The health and economic benefits of reducing intimate partner violence: an Australian example

Intimate partner violence (IPV) is an important global public health problem. It was ranked 23^rd in terms of Disability Adjusted Life Years (DALYs) arising in women in the most recent update of the Global Burden of Disease (GBD) Study 2010, following after other important risk factors such as high total cholesterol, suboptimal breastfeeding, alcohol use, physical inactivity, high blood pressure and dietary risks [1].

Globally, IPV is a major public health concern because it is the most common form of violence that women experience [2]. IPV has been defined for an international survey as women who had a current or former intimate partner and had experienced violence from that partner during their lifetime [3]. Alternate terminologies include domestic violence, family violence or relationship violence. IPV refers to violence occurring between people who are, or were formerly, in an intimate relationship and were subject to economic, psychological or emotional abuse through to physical and sexual violence [4].

There have been several attempts to make meaningful cross-country comparisons in rates of IPV [2, 5‐10]. One of the most comprehensive (including past and current exposure) to compare rates of IPV across countries was the International Violence Against Women Survey (IVAWS). One of the main aims of the IVAWS was to facilitate international comparisons by using standardised measurement tools and techniques to establish prevalence in each participating country. Johnson et al. [3] reported data for Australia and eight other countries. Prevalence of IPV was highest in Mozambique (40 %) and lowest in Hong Kong (9 %), with other countries in between: Czech Republic (37 %), Costa Rica (36 %), Australia (27 %), Denmark (22 %), Poland (16 %), Philippines (10 %) and Switzerland (10 %) [3]. In 2010, Devries and colleagues estimated women’s lifetime prevalence of IPV for 21 regions [2]. Globally, 30 % of women aged 15 and over were reported to have experienced IPV during their lifetime. In this latter study,  the greatest prevalence of IPV was found in Central Sub-Saharan Africa (66 %) and lowest in East Asia (16 %), while, Australasia had an IPV prevalence of 28 % [2]. In the most recent personal safety survey for Australia, 17 % of women surveyed in 2012 had experienced IPV since the age of 15 [11]. In Victoria (Australia), IPV contributes 3 % to the total disease burden and 8 % in women aged 15–44 [4]. In this younger age group, it is the leading contributor to death and disability, being responsible for more disease burden than other well-known risk factors, such as high blood pressure, smoking and obesity [4].

Recognizing its importance on health and production consequences, the prevention of IPV should be a priority in all countries [12]. A reduction in prevalence of IPV can have important impacts on society, including enhanced psychological well-being, which at a population level improves general health and productivity. Quantifying the economic impact of IPV is also important, since it demonstrates the extent of these societal impacts. However, few studies of the costs of IPV have been undertaken, and the methods used vary since the research questions also differ e.g. including impacts of IPV on both men and women, including a larger range of indirect or intangible costs such as counting the costs to future generations, costs of pain and suffering, consumption related costs, or transfer costs [5]. This means that these studies are generally incomparable due to the different definitions and assumptions applied (including the time horizon for counting the costs), limiting the ability to compare the economic implications of IPV with other important conditions. Previous costing studies in Australia have found IPV was responsible for increased costs to the community between AUD8.1 billion and AUD9.9 billion (2007–08 dollars) annually [5, 13] and was predicted to grow to AUD15.6 billion by 2021–22 if no action is taken [13]. These earlier studies provide estimates of the total economic costs that could be avoided if the violence was to be eliminated, rather than estimating the extent of societal benefits if the prevalence of IPV against females was to be reduced by a feasible amount.

The aim of this study was to separately quantify the health and economic benefits within the paid and unpaid production sectors and health sector that could be achieved following a feasible reduction (as opposed to a complete elimination) of IPV in the Australian female adult population. This study on IPV was part of a larger study funded by the Victorian Health Promotion Foundation (VicHealth) whereby the benefits of feasible reductions in the prevalence of long term harmful alcohol consumption, physical inactivity, high body mass index, tobacco smoking, inadequate consumption of fruit and vegetables, and IPV were estimated using a standardized approach [14]. Therefore, the advantage of our study in relation to prior research in this field is that we can reliably put the findings for IPV within a broader policy context. This is because we  highlight not just the total attributable burden based on the aspects we have measured, but also what could be gained if a realistic absolute reduction in IPV prevalence could be achieved.

Relevant diseases and injuries associated with IPV include depression, anxiety, cancer, chronic obstructive pulmonary disease (COPD), and suicide [15, 16]. To estimate the health and economic benefits to society of reducing the prevalence of IPV, the impact of an absolute reduction in the prevalence of IPV levels for the 2008 Australian female population was calculated. This was quantified as reduced: a) incident cases of preventable IPV-related diseases/injuries; b) deaths; and c) disability adjusted life years (DALYs). We simulated benefits of risk factor prevalence reductions for this one reference year. However, the time horizon for benefits estimation was based on a lifetime perspective for that 2008 female population cohort.

