Unrecorded alcohol use: a global modelling study based
on nominal group assessments and survey data
, Jakob Manthey
, Aaron Merey
, Margaret Rylett
Institute for Mental Health Policy Research, CAMH, Toronto, ON, Canada,
Institute for Clinical Psychology and Psychotherapy, TU Dresden, Dresden, Germany,
Campbell Family Mental Health Research Institute, CAMH, Toronto, ON, Canada,
PAHO/WHO Collaborating Centre for Mental Health and Addiction, Toronto,
Institute of Medical Science (IMS), University of Toronto, Toronto, ON, Canada,
Department of Psychiatry, University of Toronto, Toronto, ON,
and Dalla Lana School of Public Health, University of Toronto, Toronto, ON, Canada
Background and aims
Alcohol use is among the most important risk factors for burden of disease globally. An estimated
quarter of the total alcohol consumed globally is unrecorded. However, due partly to the challenges associated with its as-
sessment, evidence concerning the magnitude of unrecorded alcohol use is sparse. This study estimated country-speciﬁc
proportions of unrecorded alcohol used in 2015.
A statistical model was developed for data prediction using data
on the country-speciﬁc proportion of unrecorded alcohol use from nominal group expert assessments and secondary,
nationally representative survey data and country-level covariates.
Estimates were calculated for the country
level, for four income groups and globally.
A total of 129 participants from 49 countries were included
in the nominal group expert assessments. The survey data comprised 66 538 participants from 16 countries.
Experts completed a standardized questionnaire assessing the country-speciﬁc proportion of unrecorded
alcohol. In the national surveys, the number of standard drinks of total and unrecorded alcohol use was assessed for the
past 7 days.
Based on predictions for 167 countries, a population-weighted average of 27.9% [95% conﬁdence
interval (CI) = 10.4–44.9%] of the total alcohol consumed in 2015 was unrecorded. The proportion of unrecorded alcohol
was lower in high (9.4%, 95% CI = 2.4–16.4%) and upper middle-income countries (18.3%, 95% CI = 9.0–27.6%) and
higher in low (43.1%, 95% CI = 26.5–59.7%) and lower middle-income countries (54.4%, 95% CI = 38.1–70.8%). This
corresponded to 0.9 (high-income), 1.2 (upper middle-income), 3.2 (lower middle-income) and 1.8 (low-income) litres of
unrecorded alcohol per capita.
A new method for modelling the country-level proportion of unrecorded al-
cohol use globally showed strong variation among geographical regions and income groups. Lower-income countries were
associated with a higher proportion of unrecorded alcohol than higher-income countries.
Keywords Alcohol exposure, alcohol per capita, burden of disease, global health, risk modelling, unrecorded alcohol.
Correspondence to: Charlotte Probst, Institute for Mental Health Policy Research; Centre for Addiction and Mental Health (CAMH), 33 Russell Street, Toronto,
Ontario M5S 2S1, Canada. E-mail: firstname.lastname@example.org
Submitted 17 April 2017; initial review completed 24 July 2017; ﬁnal version accepted 23 January 2018
Alcohol use is among the most important risk factors for
burden of disease globally . As under-reporting of alco-
hol use in surveys can be extensive, with reported alcohol
consumption often covering only 15–60% of all alcohol
consumed in a jurisdiction [2,3], burden of disease
estimates rely heavily upon ofﬁcial data sources such as
alcohol sales, production, taxation, import and export to
estimate the total amount of alcohol consumed in a coun-
try . Alcohol assessed in such ofﬁcial data sources is also
referred to as recorded alcohol. However, with a large
variation between countries, approximately a quarter of
the total alcohol consumed globally is unrecorded .
In the 2014 Global Status Report on Alcohol and
Health from the World Health Organization (WHO) unre-
corded alcohol was deﬁned as ‘[…] alcohol that is not taxed
in the country where it is consumed because it is usually
produced, distributed and sold outside the formal channels
under government control’ (, p. 30). Unrecorded alcohol
can be classiﬁed into four broad subgroups , namely: (i)
home-made alcohols generally intended for consumption
by the producers or their immediate environment [7,8];
(ii) informally produced alcohol for commercial purposes
© 2018 Society for the Study of Addiction Addiction, 113,1231–1241