Australian alcohol-attributable harm visualisation tool

Version 2.1 (released January 2023)

How to use

Click on the relevant buttons to access each function (line chart, bar chart, map)
Select data variables you are interested in visualising
Select data visualisation options (e.g. line chart, table or both)
Click on generate button to view output (line chart is limited to 72 data points, bar chart is limited to 4 bars per state/territory)
Hover your mouse over a data point, bar or map region to view the data value
Click on download buttons to download the chart/map (.png or .jpg) and/or table (.csv)

Background

This visualisation tool was developed in a collaboration between Curtin University’s National Drug Research Institute (NDRI) and University of Victoria’s Canadian Institute for Substance Use Research (CISUR) and is maintained and regularly updated by NDRI.

It was developed as a component of the National Alcohol Indicators Project (NAIP) (Principal Investigator: Prof Tanya Chikritzhs; Chief Investigator: Dr William Gilmore; Project Staff: Paul Jones), a project federally funded and coordinated by NDRI since 1999, aimed at tracking and reporting on local, state/territory and national trends in alcohol-attributable harm across Australia.

Wholly and partially alcohol-attributable

The tool presents trends and patterns in hospitalisations and deaths for wholly (i.e. 100% alcohol caused) and partially (i.e. fractionally alcohol caused) alcohol-attributable conditions (see Box 1). For death data this is based on the underlying cause of death, and for hospitalisation data this is based on either primary diagnosis or external cause.

For example,

Wholly = Alcoholic liver cirrhosis etc.
1 alcoholic liver cirrhosis (K70) death = 1 wholly alcohol-attributable digestive disease death

Partially = Falls, assaults, chronic pancreatitis etc.
1 assault injury (X85-X89.9; Y00-Y09.9; Y87.1) hospitalisation = 0.15 partially alcohol-attributable intentional injury hospitalisation.

'Net' totals

The tool presents 'net' totals by adjusting for apparent protective effects of alcohol at low dose (i.e. deaths or hospitalisations avoided) for cardiovascular, digestive and endocrine conditions.

For example,

100 partially alcohol-attributable ischaemic stroke deaths + 50 partially alcohol-attributable ischaemic stroke deaths avoided (i.e. minus 50 deaths) = 50 ‘net’ partially alcohol-attributable cardiovascular disease deaths

Calculating partially alcohol-attributable fractions

Australia-specific alcohol-attributable fractions for all partially alcohol-attributable health conditions by age and sex were calculated in an open access online alcohol harm estimator (International Model of Alcohol Harms and Policies, InterMAHP v2.1) with detailed methods described there [1,2].

Alcohol-attributable fractions from previous NAIP publications, using data sources deemed more applicable to the Australian population, superseded those calculated in InterMAHP for road traffic crash injury (pedestrians and non-pedestrians), intentional self-harm injury, assault injury (0-14 age group). Stomach cancer is an addition that had not been included in InterMAHP.

The protective effects of alcohol at low dose for Ischaemic heart disease, Haemorrhagic stroke, Ischaemic stroke (within Cardiovascular diseases group), Acute pancreatitis (within Digestive diseases group) and Diabetes Type 2 (within Endocrine conditions group) have been applied to the data and results in deaths or hospitalisations avoided. For Ischaemic heart disease we followed the Roerecke and Rehm option in InterMAHP [1,2] that has the greatest protective effect and produces the most conservative estimates of overall alcohol-attributable harm.

Box 1: Health conditions and groupings*

Cancers:: Oral cavity and pharynx cancer, Oesophageal cancer, Liver cancer, Laryngeal cancer, Breast cancer, Pancreatic cancer, Stomach cancer, Colorectal cancer
Cardiovascular diseases:: Hypertension, Alcoholic cardiomyopathy, Atrial fibrillation and cardiac arrhythmia, Haemorrhagic stroke, Oesophageal varices, Ischaemic stroke, Unspecified stroke, Ischaemic heart disease
Digestive diseases:: Alcoholic gastritis, Acute pancreatitis, Chronic pancreatitis, Alcohol-induced pancreatitis, Liver cirrhosis, Alcoholic Liver cirrhosis
Endocrine conditions:: Diabetes mellitus (Type 2), Alcohol-induced pseudo-Cushing’s syndrome
Infectious diseases:: Tuberculosis, HIV, Lower respiratory tract infections
Injuries (intentional):: Intentional self-harm, Assault (15+), Assault (0 - 14), Intentional self-poisoning by alcohol
Injuries (unintentional):: Road traffic crashes (non-pedestrians), Road traffic crashes (pedestrians), Falls, Fires, Drowning, Aspiration, Accidental poisoning by alcohol, Occupational machine injuries
Neuropsychiatric conditions:: Alcoholic psychoses, Alcohol dependence, Alcohol abuse, Epilepsy, Alcoholic polyneuropathy, Degeneration of nervous system due to alcohol, Alcoholic myopathy

*ICD-10-AM code ranges used for each condition differ in some instances to those specified in the InterMAHP guide [1]. For ICD-10-AM codes used in this tool please click link.

