ONS - Defining alcohol-related deaths Summary of responses to

Document Sample
ONS - Defining alcohol-related deaths Summary of responses to Powered By Docstoc
					ONS - Defining alcohol-related deaths
Summary of responses to discussion paper.

Release on National Statistics website: 18 July 2006

In November 2005, ONS circulated a discussion paper on potential options for revising its
existing definition of alcohol-related deaths to individuals with relevant topic expertise in the
following organisations:

Alcohol Concern
British Liver Trust*
Department of Health
General Register Office for Scotland
Health and Social Care Information Centre
Imperial College
ISD Scotland
Northern Ireland Statistics and Research Agency
Preston Primary Care Trust
Welsh Assembly Government

*The British Liver Trust kindly arranged a coordinated response with the Royal College of Physicians
and the British Association for the Study of the Liver.

The discussion paper proposed a series of options for revising the definition currently used by ONS to
report on alcohol-related deaths:

1. Include causes of deaths with a clear causal relationship to alcohol consumption which
   have not previously been included in the ONS definition.
2. Include causes where some causal relationship with alcohol consumption has been
   demonstrated and there is enough evidence to support the inclusion of all deaths, or a
   proportion of all deaths, in the definition.
3. Report on all deaths where an alcohol-related cause is mentioned on the death certificate
   rather than just those where it is the underlying cause of death.

The paper concluded with a proposal that to monitor trends in alcohol-related deaths since 2001
ONS should continue to use its existing provisional definition using the Tenth Revision of the
International Classification of Disease (ICD-10) but with the addition of causes which are
clearly alcohol-related and had not previously been included. The discussion paper also
included a proposal to retain the definition which had been used in the Ninth Revision of the
ICD (ICD-9). It also included a proposal to continue monitoring these deaths using the
underlying cause, rather than mentions of any alcohol-related conditions.


1a. Pancreatitis - There was general agreement among respondents that causes such as
alcohol induced chronic pancreatitis, which were clearly alcohol-related but had not
previously been included in the ONS definition, should now be incorporated.

1b. Liver disease and cirrhosis- Concern was raised however by some respondents about
the inclusion by ONS of deaths from liver disease and cirrhosis when alcohol was not
mentioned on the death certificate. Such deaths would normally be coded to one of the
following ICD codes:

571 – Chronic liver disease and cirrhosis
    571.4 Chronic hepatitis
    571.5 Cirrhosis of liver without mention of alcohol
    571.6 Biliary cirrhosis
    571.8 Other chronic nonalcoholic liver disease
    571.9 Unspecified chronic liver disease without mention of alcohol

K73 - Chronic hepatitis, not elsewhere classified
    K73.0 Chronic persistent hepatitis, not elsewhere classified
    K73.1 Chronic lobular hepatitis, not elsewhere classified
    K73.2 Chronic active hepatitis, not elsewhere classified
    K73.8 Other chronic hepatitis, not elsewhere classified
    K73.9 Chronic hepatitis, unspecified

K74 - Fibrosis and cirrhosis of liver
    K74.0 Hepatic fibrosis
    K74.1 Hepatic sclerosis
    K74.2 Hepatic fibrosis with hepatic sclerosis
    K74.3 Primary biliary cirrhosis
    K74.4 Secondary biliary cirrhosis
    K74.5 Biliary cirrhosis, unspecified
    K74.6 Other and unspecified cirrhosis of liver

Very few of the deaths coded to these causes include a mention of alcohol on the death
certificate. However they are included in the current ONS definition on the assumption that in
England and Wales many of them were caused by alcohol, even if this was not explicitly
stated. Some respondents raised concerns regarding this assumption and recommended that
these codes should not be included in a definition of alcohol-related deaths.

The discussion paper detailed the reasons why ONS has traditionally included these deaths in
its definition, including the known under reporting of alcohol misuse on death certificates.
For most of the conditions listed there is a known relationship with alcohol. Evidence of how
valid the assumption that individual cases are really alcohol-related deaths appears scarce
however. The discussion paper referenced one study in England which for deaths from liver
disease of unspecified cause, case notes suggested that 67 per cent were the result of alcohol
misuse.1 This was based on a small sample however (66 deaths) between 1993-2000. A study
in the USA also reported that approximately 40% of deaths from unspecified liver disease in
the United States are attributable to heavy alcohol consumption.2

It is also possible that the percentage of these deaths which are alcohol-related will differ
between areas and will change over time as it will be affected by factors such as the
prevalence of viral hepatitis.

