The Impact of Childhood Abuse on Young Danish Adults from Families with Alcohol Problems ()
1. Introduction
Grant [1] concluded that approximately 1 in 4 children younger than 18 years in the United States had been exposed to alcohol abuse or alcohol dependence in the family. Dube and colleagues [2] found in an American community sample that 22% reported that at least one of their parents was a problem drinker or alcoholic. In 2017, based on national surveys combining adult and child household data, Lipari & van Horn [3] estimated that 10.5% of US children lived in households with at least one parent who had a past year alcohol use disorder. Manning and colleagues [4] reviewed relevant British surveys and concluded that approximately 30% of children under 16 (3.3 - 3.5 million) in the UK lived with at least one binge drinking parent. 8% lived with at least two binge drinkers and 4% lived with a lone binge drinking parent. 22% of children under 16 (2.6 million) lived with a hazardous drinker, and 6% (705,000) lived with a dependent drinker. Recently, the Danish National Institute of Public Health estimated that approximately 7.45% of all children ages of 0-18 grew up in families with alcohol problems [5].
Parental drinking is thus an issue that affects a significant number of children and young adults. Several studies have found statistically significant positive associations between parental alcohol abuse and childhood physical abuse [6]-[12]. The view that alcohol problems simply cause increased levels of child abuse has been dispelled. Studies have found that adverse childhood experiences can be viewed as a complex set of highly interrelated experiences that may include factors such as childhood abuse or neglect, parental alcohol and drug abuse, domestic violence, parental marital discord, crime [13], mental illness, and social deprivation [6] [11] [14].
The uniformity myth that suggests that AOPDs share a common heritage and the specificity myth that suggests that AOPDs share a common outcome that is not shared by others have both been disproven by research [15] [16]. Research generally finds that AOPDS are a heterogeneous population at risk for higher psychological distress [17] [18], lower self-esteem [19]-[22], and poorer social adjustment [18] [23] than non-AOPDS. Some studies have, however, not found lower self-esteem in AOPDS [24] [25] or higher levels of psychological distress among AOPDS [26].
A range of studies have pointed to features of AOPD experiences beyond parental drinking that are associated with negative outcomes, such as emotional abuse [17], interparental violence and experiencing emotional and physical abuse [9] [27], family disharmony [25], difficult family environment [14], disruption of family rituals [28], chronic family stress [29], and parental comorbidity, that is parents having alcohol problems as well as a non-alcohol or drug related psychiatric diagnosis [30] [31]. Positive outcomes have been associated with having a significantly larger number of sources of support during childhood [30] [32] and resilience [33]. Cumulative theories have suggested that levels of offspring symptoms increase depending on whether a person has experienced one, or two or more types of trauma, such as childhood physical abuse, childhood sexual abuse, and parental alcohol abuse [19] [34]. Research has thus concluded that it is not parental drinking alone that is associated with AOPDS symptoms but other features of AOPDS experiences are also significant.
These findings have led us to focus this study on the impact of certain specific aspects of childhood physical and psychological abuse, and their associations with AOPDS levels of psychological distress, PTSD, self-esteem, and work and social adjustment. We hypothesized that factors such as whether abuse was primarily physical or psychological, and the length of the abuse in childhood years, might be significant for symptoms. The level of childhood social support and more specifically who the child talked to about parental drinking, and whether persons outside the family knew about the drinking, were considered of possible significance for symptom severity. This study investigates the relationship between Adult Offspring of Problem Drinkers’ (AOPDs) childhood experiences of physical and psychological abuse and their levels of psychological distress, self-esteem, post-traumatic stress disorder (PTSD), and work and social adjustment when entering counseling.
