Spatio-Temporal Analysis of Influenza-Like Illness and Prediction of Incidence in High-Risk Regions in the United States from 2011 to 2020

The prediction of influenza-like illness using national influenza surveillance data and Baidu query data

BACKGROUND

Seasonal influenza and other respiratory tract infections are serious public health problems that need to be further addressed and investigated. Internet search data are recognized as a valuable source for forecasting influenza or other respiratory tract infection epidemics. However, the selection of internet search data and the application of forecasting methods are important for improving forecasting accuracy. The aim of the present study was to forecast influenza epidemics based on the long short-term memory neural network (LSTM) method, Baidu search index data, and the influenza-like-illness (ILI) rate.

METHODS

The official weekly ILI% data for northern and southern mainland China were obtained from the Chinese Influenza Center from 2018 to 2021. Based on the Baidu Index, search indices related to influenza infection over the corresponding time period were obtained. Pearson correlation analysis was performed to explore the association between influenza-related search queries and the ILI% of southern and northern mainland China. The LSTM model was used to forecast the influenza epidemic within the same week and at lags of 1-4 weeks. The model performance was assessed by evaluation metrics, including the mean square error (MSE), root mean square error (RMSE) and mean absolute error (MAE).

RESULTS

In total, 24 search queries in northern mainland China and 7 search queries in southern mainland China were found to be correlated and were used to construct the LSTM model, which included the same week and a lag of 1-4 weeks. The LSTM model showed that ILI% + mask with one lag week and ILI% + influenza name were good prediction modules, with reduced RMSE predictions of 16.75% and 4.20%, respectively, compared with the estimated ILI% for northern and southern mainland China.

CONCLUSIONS

The results illuminate the feasibility of using an internet search index as a complementary data source for influenza forecasting and the efficiency of using the LSTM model to forecast influenza epidemics.

Geographical characteristics and influencing factors of the influenza epidemic in Hubei, China, from 2009 to 2019

Influenza is an acute respiratory infectious disease that commonly affects people and has an important impact on public health. Based on influenza incidence data from 103 counties in Hubei Province from 2009 to 2019, this study used time series analysis and geospatial analysis to analyze the spatial and temporal distribution characteristics of the influenza epidemic and its influencing factors. The results reveal significant spatial-temporal clustering of the influenza epidemic in Hubei Province. Influenza mainly occurs in winter and spring of each year (from December to March of the next year), with the highest incidence rate observed in 2019 and an overall upward trend in recent years. There were significant spatial and urban-rural differences in influenza prevalence in Hubei Province, with the eastern region being more seriously affected than the central and western regions, and the urban regions more seriously affected than the rural region. Hubei's influenza epidemic showed an obvious spatial agglomeration distribution from 2009 to 2019, with the strongest clustering in winter. The hot spot areas of interannual variation in influenza were mainly distributed in eastern and western Hubei, and the cold spot areas were distributed in north-central Hubei. In addition, the cold hot spot areas of influenza epidemics varied from season to season. The seasonal changes in influenza prevalence in Hubei Province are mainly governed by meteorological factors, such as temperature, sunshine, precipitation, humidity, and wind speed. Low temperature, less rain, less sunshine, low wind speed and humid weather will increase the risk of contracting influenza; the interannual changes and spatial differentiation of influenza are mainly influenced by socioeconomic factors, such as road density, number of health technicians per 1,000 population, urbanization rate and population density. The strength of influenza's influencing factors in Hubei Province exhibits significant spatial variation, but in general, the formation of spatial variation of influenza in Hubei Province is still the result of the joint action of socioeconomic factors and natural meteorological factors. Understanding the temporal and spatial distribution characteristics of influenza in Hubei Province and its influencing factors can provide a reasonable decision-making basis for influenza prevention and control and public health development in Hubei Province and can also effectively improve the scientific understanding of the public with respect to influenza and other respiratory infectious diseases to reduce the influenza incidence, which also has reference significance for the prevention and control of influenza and other respiratory infectious diseases in other countries or regions.

Self-reported critical gaps in the essential knowledge and capacity of spatial epidemiology between the current university education and competency-oriented professional demands in preparing for a future pandemic among public health postgraduates in China: a nationwide cross-sectional survey

BACKGROUND

Spatial epidemiology plays an important role in public health. Yet, it is unclear whether the current university education in spatial epidemiology in China could meet the competency-oriented professional demands. This study aimed to understand the current situation of education and training, practical application, and potential demands in spatial epidemiology among public health postgraduates in China, and to assess the critical gaps in a future emerging infectious diseases (EID) pandemic preparedness and response.

METHODS

This study was divided into three parts. The first part was a comparative study on spatial epidemiology education in international public health postgraduate training. The second part was a cross-sectional survey conducted among public health professionals. The third part was a nationwide cross-sectional survey conducted among public health postgraduates at Chinese universities from October 2020 to February 2021. Data was collected by the WeChat-based questionnaire star survey system and analyzed using the SPSS software.

