Research has long found that, when compared with abstinence, moderate alcohol consumption is associated with lower risk for a diverse range of health outcomes. But determining whether alcohol is truly health protective depends on whether we can tease apart causation from association – what we call ‘causal inference’. Until recently, most studies did not use strong methods for causal inference. In our new paper, we review studies that have used newer, more robust methods to see what they can tell us about the effects of alcohol on our health.
Why is it hard to determine whether a relationship is one of cause and effect?
The ideal way to test this research question would be an experiment in which we assigned people to various levels of drinking and measured their health outcomes in the long-term. For various ethical and practical reasons though, these studies haven’t been possible. Instead, researchers rely on observational studies – simply observing how much people drink and following up on their health. But this can lead to a number of problems or ‘biases’ that make causal inference hard. Of particular concern is something called confounding, which means that certain factors, for example socio-economic status, are associated with bothhow much a person drinks and their later health outcomes. Determining whether it is these factors that are responsible for any association between drinking and health, as opposed to the effects of drinking itself, can be very challenging!
What can be done about this?
Recently developed research methods can help address this problem. One category of methods involves complex statistical approaches that allow researchers to incorporate multiple measurements of drinking, health outcomes, and confounding factors over time. Another one is a technique that makes use of genetic information – specifically, substituting a person’s actual level of alcohol consumption for their genetically-predicted level of alcohol consumption (which can’t be confounded by things like socio-economic status).
What did our study do?
We reviewed how these research methods have been applied to the study of alcohol and health outcomes to date in the scientific literature. We wanted to see if their results backed up the idea that a small amount of alcohol is truly health protective.
What did we find?
Based on the studies we reviewed, it’s likely that our previous understanding of how beneficial moderate drinking is for health is overblown. For certain health conditions, using these better methods for causal inference has overturned the idea that a little bit of alcohol reduces risk. The biggest one here was heart disease: most of the evidence we reviewed indicated that any amount of alcohol increased risk (for example, in these two large genetically-based papers). Similarly, for type 2 diabetes, there is little evidence of benefits except possibly in overweight women, and even in this group the benefit is likely smaller than previously estimated. Interestingly, our review did not rule out the possibility that a small amount of alcohol reduces risk for developing depression, but more research is required here to confirm this finding and identify the mechanism (it might be the social interaction facilitated by drinking, rather than biologically-active components of the alcohol itself).
What’s the takeaway?
It’s critical that more causally-focused research looking at alcohol and long-term health is conducted. Knowing whether there could be benefits of moderate alcohol use for certain conditions is important for individuals when making decisions about their own drinking, as well as for policymakers when generating population low-risk drinking guidelines. In fact, in Australia, modelling for the National Health and Medical Research Council indicates that if all evidence of alcohol’s health benefits was assumed to be wrong, the latest national guidelines would have recommended no more than 2.5 drinks per week – much less than the 10 drinks per week limit that became the policy.
It is also important to keep in mind that drinking at heavier levels can lead to severe health harms, and that there are a range of serious conditions, like cancer, where we know that every drink increases your risk. That’s why it’s not advisable to take up drinking if you don’t already consume alcohol, and if you do, to stick within your country’s low-risk recommendations.
Written by Dr Rachel Visontay, Postdoctoral Research Fellow, The Matilda Centre for Research in Mental Health and Substance Use, The University of Sydney.
All IAS Blogposts are published with the permission of the author. The views expressed are solely the author’s own and do not necessarily represent the views of the Institute of Alcohol Studies.