missing data especially in biomedical research. Unlike standard imputation approaches, RF-based imputation methods do not assume normality or require specification of parametric models. However, it is still inconclusive how they perform for non-normally distributed data or when there are non-linear relationships or interactions. Methods: To examine the effects of these three factors, a variety of datasets were simulated with outcome-dependent missing at random (MAR) covariates, and the performances of the RF-based imputation methods missForest and CALIBERrfimpute were evaluated in comparison with predictive mean matching (PMM). Results: Both missForest and CALIBERrfimpute have high predictive accuracy but missForest can produce severely biased regression coefficient estimates and downward biased confidence interval coverages, especially for highly skewed variables in nonlinear models. CALIBERrfimpute typically outperforms missForest when estimating regression coefficients, although its biases are still substantial and can be worse than PMM for logistic regression relationships with interaction. Conclusions: RF-based imputation, in particular missForest, should not be indiscriminately recommended as a panacea for imputing missing data, especially when data are highly skewed and/or outcome-dependent MAR. A correct analysis requires a careful critique of the missing data mechanism and the inter-relationships between the variables in the data.

Background: Excessive gestational weight gain (GWG) poses negative impact on mothers and their children. It is important to understand the modifiable lifestyle factors associated with excessive GWG during pregnancy to guide future public health practice. Aim: To investigate the association between physical activity during pregnancy and GWG of Chinese urban pregnant women. Methods: A pregnant women cohort was established between 2005 and 2007 in Changzhou, China. Physical activity levels of pregnant women were assessed using pedometer in the 2^nd and 3^rd trimester, respectively. According to step counts, pregnant women were categorized into 4 different physical activity groups: Sedentary, Low Active, Somewhat Active and Active. The pregnant women were followed for eligibility and data collection from the 2^nd trimester to delivery. Multiple linear regression and multiple binary logistic model were applied to determine the association between physical activity and GWG. Results: Physical activity levels and GWG of 862 pregnant women were assessed, among them 473 (54.9%) experienced excessive GWG. The adjusted odds ratio (OR) was 0.59 (95%CI: 0.36 ~ 0.95) for excessive GWG in the Active group during the 2^nd trimester and 0.66 (95%CI: 0.43 ~ 1.00) in the Somewhat Active group during the 3^rd trimester, compared with the Sedentary group respectively. In the last two trimesters, the Active group had 1.45 kg less GWG, than the Sedentary group. The ORs of excessive GWG decreased with the increased level of physical activity (P < 0.05). Conclusion: This study suggests that pregnant women being physically active have less weight gain during pregnancy.

Hand, foot, and mouth disease (HFMD) has become an emerging infectious disease in China in the last decade. There has been evidence that meteorological factors can influence the HFMD incidence, and understanding the mechanisms can help prevent and control HFMD. HFMD incidence data and meteorological data in Minhang District, Shanghai were obtained for the period between 2009 and 2015. Distributed lag non-linear models (DLNMs) were utilized to investigate the impact of meteorological factors on HFMD incidence after adjusting for potential confounders of long time trend, weekdays and holidays. There was a non-linear relationship between temperature and HFMD incidence, the RR of 5th percentile compared to the median is 0.836 (95% CI: 0.671–1.042) and the RR of 95th percentile is 2.225 (95% CI: 1.774–2.792), and the effect of temperature varied across age groups. HFMD incidence increased with increasing average relative humidity (%) (RR = 1.009, 95% CI: 1.005–1.015) and wind speed (m/s) (RR = 1.197, 95% CI: 1.118–1.282), and with decreasing daily rainfall (mm) (RR = 0.992, 95% CI: 0.987–0.997) and sunshine hours (h) (RR = 0.966, 95% CI: 0.951–0.980). There were significant relationships between meteorological factors and childhood HFMD incidence in Minhang District, Shanghai. This information can help local health agencies develop strategies for the control and prevention of HFMD under specific climatic conditions.