Counselling people before an HIV test has no effect on their future behaviour, says a study published in the Journal of the American Medical Association. For those of us who have worked in the field for a long time: this was actually good news. Some earlier small studies had suggested that people who are counselled take more risks than those who were tested without counselling.

Expect a spirited fight back from the AIDS industry, though. Many AIDS NGOs make a lot of money from various kinds of counselling and fight hard against efforts to make testing more widely available through family doctors and routine screening. Testing without counselling is, of course, quicker and much easier to administer.

Understanding why people test is not easy, though: take the forgotten story of one of the great commercial failures in over-the-counter diagnostic testing. About 20 years ago, Johnson & Johnson secured permission to sell an HIV test through retail outlets in the USA. The catch was that you had to mail in a piece of blotting paper with your blood sample. A few days later, you rang a toll free number to get the result. If the line went to a live person, it was bad news; if it went to a recording exhorting future safe behaviour, your test was negative. Despite lots of promotion, J&J never sold many of the kits (they were very expensive). Most interesting, though, was that most of those who had spent up to $100 buying the test did mail in their sample but never called for their results.

The world has moved on but laws still prevent the sale of simple, accurate instant OTC tests in most countries. These tests are pretty accurate and rarely give false negatives (although they do give some false positives). However, they are not perfect as the experience of the pornography industry shows. Outside a mutually monogamous relationship, there is no substitute for using condoms.