Understanding alcohol-related hospital admissions

Alcohol is England’s second biggest cause of premature deaths behind tobacco. 34 per cent of men and 28 per cent of women exceeded current consumption guidelines on at least one day in the last week. Public Health England, in partnership with the Department of Health and the Health and Social Care Information Centre, recently announced that the current alcohol-related hospital admissions indicator will be supplemented by a new indicator to be published in the Public Health Outcomes Framework. So, why do we need an additional indicator, what does it measure, and how do the two indicators differ? We will try and explain.

Alcohol-related hospital admissions are used as a way of understanding the impact of alcohol on the health of a population. The current indicator reports that the number of alcohol-related hospital admissions in England is about one million per year and has been steadily rising. An increase in alcohol harm has been observed over the last decade, but we have also become better at understanding and recording its’ impact. To reflect these changes, it has been decided that the presentation of two indicators (a broad measure and a narrow measure) will give us a more comprehensive picture of the contribution of alcohol to ill-health. To explain how the indicators differ, it is important to understand the two essential ingredients for measuring alcohol-related hospital admissions: clinical coding and alcohol-attributable fractions.

Clinical coding is at the heart of all hospital data analysis. It is done by specially trained staff and is the process whereby information written in patient notes is translated into coded data and entered into hospital information systems. The clinical notes are translated into a series of codes or condition groups that are defined within a standard framework -the International Statistical Classification of Diseases and Related Health Problems (ICD-10).The coder must identify a primary code, which could be seen as the main reason for admission but they can also record up to 19 secondary codes which describe other diagnoses that affect treatment. Additionally, the ICD-10 allows for some external cause codes to be recorded in order to help understand more about the admission. These might include codes indicating a motor accident, fall or assault. External cause codes can be listed within the 19 secondary codes but cannot be recorded as a primary code.

Alcohol-attributable fractions: Alcohol causes, or can contribute to the development of, many health conditions. Academics have been able to use high quality research evidence to estimate what proportion of cases of a health condition are alcohol-related. Conditions such as alcoholic liver disease where alcohol is the sole cause are known as alcohol-specific or wholly alcohol-attributable conditions and their alcohol-attributable fraction is 1.0 (100 per cent). For other conditions, where alcohol has a proven relationship but it is one of a range of causative factors, an estimate of the contribution alcohol makes is calculated. For example, it is estimated that alcohol plays a causative role in 25-33 per cent of cardiac arrhythmias. These are the partially alcohol-attributable conditions and the alcohol-attributable fractions would be 0.25-0.33. Fractions differ slightly for men and women. Some external cause codes also have an alcohol-attributable fraction (for example, 27 per cent of assaults are estimated to be alcohol-related and therefore the alcohol-attributable fraction is 0.27).

The total number of alcohol-related hospital admissions, as described by the indicators, is not a number of actual people or a number of actual admissions but an estimated number of admissions calculated by adding up all of the fractions we have identified. The infographic below illustrates how all the partially alcohol-attributable admissions combine to make an alcohol-related hospital admission.

ARHA graphic

It is important to remember that this is an exercise using research evidence that is applied to hospital data. There will be people who don’t drink alcohol whose admission will be included in the figures; injuries and illnesses that are entirely the result of alcohol use that are not given appropriate recognition; and circumstances where the contribution of alcohol is simply too complex to quantify (such as child malnutrition and neglect arising from parental alcohol dependence).

So what’s the difference between the original and the new supplementary indicator? The original indicator considers all codes (primary and any secondary codes) that are recorded in relation to a patient’s admission record, and if any of these codes has an alcohol-attributable fraction then that admission would form part of the alcohol-related admission total. This can be seen as a broad measure. It provides evidence of the scale of the problem but is sensitive to changes in coding practice over time.

The new indicator seeks to count only those admissions where the primary code has an alcohol-attributable fraction. Although alcohol-attributable fractions exist for external cause codes (such as 27 per cent of assaults), these cannot be recorded as a primary code so the new indicator also includes admissions where the primary code does not have an alcohol-attributable fraction but where one of the secondary codes is an external cause code with an alcohol-attributable fraction. This represents a narrower measure. Since every admission must have a primary code it is less sensitive to coding practices but also understates the part alcohol plays in the admission.

In summary, the new supplementary indicator provides a narrower measure of alcohol harm that is less sensitive to the changes that have occurred in coding over the years and therefore enables fairer comparison between levels of harm in different areas and over time. It is also more responsive to change resulting from local action on alcohol. However, the original indicator is a better measure of the total burden that alcohol has on community and health services. These indicators measure different things and are to be used for different purposes. What matters most is that they are used to develop understanding, direct action, and achieve positive change in reducing alcohol harm.


