Project SMART

Go to content


Meetings and Conference

Minutes of the European Conference on Standardizing Measurement of Alcohol Troubles (SMART). Barcelona, 19 October 2010.

< PDF document >   

The symposium of the conference devoted to alcohol survey methodology was held on 19th October 2010 in Barcelona. It brought together 31 participants from 19 European countries and even one presenter from Canada. Representatives of World Health Organization Head Quarters (WHO HQ) in Geneva and European Monitoring Centre for Drugs and Drug Abuse (EMCDDA) were also present.

After welcome made by Joan Colom on behalf of hosting institution Generalitat de Catalunya (GENCAT) and by Jacek Moskalewicz on behalf of the SMART project the floor was given to Dag Rekve (WHO HQ). He talked about the WHO Global Alcohol Strategy and role of research in monitoring its implementation.

Then, Ann Hope (Ireland) gave a brief presentation on an important product of the project which was review of the methodological literature on alcohol surveys. The review was assessed as a very comprehensive product which is currently ready for submission for publication.

Consecutive product of the project - EU Survey Review was presented by John Foster (UK). The review covered about 40 alcohol surveys carried out in more than 20 European countries.
The review benefited from the discussion which suggested replacement of research recommendations by summary and suggestions for future research.

Results of the pilot survey on alcohol consumption measures were presented by Janusz Sieroslawski (Poland). It was found that Beverage Specific Quantity Frequency (BSQF) method work well in all nine countries participating in the pilot offering in most countries higher estimates of annual alcohol consumption compared to Graduated Frequency and Last Occasion (LO) approach. Moreover, BSQF method offers similar estimations of annual frequency of drinking as generic Quantity Frequency method.

Results of the pilot survey which addressed standard instruments on alcohol abuse and dependency were shown by Ludwig Kraus (Germany). Three standard measures were used in the pilot: diagnostic instrument (M-CIDI) and screening instruments AUDIT and RAPS. The preliminary results showed that there was considerable variation across countries in the RAPS and AUDIT which raises questions on the pre selection cut-off points for these measures. More people were diagnosed using the M-CIDI than the 10% of heavy drinking in the sample. A comparison of the M-CIDI and AUDIT, showed good specificity (small number of false positives) however sensitivity was very low, in other words a large proportion were false negative. A comparison between M-CIDI and RAPS showed acceptable specificity (70.8%) and good sensitivity (89.6%). More detailed analysis is considered necessary. The conceptual mismatch between instruments raises the question of what is the purpose of this section of the questionnaire - screening or diagnosis?

The section on 'harm to others' presented by Jacek Moskalewicz as a result of someone drinking showed the proportion of known heavy drinkers were relatively similar across countries. However, those negatively affected differed across countries. The measures on impact of others drinking in the community showed as expected, the most common were of a less sever nature (annoyed, awaken at night) and the more serious (physical abused) less common. A factor analysis of the original 11 items were reduced to 9 items and showed three dimensions- factor one 'looking for trouble', factor two 'verbal level of abuse' and factor three 'public space'.

The section on public support for alcohol policies was presented by Jacek Moskalewicz. Public support for alcohol policies was measured using 12 statements. A factor analysis of the 12 items were reduced to 8 items which gave 3 dimensions- factor 1 (alcohol control measures), factor two (drink driving measures) and factor three (liberal attitudes). A question was raised as to whether the opposite phrasing of the items were also tested which could give different results.

Qualitative assessment of the SMART pilot instrument (Zsuzsanna Elekes and Alina Allaste) was undertaken through focus group discussions (18 focus groups) in six of the countries. The main purpose was to understand the cultural meaning of the major SMART concepts. People distinguished between alcohol consumption (any kind of alcohol use) and drinking (heavy drinking, social events) and the concept of what constitutes a 'drink' is more likely to be a social event or some mixed alcohol beverage like a cocktail. When speaking about a 'drink', people never see it as a unit alcohol and there are major differences across countries about the quantity of 'a drink'. Reporting on the frequency and quantity of drinking, people tended to see a year as too long, especially for those who do not regularly drink. In terms of quantity the 'usually' is what people report and rarely heavy drinking occasions. Drinking problems are associated with alcoholism and prevention measures for the general public are not seen as the solution. The social context such as better living conditions are seen as solutions to problems. People consider strict rules and regulations for reduce drinking among young people and to offer alternatives, thus change the motivation for drinking. Problems of drinking in the community are seen mainly as activity done by alcoholics.

