COVID-19 forecasting using new viral variants and vaccination effectiveness models

Abstract

Recently, a high number of daily positive COVID-19 cases have been reported in regions with relatively high vaccination rates; hence, booster vaccination has become necessary. In addition, infections caused by the different variants and correlated factors have not been discussed in depth. With large variabilities and different co-factors, it is difficult to use conventional mathematical models to forecast the incidence of COVID-19. Machine learning based on long short-term memory was applied to forecasting the time series of new daily positive cases (DPC), serious cases, hospitalized cases, and deaths. Data acquired from regions with high rates of vaccination, such as Israel, were blended with the current data of other regions in Japan such that the effect of vaccination was considered in efficient manner. The protection provided by symptomatic infection was also considered in terms of the population effectiveness of vaccination as well as the vaccination protection waning effect and ratio and infectivity of different viral variants. To represent changes in public behavior, public mobility and interactions through social media were also included in the analysis. Comparing the observed and estimated new DPC in Tel Aviv, Israel, the parameters characterizing vaccination effectiveness and the waning protection from infection were well estimated; the vaccination effectiveness of the second dose after 5 months and the third dose after two weeks from infection by the delta variant were 0.24 and 0.95, respectively. Using the extracted parameters regarding vaccination effectiveness, DPC in three major prefectures of Japan were replicated. The key factor influencing the prevention of COVID-19 transmission is the vaccination effectiveness at the population level, which considers the waning protection from vaccination rather than the percentage of fully vaccinated people. The threshold of the efficiency at the population level was estimated as 0.3 in Tel Aviv and 0.4 in Tokyo, Osaka, and Aichi. Moreover, a weighting scheme associated with infectivity results in more accurate forecasting by the infectivity model of viral variants. Results indicate that vaccination effectiveness and infectivity of viral variants are important factors in future forecasting of DPC. Moreover, this study demonstrate a feasible way to project the effect of vaccination using data obtained from other country.

Keywords: COVID-19; Deep learning; Forecasting; Vaccination effectiveness.

Graphical abstract

Fig. 1

Outline of proposed model for COVID-19 forecasting with vaccination effectiveness. (a) Initial forecasting is computed using a blend of time-series data; (b) the vaccination effect is computed using data acquired from different regions; and (c) the full model includes steps in (a) forecasting and (b) adaptation. Network detailed architecture is in Fig. 2.

Fig. 2

LSTM network architecture.

Fig. 3

Training and testing phases for the COVID-19 forecasting model. In training, different networks (A–E) are trained to forecast specific indicators. Long-term forecasting is achieved in the testing phase with concurrent data updates.

Fig. 4

Schematic explanation of the change in vaccination status with a sample population ($P$=8) over time. At $d=d_0$$N_1=N_2=N_3$=0. At $d_1$, $N_1$=3 and $N_2=N_3$=0, at $d_2$, $N_1=N_2$=2 and $N_3$=0. Finally, at $d_3$, $N_1$=1, $N_2$=3 and $N_3$=1.

Fig. 5

Map of Japan with study areas and regions used to represent downtown districts.

Fig. 6

(a) Vaccination effectiveness model (Eq. (1)) in Tel Aviv with different values of $s$ and $a_3$ along with DPC. (b) Forecasted DPC (7-day average) using different vaccination effectiveness models during the decay of the COVID-19 wave. (c) Detailed forecasted DPC data including the 95% confidence interval and associated vaccination effectiveness model. (d) Error associated with the forecasts using different vaccination effectiveness models.

Fig. 7

Example of a variant infectivity index computed using data of viral variant measures in Tokyo with associated weight values representing relative infectivity.

Fig. 8

Predicted DPC in Tokyo for the fifth wave with (a) all datasets (included in Table 1) and (b) optimized datasets (only values of 1-1, 2-4, 5-1, 6-1, 7-1 from Table 1) along with true values. Training data are from August 1, 2020 to July 30, 2021.

Fig. 9

Comparison of forecasting and adaptation models (shown in Fig. 1) for DPC in Tokyo. This demonstrates the effect of vaccination, which shows a weak spread phase and prolonged decay phase of the fifth COVID-19 wave.

Fig. 10

Forecasting of DPC, serious cases, and deaths in Tokyo, Osaka, and Aichi with considering optimized input data and both forecasting and adaptation models.