Retrospective analysis assessing the spatial and temporal distribution of paediatric acute respiratory tract infections in Ho Chi Minh City, Vietnam

Abstract

Background: Acute respiratory tract infections (ARIs) are the leading cause of morbidity and mortality in young children in low/middle-income countries. Using routine hospital data, we aimed to examine the spatial distribution, temporal trends and climatic risk factors of paediatric ARIs in Vietnam.

Methods: Data from hospitalised paediatric (<16 years) patients with ARIs residing in Ho Chi Minh City (HCMC) between 2005 and 2010 were retrieved from the two main Children's Hospitals and the Hospital for Tropical Diseases in HCMC. Spatial mapping and time series analysis were performed after disaggregating data into upper respiratory tract infections (URIs) and lower respiratory tract infections (LRIs).

Results: Over the study period, there were 155 999 paediatric patients admitted with ARIs (33% of all hospital admissions). There were 68 120 URIs (14%) and 87 879 LRIs (19%). The most common diagnoses were acute pharyngitis (28% of all ARI), pneumonia (21%), bronchitis (18%) and bronchiolitis (16%). A significant increasing trend over time was found for both URIs (mean weekly incidence per 1000 population, I=3.12), incidence rate ratio for 1-week increase in time (RR 1.0, 95% CI 1.02 to 1.17) for URI and (I=4.02, RR 1.08 (95% CI 1.006 to 1.16)) for LRI. The weekly URI incidence peaked in May-June and was significantly associated with lags in weekly URI incidence and the average humidity, rainfall and water level. The weekly LRI incidence exhibited significant seasonality (P<0.0001), with an annual peak in September-October and was significantly associated with lags in weekly LRI incidence and lags in weekly average temperature, rainfall and water level.

Conclusions: ARIs are a leading cause of childhood hospitalisation in HCMC, Vietnam. The incidence of ARIs was higher in the wet season and in specific HCMC districts. These results may guide health authorities in where and when to effectively allocate resources for the prevention and control of ARIs.

Keywords: acute respiratory infections; lower respiratory infections; pediatrics; spatial distribution; temporal patterns; upper respiratory infections.

Figure 1

Maps for empirical Bayesian estimated incidence rates of acute respiratory tract infections (ARIs) and other diseases at ward level: (A) ward-level empirical Bayesian estimated incidence rates (EBR) per 1000 population for ARIs with locations of the three study hospitals in green, (B) ward-level EBR for other diseases rather than ARIs. District names and district boundaries (grey) are shown. *Rural districts. CH1, Children Hospital 1; CH2, Children Hospital 2; HTD, Hospital for Tropical Diseases.

Figure 2

Plots of the city-wide weekly incidence time series for upper respiratory tract infections (URIs) and lower respiratory tract infections (LRIs) from the generalised additive mixed model (GAMM). (A) Smoothed seasonal cycles (red line), time trend (blue line) from the basic Poisson GAMM model, fitted values (green line) from the full Poisson GAMM model and observed values (black line) of city-wide weekly incidence (number of cases per 1000 population) of URIs. (B) Smoothed seasonal cycles (red line), time trend (blue line) from the basic Poisson GAMM model and fitted values (green line) from the full Poisson GAMM model and observed values (black line) of city-wide weekly incidence of LRIs.

Figure 3

Time series of city-wide weekly incidence of upper respiratory tract infections (URIs) and lower respiratory tract infections (LRIs) versus weekly climatic covariates. Left panel: time series of weekly URI and LRI incidence and weekly average climatic covariates. Right panel: scatter plots with fitted lines and 95% CI from univariate Poisson generalised linear model for weekly URI or LRI incidence versus each climatic covariate.