Time series analysis of hemorrhagic fever with renal syndrome in mainland China by using an XGBoost forecasting model.

BACKGROUND: Hemorrhagic fever with renal syndrome (HFRS) is still attracting public attention because of its outbreak in various cities in China. Predicting future outbreaks or epidemics disease based on past incidence data can help health departments take targeted measures to prevent diseases in advance. In this study, we propose a multistep prediction strategy based on extreme gradient boosting (XGBoost) for HFRS as an extension of the one-step prediction model. Moreover, the fitting and prediction accuracy of the XGBoost model will be compared with the autoregressive integrated moving average (ARIMA) model by different evaluation indicators. METHODS: We collected HFRS incidence data from 2004 to 2018 of mainland China. The data from 2004 to 2017 were divided into training sets to establish the seasonal ARIMA model and XGBoost model, while the 2018 data were used to test the prediction performance. In the multistep XGBoost forecasting model, one-hot encoding was used to handle seasonal features. Furthermore, a series of evaluation indices were performed to evaluate the accuracy of the multistep forecast XGBoost model. RESULTS: There were 200,237 HFRS cases in China from 2004 to 2018. A long-term downward trend and bimodal seasonality were identified in the original time series. According to the minimum corrected akaike information criterion (CAIC) value, the optimal ARIMA (3, 1, 0) × (1, 1, 0)12 model is selected. The index ME, RMSE, MAE, MPE, MAPE, and MASE indices of the XGBoost model were higher than those of the ARIMA model in the fitting part, whereas the RMSE of the XGBoost model was lower. The prediction performance evaluation indicators (MAE, MPE, MAPE, RMSE and MASE) of the one-step prediction and multistep prediction XGBoost model were all notably lower than those of the ARIMA model. CONCLUSIONS: The multistep XGBoost prediction model showed a much better prediction accuracy and model stability than the multistep ARIMA prediction model. The XGBoost model performed better in predicting complicated and nonlinear data like HFRS. Additionally, Multistep prediction models are more practical than one-step prediction models in forecasting infectious diseases.

Speckle tracking echocardiography assessment of global and regional contraction dysfunction in the mice model of pressure overload.

Speckle tracking echocardiography (STE) has been applied to the evaluation of cardiac contraction dysfunction. However, there were few studies on alteration of global and regional STE parameters in the process of myocardial hypertrophy and heart failure. In this study, STE was applied to evaluate the global and regional cardiac function under heart failure and hypertrophy in the mice model of pressure overload. Adult mice were subjected to mild or severe aortic banding with a 25 Gauge (G) or 27 G needle. After surgery, STE and conventional echocardiography were used in the sham group (n=10), mild trans-aortic banding (TAB) group (n=14) and severe TAB group (n=10) for 8 weeks. The results showed that the mice subjected to severe TAB showed a significant change in fractional shortening (FS), left ventricular (LV) mass, and left ventricular end diastolic diameter (LVEDD) (P<0.05 for each). Meanwhile, there were no significant differences in FS and LVEDD between the sham group and mild TAB group during the experimental procedures (P>0.05 for both). STE analysis revealed that longitudinal strain (LS) was significantly decreased in the severe TAB group as compared with the sham and mild TAB groups (P<0.05 for both) from the postoperative week 1. LS in the mild TAB group was reduced as compared to the sham group (P<0.05). Radial strain (RS) and circumferential strain (CS) were significantly decreased in the severe TAB group as compared to the sham group and the mild TAB group (P<0.05 for both) from the postoperative week 1 (P<0.05 for both). Compared to the sham group, CS in the mild TAB group maintained unchanged during the test period, and RS was reduced only on the postoperative week 6 (P<0.05). Finally, regional contraction dysfunction was analyzed in both hypertrophic and failing myocardium in longitudinal and radial directions. It was found that LS was largest in the apex region and RS was smallest in the apex region in the healthy and hypertrophic myocardium. It was also found that compared to the sham group, only base longitudinal strain in the mild TAB group was decreased. Each of regional strain in the severe TAB group was uniformly depressed in radial and longitudinal directions. It is concluded that STE has provided a non-invasive and highly feasible way to explore the global and regional contraction dysfunction in hypertrophic and heart failure myocardium in the murine model of pressure overload.

Application of a data-driven XGBoost model for the prediction of COVID-19 in the USA: a time-series study.

OBJECTIVE: The COVID-19 outbreak was first reported in Wuhan, China, and has been acknowledged as a pandemic due to its rapid spread worldwide. Predicting the trend of COVID-19 is of great significance for its prevention. A comparison between the autoregressive integrated moving average (ARIMA) model and the eXtreme Gradient Boosting (XGBoost) model was conducted to determine which was more accurate for anticipating the occurrence of COVID-19 in the USA. DESIGN: Time-series study. SETTING: The USA was the setting for this study. MAIN OUTCOME MEASURES: Three accuracy metrics, mean absolute error (MAE), root mean square error (RMSE) and mean absolute percentage error (MAPE), were applied to evaluate the performance of the two models. RESULTS: In our study, for the training set and the validation set, the MAE, RMSE and MAPE of the XGBoost model were less than those of the ARIMA model. CONCLUSIONS: The XGBoost model can help improve prediction of COVID-19 cases in the USA over the ARIMA model.

