I read with interest the news story, 'Schizophrenia patients face denial of new drug treatments' (page 8, 21 October), based on a leaked draft document by the national schizophrenia guidelines group.

The document is a major scientific analysis of the overall effectiveness of the new 'atypical' drugs for schizophrenia, which takes account of all the published and unpublished trials we could find.

All the new atypical drugs were compared with one older drug, haloperidol, known to produce a lot of side-effects on movement, so the comparison is loaded in favour of the new drugs. Each atypical drug was better in terms of side-effects and tolerability, but in many studies the dose of haloperidol used was very high. We therefore controlled for the dose of haloperidol used, and the advantages of the new drugs then appeared much less impressive than we expected.

Our work's significance is to make clear that the older drugs are used in doses that are too high. These doses are less effective in treating the disorder, and produce a lot of side-effects. We agree that movement side-effects are to be avoided. But we do think it is still reasonable to use low doses of an older type of drug as a first-line treatment.

These drugs suit quite a lot of people if properly used. Obviously, if they show the slightest signs of movement side-effects, or if the medication appears not to be working, that is the time to switch to the new drugs. Our approach is partly based on the belief that new drugs need a period of assessment in ordinary practice because they often have side-effects that do not become immediately apparent.

Our argument has nothing to do with cost but with caution. To pretend it is a cost issue may not be to the advantage of people suffering from this awful disorder.

Professor Paul Bebbington Co-chair National Schizophrenia Guidelines Group