The variable impact of ENSO events on regional dengue/DHF in Indonesia

Although studies have demonstrated significant associations between ENSO events and dengue fever, few have explored regional impacts on dengue fever of separate events. This study explores the impacts of two ENSO events on regional patterns of dengue/dengue haemorrhagic fever (DHF) incidence in Indonesia. Data consist of monthly cases of dengue/DHF from 1992 to 2001 for each of Indonesia's 27 provinces, and monthly figures for rainfall, rainfall anomalies, temperature, relative humidity and the Southern Oscillation Index (SOI). We conducted Pearson correlation analyses for each independent variable against dengue/DHF incidence, using a direct month-by-month correlation and applying a lag of between one and six months to each variable with respect to dengue/DHF incidence. Based on the SOI value, we identified two ENSO events between 1992 and 2001. To explore each event, we created two dummy variables and in regression analyses for eight provinces. The variance of between 12.9 per cent and 24.5 per cent in provincial dengue/DHF incidence is explained by two or three climate variables in each of the provinces (p < 0.01 to 0.1). During the 1997/98 event, the explained variance increased by between 7 per cent and 15 per cent in provinces whose climate regimes were most affected by this event. This study demonstrates that indicators of ENSO such as the SOI may assist in the forecast of potential dengue/DHF incidence and distribution in Indonesia.