Annex F: A consistent series of drug-related death numbers, based on the classification at the end of the latest year covered by the publication

F1. The standard definition of a drug-related death that National Records of Scotland (NRS) uses for its statistics is set out in paragraph A2 of Annex A. Simplifying slightly, NRS counts a death as ‘drug-related’ if:
either (a) the underlying cause of death was coded to one of certain specified categories of mental and behavioural disorders due to psychoactive substance use
or (b) the underlying cause was coded to one of certain specified categories of poisoning (or self-poisoning) and a drug listed under the Misuse of Drugs Act (1971) was known to be present in the body at the time of death.

F2. Following the definition, a note at the end of paragraph A2 adds that: If a drug's legal status changes, NRS aims to count it on the basis of its classification on the day the person died … . For example, mephedrone was banned under the Misuse of Drugs Act with effect from 00.01 on 16 April 2010. Therefore, if mephedrone was the only drug found to be present in the body, a death coded to one of the categories listed under (b) would not be counted in NRS's implementation of the 'baseline' definition if it occurred before 16 April 2010. (Other notes explain why a few deaths in the specified categories are excluded.)

F3. As the 'mephedrone' example indicates, the requirement that a drug listed under the Misuse of Drugs Act must be present for a death to be counted as drug-related (under part [b] of the standard definition) means that whether NRS will count as drug-related a death from poisoning by a drug which is now controlled depends on when the death occurred: pre- or post-control. So the 'coverage' of NRS's standard definition 'widens' every time another drug is added to the list of controlled substances, because all subsequent deaths from poisoning by that drug will be counted as drug-related. In theory, this could cause a break in the continuity of NRS's figures for drug-related deaths (using the standard definition) every time that another drug becomes controlled.

F4. In practice, changes in the classification of drugs that occurred in the years up to and including 2013 had little effect on the figures: in that period, almost all the deaths which involved substances that were uncontrolled then, but are now controlled, also involved drugs that were already controlled, and so were counted as drug-related (in terms of the standard definition). For example, the foot of Table NPS2 (in the '… in 2013' edition of this publication) showed that almost all the deaths which involved New Psychoactive Substances (as defined for the purposes of that publication) were included in NRS's standard figures for drug-related deaths (in total, over the five years from 2009 to 2013, only 11 'NPS' deaths were not included in the standard figures). This is because (for example ) there were few 'mephedrone only' deaths before it was controlled; any deaths from (say) 'mephedrone and diazepam intoxication' were counted as drug-related because (say) diazepam was present.

F5. However, changes in the classification of drugs that occurred in 2014 could have caused a noticeable break in the continuity of NRS’s figures (based on the standard definition). Tramadol became a controlled substance with effect from 10 June 2014, along with some other substances. In 2013, there were over two dozen 'poisoning' deaths which involved only tramadol, or only tramadol and one or more other substances which were not controlled at that time. Using NRS's standard definition, such deaths (and those like them in the first part of 2014) are not counted as drug-related, but their equivalents from 10 June 2014 are counted as drug-related. So tramadol being controlled with effect from 10 June 2014 could have increased the number of deaths in 2014 counted as drug-related by a few percent (compared to what would have happened without that change), and there could, in due course, have been a similar effect on the figure for 2015 (because that will be the first year for which tramadol was controlled throughout). It follows that NRS’s standard figures could give a misleading impression of changes and any trends in drug-related deaths between 2013 and 2014, and between 2014 and 2015.

F6. Therefore, in order to give more accurate indications of changes and trends, NRS developed a 'consistent series' of numbers of drug-related deaths in previous years, which is based upon the classification of each substance at the end of the latest year covered by the publication. This 'consistent series' includes all the deaths involving tramadol, mephedrone and the other substances which have become controlled in recent years, regardless of their status at the time of death. It should show changes and trends which would be unaffected by the reclassification of substances. The consistent series goes back to 2000, as that is the first year of NRS’s current drug-related deaths database.

F7. For simplicity, the consistent series is based on the classification of drugs at the end of the latest year covered by the publication (rather than, say, at the time the publication was prepared), so it does not take account of any reclassifications after the final year for which the publication gives figures. The basis of the consistent series is therefore 'as at 31 December 2014' for the 'in 2014' edition, 'as at 31 December 2015' for the 'in 2015' edition, 'as at 31 December 2016' for the 'in 2016' edition, and so on. In consequence, the consistent series’ figures for previous years may be revised retrospectively every year, following more substances becoming controlled, if those substances had been involved in deaths (registered in earlier years) which had not been counted in the consistent series before because none of the substances involved were controlled at the end of the previous year.

