Wilson I.M., Graham K., Taft A.
BMC Public Health: 2014, 14(881), p. 1–11.
Unable to obtain a copy by clicking title? Try asking the author for a reprint by adapting this prepared e-mail or by writing to Dr Wilson at email@example.com.
What constitutes ‘alcohol-related’ domestic abuse, and to what extent can interventions designed to reduce the harms of alcohol also reduce domestic abuse?
This paper reviewed the effects of alcohol interventions on intimate partner violence at the levels of the individual, relationship, community, and population.
It found weak or no evidence for restricting sales of alcohol and the number of outlets that sell alcohol, or for pricing and taxation policies. Some positive effects were observed when brief alcohol interventions were used as an add-on to perpetrator programmes, but these effects were often not sustained.
Despite evidence linking problem drinking to intimate partner violence, the potential for alcohol interventions to reduce intimate partner violence has not been adequately tested. One reason for this is that studies have not focused on those most at risk of alcohol-related intimate partner violence.
Defined by the World Health Organization as “any behaviour within an intimate relationship that causes physical, psychological or sexual harm”, intimate partner violence is estimated to affect 30% of women with partners worldwide.
While there is evidence to suggest that alcohol plays a contributing role in aggression, its role in intimate partner violence specifically is complex and contested. However, it does seem that drinking by men is likely to play a more important role in intimate partner violence perpetration than drinking by women – reflecting both the gendered nature of drinking problems and intimate partner violence.
One of the ways that drinking may contribute to the risk and severity of intimate partner violence is by impairing people’s ability to address conflicts constructively, particularly when both partners have been drinking. Some people may intentionally engage in aggression or violence toward an intimate partner because they have the expectation that their behaviour will be excused by their drinking.
Although drinking can occur without intimate partner violence and intimate partner violence without drinking, both are sufficiently linked that the World Health Organization proposed that interventions to reduce the harm caused by alcohol might help prevent intimate partner violence.
For understanding violence and violence prevention in general, the World Health Organization recommends an “ecological framework” – acknowledging factors at different levels (individual, relationship, community, and population), which may influence violent behaviour on their own and in combination with other factors.
The featured review was the first of its kind to bring together a diverse range of studies relating to alcohol and intimate partner violence to examine the effects of alcohol interventions on intimate partner violence at all levels of the World Health Organization ecological framework.
Eleven databases were searched for English language studies published between January 1992 and March 2013. This included academic papers which had been ‘peer-reviewed’ (ie, subjected to the scrutiny of other scholars), and research from organisations outside of academia including government departments.
Studies were kept for analysis if they investigated whether an intervention or policy to reduce alcohol consumption was directly or indirectly associated with a change in intimate partner violence as a primary or secondary outcome. Eligible studies included people 18 years and older, and intimate partner violence perpetrated by either sex within a current or former heterosexual or homosexual relationship.
Forty studies (44 papers) met the initial selection criteria. The researchers then assessed whether the study design and sample size allowed outcomes to be attributed, at least in part, to the intervention or policy being evaluated. Eleven studies met this second tier of criteria. Due to questions regarding the integrity of one researcher, Dr. William Fals-Stewart, studies that he lead-authored or that were based on his data were excluded. [See this paper for more information]
Because only a small number of studies met the criteria and many did not test the assumption that an intervention’s impact on intimate partner violence may be contingent on its impact on alcohol consumption, the researchers included a further ten studies that provided some evidence that variations in intimate partner violence and alcohol consumption were linked, and that both of these were associated with the intervention. However, because of the way they were designed, these studies could not rule out the influence of other factors.
Selected studies were categorised according to the level of the problem they addressed: individual, relationship, community, and population.
Limited evidence to suggest that drinking interventions targeted at individuals reduce intimate partner violence
Across three studies with sufficient numbers of participants to be included, some positive effects were found when brief alcohol interventions were used as an add-on to so-called ‘batterer treatment’ for perpetrators of intimate partner violence who are also hazardous drinkers, and when brief interventions were used with younger people who were not dependent on alcohol. However, the effects were often not sustained.
A well-designed study recruited 252 hazardous-drinking men enrolled in perpetrator programmes, 98% of whom were required by courts to participate in the programme. It compared a standard perpetrator programme combined with a personalised brief alcohol intervention, to a standard batterer programme that included one session of group therapy for substance use. The alcohol intervention was a 90-minute therapist-led motivational interview with personalised feedback based on the participant’s current drinking. Significant reductions were found among people in the brief intervention group on drinking outcomes, though these were not sustained. No significant difference was found in the frequency of physical intimate partner violence, but there were reductions in severe psychological aggression and injuries to partners in the brief intervention group.
