Intimate partner violence: bringing facts into light under a public health perspective

The contribution of violence to the overall burden of diseases and death is undeniable, yet avoidable. This recognition urges societies to make violent behavior visible and to eradicate it. Although various types of violence have been distinguished with victims and perpetrators being often found amongst the socially most vulnerable groups of the population, its pervasiveness suggests that a great amount of violent behavior is still invisible. Given its inherently private nature, intimate partner violence (IPV) is one of the hardest to grasp types of violence, but still one of the most frequently experienced, thus a challenge for the design of health initiatives and healthcare planning. The work developed in this thesis aimed to address IPV in adults from developed societies. Following a public health perspective we performed seven studies, starting with a focus on methodological aspects of IPV measurement. Acknowledging that cross-regional comparisons are often hampered by different definitions and methodological choices, we conducted a population-based study to describe the magnitude of four IPV types experienced by women and men across six different European urban centers. Then, we explored socioeconomic inequalities in the expression of IPV in each gender and measured the impact of IPV in health-related quality of life (HRQoL). Turning our focus to healthcare, we synthesized a tool to measure the quality of mental healthcare provision to socially marginalized groups throughout Europe and explored country-level factors influencing such quality. In the last study presented in this thesis we tested whether IPV could be associated with the decision to forgone or delay healthcare. To execute these series of studies, we used data gathered in the scope of two European projects. The DOVE project, a cross-sectional international multicenter study designed to measure IPV in the general population of adult men and women living in eight European cities was the basis for five studies of this thesis. The PROMO project, an international multicenter study conducted in 14 European cities designed to identify best practice in the delivery of health and social care for people with mental health problems who belong to one of six defined marginalized groups (long-term unemployed, homeless, street sex workers, asylum seekers/refugees, illegal immigrants and travelling communities) was the basis for one of this thesis’ studies. In the following paragraphs we present a brief description of the objectives, methods and results of each study performed. Study I: We aimed to map existing evidence on strategies to measure male and female IPV. To pursue this goal we conducted a scoping review of the literature. Pubmed®, ISI Web of Knowledge® and Scopus® databases were searched from inception to 2014 and the abstracted information included type of instruments, samples, prevalence estimates, psychometrics and publications’ year and region. In total 1098 studies were analyzed. The most commonly followed strategy used all over the world to assess IPV, was the creation of study specific questions (30.3%). This was the preferred option found when dealing with large and community samples. Regarding standardized instruments, the Conflict Tactics Scales (CTS) was the most frequent choice, whereas for clinical samples, the preferred tool was the Abuse Assessment Screen (AAS). Prevalence estimates were generally higher when the original versions of the CTS were used. This review showed a predominance of studies from North America, which represents more than 50% of the publications analyzed. Study II: In this methodological note we describe the design, methods, procedures and characteristics of the population involved in a multicenter international study designed to compare IPV in eight countries. Electoral roles, municipal registries, random-route and random-digit-dialing were used to sample women and men aged 18-64, living in Ghent-Belgium (n=245), Stuttgart–Germany (n=546), Athens–Greece (n=548), Budapest–Hungary (n=604), Porto-Portugal (n=635), Granada–Spain (n=138), Östersund–Sweden (n=592) and London–United Kingdom (UK) (n=571). Three methods were used to administer different sections of a common questionnaire: self-completed (IPV), face-to-face (demographics, health) or mail. Five-age strata population fractions for sex and education were computed and population fraction ratios were used to evaluate samples’ representativeness. Differences in the age distributions were found among women from Sweden and Portugal and among men from Belgium, Hungary, Portugal and Sweden. Over-recruitment of more educated respondents was noted in all sites. Study III: We aimed to assess four types of IPV among adult men and women from the general population of six different European urban centers (Athens, Budapest, London, Porto, Östersund and Stuttgart). IPV types were measured with the Revised Conflict Tactics Scales (CTS2). Sex- and city- differences in past-year prevalence were examined considering victims, perpetrators or both and considering violent acts’ severity and repetition. Male victims of psychological aggression ranged from 48.8% (Porto) to 71.8% (Athens) and female victims from 46.4% (Budapest) to 70.5% (Athens). Male and female victims of sexual coercion ranged from 5.4% and 8.9% respectively in Budapest to 27.1% and 25.3% in Stuttgart. Male and female victims of physical assault ranged from 9.7% and 8.5% respectively in Porto, to 31.2% and 23.1% in Athens. Male victims of injury were 2.7% in Östersund and 6.3% in London and female victims were 1.4% in Östersund and 8.5% in Stuttgart. IPV differed significantly across cities. Men and women predominantly experienced IPV as both victims and perpetrators with few significant sex-differences within cities. Study IV: We explored the association between socioeconomic position (SEP) and IPV considering the perspectives of men and women as victims, perpetrators and as both (bidirectional). A total of 3496 adults (18-64 years) were randomly selected from the general population living in Athens, Budapest, London, Porto, Östersund and Stuttgart. Physical IPV was measured with the CTS2. Education (primary, secondary and university), occupation (upper white, lower white and blue collar) and unemployment duration (never, ≤12 months and >12 months) were considered SEP indicators. Logistic regression models were fitted and age- and city-adjusted odds ratios, 95% confidence intervals (OR, 95%CI) computed. Past year physical IPV was declared by 17.7% of women (3.5% victims, 4.2% perpetrators and 10.0% bidirectional) and 19.8% of men (4.1% victims, 3.8% perpetrators and 