The selected feasible prevalence reduction targets for IPV in Australia were based on consensus by an expert Advisory Committee who were presented with a review of the literature. Since the literature on prevention interventions did not include readily usable estimates for feasible reductions, prevalence rates of IPV from other countries were consulted.

Comparing prevalence rates in Australia to other countries provides indirect evidence of possible achievable prevalence levels from a country with a broadly comparable social, demographic and economic profile [14]. The feasible reductions could then be modelled against attainment of lower prevalence levels of IPV as observed in a comparable country to Australia (referred to as an ‘Arcadian’ normal [17]). However, comparing across countries posed some difficulties since some prevalence estimates were based on all women, others based on only those women who have ever had a partner, and others based on women who currently have a partner. Furthermore, the definition of an ‘intimate partner’ (husband, de facto, cohabiting partner, boyfriend, etc.) and the definition of what constitutes violence differs across studies. Using data from the IVAWS study overcame many of these issues [3]. However, we acknowledge that in this study there were important differences between participating countries, including sample size, age distribution of respondents, response rate and method of interview. Among the countries with lower prevalence estimates than Australia, Poland, Philippines and Hong Kong were ruled out as comparators for being socio-culturally different. Although broadly comparable, Switzerland had a prevalence of only 10 % which seemed remarkably low, and was considered by the Advisory Committee too ambitious a target for Australia to achieve in the short term. Denmark, with a prevalence of 22 %, was selected as the most appropriate Arcadian comparator. The Advisory Committee agreed that there was the potential to reduce IPV prevalence in Australia by up to 5 percentage points, shifting the estimated current prevalence of 27 % to 22 %. A 5 percentage point absolute reduction in IPV was selected as an ideal feasible target (22 %), with a progressive target of 24.5 % also assessed.

Past and current exposure to IPV is most influential in determining the current health burden, health sector costs and production and leisure costs, but past exposure is not amenable to prevention efforts. The effects of past exposure can only be ameliorated through current health services. It is current levels of IPV that are amenable to preventive interventions and, thus gradually over time, the lifetime prevalence of IPV can be expected to fall. If a woman ceases to be exposed to current IPV, we have modelled that she will change from a high level of psychological distress to a moderate level, since she remains exposed to lifetime (past) IPV.

The demographic data and days of reduced activity for females with moderate psychological distress and high or very high distress by age and workforce status are presented in Table 1. Females with high or very high distress participated less in the workforce than females with moderate distress (Fig. 1). The females with high or very high distress in the workforce, took more days off work compared with females with moderate distress. In addition, females with high or very high distress not in the workforce or past retirement age had more days of reduced activity compared to females with moderate distress.

If the prevalence of IPV in the adult Australian population was reduced by 5 percentage points, the estimated 34,000 annual new cases of IPV-related disease could be reduced by 6000 (18 %); the 440 annual deaths attributed to IPV could be reduced by 74 deaths (17 %); and the estimated 29,000 DALYs attributed to IPV could be reduced by about 5000 (17 %) (Table 2). Nearly half of these benefits would arise if a 2.5 percentage point progressive absolute reduction in prevalence target was achieved.

The estimated health and economic benefits from reduced IPV resulted in total potential opportunity cost savings of AUD377 million, which is comprised of AUD38 million to the health sector (19 % of IPV attributable annual health sector costs), AUD94 million in workforce production (FCA), AUD98 million in leisure based production and AUD147 million in home based production (Table 3). The latter production costs represent approximately 18 % of the total attributable to this risk factor (AUD1,801 million). The largest component of these total potential opportunity cost savings would occur in household production and leisure (Fig. 2).

The primary finding of this study is that a feasible 5 percentage point absolute reduction in prevalence of IPV can lead to total potential opportunity cost savings of AUD377 million, with only 10 % of the savings arising in the health sector. The largest potential saving from reduced lifetime prevalence of IPV in the female population will occur in home based production, followed by leisure, and workforce production. These savings would be much larger (approx. AUD2 billion) if IPV was able to be totally eliminated from Australia (AUD207 million in health sector, AUD1,801 million [FCA] in production and leisure). However, our aim was to identify savings that were realistic and relevant to the setting of future public health campaigns focusing on risk factor reduction strategies.

An important aspect of the present study was the separate identification of benefits across workforce, household production and leisure time associated with reduced IPV. Earlier studies that have not accounted for home-based and leisure production underestimate potential savings in the context of IPV prevention. While the current methodology does identify the contribution of household production and leisure, it is conservative since it excludes an assessment of the additional time available for household and leisure tasks, which the cohort of the working age 2008 female population will have following retirement, through their remaining life expectancy.

We are able to compare results of reducing IPV with five other risk factors assessed using the same methodology and data sources [14]. We found that the total economic benefit of reducing IPV was ranked third behind gains following reductions in high risk alcohol consumption and tobacco consumption. This is jointly explained by the different reduction targets deemed feasible for each respective risk factor, the distribution of the population exposure to the risk factors across age groups and sexes as well as the disease and injuries associated with risk factors. While we recognise a potential equity issue by applying lower average wage rates for females compared with males, we were able to test its impact in sensitivity analysis. Adopting annual average unisex wages increased the overall estimates IPV gains by AUD25 million. However if we had analysed and reported all the risk factors with unisex wage rates, the relative ranking of the gains through a feasible reducing IPV would still rank third behind the gains following feasible reductions in alcohol consumption and tobacco use.

