- Split View
-
Views
-
Cite
Cite
Noemie Roland, Christophe Leon, Enguerrand du Roscoat, Henri Panjo, Marie-Josephe Saurel-Cubizolles, Virginie Ringa, Witnessing interparental violence in childhood and symptoms of depression in adulthood: data from the 2017 French Health Barometer, Family Practice, Volume 38, Issue 3, June 2021, Pages 306–312, https://doi.org/10.1093/fampra/cmaa127
- Share Icon Share
Abstract
Every year, it is estimated that 143 000 French children are exposed to interparental violence (IPV). This exposure may have deleterious lifelong impact on mental health.
To investigate the associations between exposure to IPV in childhood and, as adults, the presence of depressive symptoms during the past year, suicidal thoughts and lifetime suicide attempt.
Our study is based on data from the 2017 French Health Barometer, a general population cross-sectional phone survey. 25 319 adults living in Metropolitan France responded. Data were weighted to obtain a representative sample of the French population. Associations between childhood exposure to IPV and each of three outcomes in adulthood—symptoms of depression, suicidal ideation in the past year and lifetime suicide attempt—were studied by calculating odds ratios (ORs) and their 95% confidence intervals (CIs), stratifying by sex and using multivariate models (logistic regression).
After adjustment for age, social variables, negative life events in childhood and lifetime history of sexual violence, reporting witnessed IPV is significantly associated with the presence of symptoms of depression during the past 12 months (aORmen = 1.88, 95% CI = 1.49–2.38, and aORwomen = 2.00, 1.72–2.32), suicidal thoughts in the past 12 months (aORmen = 1.97, 1.48–2.64, and aORwomen = 2.35, 1.89–2.93), and at least one lifetime suicide attempt (aORmen =2.39, 1.83–3.11 and aORwomen = 2.66, 2.25–3.16).
Associations shown between a history of exposure to IPV and three mental health indicators in adulthood underline the need to study the lifelong impact of IPV.
• A history of childhood intimate partner violence (IPV) exposure was reported by one in five people aged 18–75 years.
• Exposure to IPV doubles the risk of depressive symptoms and suicidal behaviour in the past year.
• Exposure to IPV doubles the risk of having a lifetime history of suicide attempt(s).
• These associations with mental health indicators are significant for both genders.
• These associations persist regardless of social category or history of sexual violence.
Background
The term ‘child witness’ designates children who live in households where there is domestic violence between their parents. The number of child witnesses to intimate partner violence (IPV) worldwide is estimated between 133 and 275 million each year (1).
According to a survey in 2010–2015 (2,3), approximately 143 000 French children live in a household where a woman reported physical and/or sexual IPV during the preceding year. IPV exposure has major consequences on the physical and mental health of children (4). Since 2014, in accordance with the UK National Institute of Health and Clinical Excellent (5), French medical guidelines have considered children who witness IPV to be abused in their own right (6).
Family practitioners are aware about the importance of screening domestic violence to protect families (7), but they may be less aware about long-term consequences of violence exposure. Little work has taken place in France around the issue of children exposed to IPV (8), and still fewer studies have considered their outcomes in adulthood. Two French studies have questioned adults about their history of IPV exposure, but neither focused on the general population: first, the ENVEFF study (9) (national survey on violence towards women in France) in 2000 interviewed only women and, second, the SIRS study (Santé, Inégalités et Ruptures Sociales/Health and Social Inequalities and Breakdowns) (10) in 2005, which surveyed more than 3000 adult men and women, only in Paris region (11). Until now, no study has examined the mental health repercussions of IPV exposure in the French general population in adulthood. Evidence has found that children exposed to IPV are more likely to have physical and psychological health issues (12–14) than non-exposed children. Being exposed to IPV is associated with the presence of major depressive episodes (MDEs) (15), substance abuse (16) and an increased risk of mental health hospitalization (17) during childhood and adolescence. At last, exposed children show lower social competence and are more likely to experience behavioural problems than non-exposed children (18).
Our hypothesis is that adults—both men and women—who witnessed violence between their parents as children report poorer mental health than adults who were not exposed to IPV as children: more symptoms of depression and suicidal behaviour in the past year and more lifetime suicide attempts.
Methods
Study population
This study relies on data from the 2017 Health Barometer, a cross-sectional telephone survey conducted since 1992 by Santé Publique France (the National Public Health agency). Regular waves of assessments using the same methodology are carried out, each time on new samples, to measure behaviours and health determinants among the French population. Health Barometer is not intended to provide diagnoses, but the objective is to make a brief annual report of health indicators in France.
