by revere, cross-posted from Effect Measure
Just a brief note to remind everyone about the case definitions CDC is using for reporting on swine flu (or whatever name we collectively settle on). In order to make sure numbers are comparable from day to day and place to place we have to decide on criteria for knowing we have something to count. Is someone with flu-like symptoms to be counted as a case? Or do we confine it to someone with laboratory proved infection with the virus? Should there be different categories of diagnostic certainty?
For the moment, CDC is using the following definitions for suspected, probable and confirmed cases of swine flu.
To be considered under any of these categories, the person must first have an acute febrile respiratory illness. The fever has to be at least 100 degrees F. (that’s the febrile part) and accompanied by recent onset (the acute part) of at last one of these respiratory tract symptoms: rhinorrhea (runny nose) or nasal congestion, sore throat, cough (these, obviously, are the respiratory illness components of the definition).
Having an acute febrile illness at this juncture is not enough to put you into the biggest category, “suspect case.” You also have to have been in close contact with a confirmed cases within the last 7 days (the upper range of the incubation period) or have traveled in the last 7 days (in the US or internationally) someplace where there have been confirmed cases, or now, live in a US community that has cases. These are people who might be infected with this virus or could have another reason for their symptoms. Suspect cases vastly outnumber true cases at the beginning of any outbreak, and that’s true this time as well.
The index of suspicion gets ratcheted up when a suspect case is shown to be shedding influenza A virus through one of the rapid antigen tests available in a doctor’s office. These tests don’t detect all flu A, however. They are said to be about 50% sensitive, meaning they miss about half of true flu A infections. Their main use was as a way to detect the presence of flu A (or flu B) in the community. Using it as a test for an individual, as here, is not the best way to make a diagnosis, but the number of errors that will be made will be a function of the prevalence of flu A among those with acute febrile respiratory illnesses at any point in time. If they are positive by this test, the swab is supposed to be sent to a state-level public health laboratory for subtyping. If it cannot be subtyped at this level, it is sent to CDC for a specific test for the current swine flu virus involved in the outbreak. From the moment a person is flu A positive to the final disposition either at the state lab or CDC, the person is a probable case (probable pending laboratory confirmation).
Specimens that are shown to contain the genetic material of the current swine flu virus (by RT-PCR) or that yield the virus when cultured in cell culture are the tip of the diagnostic pyramid, the confirmed cases.
Probable cases are popping up all over and I wouldn’t be surprised if CDC’s lab started to back up. That will mean that decisions about closing schools or businesses will be waiting for final confirmation. The decision makers are in a difficult position, because shutting schools at the end of a school year, with proms, SAT and final exams and who knows what else creates extremely difficult and painful problems. If this (so far) mild influenza peters out they will be accused of panicking and over reacting. But if this outbreak builds up a real head of steam, the consequences could be a great deal of illness, discomfort and some deaths and almost certainly medical facilities will be overwhelmed.
There is an old adage: when public health works, nothing happens. In this case, the purpose of closing schools and canceling events is to prevent further transmission of disease. Experience indicates that the earlier social distancing and other similar interventions are applied, the more likely there will be a good effect. In this case, success will mean “nothing happens,” which is what we want.
Unfortunately it is more than likely that the decision makers who helped bring it about will be blamed for doing what needs to be done.