The discussion that followed these presentations raised three main issues. Firstly, the translation issue where some words have no translation into certain languages, for example there is no German word for 'intoxication'. The conceptual choice of measuring problems was highlighted, diagnostic measures versus screening measures. Finally, the major challenge of how to measure a standard drink was discussed. Some argued that a standard drink is far from the reality of 10g drink. The market sizes are much bigger and stronger and so the possibility of underestimation is great. Some countries try to increase reliability by asking in the national glass sizes of the country and then translating into grams of alcohol later.

After discussing experiences of the pilot survey participants were divided into six working groups discussing consecutive sections of the questionnaire. Reports from the working groups were presented:

Measuring alcohol consumption (Kim Bloomfield):
With regard to question F-1, Janusz proposed deleting the frequency item as annual frequency estimated from the most frequently drunk beverage in from BSQF questions is very close to generic frequency question. Kim brought up the reason for having a generic frequency question, i.e., that one can then calculate annual consumption with it. The representative from Spain mentioned also that it can be used as a filter to determine drinking status.
There was also discussion of the phrasing drinking "even in small amounts". It was asked whether sips and trials should be included in the question according to WHO definitions of abstention. But others in the group were for keeping it in.
There was also discussion of collapsing frequency categories.
There was also a proposal to include a question about lifetime abstention which prompted a discussion of optional questions regarding starting and quitting drinking.
Another group member referred to the ECHIM health indicator project and the indicators used for it: alcohol-attributable deaths, total per capita consumption, hazardous drinking (women 20+ gr., men 40+ gr) and RSOD. The questionnaire should be in agreement with these indicators.

Drunkenness and RSOD (Allaman Allamani)
RSOD is based on quantity and frequency whereas drunkenness is a perception. Are they complementary? Also tolerance plays a role in drunkenness which can make people feel less drunk. Which of the two items is more clear to measure. Delete either AUD3 or the RSOD question. Time span: average of 3-6 hours for the occasion but a shorter period would be more critical. A context of RSOD was considered important.

Unrecorded consumption (Alina Allaste)
The group was satisfied with the questions. But there were a couple of problems: 1) people maybe are buying illegal alcohol without knowing it (but nothing can be done about this), 2) illegal activity: people don't admit to it. Maybe a question in the third person would be better. More questions were raised in the discussion, including an issue of competitive sources of information and existence of unrecorded supply in wine countries.

Alcohol abuse/dependence (Daniela Piontek)
Conceptual discussion: is it possible to have a measure of clinical dependence or should there be a broader measure? What is the purpose of the survey? Determine what you include. CIDI for dependence, abuse/general abuse should have another measure. Is DSM the gold standard? RE: AUDIT, RAPS, cut-offs would be different across countries. We cannot recommend an instrument since we have no real basis of psychometric testing. Depends on what you want to do, e.g., what population and what to measure. This is not a validation study. In the discussion it was suggested to add few more questions on individual harm in such life areas as health and finances. Ethical considerations of screening were also raised.

Harm to others (John Foster)
Harm to people who are known vs. harm to strangers. Relationship to heavy drinker. Take into account WHO global strategy. Technical report, page 64: other kinds of harm to be included, optionally. Frequency of harms vs. yes/no occurrence. In the discussion a distinction between nuisance problems and harm problems.

Public support (Ladislav Csemy)
Number of factors was reduced in a course of factor analysis. Those which remained consists in four consistent factors. No major revisions needed except for removing "random" from "random breath testing" and few more small revisions. The group recommends items 2, 4, 5, 6, 7, 8 and 11, some with modifications.

