Thank you for your letter of 29 November.

I take your point about the reference being incorrect, rather than the figure attributed to it. I still think that the estimate of about a 1% dependence rate with benzodiazepines is damaging and misleading.

Since you last wrote I have been able to look at the WPA report, which you told me was an earlier version of the WHO (PSA) paper. Not surprisingly, the WPA task force gave no estimate of the incidence of BDZ dependence; indeed the WPA doesn't even cite the Khan paper, which would have been published at around the same time.

This suggests that WHO is responsible for the error, even if it is now not possible to establish how it occurred. That is of course a pity, because it means one cannot confidently exclude the possibility of some element of data fabrication. Certainly, the standards of verification seem poor.

Thank you for clarifying that your own involvement with the PSA was rather limited.