RESULTS

International education institutions had required public health postgraduates to master the essential knowledge and capacity of spatial epidemiology. A total of 198 public health professionals were surveyed, and they had a median of 4.00 (IQR 3.13-4.53) in demand degree of spatial epidemiology. A total of 1354 public health postgraduates were surveyed from 51 universities. Only 29.41% (15/51) of universities offered spatial epidemiology course. Around 8.05% (109/1354) of postgraduates had learned spatial epidemiology, and had a median of 1.05 (IQR 1.00-1.29) in learning degree and a median of 1.91 (IQR 1.05-2.78) in practical application degree of spatial epidemiology. To enhance professional capacity, 65.95% (893/1354) of postgraduates hoped that universities would deliver a credit-course of spatial epidemiology.

CONCLUSIONS

A huge unmet education and training demand in spatial epidemiology existed in the current education system of public health postgraduates in China. To enhance the competency-oriented professional capacity in preparedness and response to a future pandemic, it is urgent to incorporate the teaching and training of spatial epidemiology into the compulsory curriculum system of public health postgraduates in China.

Estimation of influenza incidence and analysis of epidemic characteristics from 2009 to 2022 in Zhejiang Province, China

Background

Influenza infection causes a huge burden every year, affecting approximately 8% of adults and approximately 25% of children and resulting in approximately 400,000 respiratory deaths worldwide. However, based on the number of reported influenza cases, the actual prevalence of influenza may be greatly underestimated. The purpose of this study was to estimate the incidence rate of influenza and determine the true epidemiological characteristics of this virus.

Methods

The number of influenza cases and the prevalence of ILIs among outpatients in Zhejiang Province were obtained from the China Disease Control and Prevention Information System. Specimens were sampled from some cases and sent to laboratories for influenza nucleic acid testing. Random forest was used to establish an influenza estimation model based on the influenza-positive rate and the percentage of ILIs among outpatients. Furthermore, the moving epidemic method (MEM) was applied to calculate the epidemic threshold for different intensity levels. Joinpoint regression analysis was used to identify the annual change in influenza incidence. The seasonal trends of influenza were detected by wavelet analysis.

Results

From 2009 to 2021, a total of 990,016 influenza cases and 8 deaths were reported in Zhejiang Province. The numbers of estimated influenza cases from 2009 to 2018 were 743,449, 47,635, 89,026, 132,647, 69,218, 190,099, 204,606, 190,763, 267,168 and 364,809, respectively. The total number of estimated influenza cases is 12.11 times the number of reported cases. The APC of the estimated annual incidence rate was 23.33 (95% CI: 13.2 to 34.4) from 2011 to 2019, indicating a constant increasing trend. The intensity levels of the estimated incidence from the epidemic threshold to the very high-intensity threshold were 18.94 cases per 100,000, 24.14 cases per 100,000, 141.55 cases per 100,000, and 309.34 cases per 100,000, respectively. From the first week of 2009 to the 39th week of 2022, there were a total of 81 weeks of epidemics: the epidemic period reached a high intensity in 2 weeks, the epidemic period was at a moderate intensity in 75 weeks, and the epidemic period was at a low intensity in 2 weeks. The average power was significant on the 1-year scale, semiannual scale, and 115-week scale, and the average power of the first two cycles was significantly higher than that of the other cycles. In the period from the 20th week to the 35th week, the Pearson correlation coefficients between the time series of influenza onset and the positive rate of pathogens, including A(H3N2), A (H1N1)pdm2009, B(Victoria) and B(Yamagata), were - 0.089 (p = 0.021), 0.497 (p < 0.001), -0.062 (p = 0.109) and - 0.084 (p = 0.029), respectively. In the period from the 36th week of the first year to the 19th week of the next year, the Pearson correlation coefficients between the time series of influenza onset and the positive rate of pathogens, including A(H3N2), A (H1N1)pdm2009, B(Victoria) and B(Yamagata), were 0.516 (p < 0.001), 0.148 (p < 0.001), 0.292 (p < 0.001) and 0.271 (p < 0.001), respectively.

Conclusion

The disease burden of influenza has been seriously underestimated in the past. An appropriate method for estimating the incidence rate of influenza may be to comprehensively consider the influenza-positive rate as well as the percentage of ILIs among outpatients. The intensity level of the estimated incidence from the epidemic threshold to the very high-intensity threshold was calculated, thus yielding a quantitative standard for judging the influenza prevalence level in the future. The incidence of influenza showed semi-annual peaks in Zhejiang Province, including a main peak from December to January of the next year followed by a peak in summer. Furthermore, the driving factors of the influenza peaks were preliminarily explored. While the peak in summer was mainly driven by pathogens of A(H3N2), the peak in winter was alternately driven by various pathogens. Our research suggests that the government urgently needs to address barriers to vaccination and actively promote vaccines through primary care providers.