  1. Dr. Becker

    Is there a measure for indirect health impact of alcohol abuse, i.e. neglect of other patients while caring for those that have been admitted for alcohol related reasons?

    Link to this comment Reply
    • Matt Hennessey

      Thanks for getting in touch. At present there are no standard measures specifically looking at the kind of indirect impacts you suggest. The closest measure currently in use is to cost the impact of alcohol on health services and make an assumption that this resource could have been effectively deployed elsewhere. In written evidence to the Health Select Committee the Department of Health estimated the cost to the NHS to be about £3.5billion per year at 2009-10 costs (

      Link to this comment Reply
  2. Bren

    Hello Claire and Matt,

    This is a really helpful example of measures, and how the two measures have strengths and weaknesses in their respective areas of measurement.

    What was a really powerful statement in your blog was, "What matters most is that they are used to develop understanding, direct action, and achieve positive change in reducing alcohol harm." This has to be the primary focus in all that we do.

    Thanks Claire and Matt,


    Link to this comment Reply
  3. Stuart Dodd

    I sense some spin here.... Would it be fair to say that the new narrow measure is designed to disregard the significant contribution of alcohol to chronic diseases, i.e. those conditions that will continue to increase in the population for decades to come (and to which alcohol will come to contribute more to over time).

    My suspicions will be confirmed when the broad measure is eventually dropped. However it's only by tracking the broad measure that the real cost of alcohol to the NHS will be understood.

    Link to this comment Reply
    • James Morris

      Hi Stuart, I think you may well have a point about why this change was triggered - sections of the alcohol industry clearly hated the media attention that often followed the publication of new annual stats e.g '1 million alcohol admissions a year' etc. I hope the broad measure is not eventually dropped, but also that the narrow measure is available on past years to demonstrate trends.

      Link to this comment Reply
  4. James Morris

    Can you explain a bit more about how "the new supplementary indicator ... is also more responsive to change resulting from local action on alcohol." ?

    Link to this comment Reply
  5. Matt Hennessey

    Hi James,
    Thank you for taking the time to post a question. There are several ways in which the new supplementary indicator is likely to be more responsive to local action. Local action, in this case, might include things such as managing access to alcohol through licensing controls; increased and improved specialist treatment; effective identification and brief advice; and the effective co-ordination of care, perhaps through the appointment of an alcohol specialist nurse or health worker.

    The narrower measure contains a larger proportion of acute conditions where excessive alcohol use may have played a part. It is easier to achieve a noticeable impact in respect of acute conditions in a short period of time than it is to achieve a similar impact in chronic conditions, which may take several years. It is important, however, to ensure that any impact is sustainable. The new measure also reports smaller numbers so the impact of successful interventions may be more noticeable in the data. For example, if you can intervene to avert 50 hospital admissions out of a total of 1000, it will be more apparent than 50 out of a larger figure of 100,000 which may be subject to seasonal variations that are much greater than this figure of 50.

    Finally, the broad measure includes admissions where the treatment may cost more because of the part alcohol plays but the treatment itself does not specifically address the alcohol related condition. In the broad measure a patient may be counted if they were admitted for a cataract operation but with the complication of hypertensive disease. The treatment offered in this admission is not designed to directly address the alcohol related condition - it may simply cost more because of the alcohol related condition. In the narrower measure, since the primary reason for admission is an alcohol related condition it is likely that the treatment provided will be more directly related to addressing the alcohol contribution to the condition. For example, if someone is primarily admitted for hypertensive disease then the treatment will include elements aimed at reducing any contribution from alcohol where it exists.

    Link to this comment Reply
  6. Olivea Ethan

    Alcohol is a very serious problem of any other country. You write Alcohol is England’s. You also write public health England and current alcohol related hospital admissions. This information is very important for every alcohol addicted person.

    Link to this comment Reply
  7. John Turnbull

    Dear Claire and Matt,

    The article and subsequent comments are extremely interesting and I am aware that there may be local initiatives and pilots to reduce the number of alcohol related hospital admissions. Would you be aware of any specific programmes or example of care models that may have been successful in reducing the number of hospital admissions ?

    Link to this comment Reply
  8. nobelhassasn1

    Thanks.That's give us a good idea about hospital.

    Link to this comment Reply

Leave a comment