A two-stage random survey (household and individual) is conducted via a computer-assisted telephone interview following random generation of telephone numbers (landline and cell). People living in Metropolitan France (aged 18–75) responded to this telephone questionnaire between 5 January and 18 July 2017. The global response rate was 60%. Each respondent’s data were weighted to consider their sex, age, region of residence, size of municipality, educational level and probability of being randomly selected from within the household. The methods used for the 2017 Health Barometer have been described in detail and are accessible online (19).
Outcomes
Three outcomes were chosen: the presence of depressive symptoms or suicidal thoughts in the past year and at least one lifetime suicide attempt. Health Barometer uses a phone method to select populations based on broader sampling frameworks and to question people from all the country. This survey has been carried out every year to assess the presence of depressive symptoms in the French population and its evolution over time. For that reason, the questionnaire focused on the past year to evaluate depressive symptoms and self-harming behaviours and did not take former depressive periods or treatments into account.
In Health Barometer, MDE symptoms were assessed by the Composite International Diagnosis Interview Short Form (CIDI-SF) (20). An MDE is defined by the presence for a depressed mood or loss of interest in usual activities for at least two consecutive weeks, associated with at least three secondary symptoms (insomnia or hypersomnia, fatigue or lack of energy, gain or loss of at least 5 kg, agitation or psychomotor deceleration, disorders of concentration, sense of worthlessness and recurrent thoughts of death) and disruption of the activities of daily living (21).This questionnaire was developed by the World Health Organization in 1998 and can be used to assess the occurrence over the past 12 months of seven mental disorders. CIDI-SF has a sensitivity of 89.6% and a specificity of 93.9% compared to CIDI/DSM-III-R MDE diagnose (20) and has been validated in French (22).
This questionnaire is administered by trained (but not specialist) investigators and has been validated with telephone interviews. There is a good agreement between telephone and face-to-face assessment for major depressive disorder (23).
The presence of suicidal thoughts during the past year was measured by the question: ‘During the past 12 months, have you thought about killing yourself?’ A history of suicide attempts was assessed by the question: ‘At any time during your life, have you tried to kill yourself?’
Principal explanatory variable
A history of exposure to IPV and violence was measured by the question: ‘Before your 18th birthday, did you experience serious fights or a climate of violence between your parents?’
Other explanatory variables
The other explanatory variables, which were used for adjustment, were: age (continuous), income in tertiles, work status, highest educational level, two of the following negative events in childhood: parental separation or divorce and parental serious disease or death; and any lifetime sexual aggression or rape.
Statistical analysis
The analyses were performed separately by sex, with Stata software, v.14®. There were non-responses concerning less than 0.05% of the total responses for each of the variables of interest. It was decided not to reclassify these non-responses.
The frequency of exposure was first described in a bivariate analysis with socio-demographic and health characteristics, stratified by sex. Next, logistic regressions were used to study the association between childhood exposure to IPV and each of the outcomes by calculating odds ratios (ORs) and their 95% confidence intervals (CIs), first with univariate and then multivariate models:
- -
the first model adjusted for age, income in tertiles, educational level, work status, parental separation or divorce in childhood and parental serious disease or death in childhood.
- -
And a second model that, in addition to the above adjustment variables, included lifetime history of sexual aggression and/or rape.
Results
Overall, 25 319 people aged 18–75 were questioned (48.7% men and 51.3% women). The prevalence of exposure to IPV in childhood was 21.4% (n = 5073; men: 18.4%, women: 24.3%). Table 1 describes the frequency of exposure by socio-demographic and health characteristics and Table 2 by childhood history and lifetime sexual violence. Among the participants, 9.8% had had depressive symptoms (men: 6.3%; women: 13%), 4.7% had had suicidal thought during the past year (4.0% and 5.4%, respectively) and 7.2% had attempted suicide at least once during their life (4.4% and 9.9%).