Study on the prediction effect of a combined model of SARIMA and LSTM based on SSA for influenza in Shanxi Province, China

BACKGROUND

Influenza is an acute respiratory infectious disease that is highly infectious and seriously damages human health. Reasonable prediction is of great significance to control the epidemic of influenza.

METHODS

Our Influenza data were extracted from Shanxi Provincial Center for Disease Control and Prevention. Seasonal-trend decomposition using Loess (STL) was adopted to analyze the season characteristics of the influenza in Shanxi Province, China, from the 1st week in 2010 to the 52nd week in 2019. To handle the insufficient prediction performance of the seasonal autoregressive integrated moving average (SARIMA) model in predicting the nonlinear parts and the poor accuracy of directly predicting the original sequence, this study established the SARIMA model, the combination model of SARIMA and Long-Short Term Memory neural network (SARIMA-LSTM) and the combination model of SARIMA-LSTM based on Singular spectrum analysis (SSA-SARIMA-LSTM) to make predictions and identify the best model. Additionally, the Mean Squared Error (MSE), Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) were used to evaluate the performance of the models.

RESULTS

The influenza time series in Shanxi Province from the 1st week in 2010 to the 52nd week in 2019 showed a year-by-year decrease with obvious seasonal characteristics. The peak period of the disease mainly concentrated from the end of the year to the beginning of the next year. The best fitting and prediction performance was the SSA-SARIMA-LSTM model. Compared with the SARIMA model, the MSE, MAE and RMSE of the SSA-SARIMA-LSTM model decreased by 38.12, 17.39 and 21.34%, respectively, in fitting performance; the MSE, MAE and RMSE decreased by 42.41, 18.69 and 24.11%, respectively, in prediction performances. Furthermore, compared with the SARIMA-LSTM model, the MSE, MAE and RMSE of the SSA-SARIMA-LSTM model decreased by 28.26, 14.61 and 15.30%, respectively, in fitting performance; the MSE, MAE and RMSE decreased by 36.99, 7.22 and 20.62%, respectively, in prediction performances.

CONCLUSIONS

The fitting and prediction performances of the SSA-SARIMA-LSTM model were better than those of the SARIMA and the SARIMA-LSTM models. Generally speaking, we can apply the SSA-SARIMA-LSTM model to the prediction of influenza, and offer a leg-up for public policy.

Spatiotemporally comparative analysis of three common infectious diseases in China during 2013-2015

BACKGROUND

Dengue fever (DF), influenza, and hand, foot, and mouth disease (HFMD) have had several various degrees of outbreaks in China since the 1900s, posing a serious threat to public health. Previous studies have found that these infectious diseases were often prevalent in the same areas and during the same periods in China.

METHODS

This study combined traditional descriptive statistics and spatial scan statistic methods to analyze the spatiotemporal features of the epidemics of DF, influenza, and HFMD during 2013-2015 in mainland China at the provincial level.

RESULTS

DF got an intensive outbreak in 2014, while influenza and HFMD were stable from 2013 to 2015. DF mostly occurred during August-November, influenza appeared during November-next March, and HFMD happened during April-November. The peaks of these diseases form a year-round sequence; Spatially, HFMD generally has a much higher incidence than influenza and DF and covers larger high-risk areas. The hotspots of influenza tend to move from North China to the southeast coast. The southeastern coastal regions are the high-incidence areas and the most significant hotspots of all three diseases.

CONCLUSIONS

This study suggested that the three diseases can form a year-round sequence in southern China, and the southeast coast of China is a particularly high-risk area for these diseases. These findings may have important implications for the local public health agency to allocate the prevention and control resources.

Epidemiological Trends of Malaria in Five Years and under Children of Nsanje District in Malawi, 2015-2019

Background: Malaria continues to be a major public health problem in Malawi and the greatest load of mortality and morbidity occurs in children five years and under. However, there is no information yet regarding trends and predictions of malaria incidence in children five years and under at district hospital level, particularly at Nsanje district hospital. Aim: Therefore, this study aimed at investigating the trends of malaria morbidity and mortality in order to design appropriate interventions on the best approach to contain the disease in the near future. Methodology: Trend analysis of malaria morbidity and mortality together with time series analysis using the SARIMA (Seasonal Autoregressive Integrated Moving Average) model was used to predict malaria incidence in Nsanje district. Results: The SARIMA model used malaria cases from 2015 to 2019 and created the best model to forecast the malaria cases in Nsanje from 2020 to 2022. An SARIMA (0, 1, 2) (0,1,1)₁₂ was suitable for forecasting the incidence of malaria for Nsanje. Conclusion: The mortality and morbidity trend showed that malaria cases were growing at a fluctuating rate at Nsanje district hospital. The relative errors between the actual values and predicted values indicated that the predicted values matched the actual values well. Therefore, the model proved that it was adequate to forecast monthly malaria cases and it had a good fit, hence, was appropriate for this study