FAQ - More Information and Notes

+ Why ask about consumption in the last 12 hours?

Alcohol consumption in the last 12 hours was seen as the best available timeframe by the investigative team when designing the Last Drinks questions.

Other timeframes – both shorter and longer – were considered, and rejected due to the likelihood of missing episodes of alcohol consumption if the patient had slept prior to ED attendance, or overestimating harm by capturing alcohol consumption from a previous session.

+ How reliable is self-report of alcohol use?

Any self-reported data is vulnerable to some biases. For example, respondents may engage a social desirability bias (i.e. modifying their response to fit what they see as a more socially acceptable response). Responses can also be affected by recall biases (i.e. remembering incorrectly).

However, recent peer-reviewed studies conducted in Emergency Departments have found that self-report of alcohol use in the ED is remarkably accurate, averaging 92% accuracy when the self-report was validated with a breathalyzer. Further information on this study by Cherpitel et al. is available here via the journal Drug and Alcohol Review.

+ What about the harm of alcohol to others?

The Last Drinks questions ask about the patient’s own alcohol consumption, and their own experiences of injury or illness which have led to their visit to ED.

Therefore, the Last Drinks questions do not capture:

  • Harm caused to others due to the patients’ alcohol use, and
  • Harm caused to the patient due to another persons’ alcohol use.

Previous snapshot data collected in Australian EDs suggests that the harm of alcohol to others (children, intimate partners, family members, friends’ acquaintances and strangers) accounts for a very large portion of alcohol’s total burden of harm. Further information about this study can be found here via the journal Addiction.

+ Does the “Last Drinks” method over-estimate the burden of alcohol?

It is highly unlikely that the Last Drinks method overestimates the burden of alcohol, for two main reasons:

  1. The Last Drinks questions do not capture harm to others (see previous point, “What about the harm of alcohol to others”)

  2. At least 80% of people who presented at ED after drinking alcohol in the past 12 hours can have their attendance traced to an alcohol attributable injury or condition. More specifically: a review of attendees’ ICD-10 (disease classification or diagnoses codes set by the World Health Organisation) indicates that at least 80% of all attendances who had consumed alcohol in the past 12 hours were directly attributable, or can be feasibly attributed to, alcohol consumption. This method includes F codes attributable to acute alcohol intoxication, toxicity, or alcohol dependence; F codes attributable to mental or behavioural disturbance due to alcohol intoxication, and suicidal ideation, self-harm, depressive episodes or anxiety related presentations co-occurring with alcohol consumption; R codes for symptoms and signs involving cognition, perception, emotional state and behaviour, S & T codes for injury, poisoning and certain other consequences of external causes. These coding structures have seen widespread use as broad estimation methods for alcohol attributable harm in health care systems.

Therefore, while the Last Drinks model offers unprecedented capture of the burden of harm presenting to Australian ED’s, it is likely that the model underestimates the total burden of alcohol harm to the community.

+ Is everyone who answers “yes” intoxicated?


Participants may have only had one or two drinks in the 12 hours before attending ED, or may have consumed many drinks and “sobered up” before coming to ED. Alternately, participants may be intoxicated to a life-threatening degree, and be unable to answer the questions until many hours after they arrive at ED.

As such, no patients’ attendance is ever analysed or reported in isolation – all results are aggregated data drawn from thousands of presentations over a number of months. This allows for the identification of trends, and avoids over-reliance on any one specific case.

+ Do you measure BrAC? (Use a breathalyser?)


While it would be ideal from a research perspective, taking a measure of every patient’s BrAC (Breath Alcohol Concentration) would be prohibitively expensive and extremely time consuming, and therefore not feasible in a busy ED environment.

Emergency physicians will administer BrAC or BAC (Blood Alcohol Concentration) tests on occasion for medical purposes, or if they believe that alcohol toxicity may need immediate intervention. However, the results of these tests are not provided to the Last Drinks team.