. | Men and women who witnessed IPV in childhood % weighted n = 5073 21.4% . | Men who witnessed IPV % weighted n = 2035 18.4% . | Women who witnessed IPV % weighted n = 3038 24.3% . |
---|---|---|---|
Age (years) | |||
18–24 | 24.7 | 22.6 | 26.9 |
25–34 | 24.8 | 21.4 | 28.0 |
35–44 | 24.6 | 21.4 | 27.6 |
45–54 | 22.1 | 18.5 | 25.7 |
55–64 | 18.0 | 13.6 | 22.1 |
65–75 | 14.8 | 13.7 | 15.7 |
Educational level | |||
No diploma | 26.6 | 22.2 | 30.6 |
Less than Bac | 21.6 | 18.8 | 24.5 |
Bac or equivalent | 21.2 | 18.7 | 23.4 |
Bac +2 years | 20.5 | 16.5 | 24.4 |
Bac +3/+4 years | 20.1 | 16.5 | 22.2 |
Bac +5 or more | 17.1 | 15.7 | 18.7 |
Income | |||
First tertile (low) | 25.5 | 21.4 | 28.8 |
Second tertile | 21.2 | 19.4 | 23.0 |
Third tertile (high) | 17.5 | 15.1 | 20.6 |
Work status | |||
Working | 20.8 | 18.0 | 23.7 |
Student | 23.9 | 20.9 | 26.8 |
Unemployed | 29.7 | 26.4 | 32.8 |
Retired | 15.8 | 13.8 | 17.6 |
Other not in the labour force | 29.8 | 25.1 | 31.5 |
Tobacco use | |||
Non-smoker | 18.5 | 16.0 | 20.7 |
Occasional smoker | 21.8 | 17.6 | 26.6 |
Daily smoker | 28.6 | 23.7 | 34.4 |
Drinks alcohol daily | |||
No | 21.6 | 18.4 | 24.3 |
Yes | 19.8 | 18.5 | 23.6 |
BMI (kg/m2) | |||
Low or normal | 22.3 | 19.9 | 24.2 |
Overweight | 19.5 | 16.8 | 23.3 |
Obese | 22.2 | 17.6 | 26.3 |
General health status | |||
Good | 21.0 | 18.0 | 23.8 |
Poor | 28.9 | 26.0 | 31.2 |
. | Men and women who witnessed IPV in childhood % weighted n = 5073 21.4% . | Men who witnessed IPV % weighted n = 2035 18.4% . | Women who witnessed IPV % weighted n = 3038 24.3% . |
---|---|---|---|
Age (years) | |||
18–24 | 24.7 | 22.6 | 26.9 |
25–34 | 24.8 | 21.4 | 28.0 |
35–44 | 24.6 | 21.4 | 27.6 |
45–54 | 22.1 | 18.5 | 25.7 |
55–64 | 18.0 | 13.6 | 22.1 |
65–75 | 14.8 | 13.7 | 15.7 |
Educational level | |||
No diploma | 26.6 | 22.2 | 30.6 |
Less than Bac | 21.6 | 18.8 | 24.5 |
Bac or equivalent | 21.2 | 18.7 | 23.4 |
Bac +2 years | 20.5 | 16.5 | 24.4 |
Bac +3/+4 years | 20.1 | 16.5 | 22.2 |
Bac +5 or more | 17.1 | 15.7 | 18.7 |
Income | |||
First tertile (low) | 25.5 | 21.4 | 28.8 |
Second tertile | 21.2 | 19.4 | 23.0 |
Third tertile (high) | 17.5 | 15.1 | 20.6 |
Work status | |||
Working | 20.8 | 18.0 | 23.7 |
Student | 23.9 | 20.9 | 26.8 |
Unemployed | 29.7 | 26.4 | 32.8 |
Retired | 15.8 | 13.8 | 17.6 |
Other not in the labour force | 29.8 | 25.1 | 31.5 |
Tobacco use | |||
Non-smoker | 18.5 | 16.0 | 20.7 |
Occasional smoker | 21.8 | 17.6 | 26.6 |
Daily smoker | 28.6 | 23.7 | 34.4 |
Drinks alcohol daily | |||
No | 21.6 | 18.4 | 24.3 |
Yes | 19.8 | 18.5 | 23.6 |
BMI (kg/m2) | |||
Low or normal | 22.3 | 19.9 | 24.2 |
Overweight | 19.5 | 16.8 | 23.3 |
Obese | 22.2 | 17.6 | 26.3 |
General health status | |||
Good | 21.0 | 18.0 | 23.8 |
Poor | 28.9 | 26.0 | 31.2 |
Bac, baccalaureate (secondary school leaving examination and diploma); BMI, body mass index.
