Patching together a policy study

Robin Room

An irony is built into the substantial tradition of study of the effects of Nordic alcohol policies, as we remarked in reviewing the tradition: "in the majority of studies, testing and demonstrating the effectiveness of the control measures has been possible because the particular control measure was being abandoned" (Olsson, Ólafsdóttir Room, 2002:7).

What from the point of view of alcohol policy may be extremely problematic is thus from the point of view of research a great opportunity. First, there is no more direct way to study the effects of policies than to study what happens when they change. This is particularly the case when not every comparable location changes at once, and there is the chance to use stable locations as control sites. Second, such policy changes also offer an unusual chance to study, in the context of ordinary life, changes and causal connections at the individual level. If the policy change does have a measurable effect, whose behaviour is changed more - young or old, men or women, heavier drinkers or lighter, those who drink in public or those who drink at home? What goes along with the changes in the amount or pattern of drinking at the individual level? For instance, are there connected changes in the occurrence of problems in the family from drinking? Is there a measurable impact on potentially connected behaviours such as smoking or gambling?

The primary emphasis in Nordic alcohol policy research has been on the first of these frames, the effect of policy changes on the total population, and it was partly with the aim of redressing the balance with an emphasis on the second -- looking for differential effects in subgroups of the population -- that a collaborative Nordic group of researchers had reviewed and in some instances reanalyzed studies in the Nordic alcohol policy research tradition (Room, 2002), in a study financed by the Swedish Council for Social Research (SFR; now Forskningsrådet för Arbetsliv och Socialvetenskap, FAS) in 2000-2002.

When the Danish government announced that it would be reducing its spirits taxes by 45% on 1 October, 2003, then, there was a group of researchers primed to regard this as an opportunity to study effects not only at a population level, with aggregate data such as monopoly store sales statistics and mortality, but also at the individual level with population surveys. We knew also that another change would be happening -- the final increase in cross-border traveler's allowances required of Sweden, Denmark and Finland by the European Union, to take effect on 1 January, 2004. Both this and the Danish tax change, we knew from analyses of earlier changes in traveler's allowances, would have a special impact in Skåne and other parts of southern Sweden (see papers by Norström and Trolldal in Holder, 2000), with the effects dropping off further from Denmark.

The initial planning for the study involved Denmark and Sweden. We started with some bases of data collection on which we could build. Håkan Leifman at SoRAD was directing the Monitoring survey which every month interviewed 1500 Swedish adults about their unrecorded alcohol purchases and consumption, with some data also on drinking patterns and problems (Leifman Gustafsson, 2003). This survey had been funded by the Swedish Ministry of Social Affairs and Health to provide more reliable data than the brewers' estimates of unregistered consumption in Sweden. In Denmark, Kim Bloomfield of Southern Denmark University had already applied for Danish funds for a national drinking survey, and was willing to add questions relevant to the policy changes to an already crowded questionnaire.

To study changes at the individual level, we needed to collect longitudinal data, where the same respondents are followed over time. Such follow-up data lay outside the scope of the Monitoring survey or the plans at the time for Bloomfield's survey, and there was a need also for resources to supplement the Swedish Monitoring questionnaire and for staff to analyze the data. A proposal was written and submitted in January, 2003 to the Nordic joint social science research council, NOS-S (Nordiska samarbetsnämnden för samhällsforskning) for funds for these purposes for three years (including two annual "after" data waves) starting January 2004. The funds requested were above their usual ceiling, although they were still skimpy for what was needed for the study. Later in 2003, we learned that we were funded by NOS-S at their usual ceiling level.

The immediate problem for the nascent study was to fund the "before" data collection in Sweden, before the first and then the second policy change took effect (two such "before" waves were needed if the effects of the two policy changes were to be distinguished). The design was to use northern Sweden as a control site, in comparison with southern Sweden and with Denmark, so there was also a need for twice the sample size in Sweden as in Denmark. Additional funding from official sources was unobtainable, though Leifman did succeed in getting some resources from Systembolaget to support the data collection. But a substantial part of the funding for the "before" data collection in Sweden came from SoRAD's own reserves.