Another trial assessed a motivational intervention delivered by telephone with substance-using perpetrators of intimate partner violence recruited from the community, and not receiving counselling or serving a legally-imposed sentence. The intervention was based on a personalised assessment of intimate partner violence and substance use behaviours, and was compared with a control group in which participants received education materials by mail. Less than half of people included (43%) had a diagnosed substance use disorder. At the 30-day follow-up, men in the treatment condition reported engaging in less violence and consumed fewer drinks per week. The authors did not report whether reductions in alcohol consumption were associated with reductions in intimate partner violence.
The third study tested the effects of an integrated substance use and intimate partner violence treatment approach, which offered cognitive-behavioural therapy to men classed as “alcohol dependent”. Compared with participants receiving substance use-only therapy [12-step facilitation therapy, chosen as it “closely represented standard interventions in community clinics that solely target substance use in this population”], there was a greater reduction in intimate partner violence and significantly more days’ abstinence. However, there were no significant differences at six months for either drinking or physical intimate partner violence.
Reductions in alcohol consumption and reductions in intimate partner violence observed, but the designs of these studies prevent or rule out the changes being attributed to couple-based treatment.
Couple-based alcohol treatment interventions have been shown to be effective for reducing alcohol consumption and improving relationships among treatment-seeking people with alcohol and drug problems who are in a married or cohabiting relationship. The next step from this is testing the assumption that couple-based interventions which reduce problem drinking and improve relationship functioning in relationships where there is violence may also reduce intimate partner violence.
Five studies evaluated alcohol interventions involving couples. Of these, only one – a trial of a brief intervention that addressed both intimate partner violence and drinking – met criteria for inclusion in the review. Among 49 couples at university who were ‘dating’, this study assessed the effects of feeding back an assessment of their individual risk factors for aggression and intimate partner violence in a motivational interviewing style. There was a greater decrease in harmful alcohol consumption and physical aggression in the intervention group compared with people receiving minimal non-motivational feedback. However, analyses showed that the reductions in drinking and physical aggression were not related.
Four studies did not meet criteria for inclusion but provided some further context. Three evaluations of behavioural couples-based treatments to address drinking problems in one partner reported significant reductions in male-perpetrated violence and verbal aggression against their female partners (1 2 3 4 5), and another found decreases in male- and female-perpetrated violence where the female partners were dependent on alcohol. While this provides some evidence that alcohol consumption could be related to intimate partner violence, due to the design of the studies other factors could not be ruled out.
Weak evidence of an association between community-level policies or interventions (eg, reducing hours of sale, limiting the number of alcohol outlets in a given area) and intimate partner violence.
Alcohol policies such as restricting retail hours, or the numbers of alcohol outlets within a geographical area (known as ‘outlet density’) aim to decrease consumption and related harms by increasing the effort to obtain alcohol. At the community level, the assumption can be tested that decreasing drinking opportunities and overall consumption among those who perpetrate alcohol-related intimate partner violence will reduce intimate partner violence.
One study examined the effect of a city-wide bar closing time of 11pm in a mid-sized Brazilian city with high rates of alcohol and violence. Analysing homicide rates over a 10-year period and assaults against women over a 5-year period, this study found that earlier bar closing was associated with a significant reduction in homicides in the first three years after the restriction, and a non-significant reduction in assaults against women. Conclusions were limited by the different time periods, and ‘assaults against women’ including but not being limited to intimate partner violence.
Analysing licensing data and police-recorded intimate partner violence incidents in Melbourne (Australia) over 10 years, another study found a positive association between intimate partner violence and outlet density, with a large and significant effect found for places where alcohol is sold but cannot be consumed such as convenience stores (known as ‘off-licensed premises’). An increase in one outlet per 1,000 residents translated into a 29% increase in the average rate of intimate partner violence.
A Californian study using two police-recorded measures of intimate partner violence (reported crimes and calls to the police) also found an association with the number of off-licensed premises, but not with ‘on-licensed premises’ (places that are licensed for people to consume alcohol). However, a second study by the same authors using data over a shorter period of time found that the greater the numbers of on-licensed premises in a given area, the greater the risk of emergency department visits related to intimate partner violence.
A Western Australian study found a significant association between the number of sales in off-licensed premises and assaults in private residences, suggesting that the amount of alcohol sold/consumed (not just the number of outlets) could influence intimate partner violence.