. | Men and women who witnessed IPV in childhood % weighted n = 5073 21.4% . | Men who witnessed IPV % weighted n = 2035 18.4% . | Women who witnessed IPV % weighted n = 3038 24.3% . |
---|---|---|---|
Age (years) | |||
18–24 | 24.7 | 22.6 | 26.9 |
25–34 | 24.8 | 21.4 | 28.0 |
35–44 | 24.6 | 21.4 | 27.6 |
45–54 | 22.1 | 18.5 | 25.7 |
55–64 | 18.0 | 13.6 | 22.1 |
65–75 | 14.8 | 13.7 | 15.7 |
Educational level | |||
No diploma | 26.6 | 22.2 | 30.6 |
Less than Bac | 21.6 | 18.8 | 24.5 |
Bac or equivalent | 21.2 | 18.7 | 23.4 |
Bac +2 years | 20.5 | 16.5 | 24.4 |
Bac +3/+4 years | 20.1 | 16.5 | 22.2 |
Bac +5 or more | 17.1 | 15.7 | 18.7 |
Income | |||
First tertile (low) | 25.5 | 21.4 | 28.8 |
Second tertile | 21.2 | 19.4 | 23.0 |
Third tertile (high) | 17.5 | 15.1 | 20.6 |
Work status | |||
Working | 20.8 | 18.0 | 23.7 |
Student | 23.9 | 20.9 | 26.8 |
Unemployed | 29.7 | 26.4 | 32.8 |
Retired | 15.8 | 13.8 | 17.6 |
Other not in the labour force | 29.8 | 25.1 | 31.5 |
Tobacco use | |||
Non-smoker | 18.5 | 16.0 | 20.7 |
Occasional smoker | 21.8 | 17.6 | 26.6 |
Daily smoker | 28.6 | 23.7 | 34.4 |
Drinks alcohol daily | |||
No | 21.6 | 18.4 | 24.3 |
Yes | 19.8 | 18.5 | 23.6 |
BMI (kg/m2) | |||
Low or normal | 22.3 | 19.9 | 24.2 |
Overweight | 19.5 | 16.8 | 23.3 |
Obese | 22.2 | 17.6 | 26.3 |
General health status | |||
Good | 21.0 | 18.0 | 23.8 |
Poor | 28.9 | 26.0 | 31.2 |
. | Men and women who witnessed IPV in childhood % weighted n = 5073 21.4% . | Men who witnessed IPV % weighted n = 2035 18.4% . | Women who witnessed IPV % weighted n = 3038 24.3% . |
---|---|---|---|
Age (years) | |||
18–24 | 24.7 | 22.6 | 26.9 |
25–34 | 24.8 | 21.4 | 28.0 |
35–44 | 24.6 | 21.4 | 27.6 |
45–54 | 22.1 | 18.5 | 25.7 |
55–64 | 18.0 | 13.6 | 22.1 |
65–75 | 14.8 | 13.7 | 15.7 |
Educational level | |||
No diploma | 26.6 | 22.2 | 30.6 |
Less than Bac | 21.6 | 18.8 | 24.5 |
Bac or equivalent | 21.2 | 18.7 | 23.4 |
Bac +2 years | 20.5 | 16.5 | 24.4 |
Bac +3/+4 years | 20.1 | 16.5 | 22.2 |
Bac +5 or more | 17.1 | 15.7 | 18.7 |
Income | |||
First tertile (low) | 25.5 | 21.4 | 28.8 |
Second tertile | 21.2 | 19.4 | 23.0 |
Third tertile (high) | 17.5 | 15.1 | 20.6 |
Work status | |||
Working | 20.8 | 18.0 | 23.7 |
Student | 23.9 | 20.9 | 26.8 |
Unemployed | 29.7 | 26.4 | 32.8 |
Retired | 15.8 | 13.8 | 17.6 |
Other not in the labour force | 29.8 | 25.1 | 31.5 |
Tobacco use | |||
Non-smoker | 18.5 | 16.0 | 20.7 |
Occasional smoker | 21.8 | 17.6 | 26.6 |
Daily smoker | 28.6 | 23.7 | 34.4 |
Drinks alcohol daily | |||
No | 21.6 | 18.4 | 24.3 |
Yes | 19.8 | 18.5 | 23.6 |
BMI (kg/m2) | |||
Low or normal | 22.3 | 19.9 | 24.2 |
Overweight | 19.5 | 16.8 | 23.3 |
Obese | 22.2 | 17.6 | 26.3 |
General health status | |||
Good | 21.0 | 18.0 | 23.8 |
Poor | 28.9 | 26.0 | 31.2 |
Bac, baccalaureate (secondary school leaving examination and diploma); BMI, body mass index.