The correlation between temperature and the incidence of COVID-19 in four first-tier cities of China: a time series study.

The COVID-19 outbreak emerged in Wuhan, China, and was declared a global pandemic in March 2020. This study aimed to explore the association of daily mean temperature with the daily counts of COVID-19 cases in Beijing, Shanghai, Guangzhou, and Shenzhen, China. Data on daily confirmed cases of COVID-19 and daily mean temperatures were retrieved from the 4 first-tier cities in China. Distributed lag nonlinear models (DLNMs) were used to assess the association between daily mean temperature and the daily cases of COVID-19 during the study period. After controlling for the imported risk index and long-term trends, the distributed lag nonlinear model showed that there were nonlinear and lag relationships. The daily cumulative relative risk decreased for every 1.0 °C change in temperature in Shanghai, Guangzhou, and Shenzhen. However, the cumulative relative risk increased with a daily mean temperature below - 3 °C in Beijing and then decreased. Moreover, the delayed effects of lower temperatures mostly occurred within 6-7 days of exposure. There was a negative correlation between the cumulative relative risk of COVID-19 incidence and temperature, especially when the temperature was higher than - 3 °C. The conclusions from this paper will help government and health regulators in these cities take prevention and protection measures to address the COVID-19 crisis and the possible collapse of the health system in the future.

Spatiotemporal cluster analysis of COVID-19 and its relationship with environmental factors at the city level in mainland China.

This study sought to identify the spatial, temporal, and spatiotemporal clusters of COVID-19 cases in 366 cities in mainland China with the highest risks and to explore the possible influencing factors of imported risks and environmental factors on the spatiotemporal aggregation, which would be useful to the design and implementation of critical preventative measures. The retrospective analysis of temporal, spatial, and spatiotemporal clustering of COVID-19 during the period (January 15 to February 25, 2020) was based on Kulldorff's time-space scanning statistics using the discrete Poisson probability model, and then the logistic regression model was used to evaluate the impact of imported risk and environmental factors on spatiotemporal aggregation. We found that the spatial distribution of COVID-19 cases was nonrandom; the Moran's I value ranged from 0.017 to 0.453 (P < 0.001). One most likely cluster and three secondary likely clusters were discovered in spatial cluster analysis. The period from February 2 to February 9, 2020, was identified as the most likely cluster in the temporal cluster analysis. One most likely cluster and seven secondary likely clusters were discovered in spatiotemporal cluster analysis. Imported risk, humidity, and inhalable particulate matter PM2.5 had a significant impact on temporal and spatial accumulation, and temperature and PM10 had a low correlation with the spatiotemporal aggregation of COVID-19. The information is useful for health departments to develop a better prevention strategy and potentially increase the effectiveness of public health interventions.

Characterization of voltage-and Ca2+-activated K+ channels in rat dorsal root ganglion neurons.

Auxiliary beta-subunits associated with pore-forming Slo1 alpha-subunits play an essential role in regulating functional properties of large-conductance, voltage- and Ca(2+)-activated K(+) channels commonly termed BK channels. Even though both noninactivating and inactivating BK channels are thought to be regulated by beta-subunits (beta1, beta2, beta3, or beta4), the molecular determinants underlying inactivating BK channels in native cells have not been extensively demonstrated. In this study, rbeta2 (but not rbeta3-subunit) was identified as a molecular component in rat lumbar L4-6 dorsal root ganglia (DRG) by RT-PCR responsible for inactivating large-conductance Ca(2+)-dependent K(+) currents (BK(i) currents) in small sensory neurons. The properties of native BK(i) currents obtained from both whole-cell and inside-out patches are very similar to inactivating BK channels produced by co-expressing mSlo1 alpha- and hbeta2-subunits in Xenopus oocytes. Intracellular application of 0.5 mg/ml trypsin removes inactivation of BK(i) channels, and the specific blockers of BK channels, charybdotoxin (ChTX) and iberiotoxin (IbTX), inhibit these BK(i) currents. Single BK(i) channel currents derived from inside-out patches revealed that one BK(i) channel contained three rbeta2-subunits (on average), with a single-channel conductance about 217 pS under 160 K(+) symmetrical recording conditions. Blockade of BK(i) channels by 100 nM IbTX augmented firing frequency, broadened action potential waveform and reduced after-hyperpolarization. We propose that the BK(i) channels in small diameter DRG sensory neurons might play an important role in regulating nociceptive input to the central nervous system (CNS).