The distribution of opportunity cost savings differs for IPV relative to other risk factors we have evaluated using the same methodology and data sources [14]. For all others, the savings in the health sector represent the larger proportion of total savings (Fig. 2). Health sector cost savings due to reduced prevalence of IPV were only found to be 10 % of the total savings. In all other risk factors we evaluated (except physical inactivity), the opportunity cost savings to the health sector were as high as the productivity and leisure costs, if not two to three times greater [14]. This arises primarily due to four main reasons: (i) the benefits of reducing IPV apply mostly to the subset of females who are providing the child bearing and child raising roles during their productive years; (ii) there are disproportionately lower rates of treatment of mental illness associated with IPV compared to physical impacts of cancers and heart disease associated with other major risk factors, (iii) only 41 % of women with any mental health problem (assessed as lifetime prevalence and 12 months symptoms prior to the survey date) sought help for their mental health problem as reported in the most recent National Mental Health Survey [35]; and (iv) a General Practitioner was likely to be the service provider most frequently sought by females with a mental health problem, hence there is an absence of expensive hospital costs in the treatment of anxiety disorders and depression [35].

Comparisons with other IPV research in this field are difficult because of differences in methods, assumptions, populations included and the scope of the costs analyzed. However, earlier assessment of the costs of domestic violence in Australia also highlighted that household, leisure and production impacts are where the greatest costs attributable to IPV may be found [5].

Our estimates are very conservative in comparison to other studies where estimates of potential cost savings to future generations of children in families experiencing domestic violence are included. In the studies reported by Access Economic [5] and KPMG, second generation costs of AUD28 million were to be expected from a 10 % reduction in IPV prevalence [13]. This category includes short-term costs of providing protection and other services (such as child protection services, childcare and remedial/special education) to children of relationships where there is domestic violence, and longer-term costs imposed on society by these children as they grow older (such as increased crime and future use of government services). Our estimates, which take a narrower perspective and address realistic reductions in prevalence, are an underestimate of total current and future generation societal costs. In addition, by excluding males subjected to IPV (in 2012 estimated to be 5.3 % of Australian males experiencing physical or sexual abuse since the age of 15 by a partner) [11], we have understated the potential economic impacts of reducing the prevalence of IPV.

Several other limitations of the current study must be acknowledged. Since women are not directly questioned in the NHS on the issue of IPV, we have relied on surveyed levels of psychological distress amongst surveyed females as a proxy. We acknowledge that not all women with psychological distress, as measured on the Kessler 10, are IPV victims, and nor are women with psychological distress confirmed as having psychological conditions. Since direct evidence was unavailable from the NHS for women with IPV, we chose to use high psychological distress as a reasonable proxy, however this may have led to sources of under- or overestimation, of workforce impacts, since it remains unclear if women who experience IPV have similar patterns of psychological distress as women who do not. The Burden of Disease studies have provided evidence  that depression and anxieties are the largest components of illness attributable to IPV [36]. For this study, the levels of depression and anxieties was adequately captured by the Kessler 10 score used in the NHS [20]. To-date, there remains a lack of consistent and direct evidence for IPV in Australia and the ABS has only recently released a national data framework for enabling a standardized approach to capturing data on family, domestic and sexual violence, so in the future more direct evidence may be available for these studies [37]. Another potential limitation is that we have not controlled for the presence of other risk factors, or socioeconomic status, in the analysis and this may be a source of potential over- or under-estimation of benefits. We have assumed risk reversibility and risk reduction arising from absence of current IPV, and that the cross sectional evidence in the NHS related to workforce behaviours would be associated with a future change in IPV status of women. Longitudinal cohort data are missing at this point to confirm our assumptions. Despite these limitations, we attempted to address such sources of under- or over-estimation by undertaking detailed sensitivity and uncertainty analyses.

Investment in disease prevention and health promotion in Australia is dwarfed by avoidable spending on disease treatment. The findings of this study contribute important new knowledge about the major impact of IPV on the productivity of females in both the paid and unpaid sectors, as well as health sector expenditure. Our findings reinforce the argument that greater investment in IPV reduction strategies is required and economically justified, particularly in young women. Importantly, our study highlights the need for better data in this field so that the economic impacts can be assessed using direct evidence. We applaud the ABS and the Australian Government who have now established a national data framework to address these methodical issues [38]. While anxiety and depression represent the greatest proportion of the disease burden attributable to IPV, women subjected to IPV tend to have other behavioural risk factors for chronic disease and the impacts extend to the whole family, in particular children who can also suffer health impacts from witnessing IPV. Our findings provide a much needed business case to support policy makers in addressing this important societal issue.