. | Men and women who witnessed IPV in childhood % weighted n = 5073 21.4% . | Men who witnessed IPV % weighted n = 2035 18.4% . | Women who witnessed IPV % weighted n = 3038 24.3% . |
---|---|---|---|
Did your parents separate or divorce before your 18th birthday? Yes | 48.7 | 44.9 | 51.9 |
Did one of your parents have a major health problem or die before your 18th birthday? Yes | 29.9 | 26.1 | 33.1 |
In your lifetime, have you been forced to endure or perform sexual or touching or to have sexual relations against your will? Yes | 52.0 | 47.5 | 52.9 |
. | Men and women who witnessed IPV in childhood % weighted n = 5073 21.4% . | Men who witnessed IPV % weighted n = 2035 18.4% . | Women who witnessed IPV % weighted n = 3038 24.3% . |
---|---|---|---|
Did your parents separate or divorce before your 18th birthday? Yes | 48.7 | 44.9 | 51.9 |
Did one of your parents have a major health problem or die before your 18th birthday? Yes | 29.9 | 26.1 | 33.1 |
In your lifetime, have you been forced to endure or perform sexual or touching or to have sexual relations against your will? Yes | 52.0 | 47.5 | 52.9 |
. | Men and women who witnessed IPV in childhood % weighted n = 5073 21.4% . | Men who witnessed IPV % weighted n = 2035 18.4% . | Women who witnessed IPV % weighted n = 3038 24.3% . |
---|---|---|---|
Did your parents separate or divorce before your 18th birthday? Yes | 48.7 | 44.9 | 51.9 |
Did one of your parents have a major health problem or die before your 18th birthday? Yes | 29.9 | 26.1 | 33.1 |
In your lifetime, have you been forced to endure or perform sexual or touching or to have sexual relations against your will? Yes | 52.0 | 47.5 | 52.9 |
. | Men and women who witnessed IPV in childhood % weighted n = 5073 21.4% . | Men who witnessed IPV % weighted n = 2035 18.4% . | Women who witnessed IPV % weighted n = 3038 24.3% . |
---|---|---|---|
Did your parents separate or divorce before your 18th birthday? Yes | 48.7 | 44.9 | 51.9 |
Did one of your parents have a major health problem or die before your 18th birthday? Yes | 29.9 | 26.1 | 33.1 |
In your lifetime, have you been forced to endure or perform sexual or touching or to have sexual relations against your will? Yes | 52.0 | 47.5 | 52.9 |
After adjustment for age, income in tertiles, educational level, work status, negative life events in childhood and IPV exposure during childhood was significantly associated with depressive symptoms in the past year [aOR men = 1.98 (1.57–2.50), aOR women = 2.25 (1.95–2.60)], suicidal thoughts in the past year [aOR men = 2.12 (1.59–2.83), aOR women = 2.96 (2.42–3.62)] and at least one lifetime suicide attempt [aOR men = 2.60 (2.00–3.38), aOR women = 3.25 (2.76–3.83)]. When we added to the adjustment variables a lifetime history of sexual violence, witnessing IPV during childhood remained significantly associated with depressive symptoms in the past year [aOR men = 1.88 (1.49–2.38), aOR women = 2.00 (1.72–2.32)], suicidal thoughts in the past year [aOR men = 1.97 (1.48–2.64), aOR women = 2.35 (1.89–2.93)] and at least one lifetime suicide attempt [aOR men = 2.39 (1.83–3.11), aOR women = 2.66 (2.25–3.16)] (P < 0.001 for all of these aOR; Tables 3 and 4).
Men . | Number of participants, crude (% weighted) n = 11 596 . | Crude OR . | adjusted OR 1a . | adjusted OR 2b . |
---|---|---|---|---|
Symptoms of depression | 717 (6.3) | 2.29 [1.87–2.80]* | 1.98 [1.57–2.50]* | 1.88 [1.49–2.38]* |
Suicidal thoughts | 428 (4.0) | 2.66 [2.07–3.43]* | 2.12 [1.59–2.83]* | 1.97 [1.48–2.64]* |
Suicide attempts | 484 (4.4) | 3.31[2.64–4.16]* | 2.60[2.00–3.38]* | 2.39[1.83–3.11]* |
Men . | Number of participants, crude (% weighted) n = 11 596 . | Crude OR . | adjusted OR 1a . | adjusted OR 2b . |
---|---|---|---|---|
Symptoms of depression | 717 (6.3) | 2.29 [1.87–2.80]* | 1.98 [1.57–2.50]* | 1.88 [1.49–2.38]* |
Suicidal thoughts | 428 (4.0) | 2.66 [2.07–3.43]* | 2.12 [1.59–2.83]* | 1.97 [1.48–2.64]* |
Suicide attempts | 484 (4.4) | 3.31[2.64–4.16]* | 2.60[2.00–3.38]* | 2.39[1.83–3.11]* |
aAdjustment for age, income in tertiles, educational level, work status and negative life events in childhood (parental death or serious disease or divorce or separation).
bAdjustment for age, income in tertiles, educational level, work status, negative life events in childhood (parental death or serious disease or divorce or separation) and for lifetime history of sexual violence.
