Reacting to partial intelligence is like diagnosing cancer without testing

It now seems that every government involved in the war on Iraq has begun to investigate the quality of the military intelligence which informed, or misinformed, its decision to launch the invasion.

I am no expert on military affairs; but during my career I have become familiar with medical testing and screening. I am also intrigued by both the making and breaking of codes, and the analysis and perception of risk. These other areas may hold some valuable lessons about the first.

Any military action will generate some retrospective analysis of the quality of the intelligence which preceded it. But politicians, and the public, should also be asking “What was the quality of the advice?” These are different questions which generate different answers.

Let me illustrate with a medical analogy.

An intelligence officer trying to find out what’s going on in some foreign and distant country shares many of the problems of the doctor who is trying to figure out what is going on inside a patient’s body. Both people must make important decisions on the basis of limited information, and both must assess the relative risks of a false positive and a false negative in making their diagnosis.

Imagine you have gone to your medical centre about a pain in your chest. Your doctor analyses the results of external examinations, scans, and laboratory tests. After the results are available, you are invited to come in for a chat. Your doctor tells you that the tests indicate that you might have a terminal cancer, but adds that the results are unclear, and it is in fact more likely to be something benign. The only way to be sure is to open you up for a closer look.

You dislike the idea of an operation, and ask for an assessment of the risk of it being the deadly cancer. Your doctor’s response is frank: “You would die – but at this stage we don’t believe this is a serious cancer.” You then ask about the risk associated with the operation itself, which the doctor assures you is “virtually nil”.

You are now facing the problem of uncertain “intelligence”. The health risk attached to a false negative (you reject the diagnosis and decide not to operate, when it is in fact a life-threatening cancer) is extremely high, while the risk of the false positive (you proceed to operate and find it’s benign) is very low.

What do you do?

I suggest that in these circumstances you would authorise the surgical “invasion”. You wake up from your operation to hear your surgeon confirm that the test was a false positive. You did not need the operation but you offer grateful thanks for “setting your mind at rest”, because it was worthwhile to learn that the feared tumour was never really there.

On the other hand, if your doctor had dismissed the test and not proceeded with the invasion of surgery, and as a result you died, your loved ones would probably sue.

Our newspapers are full of complaints about false positives. You hardly ever hear complaints about false negatives.