The first, "alternative explanation" the statistician in me wants to mention is perhaps the way influenza data is gathered has changed.
"Only people who get tested for influenzalike illnesses—typically about 5 percent of individuals who fall ill—are tallied."
If there were a change in the way these tests were administered, for example a blaring medical bias toward another disease, that would present a significant sampling problem. That same statistician also wanted me to mention the simplest and most boring answers are usually closest to the truth.
"Only people who get tested for influenzalike illnesses—typically about 5 percent of individuals who fall ill—are tallied."
If there were a change in the way these tests were administered, for example a blaring medical bias toward another disease, that would present a significant sampling problem. That same statistician also wanted me to mention the simplest and most boring answers are usually closest to the truth.