*P < 0.001.
Men . | Number of participants, crude (% weighted) n = 11 596 . | Crude OR . | adjusted OR 1a . | adjusted OR 2b . |
---|---|---|---|---|
Symptoms of depression | 717 (6.3) | 2.29 [1.87–2.80]* | 1.98 [1.57–2.50]* | 1.88 [1.49–2.38]* |
Suicidal thoughts | 428 (4.0) | 2.66 [2.07–3.43]* | 2.12 [1.59–2.83]* | 1.97 [1.48–2.64]* |
Suicide attempts | 484 (4.4) | 3.31[2.64–4.16]* | 2.60[2.00–3.38]* | 2.39[1.83–3.11]* |
Men . | Number of participants, crude (% weighted) n = 11 596 . | Crude OR . | adjusted OR 1a . | adjusted OR 2b . |
---|---|---|---|---|
Symptoms of depression | 717 (6.3) | 2.29 [1.87–2.80]* | 1.98 [1.57–2.50]* | 1.88 [1.49–2.38]* |
Suicidal thoughts | 428 (4.0) | 2.66 [2.07–3.43]* | 2.12 [1.59–2.83]* | 1.97 [1.48–2.64]* |
Suicide attempts | 484 (4.4) | 3.31[2.64–4.16]* | 2.60[2.00–3.38]* | 2.39[1.83–3.11]* |
aAdjustment for age, income in tertiles, educational level, work status and negative life events in childhood (parental death or serious disease or divorce or separation).
bAdjustment for age, income in tertiles, educational level, work status, negative life events in childhood (parental death or serious disease or divorce or separation) and for lifetime history of sexual violence.
*P < 0.001.
Women . | Number of participants, crude (% weighted) n = 13 723 . | Crude OR . | Adjusted OR 1a . | Adjusted OR 2b . |
---|---|---|---|---|
Symptoms of depression | 1703 (13.0) | 2.60 [2.27–2.98]* | 2.25 [1.95–2.60]* | 2.00 [1.72–2.32]* |
Suicidal thoughts | 720 (5.4) | 3.52 [2.90–4.28]* | 2.96 [2.42–3.62]* | 2.35 [1.89–2.93]* |
Suicide attempts | 1258 (9.9) | 3.88 [3.34–4.51]* | 3.25 [2.76–3.83]* | 2.66 [2.25–3.16]* |
Women . | Number of participants, crude (% weighted) n = 13 723 . | Crude OR . | Adjusted OR 1a . | Adjusted OR 2b . |
---|---|---|---|---|
Symptoms of depression | 1703 (13.0) | 2.60 [2.27–2.98]* | 2.25 [1.95–2.60]* | 2.00 [1.72–2.32]* |
Suicidal thoughts | 720 (5.4) | 3.52 [2.90–4.28]* | 2.96 [2.42–3.62]* | 2.35 [1.89–2.93]* |
Suicide attempts | 1258 (9.9) | 3.88 [3.34–4.51]* | 3.25 [2.76–3.83]* | 2.66 [2.25–3.16]* |
aAdjustment for age, income in tertiles, educational level, work status and negative life events in childhood (parental death or serious disease or divorce or separation).
bAdjustment for age, income in tertiles, educational level, work status, negative life events in childhood (parental death or serious disease or divorce or separation) and for lifetime history of sexual violence.
*P < 0.001.
Women . | Number of participants, crude (% weighted) n = 13 723 . | Crude OR . | Adjusted OR 1a . | Adjusted OR 2b . |
---|---|---|---|---|
Symptoms of depression | 1703 (13.0) | 2.60 [2.27–2.98]* | 2.25 [1.95–2.60]* | 2.00 [1.72–2.32]* |
Suicidal thoughts | 720 (5.4) | 3.52 [2.90–4.28]* | 2.96 [2.42–3.62]* | 2.35 [1.89–2.93]* |
Suicide attempts | 1258 (9.9) | 3.88 [3.34–4.51]* | 3.25 [2.76–3.83]* | 2.66 [2.25–3.16]* |
Women . | Number of participants, crude (% weighted) n = 13 723 . | Crude OR . | Adjusted OR 1a . | Adjusted OR 2b . |
---|---|---|---|---|
Symptoms of depression | 1703 (13.0) | 2.60 [2.27–2.98]* | 2.25 [1.95–2.60]* | 2.00 [1.72–2.32]* |
Suicidal thoughts | 720 (5.4) | 3.52 [2.90–4.28]* | 2.96 [2.42–3.62]* | 2.35 [1.89–2.93]* |
Suicide attempts | 1258 (9.9) | 3.88 [3.34–4.51]* | 3.25 [2.76–3.83]* | 2.66 [2.25–3.16]* |
aAdjustment for age, income in tertiles, educational level, work status and negative life events in childhood (parental death or serious disease or divorce or separation).
bAdjustment for age, income in tertiles, educational level, work status, negative life events in childhood (parental death or serious disease or divorce or separation) and for lifetime history of sexual violence.
