H1N1 Swine Flu Forecast: May 15 Update and Verification

.
So how did my forecasts do from last week ?

Shown below is the global cases with forecast dated May 10th along with the actual cases shown in orange. The global forecast did quite well, the actual were slightly below the forecast line but overall the forecast looks to picking up well the deceleration in the cases over time.




Below is the U.S. cases with the May 10th forecast in blue and the actual in orange. The U.S. actual cases were well below the forecast which kind of goes with the skepticism I had last week for this forecast. I continue to have a hard time forecasting the U.S. cases mostly likely due to the smaller sample size compared to the global cases.




Updated H1N1 Swine Flu Forecasts:

Below is the global swine flu cases forecast updated through May 15th. This forecast is very consistent with the forecast last week. This is good, you would like to see model continuity meaning that it is a good sign when a forecast model agrees with itself from the previous forecast, it tells you the forecast is consistently picking up on the underlying behavior in the data. The only difference is that the forecasted plateau is down slightly from last week but I would not read too much into that. It is still forecasting the global cases to plateau around 10,000 cases.



Now here is the U.S. cases forecast. It is similar to last week's forecast showing a continued increase beyond 10,000 soon which is not consistent with the global forecast above since if the U.S. model is correct and we see 10,000 U.S. cases then the global model can't plateau at 10,000. Note the uncertainty here by the confidence intervals which is very wide this is somewhat due to the different forecasting model that I needed to fit this data but all of that is due to the noise in the U.S. cases day to day.



So do I actually think that the H1N1 swine flu pandemic is really over?

Not necessarily, but maybe, for now... As I have mentioned a few times, what I am tracking is laboratory confirmed cases. That is good and bad. It is good because it is confirmed data rather than suspected or estimated which could be overcounted for a variety of reasons. It is bad because as any pandemic continues, the ability to laboratory test every case gets more and more difficult. So eventually, these confirmed numbers are sure to reach a limit of testing where the CDC and WHO simply would not be trying to test every suspected case.

It is easy for me to stick to the numbers and say that I am forecasting and tracking consistently what I said I would all along, the confirmed cases. But like everyone, I am curious whether the real pandemic is ending or not. I am really not sure, I am certainly not an infectious disease specialist. But one thing I have wondered is whether the swine flu numbers will fade during the north hemisphere summer since that is the typical low spot for flu infections and then come back in the fall and winter as the northern hemisphere gets back into flu season (I am not the only one to have had this idea of course). I think that might be what we are seeing in the flattening of these numbers. And of course any effort to get this strain into the fall flu vaccines may have a big impact on minimizing the risk of a widespread infection rate in the northern hemisphere.

Kinda makes you wonder, given the significant change in healthcare, our ability to detect a new strain like this early in its evolution, and then potentially get it into a vaccine, are we really as vulnerable as some have thought to a worldwide flu pandemic from a novel strain? Of course I have no clue, just speculating and being optimistic.
...