11.9% bidirectional). In women, low educational level (primary vs. university) was associated with victimization (OR, 95%CI: 3.0, 1.2-7.5) and with bidirectional IPV (4.1, 2.4-7.1). Blue collar occupation (vs. upper white) in women was associated with victimization (2.1, 1.0-4.5), perpetration (3.1, 1.4-6.8) and bidirectional IPV (3.9, 2.3-6.8). Unemployment duration was associated with male perpetration (OR, 95%CI, in perpetrators with >12 months of unemployment vs. never unemployed: 3.4, 1.5-7.7) and with bidirectional IPV in both sexes (women: 1.8, 1.2-2.8; men: 1.7, 1.0-2.8). Study V: We assessed HRQoL in the presence of physical and sexual IPV considering sex and violence directionality. Adult men and women (n=3496), randomly selected from the general population of six European cities were assessed using the CTS2 and the Medical-Outcomes-Study 36-item Short-Form Health Survey (SF-36). Mean scores[standard error] of SF-36 physical and mental health summary scales in victims, perpetrators, bidirectional cases and those not involved in past-year physical assault and/or sexual coercion were compared using age-, education- and city-adjusted linear regression. We found that the HRQoL physical dimension was significantly lower in women involved in bidirectional physical assault (48.00[0.58]) compared to those declaring no physical assault (49.75[0.26]). For the mental dimension, women involved in physical assault (as victims, perpetrators or both) presented significantly lower mean scores than women reporting no physical assault. Women victims-only of sexual coercion (44.74[0.86]) and victims or involved in bidirectional concomitant physical and sexual IPV (41.43[2.36] and 43.34[1.30], respectively) also presented lower mental mean scores. In men, significantly lower mental mean scores were found in the bidirectional physical assault group (46.34[0.78]) and among those involved bidirectionally in both physical and sexual IPV (46.17[1.30]). Study VI: We assessed the organizational characteristics of services providing mental healthcare for marginalized groups in European capital cities and explored the associations between organizational quality, service features and country-level characteristics. A total of 617 services were assessed in two highly deprived areas in 14 European capital cities. A Quality Index of Service Organization (QISO) was developed and applied across all sites. Service characteristics and country level socioeconomic indicators were tested and related with the Index using linear regressions and random intercept linear models. The mean (standard deviation) of the QISO score (minimum= 0; maximum= 15) varied from 8.63 (2.23) in Ireland to 12.40 (2.07) in Hungary. The number of different programs provided was the only service characteristic significantly correlated with the QISO (p<0.05). The national Gross Domestic Product (GDP) was inversely associated with the QISO. Nearly 15% of the variance of the QISO was attributed to country-level variables, with GDP explaining 12% of this variance. Study VII: We examined in a sample of European adult (18-64 years) men and women the relation between forgone healthcare and involvement in intimate partner violence (IPV) as victims, perpetrators or both (reciprocal or bidirectional violence). We evaluated 3496 participants, randomly sampled from the general non-institutionalized population of six European cities (Athens, Porto, London, Budapest, Östersund and Stuttgart) who responded to a common questionnaire about IPV and health related characteristics. IPV was assessed with the CTS2 and forgone healthcare considered using the question “Have you been in need of a certain care service in the past year, but did not seek any help?”. To measure the association between forgoing healthcare and IPV, chi- square test was used in bivariate analysis and odd ratios and 95% confidence intervals (OR, 95%CI) were calculated fitting logistic regression models and considering potential confounders. Participants who experienced any act of past year IPV reported more often to forgone healthcare compared to those not involved in violence (18.6% vs. 15.3%, p=0.016). Declaring to have been both a victim and a perpetrator of any violent act was associated with forgone healthcare, independently of having chronic diseases, their self-assessed health status or having felt financial strain (adjusted odds ratio, 95% confidence interval: 1.41, 1.09- 1.81). An association of similar magnitude was observed among victims, although statistically non-significant (1.35, 0.89-2.04). The main conclusion of these studies can be summarized as follows: There are trends in the choice for a particular IPV measurement instrument according to the method of administration and setting of application. Clinical practice and research are hindered by lack of comprehensive evaluation of existing IPV screening tools and studies replicating associations between violence and health outcomes using similar measures of exposure are, therefore, needed. In order to conduct a multicenter study on IPV, a number of distinct ethical and logistical challenges must be addressed. Limitations to the establishment of probabilistic samples and different methods of administration are plausible explanations for demographic differences observed across sites where such endeavor is performed. However, through the utilization of a common research protocol with the same structured questionnaire, accurate estimates of IPV frequency in the general population are possible to obtain. Across the general population of adults residing in Athens, Budapest, London, Porto, Östersund and Stuttgart, the 12-month prevalence of psychological, physical, sexual and injury as forms of IPV, varies significantly, although few sex-differences are observed within cities. Most IPV is bidirectional or reciprocal, i.e., most subjects report they have been both victims and perpetrators of violent acts within their intimate relationships. Furthermore, physical IPV is associated with a disadvantaged socioeconomic position and physical and sexual IPV negatively influence HRQoL, with lower scores in the mental component being evident among female victims and among males and females involved in IPV bidirectionally. This supports the need for a gender-inclusive approach to IPV that considers the perspectives of both victims and perpetrators. Looking from the perspective of healthcare provision, socioeconomic contextual factors, in particular the national GDP are likely to influence the organizational quality of services providing mental healthcare for marginalized or vulnerable groups. Also, the influence of IPV on forgone healthcare stresses the need to address IPV amongst barriers in the access to healthcare.