Another study based on self-reported intimate partner violence from a national survey in the United States found a stronger relationship between the numbers of alcohol outlets within a geographical area and male-to-female physical intimate partner violence for couples who had drinking problems than for couples without. A further study in the United States district of Columbia found an association between police call-outs for intimate partner violence and the number of off-licensed premises. The risk was greater in areas with a high volume of off-licensed premises on weekends when heavier drinking was more likely to occur.
Weak or indirect evidence that increasing the price of alcohol reduces intimate partner violence.
Four studies evaluated the relationship between alcohol pricing or taxation policies and intimate partner violence, and of these, three studies met the design criteria to be included in the review.
Only one study, conducted in the United States, reported a significant relationship between the price of alcohol and intimate partner violence. Modelling the effects of changes in price on the probability of self-reported “husband abuse” and “wife abuse” from a 1985 national family violence survey and two annual follow-ups, the study found that a 1% increase in price was associated with a reduction of 3.1–3.5% in wife abuse. No such association was found for husband abuse. The study did not include measures of alcohol consumption, so it was not possible to assess the extent to which changes in consumption accounted for the findings.
A longitudinal study examined the relationship between changes in alcohol taxes, alcohol consumption and female homicide rates (where most of the women killed had been killed by an intimate partner) across 46 US states between 1990 and 2004. It found a significant association between increases in alcohol tax and reductions in consumption, and furthermore between reductions in consumption and reductions in intimate partner violence. However, the direct relationship between increased alcohol taxes and reduced intimate partner violence was not statistically significant. In explaining their results, the authors questioned the extent to which those who consume alcohol and commit homicide are sensitive to price.
The third and final study assessed the impact of a range of interventions (including changes to local and regional beer taxes) on intimate partner homicide and intimate partner homicide involving firearms. The analysis covered 46 of the largest US cities over 24 years. No relationship was found between increased beer taxes and reduced intimate partner homicide. While the study did not include measures of alcohol consumption, the authors suggested that tax increases may have been too small to affect drinking to the extent needed to influence intimate partner homicide. The outcome measure included all victims of intimate partner homicide regardless of gender, though evidence shows that women are the overwhelming majority of victims of homicide by an intimate partner and alcohol is more likely to be involved in male-to-female intimate partner violence.
Despite evidence linking problem drinking to intimate partner violence, the potential for alcohol interventions to reduce intimate partner violence has not been adequately tested. This is possibly because studies have not focused on those most at risk of alcohol-related intimate partner violence.
Alcohol-related intimate partner violence is a complex, multi-dimensional problem, much neglected in intervention and prevention research. Despite the consistent link between alcohol consumption and intimate partner violence, and evidence that drinking contributes to increased risk and severity of intimate partner violence, the featured review found few studies of the effects of alcohol interventions and alcohol policy interventions on intimate partner violence where the design allowed changes in intimate partner violence to be clearly attributed to the intervention.
Research is urgently needed to investigate the potential impact of alcohol interventions on intimate partner violence at the levels of the individual, relationship, community, and population. This should include more reliable measures distinguishing alcohol-related intimate partner violence from intimate partner violence not involving alcohol.
commentary While contemporary UK national policy has focussed on reducing the harms associated with excessive drinking – particularly violence and disorder – comparatively less attention has been given to tackling ‘hidden’ alcohol-related violence outside of nightlife settings and urban centres such as intimate partner violence (otherwise known as domestic abuse).
Investigating whether interventions aimed at reducing alcohol consumption could also reduce intimate partner violence, the featured review tested two assumptions: firstly that those alcohol interventions would indeed reduce consumption; and secondly that there is a link between consumption and intimate partner violence.
Laying the groundwork, the authors pointed to “clear and consistent evidence of an association between alcohol consumption and [intimate partner violence]” as well as “evidence that alcohol consumption by one or both partners is associated with increased severity of [intimate partner violence]”. However, they said the review itself was ultimately limited by a lack of studies designed to allow changes in intimate partner violence to be clearly attributed to the interventions, and by a lack of studies focusing on those most at risk of “alcohol-related intimate partner violence”.
The authors alternated between talking about intimate partner violence and alcohol-related intimate partner violence throughout the paper. While alcohol-related intimate partner violence would understandably have been difficult to distinguish from all other intimate partner violence at a community or population level, the authors were also unable to define it at a theoretical level. The featured review could only say with certainty that the link between problem drinking and domestic abuse is “complex and contested”, meaning that even if a study could focus in on cases where domestic abuse and heavy or dependent drinking co-existed, alcohol could be a causal factor, a contributory factor, or possibly an unrelated factor.