One condition in the above list however has no known link to alcohol, biliary cirrhosis, which
is an autoimmune disorder. One respondent raised a concern at the inclusion of these deaths.
The bridgecoded data from 1999 shows that there were 175 deaths where the underlying
cause was ICD-9 571.6 for that year. Using ICD-10 there were 190 deaths coded to K74.3-
K74.5 (the additional 15 almost all had an underlying cause of bronchopneumonia in ICD-9).
ONS does not consider that these should be retained in the count of alcohol-related deaths.

Removing all liver disease deaths where alcohol was not mentioned would have excluded
1,867 deaths in 1999 using ICD-9 and 1,781 using ICD-10 (34 per cent and 33 per cent
respectively of previously published figures for total alcohol-related deaths in 1999).3 For
2004, the most recent year, ONS would have reported 1,979 fewer alcohol-related deaths in
England and Wales (4,784 rather than 6,581). Given the uncertainty about the number of
cases in which alcohol may have contributed ONS considers that the current position should
be maintained. All liver disease deaths where alcohol is not mentioned (other than biliary
cirrhosis) should contribute to the count of alcohol-related mortality.

2. Attributable fractions - Several respondents acknowledged the difficulties that would be
involved in developing a definition that accurately reflected the contribution of alcohol to
other diseases, (causes where the relationship with alcohol is less clear) such as certain
cancers and stroke. Some respondents also noted that the current definition underestimates
the true burden of alcohol by excluding most external causes, in particular suicides, accidents
and sudden deaths.

One comment was also made on the emphasis on cirrhosis in the definition. (In 2004 61% of
deaths in the current ONS cause list were from alcoholic liver disease and a further 26% were
from fibrosis and cirrhosis of liver.) It was noted that this will lead to an inevitable time lag
between policy changes and mortality statistics. It was felt that a means of reporting on
alcohol-related deaths which included external causes would not only provide a more
accurate estimate of the number of deaths where alcohol was the single major factor but
would also provide a more timely mechanism of monitoring the potential impact of policy
changes on mortality.

Several respondents suggested that a review should be initiated to consider the feasibility of
developing alcohol-attributable aetiological fractions to address the issue of under-reporting.
Applied to consumption data these fractions could give a more complete estimate of alcohol-
related deaths.

ISD have informed ONS that as part of a project to develop aetiological fractions for use in
Scotland they plan to undertake a literature review of existing studies, such as those
published by English4 and Corrao.5 ONS have been invited to sit on the advisory group for
this undertaking and the feasibility of extrapolating the findings to elsewhere in the UK will
be examined as part of the project’s inception.

ONS accepts the benefits of establishing a method which would enable a truer picture of the
burden of alcohol misuse to be reported. It should be noted however that although

aetiological fractions have successfully been applied to smoking prevalence data to provide
estimates of smoking-related deaths (such as in the recent HDA report, The smoking epidemic
in England 6) the challenges of establishing prevalence of alcohol consumption are more

(A National Statistics methodology report details the problems in obtaining accurate
information on drinking through routine surveys7 and a further report highlights the fact that
surveys designed for different purposes may produce quite diverse estimates of alcohol
consumption.8 There may be a particular challenge in establishing a consistent method for
application across the UK, particularly as prevalence of alcohol consumption also varies
between areas and social groups.)

3. Mentions of alcohol-related diseases - There was support from some respondents for the
suggestion of basing statistics on deaths where an alcohol-related condition is mentioned
anywhere on the death certificate even if it is not the underlying cause. ISD also confirmed
that this is the method they currently use. For ONS however almost all analysis and reporting
of mortality data is based on the underlying cause of death as this is the disease or injury
which initiated the train of events directly leading to death, and is generally the most useful
single cause for public health purposes. We understand that the same principle is also true for
GRO-Scotland and NISRA.

Presenting trend data based on mentions would also be difficult for the reasons outlined in the
discussion paper with the change from ICD-9 to ICD-10. Long term trends would also be
limited to when each office started routinely coding every cause mentioned on the death
certificate. For ONS this was only from 1993 onwards.

Several respondents did however suggest that while it might not be appropriate to routinely
report mentions of alcohol-related causes, a periodic analysis of these deaths would
contribute to providing a truer picture of the burden of alcohol-related disease.

ONS supports the view that an initial analysis of deaths where an alcohol-related cause is
mentioned anywhere on the death certificate could produce a report which would usefully
provide additional information to routine reporting based on the underlying cause. ONS will
discuss the options for such an analysis project with colleagues in Scotland and Northern


As noted in the original discussion document ONS requires a single, statistically and
medically defensible definition, of alcohol-related mortality to monitor the public health
burden. There is a need to report on trends and geographic variations in these deaths, to
answer Parliamentary Questions, and to respond to media and other ad-hoc requests for
information. ONS also plans to report on trends and variations in alcohol-related deaths
across the UK. For this purpose it needs to agree a definition with devolved administrations.
To be able to provide reliable age and sex specific rates at local authority level, and
reasonably consistent trends over time, ONS believes that it is restricted to reporting on only
those deaths with a clear association with alcohol, as it has traditionally done.

