Annex G: Drug-related Deaths – comparison with other countries
G1. This Annex uses figures for the latest year (at the time of writing) for which other
countries’ statistics were available from a European Monitoring Centre for Drugs and
Drug Addiction (EMCDDA) report. It explains that Scotland’s drug-related death
figures imply a drug-death rate (relative to the number of people aged 15 to 64,
inclusive) which is higher than those that have been reported by all the EU countries
(although a footnote in the EMCDDA’s table states that caution is required when
comparing drug-induced deaths due to issues of coding, coverage and
under-reporting in some countries). This Annex concludes by showing that the
normally-published figures for Scotland imply a drug-death rate (relative to the size of
the population of all ages) roughly two and a half times that of the UK as a whole.
G2. When using the EMCDDA’s figures for other countries, it must be remembered that
the EMCDDA stated that difference in the national practices of coding the causes of deaths implies that
direct comparisons between countries in the numbers or rates of DRDs should
be made with caution (in the ‘Limitations’ section of its note on ‘Methods and Definitions’, which is available
on the EMCDDA’s website).
G3. That note gives some examples of differences between countries, such as (Note: the
points in square brackets have been added by NRS):
… differences in which codes are applied. In particular, in some countries ‘T’
codes [which, for deaths from poisoning, identify the types of substances that
were involved] are never or rarely used, whereas in others they are more
frequently used. Where ‘T’ codes are not applied, the number of drug-induced
deaths [refer to paragraph G5] would be an underestimate;
… differences between countries in procedures for recording cases, and in the
frequency of post-mortem toxicological investigations; and
information exchange between General Mortality Registers [GMRs, such as
NRS] and Special Registers (forensic or police) is insufficient or lacking in some
countries, which may compromise the completeness of the information.
G4. It has been suggested that better identification and recording of such deaths may be a
reason for the drug-death rate appearing to be higher in the UK (and, hence,
Scotland) than in several other countries. For example:
NRS normally allocates a ‘T’ code for every substance that was reported as
being present in the deceased’s body - so there should be very little (if any)
under-estimation in the figures for Scotland. NRS understands that the UK’s
other GMRs also make good use of ‘T’ codes, so the UK’s figures should not be
underestimates;
Scotland has a good exchange of information, as forensic pathologists provide
NRS with details of many drug-deaths (using the form which is shown in Annex
D) - so the data for Scotland should be more-or-less complete. NRS understands
that the UK’s other GMRs are usually told, by coroners, which drugs caused
each death, so (again) the UK’s figures should not be underestimates.
G5. Table EMCDDA gives the number of ‘drug-induced’ deaths aged 15 to 64 inclusive,
and the resulting rate per million population of that age, for various countries. These
numbers were copied from Table A6 of the EMCDDA’s ‘European Drug Report 2017’,
available on the EMCDDA website. ‘Drug-induced deaths’ is the EMCDDA’s term for
deaths directly caused by illegal drugs, which it defines in terms of particular codes for
the underlying causes of death, in some cases in combination with certain codes for the types of substance involved. The EU countries are listed in order of the native language versions of their names - for example, Germany appears between Denmark
and Estonia; and Austria is between Netherlands and Poland. Although the report
refers to the latest statistics being for 2015, some countries’ data are for earlier years,
and Public Health England (which supplies the EMCDDA with the statistics for the UK)
has confirmed that the UK’s figure in the EMCDDA’s Table A6 is (broadly speaking)
the number of such deaths which occurred in 2014. (The Scottish component of the
UK figure is the number of such deaths that were registered in Scotland in that year.
Because deaths in Scotland are normally registered within a few days, the number
that were registered in Scotland in any given year will be similar to the number that
occurred in Scotland in that year.)
G6. The corresponding figures for Scotland for 2014 have been added at the foot of the
table. They were produced as follows:
563 drug-induced deaths (using the EMCDDA definition) aged 15 to 64 inclusive
were registered in that year – extracted from NRS’s drug-related deaths
database. This is slightly fewer than the 574 deaths on the basis of the EMCDDA
‘general mortality register’ definition (shown in Table X), because the latter figure
includes deaths at ages 0-14 and 65+;
the drug-induced death rate (aged 15-64) per million population was then
calculated by dividing the 563 drug-induced deaths aged 15-64 by the
corresponding mid-year population estimate, of 3,526,673.
G7. The resulting drug-induced death rate (aged 15-64) for Scotland is 160 per million
population. This appears to be higher than for any of the countries shown in the
EMCDDA table. The next highest rates are for Estonia (103 per million) and Sweden
(100 per million). Scotland’s drug-induced death rate is much higher than that of the
UK as a whole (60 per million) – but that is what one would expect, given Scotland’s
share of the UK’s drug-related/’misuse’ deaths (refer to paragraphs G9 and G10,
below).
