Search

BackgroundChangeful seasonal influenza activity in subtropical areas such as Taiwan causes problems in epidemic preparedness. The Taiwan Centers for Disease Control has maintained real-time national influenza surveillance systems since 2004. Except for timely monitoring, epidemic forecasting using the national influenza surveillance data can provide pivotal information for public health response. ObjectiveWe aimed to develop predictive models using machine learning to provide real-time influenza-like illness forecasts. MethodsUsing surveillance data of influenza-like illness visits from emergency departments (from the Real-Time Outbreak and Disease Surveillance System), outpatient departments (from the National Health Insurance database), and the records of patients with severe influenza with complications (from the National Notifiable Disease Surveillance System), we developed 4 machine learning models (autoregressive integrated moving average, random forest, support vector regression, and extreme gradient boosting) to produce weekly influenza-like illness predictions for a given week and 3 subsequent weeks. We established a framework of the machine learning models and used an ensemble approach called stacking to integrate these predictions. We trained the models using historical data from 2008-2014. We evaluated their predictive ability during 2015-2017 for each of the 4-week time periods using Pearson correlation, mean absolute percentage error (MAPE), and hit rate of trend prediction. A dashboard website was built to visualize the forecasts, and the results of real-world implementation of this forecasting framework in 2018 were evaluated using the same metrics. ResultsAll models could accurately predict the timing and magnitudes of the seasonal peaks in the then-current week (nowcast) (ρ=0.802-0.965; MAPE: 5.2%-9.2%; hit rate: 0.577-0.756), 1-week (ρ=0.803-0.918; MAPE: 8.3%-11.8%; hit rate: 0.643-0.747), 2-week (ρ=0.783-0.867; MAPE: 10.1%-15.3%; hit rate: 0.669-0.734), and 3-week forecasts (ρ=0.676-0.801; MAPE: 12.0%-18.9%; hit rate: 0.643-0.786), especially the ensemble model. In real-world implementation in 2018, the forecasting performance was still accurate in nowcasts (ρ=0.875-0.969; MAPE: 5.3%-8.0%; hit rate: 0.582-0.782) and remained satisfactory in 3-week forecasts (ρ=0.721-0.908; MAPE: 7.6%-13.5%; hit rate: 0.596-0.904). ConclusionsThis machine learning and ensemble approach can make accurate, real-time influenza-like illness forecasts for a 4-week period, and thus, facilitate decision making.

Abstract Background Few published reports have evaluated the clinical features and outcome of serogroup W135 meningococcal disease. In Taiwan, W135 is the second most prevalent meningococcal disease serogroup. Method A nationwide study was conducted to retrospectively analyze epidemiologic data from 115 patients with laboratory confirmed meningococcal disease that occurred from 2001 through 2003. Results Serogroup W135 accounted for 26% of all cases and most (76.7%) were older than 20 years. There were no cases of serogroup W135 meningococcal disease associated with Hajj pilgrims, and all cases were sporadic. In 88 patients with complete case records, we compared the presenting symptoms, signs, laboratory data, and outcomes between W135 and non-W135 patients. There were no differences in presenting symptoms except for the higher prevalence of pneumonia found in W135 patients (23.8% vs. 1.5%; OR: 20.6; 95%CI: 2.3–189.0; p = 0.003). The distribution of inflammatory cells in CSF in patients with meningitis was also different between W135 and non-W135 patients. W135 patients had a trend toward more intubations and shock but it did not achieve statistical significance. In multivariate analysis of factors associated with death, three independent factors were found: bacteremia without meningitis, altered mental status, and petechiae or purpura on admission. Conclusion Sporadic serogroup W135 meningococcal disease is an important component of the meningococcal disease burden in Taiwan, but it is not directly associated with Hajj pilgrims. Compared with patients infected by other serogroups of meningococci, patients with serogroup W135 were older and more likely to have extrameningeal involvement such as pneumonia.

Objectives: Several respirable hazards, including smoking and indoor air pollution from biomass, were suggested to increase the risk of tuberculosis. Few studies have been conducted on ambient air pollution and tuberculosis. We investigated the association between exposure to ambient air pollution and incidence of active tuberculosis.Methods: We conducted a cohort study using 106,678 participants of a community-based screening service in Taiwan, 2005-2012. We estimated individual exposure to air pollution using data from the nearest air quality monitoring station and the road intensity within a 500 m buffer zone. The incidence of tuberculosis was ascertained from the national tuberculosis registry.Results: After a median follow-up of 6.7 years, 418 cases of tuberculosis occurred. Exposure to fine particulate matter (PM2.5) was associated with increased risk of active tuberculosis (adjusted HR: 1.39/10 μg/m3 (95% CI 0.95 to 2.03)). In addition, traffic-related air pollution including nitrogen dioxide (adjusted HR: 1.33/10 ppb; 95% CI 1.04 to 1.70), nitrogen oxides (adjusted HR: 1.21/10 ppb; 95% CI 1.04 to 1.41) and carbon monoxide (adjusted HR: 1.89/ppm; 95% CI 0.78 to 4.58) was associated with tuberculosis risk. There was a non-significant trend between the length of major roads in the neighbourhood and culture-confirmed tuberculosis (adjusted HR: 1.04/km; 95% CI 0.995 to 1.09).Conclusions: Our study revealed a possible link between ambient air pollution and risk of active tuberculosis. Since people from developing countries continue to be exposed to high levels of ambient air pollution and to experience high rates of tuberculosis, the impact of worsening air pollution on global tuberculosis control warrants further investigation. [ABSTRACT FROM AUTHOR]