Health & body · 2012
Could temperature and rainfall data warn of dengue outbreaks in Singapore 16 weeks ahead?
Yes—in a historical validation, a weather-based statistical model provided dengue outbreak warnings 16 weeks ahead in Singapore. Researchers compared different forecast lead times and found 16 weeks optimal, then tested forecasts from early 2011 through week 16 of 2012 using weather data alone. Performance in other places or future years was not established.
Details
Researchers developed a statistical forecasting model using Singapore data from 2000–2010, linking weekly average temperature and accumulated rainfall with reported dengue cases. They considered seasonal patterns, longer-term trends and relationships with earlier case counts. Comparing different delays between weather conditions and dengue cases, they found that 16 weeks was the optimal forecast period. They then validated the model by forecasting cases from week 1 of 2011 through week 16 of 2012 using weather data alone.
The study reported that the model distinguished outbreak from non-outbreak periods at 96% in 2004–2010, with an uncertainty range of 93–98%, and at 98% in 2011, with an uncertainty range of 95–100%. These ranges are the study’s reported confidence intervals. The model also accurately predicted the 2011 outbreak, with a reported false-alarm probability below 3%.
These historical results support the model’s potential to help plan outbreak prevention 16 weeks ahead. However, this was a forecasting study in Singapore, not a trial showing that acting on warnings reduced dengue cases. Performance outside the studied place and period was not established.
What do you think?
Good conversations begin with curiosity.
You can read comments without an account. Sign in to write.
Loading comments…