G8. It must be remembered that the figures for some countries may not be truly
comparable with those for Scotland (or the UK as a whole), for reasons like those
given in paragraphs G2 to G4. Table A6 in the the EMCDDA’s ‘European Drug Report
2017’ includes the following footnote:
"Caution is required when comparing drug-induced deaths due to issues of
coding, coverage and under-reporting in some countries"
Because some countries’ figures may be affected by (say) under-reporting, one
cannot say that Scotland has a drug-induced death rate (aged 15-64) which is
definitely ‘X’ times the level for the EU as a whole, or higher than that of exactly ‘Y’ EU
countries. However, it appears certain that Scotland’s rate is well above the level of
most (if not all) of the EU countries.
G9. Scotland’s drug-related death rate is also seen to be much higher than that of the UK
as a whole when the comparison uses the kind of drug-death figures that are normally
published for Scotland, England and Wales, and Northern Ireland. As an example, in
terms of the definition that is used for most of the statistics in this report (that
introduced in 2001 for the ‘baseline’ figures for the UK Drugs Strategy), the following
numbers of drug-related deaths were registered in 2014:
614 in Scotland – the ‘standard definition’ figure in Table 1;
2,248 in England and Wales – ‘drug misuse’ deaths (that being ONS’s term for
the number of deaths based on the ‘Drug Strategy’ definition) – more information can be found in the deaths related to drug misuse section on the ONS website;
and
So, the UK had a total of 2,950 drug-related/’misuse’ deaths registered in 2014, of
which around 21% were registered in Scotland. As Scotland accounts for only about
8% of the population of the UK, Scotland’s drug-death rate (per head of population)
appeared to be roughly two and a half times that of the UK as a whole.
G10. It should be noted that how information about drug-related/’misuse’ deaths is collected
differs between Scotland and other parts of the UK. In particular:
in England and Wales, almost all drug-related deaths are certified by a coroner
following an inquest, and cannot be registered until that is completed. As a result,
about half of their drug-related deaths registered in (say) 2014 occurred in a
previous year – more information can be found in the Impact of registration
delays on drug-related deaths section of the ONS website. Very crudely, the
England and Wales figures for (say) 2014 can be thought of as representing the
deaths which occurred between (say) mid-2013 and mid-2014, so are less ‘up to
date’ than Scottish figures for 2014, which can be thought of as representing the
deaths which occurred in the whole of that year (as almost all Scottish deaths are
registered within a few days of occurring);
there is no English equivalent of the form (shown in Annex D) which is used by
forensic pathologists in Scotland to provide details of deaths to NRS.
The UK’s other GMRs are usually told, by coroners, which drugs caused
each death, but not about all the substances that were found in the body. It
follows that some deaths could (in theory) be counted differently in, say,
Scotland and England. For example, a death from intentional self-poisoning
by an uncontrolled substance would be counted in Scotland (but not in
England) if a controlled substance was present in the body but was not
believed to have contributed to the death (because the presence of the
controlled substance would not be recorded in the data for England)
NRS is more likely than ONS to be told which drugs caused a death. In
Spring 2017, ONS said that:
in around 1 in 8 cases, it receives only a very generic description of
the death, such as ‘drug overdose’ or ‘drug-related death’. In contrast,
Scotland had only about 8 drug-deaths per year (on average, from
2008 to 2015) for which NRS was not told which drugs caused them.
in around 10% of opiate deaths, ONS is not told which opiate was
involved. In contrast, Scotland had an average of only about 4
drug-deaths per year caused by opiates (possibly in combination with
other substances) for which NRS was not told which particular opiates
were involved.
Such differences may affect the comparability of drug-death rates for Scotland and the
UK as a whole, but are unlikely to account for the majority of the difference between
those rates. For example:
if the numbers of drug-related deaths were rising at 10% per year, their being
registered (on average) six months earlier in Scotland than in England would
increase the Scottish drug-death rate by only 5% (relative to the English one), all
else being equal (because one would be comparing the Scottish number of
deaths which occurred, broadly speaking, in [say] 2014 with the English number
of deaths which occurred, broadly speaking, between [say] mid-2013 and
mid-2014 – a period when drug-death rates were lower).
on average, Scotland had only around 3 deaths per year from intentional
self-poisoning by an uncontrolled substance for which a controlled substance
was present in the body but was not believed to have contributed to the death.
Such deaths are included in the drug-related death figures for Scotland, but not
for England – but are too few in number to have much effect on the comparability
of drug-death rates.
‘drug overdose’ and ‘opiate’ deaths in England are counted as
drug-related/’misuse’ deaths, so the lack of information about which drugs were
involved does not affect the comparability of the overall drug-death rates.
(However, it could have a noticeable effect on any comparison of figures for
deaths which were caused by particular drugs, of course.)
It follows that the Scottish rate could well be at least double that of the UK as a whole
even if there were no methodological differences.