Describing the potential role of drinking on intimate partner violence, the authors concentrated on aggression and conflict – how aggression may be heightened and how partners may be less able to address conflict after or when drinking. ‘Conflict’ is a loaded term in domestic abuse research, and it pitches this review on one side of a long debate about whether it is best to understand domestic abuse as being about ‘conflict’ or ‘control’ – the things people do in a currently unhealthy relationship to deal with conflict, or the violent and abusive behaviours that people may use to control a partner. In the former the language has the potential to disperse responsibility in a couple, whereas the latter thinks more definitively in terms of a perpetrator and victim.
There may be no single way of understanding the dynamics in a relationship that can create a climate where domestic abuse occurs, but focusing on the role of conflict to the exclusion or minimisation of other factors, or conflating conflict with abusive behaviour, is problematic. To see how this can pan out, click to unfold the supplementary text.Domestic abuse can include, but is not limited to, emotional or psychological abuse, physical abuse, sexual abuse, financial abuse, harassment and stalking, and ‘coercive control’ (“a pattern of intimidation, degradation, isolation and control with the use or threat of physical or sexual violence”). And while alcohol may have a causal or contributory role in some types of domestic abuse for some people, it may also feature (and certainly complicate matters) but not be directly related.
More certain is that the presence of heavy or dependent drinking can have implications for the way that health, substance use, and criminal justice professionals view domestic abuse situations:
• if the victim of domestic abuse has been drinking, their chaotic or erratic behaviour may be blamed (and by extension they may be blamed) for ‘provoking’ abusive or violent behaviour in their partner;
• if the victim has been drinking, any claims of abusive or violent behaviour may not be deemed credible;
• and because disagreement and conflict may almost be expected in families dealing with substance use, professionals observing the situation may not immediately see domestic abuse.
In the Effectiveness Bank, a hot topic called Focus on the families includes a section on domestic abuse, breaking down some of the key issues for substance use practitioners who frequently encounter people with overlapping issues such as problem drinking and drug use, domestic abuse, and mental illness. Headlining this was the importance of professionals knowing and feeling confident enough to distinguish between family conflict and domestic abuse, and being aware of the risks of assuming either will automatically be alleviated when substance use problems are resolved. Click here to read more.
The featured paper was structured around the World Health Organization’s ecological framework (unfold supplementary text), which is based on evidence that “no single factor can explain why some people or groups are at higher risk of interpersonal violence, while others are more protected from it”. Despite this, it took surprisingly little account of the other factors (ie, beyond alcohol) that increase people’s risk to domestic abuse at the individual, relationship, community, and societal levels, or affect people’s sensitivity to alcohol interventions, such as the expected impact of raising the cost of alcohol among different income levels.
To take gender as an example, in the introduction the review authors acknowledged the greater likelihood of women being victimised, the “unequal power relationships between men and women”, the greater effect of alcohol on men’s aggression, and the gendered nature of both problem drinking and intimate partner violence, but neglected to reflect on this when discussing the implications of their findings, even though many of the studies predominantly or solely featured male perpetrators and female victims.
Emphasising the importance of investigating alcohol-based interventions – but simultaneously seeming to downplay the potential of gender-based or gender-sensitive domestic abuse interventions – the authors said in their introduction:
Further investigation of the effects of alcohol prevention on [intimate partner violence] is important because direct interventions addressing violence against women have been shown to have limited impact.”
The source on which this claim was based did not appear to draw the same conclusion. The main conclusion was that “Much has been learned in recent years about the [incidence] of violence against women, yet information about evidence-based approaches in the primary care setting for preventing intimate partner violence is seriously lacking” – indicating that the conclusions that can be drawn are limited, not the potential of the interventions themselves. Incidentally, this paper was published just over a decade before the featured paper, so it would be reasonable to ask whether there would have been further evidence to take into account by this point; and it considered only interventions in a primary care context, not, for example, a criminal justice context.
To see a briefing paper from the Institute of Alcohol Studies on alcohol, domestic abuse and sexual assault, click here. And for access to articles analysed for the Effectiveness Bank about some of the issues raised in this entry, click to view the following Effectiveness Bank collections: focus on women, couples therapies, and alcohol and families.
Last revised 12 March 2018. First uploaded 14 February 2018
DOCUMENT 2012 The government's alcohol strategy
HOT TOPIC 2017 Controlling alcohol-related crime and disorder
REVIEW 2015 Prevention of addictive behaviours