The decision was made to write a proposal in May, 2003 to the US National Institute on Alcohol Abuse and Alcoholism (NIAAA) for further funds to support the study. By this time, it had become likely that Finland, too, would reduce its taxes to forestall the border trade which would result from Estonia's accession to the EU, and colleagues from STAKES (Pia Mäkelä and Esa Österberg) joined the study team. The Finns had been able to finance their own "before" survey from institutional resources, but were interested in using northern Sweden as a control and in securing some resources from NIAAA for the follow-up questionnaires. The proposal to NIAAA also added a third follow-up year to the data collection, as well as providing resources for data analysis. Two colleagues who had previously succeeded with a NIAAA grant proposal to study a Swiss alcohol tax reduction (Heeb et al., 2003; Kuo et al., 2003; Mohler-Kuo et al., 2004), Jürgen Rehm from Canada and Gerhard Gmel from Switzerland, were added to the study team, helping substantially with the analysis plan. In view of ambitions to use the study for empirical tests of theoretical propositions about alcohol consumption (the collectivity of drinking cultures and the theory of rational addiction), Ole-Jørgen Skog from Norway, originator of the collectivity theory, and Philip Cook, an American economist, also became involved as consultants. As is common with US grant proposals these days, the proposal was not funded on its first submission, but a resubmission has received a very good priority score and is expected to be funded in the coming weeks - too late, however, for most of the first follow-up data collection. Danish research council money and STAKES resources have supported the 2004 fieldwork in those countries, while the Swedish fieldwork is supported by NOS-S funds and some further SoRAD reserves.

Presuming the NIAAA money comes through, a fully viable project has thus been mounted. It is actually the first collaborative cross-Nordic alcohol policy impact study with coordinated data collection in more than one country - all previous studies have been of a single country. But the process of mounting it revealed a number of weaknesses in Nordic research funding for studies of the effects of policy changes. Research council funds operate on too long a turnaround time to provide resources for collecting "before" data, unless changes are announced long ahead. The amount of money available from national or Nordic research council funds is insufficient for fieldwork for a substantial study collecting primary data. Funds directly from ministries were not available at the time of the project's first two data collections. At the moment, the European Union is no help at all in terms of funding such policy or social research. The outside funds which made the project a full study have thus had to be attracted from the U.S., which was willing to fund the study as a scientific endeavour with findings for social policy potentially applicable in the U.S. Clearly, also, the study would not have been possible without the existence of STAKES and SoRAD as national research centres, with intellectual and material resources that could be drawn on for an important opportunity.

If Nordic alcohol and drug policymakers are serious about basing their policies on evidence, there is a need to put more resources into policy-relevant research. Funding the cost of a full evaluation study should routinely be built into decisions to change social policy.

There are further opportunities to be seized. On 1 July, 2005, for instance, smoking will be banned in Swedish restaurants and bars. Apart from studying the effects of this on smoking and nicotine use, a study of this change offers an opportunity to study the interlinking of smoking with both drinking and gambling (gambling machines - "Jack Vegas" machines - are primarily located in bars in Sweden). The interlinked effects can be surprisingly strong: banning smoking around gambling machines in the Australian state of Victoria resulted in revenue loss from the machines of 11.5% in the first year (Harper, 2003). A ban on smoking in restaurants and bars already took effect in Norway in 2004 which means the possibility for another cross-national study.

References

Harper, T. (2003) Smoking and gambling: a trance inducing ritual. Tobacco Control 12:231-233.

Heeb, J.-L., Gmel, G., Zurbrügg, C., Kuo, M., Rehm, J. (2003). Changes in alcohol consumption following a reduction in the price of spirits: a natural experiment in Switzerland. Addiction 98:1433-1446.

Holder, H.D., ed. (2000) Sweden and the European Union: Changes in National Alcohol Policy and Their Consequences. Stockholm: Almqvist Wiksell International.

Kuo, M., Heeb, J.-L., Gmel, G., Rehm, J. (2003). Does price matter? The effect of decreased price on spirits consumption in Switzerland. Alcoholism: Clinical and Experimental Research 27:720-725.

Leifman, H. Gustafsson, N.-K. (2003) En skål för det nya milleniet: en studie av svenska folkets alkoholkonsumtioneni början av 2000-talet. Stockholm: SoRAD, Forskningsrapport nr. 11. http://www.sorad.su.

Mohler-Kuo, M., Rehm, J., Heeb, J.-L., Gmel, G. (2004). Decreased taxation, spirits consumption and alcohol-related problems in Switzerland. Journal of Studies on Alcohol 65:266-273.

Olsson, B., Ólafsdóttir, H. Room, R. (2002) Introduction: Nordic traditions of studying the impact of alcohol policies. In: Room, R., ed., The Effects of Nordic Alcohol Policies: What Happens to Drinking and Harm When Alcohol Policies Change? pp. 5-16. Helsinki: Nordic Council for Alcohol and Drug Research, NAD Publication 42.

Room, R., ed. (2002) The Effects of Nordic Alcohol Policies: What Happens to Drinking and Harm When Alcohol Policies Change? Helsinki: Nordic Council for Alcohol and Drug Research, NAD Publication 42.

Robin Room