Effect Measure is worried:
A nagging question about H5N1 infection in Vietnam is whether there is considerably more mild, clinically undetected or inapparent infection than accounted for by only considering the most seriously ill cases admitted to the hospital. Since the most seriously ill are the tip of the iceberg in most human viral diseases this is a plausible concern. Its significance would be that there is more transmission of the virus either from poultry to humans or from human to human than has been conceded at present.
Equivocal evidence to suggest this has apparently been obtained by a joint Vietnamese-Canadian research team working in Hanoi. Using a method called a Western blot, the researchers tested hundreds of stored blood samples to see if they showed evidence of antibodies to H5N1. Rumors are that "scores of samples came back positive."
Recombinomics is yet more pessimistic and suggests that phase six has probably started and only awaits confirmation:
Investigators suspect that the human-to-human transmission is extensive and sustained, signaling phase 6. However, most of the cases in northern and central Vietnam are mild relative to the H5N1 bird flu cases in the south, so the announcement of the large number of positives has been delayed for months as more excuse for not reporting the data are made.
Neither of these reports should be considered final, but yes they do cause me to revise my priors in an unfortunate direction. And let us repeat the call for the WHO to be more forthright, transparent, and accountable than it has been proving to date.