Discussion

The proposed revisions to the alcohol unit conversion factors result in higher alcohol consumption estimates.

It is important to emphasise that these differences are reflective of changes in measurement, rather than changes in reported consumption. As conversion factors were revised upwards for some drinks and no change was applied to other drinks, estimates of alcohol consumption were higher for men, women and across all age groups.

Using the revised conversion factors in the HSE 2019 data had the following impact:

• the average (mean) alcohol units consumed in a typical week (among adults that drank alcohol) was 0.9 units higher than published estimates
• the proportion of all adults classified as drinking at increasing or higher risk of harm was one percentage point higher
• the proportion of all adults drinking more than three units (women) or four units (men) on any day in the last week was four percentage points higher

Similar differences were observed for HSE 2021 data. Revised conversion factors had the following impact:

• the average (mean) alcohol units consumed in a typical week (among those that drank alcohol) was 0.8 units higher than the published estimates
• the proportion of all adults classified as drinking at increasing or higher risk of harm was two percentage points higher
• the proportion of all adults drinking more than three units (women) or four units (men) on any day in the last week was five percentage points higher

The revision had a larger impact on alcohol consumption estimates for men than women. The percentage increase in unit conversion factors was the largest for large cans of strong beer (+17%), and pints of normal strength beer (+15%), compared with other beverage types, which may explain some of this difference, since men drink higher quantities of beer than women (source: Appendix B and Mäkelä et al).

The revision to conversion factors had a larger impact on estimates derived from data relating to the maximum amount of alcohol consumed on any day in the previous week than estimates derived from data relating to drinking in a typical week. This may be due to the relative impact of the change being greater with a smaller daily unit recommend limit (three/four units of alcohol) than the weekly recommended limit (14 units of alcohol). This may have also contributed to larger differences among older adults, who tend to drink lower volumes on single occasions than younger adults (source: Table 3 and Table 9 in the HSE 2019 Adults’ health-related behaviours data tables), meaning potentially a greater number of older adults at the periphery of the daily recommended limit. Differences in beverages typically consumed by older and younger adults may also explain some of these differences.

Despite some variability in discrepancies following a change in conversion factors, patterns of how alcohol consumption varied among sex and age group remained similar following the revision for both weekly and daily measures. The revised conversion factors applied to a range of beverages that are consumed by men and women and different age groups, so the impact of the revision did not appear solely concentrated among a sub-population. We expect that in HSE 2022, the impact of using the revised conversion factors will be similar, resulting in slightly higher alcohol consumption overall, but with little change to underlying patterns.

The change in conversion factors affects assessment of yearly change in alcohol consumption. The revised conversion factors due to be implemented from HSE 2022 onwards will be clearly indicated in the trend tables, as was the case when a revision occurred in 2006, with the original and revised estimates presented alongside each other that year. Footnotes will clearly reference the change and point to the HSE 2022 Methods chapter that includes details on the revision. There is no expectation that estimates in previous reports will be revised. Researchers using the HSE who aim to assess changes over time should ensure the same conversion factors are used across years.