The influence of demographic and meteorological factors on temporal patterns of rotavirus infection in Dhaka, Bangladesh

Abstract

To quantify the potential impact of rotavirus vaccines and identify strategies to improve vaccine performance in Bangladesh, a better understanding of the drivers of pre-vaccination rotavirus patterns is required. We developed and fitted mathematical models to 23 years (1990-2012) of weekly rotavirus surveillance data from Dhaka with and without incorporating long-term and seasonal variation in the birth rate and meteorological factors. We performed external model validation using data between 2013 and 2019 from the regions of Dhaka and Matlab. The models showed good agreement with the observed age distribution of rotavirus cases and captured the observed shift in seasonal patterns of rotavirus hospitalizations from biannual to annual peaks. The declining long-term trend in the birth rate in Bangladesh was the key driver of the observed shift from biannual to annual winter rotavirus patterns. Meteorological indices were also important: a 1°C, 1% and 1 mm increase in diurnal temperature range, surface water presence and degree of wetness were associated with a 19%, 3.9% and 0.6% increase in the transmission rate, respectively. The model demonstrated reasonable predictions for both Dhaka and Matlab, and can be used to evaluate the impact of rotavirus vaccination in Bangladesh against changing patterns of disease incidence.

Keywords: birth rate; demography; meteorological factors; rotavirus seasonality; rotavirus transmission.

Figure 1.

Schematic of the mathematic model for rotavirus transmission dynamics. Model parameters are described in electronic supplementary material, table S2.

Figure 2.

Time series of observed rotavirus cases and demographic and meteorological model inputs. (a) Weekly rotavirus cases in Dhaka between 1990 and 2019. The vertical lines indicate the three time periods (part 1: 1990–2001; part 2: 2003–2012 and part 3: 2013–2019). (b) The local and global wavelet spectrum of the weekly rotavirus cases. For the local wavelet power spectrum, blue corresponds to lowest intensity and yellow to the highest intensity. (c) Stacked bar plot showing the seasonal distribution of rotavirus cases for the complete and sub-divided datasets. (d) Weekly interpolated birth rate, corresponding seasonality in the birth rate, and normalized meteorological indices used in the model: diurnal temperature range (dtr), surface water presence (wpre) and degree of wetness (dow).

Figure 3.

Comparison of model-simulated and observed rotavirus incidence in Dhaka for the complete dataset. (a) The weekly time series are plotted; the coloured lines correspond to the best-fit models, while the grey line corresponds to the observed data. (b) Comparison of the seasonal distribution of model-simulated rotavirus cases in part 1 and part 2. (c) Age distribution of rotavirus cases for the observed data (grey bars) and best-fit models (coloured bars).

Figure 4.

Comparison of models fitted to observed weekly rotavirus cases for during 1990–2001 and 2003–2012 and out-of-sample model validation for 2002. The observed (grey) and model-fitted (coloured lines) number of weekly rotavirus cases was plotted for (a) part 1 (1990–2001) and (b) part 2 (2003–2012) time series. (c) Observed and predicted weekly rotavirus cases in 2002 for the best-fit models (for part 1 and part 2). The comparison of the observed and model-fitted age distributions for part 1 and part 2 are in the electronic supplementary material, figure S1.

Figure 5.

Comparison of model-predicted and observed weekly rotavirus time-series for external model validation. (a) Out-of-sample model validation using data from Dhaka. (b) External model validation using data from Matlab. (c) Comparison of the best-fit model-simulated and observed seasonal distribution in rotavirus cases for both Dhaka and Matlab. In both instances, the best-fit model (model B_sbr) was used. (Online version in colour.)