Seasonal Influenza Forecasting in Real Time Using the Incidence Decay With Exponential Adjustment Model

Predicting regional influenza epidemics with uncertainty estimation using commuting data in Japan

Obtaining an accurate prediction of the number of influenza patients in specific areas is a crucial task undertaken by medical institutions. Infections (such as influenza) spread from person to person, and people are rarely confined to a single area. Therefore, creating a regional influenza prediction model should consider the flow of people between different areas. Although various regional flu prediction models have previously been proposed, they do not consider the flow of people among areas. In this study, we propose a method that can predict the geographical distribution of influenza patients using commuting data to represent the flow of people. To elucidate the complex spatial dependence relations, our model uses an extension of the graph convolutional network (GCN). Additionally, a prediction interval for medical institutions is proposed, which is suitable for cyclic time series. Subsequently, we used the weekly data of flu patients from health authorities as the ground-truth to evaluate the prediction interval and performance of influenza patient prediction in each prefecture in Japan. The results indicate that our GCN-based model, which used commuting data, considerably improved the predictive accuracy over baseline values both temporally and spatially to provide an appropriate prediction interval. The proposed model is vital in practical settings, such as in the decision making of public health authorities and addressing growth in vaccine demand and workload. This paper primarily presents a GCN as a useful means for predicting the spread of an epidemic.

Effects of weather-related social distancing on city-scale transmission of respiratory viruses: a retrospective cohort study

Background

Unusually high snowfall in western Washington State in February 2019 led to widespread school and workplace closures. We assessed the impact of social distancing caused by this extreme weather event on the transmission of respiratory viruses.

Methods

Residual specimens from patients evaluated for acute respiratory illness at hospitals in the Seattle metropolitan area were screened for a panel of respiratory viruses. Transmission models were fit to each virus to estimate the magnitude reduction in transmission due to weather-related disruptions. Changes in contact rates and care-seeking were informed by data on local traffic volumes and hospital visits.

Results

Disruption in contact patterns reduced effective contact rates during the intervention period by 16 to 95%, and cumulative disease incidence through the remainder of the season by 3 to 9%. Incidence reductions were greatest for viruses that were peaking when the disruption occurred and least for viruses in an early epidemic phase.

Conclusion

High-intensity, short-duration social distancing measures may substantially reduce total incidence in a respiratory virus epidemic if implemented near the epidemic peak. For SARS-CoV-2, this suggests that, even when SARS-CoV-2 spread is out of control, implementing short-term disruptions can prevent COVID-19 deaths.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-021-06028-4.

The performance of phenomenological models in providing near-term Canadian case projections in the midst of the COVID-19 pandemic: March - April, 2020

BACKGROUND

The COVID-19 pandemic has had an unprecedented impact on citizens and health care systems globally. Valid near-term projections of cases are required to inform the escalation, maintenance and de-escalation of public health measures, and for short-term health care resource planning.

METHODS

Near-term case and epidemic growth rate projections for Canada were estimated using three phenomenological models: the logistic model, Generalized Richard's model (GRM) and a modified Incidence Decay and Exponential Adjustment (m-IDEA) model. Throughout the COVID-19 epidemic in Canada, these models have been validated against official national epidemiological data on an ongoing basis.

RESULTS

The best-fit models estimated that the number of COVID-19 cases predicted to be reported in Canada as of April 1, 2020 and May 1, 2020 would be 11,156 (90 % prediction interval: 9,156-13,905) and 54,745 (90 % prediction interval: 54,252-55,239). The three models varied in their projections and their performance over the first seven weeks of their implementation. Both the logistic model and GRM under-predicted cases reported a week following the projection date in nearly all instances. The logistic model performed best at the early stages, the m-IDEA model performed best at the later stages, and the GRM performed most consistently during the full period assessed.

CONCLUSIONS

All three models have yielded qualitatively comparable near-term forecasts of cases and epidemic growth for Canada. Under or over-estimation of projected cases and epidemic growth by these models could be associated with changes in testing policies and/or public health measures. Simple forecasting models can be invaluable in projecting the changes in trajectory of subsequent waves of cases to provide timely information to support the pandemic response.

Robust two-stage influenza prediction model considering regular and irregular trends

Influenza causes numerous deaths worldwide every year. Predicting the number of influenza patients is an important task for medical institutions. Two types of data regarding influenza-like illnesses (ILIs) are often used for flu prediction: (1) historical data and (2) user generated content (UGC) data on the web such as search queries and tweets. Historical data have an advantage against the normal state but show disadvantages against irregular phenomena. In contrast, UGC data are advantageous for irregular phenomena. So far, no effective model providing the benefits of both types of data has been devised. This study proposes a novel model, designated the two-stage model, which combines both historical and UGC data. The basic idea is, first, basic regular trends are estimated using the historical data-based model, and then, irregular trends are predicted by the UGC data-based model. Our approach is practically useful because we can train models separately. Thus, if a UGC provider changes the service, our model could produce better performance because the first part of the model is still stable. Experiments on the US and Japan datasets demonstrated the basic feasibility of the proposed approach. In the dropout (pseudo-noise) test that assumes a UGC service would change, the proposed method also showed robustness against outliers. The proposed model is suitable for prediction of seasonal flu.