ILI Forecasts for 2017-2018-Epidemic Week 48

This is week six of the CDC Flu Challenge.  During the previous epidemic week (EW 47) there was a significant slowdown in the increase of %ILI in Regions 1, 2, 7, and 8.  The %ILI stayed at about the same level in Regions 3 and 5 and it decreased in Regions 4, 9 and 10.  The only Region to exhibit a large increase in %ILI during EW 47 was Region 6. Consequently, the national %ILI profile (red line in top left panel below) showed only a minor increase during the past week.  This slowing/leveling/decrease in %ILI is often observed a week or two after the Thanksgiving holiday and is likely a result of the large increase in %ILI during the holiday week itself when people tend to go to the doctor only if they are truly sick.

Currently, four Regions (1, 4, 6 and 7) and the National have been at, or above their respective onset values for the past 1-4 weeks.  Our forecast for the national onset week is therefore peaked at EW 48 (second panel from left in top row below).  The Thanksgiving holiday ‘slowdown’ is often followed by a large increase in the number of ILI cases in subsequent weeks and this has guided our selection for the national (top left panel and bottom row panels) and Regional 4-week forward forecast. Given that the %ILI profiles for most Regions and the National are still significantly above their historic values for this time of year, our prediction for the national peak timing (second panel from right in top row) is about evenly split between an early (EW52) and late (EWs 6-7) times.


As noted above, Region 6 was the only region to have a large increase in the previous week and it has crossed its %ILI onset value (dashed grey horizontal line in top left panel below).  Our forecast for onset (second panel from left in top row) is therefore centered at the current epidemic week, with a minimal inclusion of the historic NULL model which allows for the possibility of ‘downward’ data update (and consequently a later onset week).   Given the elevated  %ILI values for this region we predict (second panel from right in top row)  that this region is more likely to peak early in the season (EW 51-52  ~50%) than later (EW 6-7 15%). Our forecast for the next four weeks (top left panel and bottom row) is based on a school vacation and specific humidity time-dependent force of infection, and a data augmentation procedure which uses %ILI data from previous seasons to inform the prior distribution used in the MCMC procedure.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s