Although some respondents did raise concerns about the inclusion of liver disease deaths
where alcohol was not explicitly mentioned in this definition, ONS believes that for the
purposes of monitoring public health the exclusion of these cases would lead to a serious
underestimate of drinking-related mortality. Respondents agreed though that additional
causes which are clearly alcohol -related should be added to the current definition in ICD-10.
ONS also proposes now that the definition should be further refined by excluding deaths from
biliary cirrhosis. The new ONS definitions of alcohol-related deaths in ICD-9 and ICD-10 are
included in the Annex below. For 2004 these revisions would mean that ONS reports a total
of 6,488 alcohol-related deaths rather than the 6,581 records using the provisional ICD-10

For routine reporting purposes, and for the planned UK paper on trends and geographic
variations, ONS proposes to include only deaths where an alcohol-related cause is the
underlying cause of death. It will however consider a future analysis project on deaths with
any mention of an alcohol-related cause and will discuss the options for reporting on these
deaths with colleagues in Scotland and Northern Ireland.

ONS have been invited to sit on the advisory group for the project which ISD is leading to
establish alcohol-attributed aetiological fractions for use in Scotland, and will be considering
how these findings could be applied elsewhere in the UK

We are grateful to the respondents who provided valuable comments on the discussion paper.
This has helped the process of improving future reporting of alcohol-related deaths.

Fisher N, Hanson J, Philips A, Rao J and Swarbrick E (2002). Mortality from liver disease in the West
Midlands, 1993-2000: observational study. BMJ 325, pp 312-313.

 Parrish K M, Dufour M C, Stinson F S, Harford TC. Average daily alcohol consumption during adult life
among decedents with and without cirrhosis: the 1986 National Mortality Follow-back Survey. J Stud Alcohol
Baker A and Rooney C (2003). Recent trends in alcohol-related mortality, and the impact of ICD-10 on the
monitoring of these deaths in England and Wales. Health Statistics Quarterly 17, pp 5-14
 English DR, Holman CDJ, Milne E, Winter MG, Hulse GK, Codde JP, et al. The quantification of
drug caused morbidity and mortality in Australia 1995. Canberra: Commonwealth Department of
Human Services and Health, 1995.
 Corrao G, Bagnardi V, Zambon A, Arico S. Exploring the dose-response relationship between
alcohol consumption and the risk of several alcohol-related conditions: a meta-analysis. Addiction
1999; 94: 1551-1573
 Twigg L, Moon G, Walker S. The smoking epidemic in England. Health Development Agency,
London, 2004.
 Goddard E. Obtaining information about drinking through surveys of the general population.
National Statistics Methodology Series No 24. ONS online report 2001:
 Goddard E and Compton S. Evaluation of the use of Household Survey data on spirits consumption
for estimating the level of spirit fraud. ONS online report 2005:


Causes of death related to alcohol consumption,
International Classification of Diseases, Ninth Revision

ICD-9 Code           Description

291                  Alcoholic psychoses
303                  Alcohol dependence syndrome
305.0                Non-dependent abuse of alcohol
425.5                Alcoholic cardiomyopathy
571                  Chronic liver disease and cirrhosis
  571.0          Alcoholic fatty liver
  571.1          Acute alcoholic hepatitis
  571.2          Alcoholic cirrhosis of liver
  571.3          Alcoholic liver damage, unspecified
  571.4          Chronic hepatitis
  571.5          Cirrhosis of liver without mention of alcohol
  571.8          Other chronic nonalcoholic liver disease
  571.9          Unspecified chronic liver disease without mention of alcohol
E860                  Accidental poisoning by alcohol

Causes of death related to alcohol consumption,
International Classification of Diseases, Tenth Revision

ICD-10 Code           Description

F10               Mental and behavioural disorders due to use of alcohol
G31.2              Degeneration of nervous system due to alcohol
G62.1              Alcoholic polyneuropathy
I42.6             Alcoholic cardiomyopathy
K29.2             Acoholic gastritis
K70               Alcoholic liver disease
K73               Chronic hepatitis, not elsewhere classified
K74               Fibrosis and cirrhosis of liver (Excluding K74.3-K74.5 – Biliary cirrhosis)
K86.0             Alcohol induced chronic pancreatitis
X45               Accidental poisoning by and exposure to alcohol
X65               Intentional self-poisoning by and exposure to alcohol
Y15               Poisoning by and exposure to alcohol, undetermined intent


Shared By:
Description: ONS - Defining alcohol-related deaths Summary of responses to