Mr J Dobbin
House of Commons
London
SW1A 0AA
Dear Mr Dobbin
You will no doubt remember that we wrote to you (during the summer recess) about the concerns that have been expressed about our report of specific analyses we had undertaken of the available national data on benzodiazepine and z-drug prescriptions. We have read and considered the criticisms, some of which we think can be addressed by extra analyses and some of which we believe do not relate to the data analyses we undertook and relate to the separate report from the NTA (with whose writing we were not involved) and the responsibilities on the Department of Health (which is obviously a separate matter).
This includes the question as to whether the exclusion of clobazam and clonazepam (because of clinical indication of epilepsy) had been correct: in our opinion there is probably value in doing the analyses both ways, and we are fairly certain that the analyses can be re-run with these two drugs included – and we are currently in discussion about this with Professor Ashton before we proceed. Other questions are more difficult to address such as the question of how many different people the annual number of prescriptions relates to: nevertheless a speculative calculation (assuming an average duration of prescription of one month, which may or may not be correct) can be estimated from Figure 2.19 and Table 2.15 on pages 30 and 31 of our report – and if we look at the situation from 2000 onwards as an illustration, with all sort of cautionary notes about such calculations, this shows a progressive slow annual decline of numbers of benzo prescriptions over this period from approx 12.9 million prescriptions in 2000 to approx 10.6 million in 2009, and our back-of-envelope calculations would mean that this equated to approx 1.1m actual people taking benzos at any point in time in 2000, reducing to approx 0.9m in 2009. But, over the same period and using the same assumptions, the number of z-drug prescriptions annually increased from approx 3m in 2000 to approx 5.7m in 2009 which, applying the above back-of-envelope calculation, indicates the number of people taking z-drugs increasing from an estimated quarter of a million in 2000 to nearly half a million in 2009. However these figures have too many assumptions around then – hence our recommendation for better analyses using the GPRD database (recommendation 1 in section 6 of our review), since this figure could be worked out more accurately through study of the GP Research Database (GPRD) (but which would need to be done as a new special study which we are not currently authorized or funded to undertake).
We are considering proposing to the Department of Health that there should be fuller study of the GP Research Database (the database where we were only able to give a demonstration of feasibility and potential value in the recent review). However this would only be sensible and feasible if we were sure that the right questions were being asked and if the resources and the data were available to allow the analyses. We have discussed this with Professor Malcolm Lader and also with Professor Heather Ashton, as well as with Professor Matt Hickman, and it is possible/probable that we will be suggesting further study of the GPRD database, specifically for a population to whom, over the years covered by the data, benzodiazepines and z-drugs have been prescribed. But before proceeding further down this pathway, we would usually need to be sure that the questions we were asking were both meaningful as well as scientifically valid. Professor Ashton has already contributed her own comments directly, which have been useful, as has Professor Lader — and these discussions are ongoing. However I am also aware that both you and Mr Perrott have identified issues and, if these are different from those identified by Professor Ashton, then it would be useful to explore these with you.
If appropriate, we would be pleased to meet with you to discuss this, and we will be guided by you as to whether it would also be useful for Mr Perrott to join in this discussion (perhaps by Skype or teleconference, if travel is not possible). We look forward to hearing from you.
Yours sincerely
John Strang and Kylie Reed