*P < 0.001.
Discussion
This study, based on data from the 2017 Health Barometer of a representative sample of 25 319 individuals aged 18–75 in France showed that being exposed to IPV in childhood is significantly associated, in both men and women, with symptoms of depression during the past year, suicidal thoughts over the same period and at least one lifetime suicide attempt. Our results show that family physicians have a decisive part to play not only to protect the children but also to manage the care of adults who suffered from IPV exposure in childhood.
IPV exposure in childhood
The prevalence of exposure to IPV during childhood was 21%, a figure notably higher than the French figures reported for the women questioned in the ENVEFF survey in 2000 (15%) (8) and the Paris residents in the SIRS survey in 2003 (16%) (9). This difference in prevalence rates is evidence of the difficulty in accurately measuring the rate of children witnessing this violence. The methods for quantifying IPV are hampered by the inadequacy of the tools for identifying and the multiplicity of players involved in family violence (24). These issues lead to different and fragmented figures. Nonetheless, it is highly probable that this is a massive phenomenon (1). For example, in 2011, 18% of children in a US survey were exposed to physical IPV during their lifetime and 16% to psychological and/or emotional IPV (25). If we count all forms of violence, it is estimated that 26% of American children are exposed to a climate of domestic violence.
This topic is constantly subjected to minimization of the facts. In retrospective studies of traumatic events of childhood, the prevalence of these events is known to be systematically underestimated (26).
Exposure to violence, such as IPV, may compromise child development. Consequences of children’s exposure to violence have been inventoried: externalizing behaviours, such as aggression problems, and internalizing behaviours, such as depression, delayed cognitive development, poor academic functioning and peer relationships difficulties (27). Neuroscientists have documented the deleterious effects of exposure to violence in childhood on psychological development and subsequent psychological health (28).
Symptoms of depression during the past year
In our study, 10% of respondents presented symptoms of depression in the previous year (6% of men and 13% of women) (29). Our work found a significant association between symptoms of depression in the previous year and IPV exposure in childhood. The US Adverse Childhood Experiences (ACE) cohort study showed in 1998 that exposure to trauma in childhood, including exposure to ‘violence against one’s mother’, was a risk factor for several of the principal causes of death in adulthood, as well as for the onset of depression (30). In France, the SIRS cohort team showed that, after adjustment for current and childhood socio-economic conditions, parental mental health and family breakdowns, experiencing IPV in childhood remained associated with depression (9). The adults in the SIRS cohort exposed to IPV in childhood had a greater risk of developing a psychological disorder in adulthood even after adjustment for the other forms of violence against the child.
We found higher OR for these risks among the women in our study. A meta-analysis did not show any difference in the psychological effects of IPV on children by gender (31). The only difference identified in an interventional study concerned the response to the quality of environmental support: the rate of depression among school-aged girls diminished significantly after the establishment of social support after exposure to IPV, while it remained stable in boys (32).
Existence of suicidal thoughts in the past year and a lifetime history of at least one suicide attempt
In the 2017 Health Barometer, 5% of the respondents reported suicidal thoughts in the previous year. The association between the presence of suicidal thoughts and IPV exposure in childhood is statistically significant in our sample and does not differ between men and women. No study has thus far looked for an association between suicidal thoughts and exposure to IPV in the general population. Nonetheless, a US survey that questioned more than 30 000 participants who had ever presented post-traumatic stress disorder observed that 43% of people exposed to ‘serious fights at home’ in childhood had suicidal ideation (33).
An earlier article from the 2017 Health Barometer showed that exposure to IPV in childhood doubled the lifetime risk of attempted suicide among both men and women (34). Many international studies have examined the association between IPV exposure and one or more lifetime suicide attempts; these include the ACE study in 1998 (19) and the study by Rico et al. in 2011 (35). This study showed that those exposed to IPV in childhood had a risk 1.8 times greater than those not so exposed to attempt suicide later in life. Conversely, in the French SIRS study, the association between IPV exposure and lifetime history of suicide attempts was not significant after adjustment for family and social stress factors in childhood (10).
