Drug-related deaths in Scotland in 2009

2 Data sources

2.1 The General Register Office for Scotland (GROS) holds details of all deaths which are registered in Scotland. By convention, deaths are counted on the basis of the calendar year in which they are registered rather than the year of occurrence (as the latter might not be known). GROS closes its database for a calendar year around the end of the following June, so the statistics for 2009 are based upon the information which GROS had obtained by June 2010. GROS classifies the underlying cause of each death using International Classification of Diseases (ICD) codes, based on what appears in the medical certificate of the cause of death together with any additional information which is provided subsequently by (e.g.) certifying doctors, the Crown Office, pathologists or Procurators Fiscal.

2.2 Drug-related deaths are identified using details from the death registrations supplemented by information from a specially-designed questionnaire, which is completed by forensic pathologists. GROS requests this information for all deaths involving drugs or persons known, or suspected, to be drug-dependent. Additionally, GROS follows up all cases of deaths of people where the information on the death certificate is vague or suggests that there might be a background of drug abuse. A copy of the questionnaire used with effect from 2008 is in Annex D. This enhancement to the data collection system was described in a paper published by GROS in June 1995 (see Annex C: References).

2.3 The questionnaire was revised for 2008, in order to collect more complete information about the substances present in the body. This caused a break in the series of figures for "drugs reported". The discontinuity arose because:

pre-2008, the form asked about the "principal drug or solvent found in a fatal dose" and about "any other drugs or solvents involved in this death" - so some pathologists reported only the substances which, they believed, contributed directly to each death;
now the form asks about the drugs or solvents "implicated in, or which potentially contributed to, the cause of death" and about "any other[s] … which were present, but which were not considered to have had any direct contribution to this death" - so some pathologists now report substances which they would not have mentioned previously.

2.4 When GROS received the completed questionnaires for 2008 deaths, it simply recorded all the drugs which had been reported, without making any distinction between (a) drugs or solvents implicated in, or which potentially contributed to, the cause of death and (b) any other drugs which were present, but which were not considered to have had any direct contribution to this death. So, when GROS produced the "drugs reported" figures for 2008 that were published in August 2009, it did so by counting all the drugs which had been reported as having been found to be present in the body. It is thought that the change in the information collected using the questionnaires accounted for most (if not all) of the apparent large increases, between 2007 and 2008, in the figures for (e.g.) benzodiazepines, diazepam and alcohol that were published in August 2009.

2.5 At its meeting in September 2009, the National Forum on Drug-related Deaths discussed the basis of GROS's figures for deaths for which particular drugs were reported. GROS considered the comments that were made, and prepared proposals for changing its method of producing these statistics, which were put to the National Forum's meeting in February 2010, revised in the light of the views expressed there, and subsequently implemented when GROS produced the figures given in this edition of the publication. These are the main points to note about the change in the basis of GROS's figures for deaths involving particular drugs:

"Drugs which were implicated in, or which potentially contributed to, the cause of death" is now the standard basis for the figures for 2008 onwards that GROS produces for individual drugs.
GROS reprocessed the data from the questionnaires for 2008, in order to distinguish between (a) drugs which were implicated in, or which potentially contributed to, the cause of death and (b) any other drugs which were present, but which were not considered to have had any direct contribution to the death. This distinction is also made when processing the questionnaires for 2009 onwards.
As a result, GROS can now produce figures for 2008 onwards:

(i) on the new standard basis - i.e. counting only drugs which were reported under (a); and
(ii) on the "all drugs which were found to be present in the body" basis - i.e. covering drugs which were reported under either (a) or (b).

2.6 It should be noted that:

this change has not affected the overall total number of drug-related deaths - it has just reduced the number of drugs which are counted, for the purpose of the standard figures, for some deaths (e.g. the change reduces markedly the figure for "benzodiazepine deaths" for 2008); and
while the change has reduced the size of the discontinuity, between 2007 and 2008, in the figures for individual drugs, there is still a break in the series due to the introduction of the new questionnaire. This is because, in 2007 and earlier years, some pathologists reported, in the old questionnaire, all the drugs that they found (i.e. not just the drugs that they believed were implicated in, or contributed to, the cause of death) - so they provided information on the "all drugs which were found to be present in the body" basis (i.e. not on the new standard basis). GROS cannot produce figures for 2007 (or any earlier year) on the new standard basis unless pathologists complete copies of the new questionnaire in respect of the relevant deaths, which is impractical: pathologists would have to refer to old records, which may no longer exist, or which might not contain the information that the pathologists would need to make the distinction between (a) drugs which were implicated in, or which potentially contributed to, the cause of death and (b) any other drugs which were present, but which were not considered to have had any direct contribution to the death.

2.7 Following the change in the standard basis of GROS's figures for individual drugs:

in this publication, Tables 3 and Y provide figures for both 2008 and 2009 on the new standard basis, Tables HB3 and C3 give figures for 2009 on the new standard basis, and Table 6 shows the figures for 2009 on the two bases;
revised versions of Tables HB3 and C3, giving figures for 2008 on the new standard basis, are available from the GROS web site; and
alternative versions of Tables HB3 and C3, giving figures for 2009 on the "all drugs which were found to be present in the body" basis, are available on the GROS web site (which makes it clear that they are not on the standard basis).

2.8 The statistics of drug-related deaths may be affected by other differences, between years and/or between areas, in the way in which the information was produced. For example:

technical advances may enable the detection of small quantities of substances that could not have been found in the post-mortems that were performed several years ago;
the range of substances for which tests are conducted may change - e.g. for a number of years, a laboratory did not routinely test for the presence of cannabis (because the view was that, in general, it did not contribute to causing deaths), but now does so more often, because Procurators Fiscal are now more likely to want to know whether the deceased had been using it. More generally, advice is that there is a demand to obtain more complete and thorough toxicology on all cases tested for drugs, which includes fuller examinations for, and hence a greater possibility of finding, more drugs; and
pathologists in one area report any findings of benzodiazepines by referring to that group of drugs unless they are sure that only one particular benzodiazepine (e.g. diazepam) was used, so the areas which they serve appear to have low proportions of deaths for which diazepam is mentioned (compared to areas where diazepam is more likely to be reported, and where there are proportionately fewer reports of benzodiazepines as a group).

previous | contents | next