Two last presentations dealt with comparability of survey results across different countries. Kevin Shield presented a paper co-authored by Jόrgen Rehm on prospects of standardizing alcohol consumption estimated in surveys by recorded or overall alcohol consumption. By modelling alcohol consumption using a gamma distribution, they developed new methodology to account for the undercoverage of alcohol consumption which was observed using other methodologies. This new method was developed as part of the ongoing Global Burden of Disease (GBD) Comparative Risk Assessment and the US Burden of Disease Study.

Robin Room in his presentation on cross-cultural applicability of alcohol survey showed a long tradition of comparative surveys reaching back to the 1960's. He recalled the WHO Community Response Study run in Zambia, Mexico, Scotland and California, in the 1970's. More recent traditions include WHO affiliated GENESIS study as well as World Mental Health Survey Initiative on mental health epidemiology which include also large section on alcohol. More recent studies include Nordic initiatives seven-country ECAS project, as well as EUROBAROMETER study sponsored by European Commission. There are also two studies targeting school children: ESPAD and HBSC.

Room pointed to methodological controversies in cross-national surveys including cultural incompatibility at the conceptual level, questions on translation, difficulties with diagnosis- based questions as well as dramatic fall in response rate. The presentation was followed by vivid discussion.

In the discussion on feasibility representative of all countries expressed their commitment to consider prospects of participating in a future comparative study and even to raise funds from national sources.

Jacek Moskalewicz in his wrap-up stressed great contribution of all participants which make us think of improvement of the questionnaire itself but also of expansion of methodological context of a potential European comparative alcohol survey.

Despite certain amount of skepticism there is a need of comparative cross-European surveys as EU becomes increasingly more integrated and its alcohol strategy is very likely to be elaborated and need monitoring.

A question of non-response rates should be challenged in different ways. Major issue is not a low respond rate but to what extent dropouts are random. Therefore, careful attention should be given to those who dropout from the sample. Their social demographic characteristics as well as reasons for non-response have to be analyzed.

It seems beyond capacity and mandate of the project to produce final complete questionnaire. The project should rather propose a number of core questions whose implementation should not take more than 15 minutes and a series of optional questions.

Allaman Allamani – ASF – Azienda, Italy
Airi-Alina Allaste – Institute for International Social Studies, Tallin Univ., Estonia
Sharon Arpa – Foundation for Social Welfare Services, Malta
Kim Bloomfield – Arhus University, Denmark
Elin k. Bye – SIRUS, Norway
Andrei Botescu – Romanian National Antidrug Agency, Romania
Joan Colom – Generalitat Catalunya, Spain
Ladislav Csemy – NIPH - National Inst. Public Health, Czech
Zsuzsanna Elekes – Corvinus University Budapest, Hungary
Fernanda Feijao – Institute on Drugs & Drug Addictions (IDT), Portugal
John Foster – UK
Lydia Gisle – Scientific Institute of Public Health, Belgium
Ann Hope – Health Service, Ireland
Oyvind Horverak – SIRUS, Norway
Marke Jaaskelainen – The National Institute for Health and Welfare, Finland
Thomas Karlsson – National Institute for Health and Welfare, Finland
Ludwig Kraus – IFT, Germany
Paul Lemmens – University of Maastricht, Holland
Jacek Moskalewicz – Institute of Psychiatry & Neurology, Poland
Daniela Piontek – IFT, Germany
Dag Rekve – WHO Headquarters
Robin Room – SORAD, Sweden
Emanuele Scafato – Istituto Superiore di Sanita’, Italy
Lidia Segura – Generalitat Catalunya, Spain
Kevin Shield – CAMPH, Canada
Janusz Sieroslawski – Institute of Psychiatry & Neurology, Poland
Hana Sovinova – National Institute of Public Health, Czech
Alfred Uhl – Anton Prosksch Institut (API), Austria
Momtchil Vassilev – National Centre for Addictions, Bulgaria
Juliαn Vicente – EMCDDA

Back to content | Back to main menu