History of sexual violence
In our sample, 6% of the respondents reported forcible sexual touching and/or rape (2% of men and 10% of women) (36). According to the VIRAGE survey in 2015 (cross-sectional survey in France of a representative sample of 28 000 individuals aged 18–69 years), 14% of the women and 4% of the men had experienced at least one act of sexual violence in their lifetime (37). This experience is a major risk factor for again becoming a victim of violence (38) and for morbidity and mortality (39). These effects on the physical and mental health of the victims of sexual violence can induce confounding of the measures of our statistical associations. Nonetheless, after adjusting for a history of sexual violence, the association between IPV exposure in childhood and these three adult mental health indicators in adulthood remained significant.
Strengths and limitations
The first limitation in our study is that there is no French definition of IPV nor any specific tool or validated question for measuring it. Nonetheless the same question used in the Health Barometer was asked in the ENVEFF survey (8), but this question remained subject to the respondents’ interpretation. The US National Survey of Children’s Exposure to Violence (40) used a validated questionnaire to overcome this limitation in a telephone survey conducted among 4549 minors; they asked six questions that described precisely the different forms of IPV that children might have witnessed. Regarding MDE identification, it is important to recall that CIDI-SF is not a diagnostic test but a screening tool (41). Depression is a psychiatric disorder, which requires adequate differential diagnoses and wider clinical examination. This phone questionnaire, administered by non-clinician investigators, collects self-rated symptoms of depression that could lead to reporting bias. In addition, possible previous treatment and recovering from previous depressive episodes were not considered. However, our study could serve as an interesting basis for further analysis and primary care researches. A second limitation involves memory bias because the questions were asked of adults about their childhood, different amounts of time ago, depending on the respondents’ ages.
Health Barometer did not ask respondents if they had experienced other forms of abuse in childhood. It was, thus, impossible to adjust for the existence of ‘direct abuse’ coexisting with IPV exposure (although this coexistence is often described). Residual confounding, thus remains possible.
Finally, this is a cross-sectional rather than a longitudinal study, which makes it impossible to conclude that a causal link exists. Cross-sectional sample in adulthood is a major limitation when studying traumatic events from childhood and their connection to adult health outcomes. Traumatizing environment during the childhood and youth development is associated with various health problems of which depression is only one of several major entities.
Besides these limitations, it is important to underline the originality of this work, which deals with the association between mental health indicators and a history of exposure to IPV, in a larger general population than that considered in the previous French work on this subject. Next, the rarity of the existing data on IPV exposure supports the importance of analyzing the data that do exist and shows the potential impact of this exposure throughout the lifetimes of the children concerned. Finally, the large sample size enabled us to stratify this analysis by sex with satisfactory statistical power. The post-stratification weighting of the data allowed us to correct the sample for multiple adjustment criteria (sex, age, educational level, municipality population, etc.). This reweighting assumes that, in each category defined by the adjustment variables, respondents and non-respondents are on average similar. Our study population was, thus, representative, and our results can be extrapolated to the general population.
Conclusion
The associations shown in this study between a history of exposure to IPV and three mental health indicators in adulthood underline the need to study the effect of IPV throughout the lifetimes of the subjects who witnessed it and to help overcome the trivialization of this violence at both the medical and societal levels. Our study shows that exposure to IPV during childhood doubles the risk of symptoms of depression and suicidal behaviour during adulthood regardless of gender and social category.
This could be recognized as a major public health problem, considering that, in the data analysis of the 2017 Health Barometer, the rate of people reporting that they witnessed IPV is high. Family practitioners must be aware of this worrisome prevalence rate and encouraged to ask adults about a potential IPV exposure in childhood. Our results highlight the need of the biopsychosocial approach taken by French GP to break the cycle of domestic violence and its consequences.
Acknowledgement
We thank Jo Ann Cahn for her help in correcting and clarifying the manuscript.
Declaration
Funding: this study was funding from Institut National de la Santé et de la Recherche Médicale (INSERM, Institute of Medical Research): scholarship for internship for Noémie Roland (3000 euros) and translation budget; data came from the 2017 ‘Baromètre santé’, a study conducted with financial support from the French Public Health Agency (Santé publique France).
Ethical approval: no necessary ethical approval.
Conflict of interest: none of the authors has any financial or non-financial competing interests concerning the present study.
Data availability
The authors cannot publicly release the data from the 2017 Health Barometer. Health Barometer database availability is subject to authorization via an agreement with Santé Publique France.