F8. The consistent series appears in Table 1 in order to show the underlying trends for Scotland (comments on those figures can be found in Section 3.1). In addition, Tables CS1 and CS2 provide the consistent series’ numbers of 'extra' deaths in each year (that is the deaths which have been added retrospectively to the numbers that were originally produced using the standard definition), broken down by the names of the relevant drugs (that is the drugs for which the change in classification has caused deaths which were not counted as drug-related at the time to be included in the consistent series) and by sex and age-group. Finally, the numbers of ‘extra’ deaths counted in the consistent series for NHS Board areas appper in Table HB1, in order to show their scale (comments on those figures can be found in Section 4). The consistent series and the numbers of 'extra' deaths do not appear in any other tables, because a proliferation of additional figures could cause confusion – especially as the consistent series figures may, in theory, be revised every year (for the reason given in the previous paragraph).

F9. Table CS1 shows how the number of 'extra' deaths, based on the classification of drugs at the end of the latest year covered by this edition, varied from year to year. It should be noted that the total number of 'extra' deaths could be less than the sum of the figures for the individual drugs, due to deaths which involved more than one of the drugs. For example, a death in (say) 2013 for which the cause was given as 'tramadol and zopiclone intoxication' would be counted in the figures for both of those drugs, but only once in the total number of 'extra' deaths.

F10. The number of 'extra' deaths for 2014 (5) is not on the same basis as the figure for 2013 (30), because the figure for 2014 includes (for example ) ‘tramadol only’ deaths only for the period up to 9 June 2014 whereas the figure for 2013 includes such deaths for the whole of the year. 'Tramadol only' deaths in the rest of 2014 are included in the standard definition (and are therefore not counted as 'extra' deaths) because tramadol became a controlled substance with effect from 10 June 2014.

F11. The fact that the consistent series has only five 'extra' deaths for 2014 indicates that the drug classification changes in 2014 had less effect on the figures than one would have expected from the previous years' numbers of (for example .) 'tramadol only' deaths. With between 22 and 30 'extra' deaths in each of the previous five years, one would have expected a dozen or so between 1 January and 9 June 2014 (assuming that, say, 'tramadol only' deaths continued at the same rate, a dozen or so would be the 'pro rata' number for the part of 2014 for which they would not be counted in the standard definition). However, as it turned out, 1 January to 9 June 2014 had few (for example ) 'tramadol only' deaths, so the consistent series has only five 'extra' deaths for 2014. (It will be seen from Table Y that tramadol was implicated in, or potentially contributed to, the cause of 38 deaths in 2014: markedly fewer than the 64 in 2013. Note: these figures cover both 'tramadol only' deaths and those for which tramadol and one or more other drugs were implicated in, or potentially contributed to, the cause of death.)

F12. The table shows that a majority of the 'extra' deaths involved tramadol, and most of the rest involved zopiclone (which has also been controlled from 10 June 2014). Three 'extra' deaths involved mephedrone, with none after 2010 because it has been controlled from 16 April 2010; similarly, there were no 'extra' deaths involving phenazepam after it became controlled on 13 June 2012. A few of the 'extra' deaths involved other substances, not controlled at the time, which were controlled by the end of the period covered by this edition, but none of tramadol, zopiclone, mephedrone or phenazepam.

F13. It can be seen, from Table CS2 that women tend to account for a higher proportion of the 'extra' deaths than of the deaths which are counted in the standard definition: in some years, there were more 'extra' deaths of women than of men. The table also shows the number of 'extra' deaths in each of five age-groups: in some recent years, this has tended to be highest for the '55 and over' age-group (in contrast to the standard figures for drug-related deaths, which are much higher for '25-34', '35-44' and '45-54' than for '55 and over' – refer to Table 4).

F14. NRS data for the years 2000 to 2013 combined (which do not appear in a table) show that the vast majority of the 'extra' deaths which involved tramadol were of people who were aged 35 and over, and that, of all the age-groups, 55+ was the one which had the largest number (around a third) of the 'extra' deaths which involved tramadol. This was the case for both males and females. The position was broadly similar for the 'extra' deaths which involved zopiclone. The numbers of extra deaths involving other substances were too small for such analysis.