Data sources

De-identified unit record hospitalisation data were requested and provided by the Australian Institute of Health and Welfare from the National Hospital Morbidity Database.

De-identified unit record death data were provided by the Australian Coordinating Registry (ACR) at the Queensland Registry of Births, Deaths and Marriages who act on behalf of the custodians of the data, which include staff in the eight jurisdictional Registries of Births, Deaths and Marriages, eight jurisdictional Coroners and the National Coronial Information System.

Estimated Resident Population data for calculating rates were sourced from the Australian Bureau of Statistics.

Per capita alcohol consumption data and alcohol consumption pattern data used in calculating alcohol-attributable fractions were sourced from the Australian Bureau of Statistics (Apparent Consumption of Alcohol, Australia) and Australian Institute of Health and Welfare (National Drug Strategy Household Survey) respectively.

Caveats

Counts and rates

Counts of alcohol-attributable hospitalisations and deaths have been included for transparency and for use by others but should not be used for direct comparison between years, states/territories, region types, sexes or age groups due to changes and differences in the underlying population at risk.

Age-specific (15-34, 35-64, 65+) and directly age-standardised rates (calculated for the All (15+) category only) allow for better comparison of overall rates between states/territories, by adjusting for their different age structures.

Year

In the hospitalisation data, ‘Year’ is the financial year of the hospital separation e.g. 2020 is FY2019/20. In the death data, ‘Year’ is the calendar year of death registration.

State/territory and Region type

In the hospitalisation data, both ‘State/territory’ and ‘Region type’ are based on place of residence except for 2014-2018 when ‘State/territory’ is the state/territory of hospitalisation. We don’t expect this to affect the trends as inhouse analyses found 98% of 2010 hospitalisations were in the state/territory of ‘residence’. In the death data, both ‘State/territory’ and ‘Region type’ are based on place of residence.

Metro and Non-metro areas are defined differently for hospitalisations (using ABS Remoteness Areas classification) and deaths (using ABS Greater Capital City Statistical Areas). In both hospitalisation and death data the Australian Capital Territory has been classified as Metro only and the Northern Territory and Tasmania have been classified as Non-metro only. Due to some unit records missing sub-state residence information, ‘Metro’ plus ‘Non-metro’ counts do not always equal ‘All’.

We excluded records where state/territory was either an overseas territory or missing.

Health conditions

Health conditions related to the impacts of maternal consumption on the foetus, such as foetal alcohol spectrum disorder, have been excluded from this analysis. Counts of these conditions are low in hospitalisation and death data and would need to be suppressed for confidentiality reasons. They are also likely to be an underestimate due to difficulties in diagnosis.

Sex

We excluded records where sex was either ‘Other’ or missing as this prevented calculation of attributable fractions and calculation of population rates.

Age groups

The data presented is for Australian ‘adults’ aged 15 years and over, except in the case of Assault injury (within intentional injuries group) where we have included data for the 0 to 14 age group.

We excluded records where age was missing as this prevented calculation of attributable fractions and calculation of population rates.

Confidentiality

Combinations where the actual hospitalisation or death counts (before alcohol-attributable fractions applied) are less than or equal to 5 are suppressed for confidentiality reasons.

Coronial inquiry

Death data (2012-2020) were preliminary at time of data request, and so our results may underreport alcohol-attributable deaths where cause of death was not confirmed until a coronial inquiry was completed.

Comparisons with previous NAIP publications

Due to continuing revisions in data sources and updated methods of estimating attributable fractions, comparisons with figures released in previous NAIP publications should be done with a high degree of caution. The data in the current online version of the tool should be taken as the latest estimates.

Funding

The National Drug Research Institute at Curtin University is supported by funding from the Australian Government under the Drug and Alcohol Program.

Acknowledgements

We thank the ABS, AIHW, ACR and all jurisdictional departments for access to data.

We thank the team who developed the open access alcohol-attributable fraction estimator, InterMAHP.

To acknowledge everyone’s contributions we suggest that you please use the following citation when publishing data from this tool.

Suggested citation:

Gilmore, W., Lensvelt, E., Jones, P., Dorocicz, J., Sherk, A., Churchill, S., Stockwell, T., Chikritzhs, T. (2023). Australian alcohol-attributable harm visualisation tool version 2.1. National Drug Research Institute, Curtin University and Canadian Institute for Substance Use Research, University of Victoria. www.alcoholharmtool.info

Resources

References

Sherk, A., Stockwell, T., Rehm, J., Dorocicz, J., Shield, K.D. (2017). The International Model of Alcohol Harms and Policies (InterMAHP): A comprehensive guide to the estimation of alcohol-attributable morbidity and mortality. Version 1.0: December 2017. Canadian Institute for Substance Use Research, University of Victoria, British Columbia, Canada.
Sherk, A., Stockwell, T., Rehm, J., Dorocicz, J., Shield, K.D. (2019). The International Model of Alcohol Harms and Policies (InterMAHP). Version 2.1: 2019. Canadian Institute for Substance Use Research, University of Victoria, British Columbia, Canada.