Friday, June 09, 2006

Reader beware

When you read news reports, you have to be circumspect. There are some excellent reporters on flu but most don't know much about it and have a tendency to transcribe whatever some official spokesperson says. Even large news organizations depend on local stringers who send in bits and pieces they think might be of interest.

A good example comes this morning in the form of two short filler pieces, one from Associated Press (AP) and one from Bloomberg News Service. Both report the failure of a WHO reference lab (neither tells us which one) to confirm H5N1 in a seven year old girl who died June 1 near Jakarta. Her ten year old brother had died three days earlier with similar symptoms but no specimens were obtained. Contact with sick poultry was alleged. "Local tests" of the little girl were positive for H5N1. Who does the local tests is not stated, but it is my impression the US Naval Lab NAMRU2 is involved and they are technically expert and all previous local tests have been confirmed for H5N1 by WHO reference laboratories. Apparently that isn't the case this time.

Here is the AP version, run under the headline,"Indonesian girl's flu death questioned," (RSS headline, "WHO lab indicates Indonesian girl did not die of bird flu"). The piece cited the Indonesian Ministry of Public Health:
A lab approved by the World Health Organization said a seven-year-old Indonesian girl who tested positive locally for bird flu did not have the virus, a senior Health Ministry official said Friday.


"This is the first time local tests came back positive and Hong Kong laboratory tests negative," said Kandun, adding WHO needs to carry out new tests to reconfirm its findings. (AP via CANOE Network, Canada)
Believing the Indonesian Ministry of Health on bird flu these days is like believing Donald Rumsfeld on torture.

Here is the Bloomberg version, run under the headline "Bird flu tests are inconclusive on Indonesian girl: WHO":
Confirmatory tests for bird flu on a 7-year-old girl in Indonesia were inconclusive and more specimens will need to be tested, a World Health Organization spokeswoman said Friday.

"Test results are pending further analysis, and more samples are to be collected" for testing by a laboratory in Hong Kong, said Sari Setiogi, a spokeswoman with the WHO in Jakarta. (Bloomberg via Jakarta Post)
In this case it is pretty clear which is the more accurate report. But if you'd only seen the AP report and didn't read it carefully (it is only a small filler piece) you could easily have gotten the impression that WHO had determined the Jakarta case was not bird flu. So far, that hasn't happened (although it may). There are many reasons why a truly positive test may be inconclusive.

The internet brings us information of epidemiologic interest at unprecedented speed nowadays. But the usual cautions apply: caveat lector.