2. Method
2.1. Recruitment and Participants
All clients (N = 506) over 18 seeking treatment at a Danish national counseling service for young AOPDS in a period of 18 month were asked to participate. The counseling service was a private specialist service (TUBA) that offered counseling to young persons from families with alcohol problems. The counseling was funded primarily by the Danish State and was free for clients. Clients did not require a referral from a doctor or other instance to receive counseling. The counseling service addressed clients’ emotional and relational problems. Typically, clients at the service required help to understand and deal with their emotions, thoughts and reactions; help to maintain their own perspective when facing problems; and advice about how to deal with a wide range of life problems, as well the specific problem of having an alcohol dependent parent. Respondents were surveyed when they enrolled to receive TUBA’s treatment. The anonymity of clients was guaranteed. Participation in the study was voluntary and did not impact upon the clients’ access to counseling. Clients were told that they could miss questions out if they found them hard to answer. Clients were encouraged to address issues that completing questions raised, with their counselors. Counselors were prepared for this eventuality and offered support in how to deal with such questions. Generally, clients found the questions highly relevant, offering them an additional way of reflecting about their childhood experiences. The clients came from 11 centres across Denmark. Four did not register their gender. Six clients were excluded from the study due to missing data. Thus, 496 clients participated, 100 (20.1%) men and 396 (79.8%) women. Clients who had alcohol or drug problems themselves were referred on to specialist services and data about them were therefore not collected.
2.2. Procedures and Ethical Considerations
Clients were asked to fill out a range of measures and survey instruments in connection with their first counseling session on a computer at the counseling service. The counselors were not present when this took place. Data about the individual clients were not given to the counselors. Participation in the study was voluntary and did not impact upon the clients’ access to counseling. Clients were told that they could miss questions out if they found them hard to answer. Clients were encouraged to address issues that completing questions raised, with their counselors. Counselors were prepared for this eventuality and offered support in how to deal with such questions. Generally, clients found the questions highly relevant, offering them an additional way of reflecting about their childhood experiences. The research project was registered with the Danish Agency for Data Protection (RIO #12.203). The Danish regional science ethics committees approve drug trials but there are in no ethics committees in Denmark that approve social science studies such as this study.
2.3. Measures/Instruments
Childhood abusive experiences, years of parental drinking, disclosure were surveyed using the ACATI [35]. The ACATI contains 159 items and takes up to 30 minutes to complete. The ACATI has good test-retest reliability and good convergent validity with the Family Tree Questionnaire [36]. There are sections on the parents’ (including stepparents’) drinking and behavior mediating factors, physical, psychological abuse, and sexual abuse, experiences of suicide behavior, parental divorce, and possible traumatic confounders. The ACATI is written in a simple and clear style. For each parent, there is a timeline for the age range 0 - 18, with the instruction “in which period in your childhood (0 - 18 years) did your mother/father/stepparent drink too much?”
The Crisis Support Scale [37] measures received social support after a traumatic event. The 7 items are rated on a 7-point Likert scale ranging from “never” to “always”, and relate to perceived available support (1), contact with other bereaved (2), ability to express thoughts and emotions related to the loss (3), received sympathy and emotional (4) and practical (5) support, the degree to which one feels let down (6) and finally general satisfaction with social support (7). The scale has good internal consistency and discriminatory power as well as good psychometrical reliability and validity (ibid.). The alpha for the total scale was 0.82. Because several studies have shown that the negative social support (item #6) had a strong predictive value, we chose that to be included in the regression analysis instead of the full scale.
General psychological distress was measured using the Clinical Outcomes in Routine Evaluation outcome measure (CORE-34) [38]. The CORE was designed as a non-proprietary measure of psychological distress, comprising domains for subjective well-being, problems/symptoms, functioning, and risk to self or others. The alpha for the total scale was 0.94.
Levels of posttraumatic stress were measured with PTSD-8 [39], a short screening measure that registers the symptoms of PTSD. PTSD-8 has shown to have good psychometric properties in three independent samples and good test-rest reliability. The alpha for the total scale was 0.79.
The Work and Social Adjustment Scale [40], a valid and reliable measure was employed to monitor clients functioning impairment. Self-esteem was measured using the Rosenberg Self-Esteem Scale [41], an established and validated measure to assess global self-esteem [42]. The alpha for the total scale was 0.79.
2.4. Statistical Procedures and Analyses
Descriptive statistics were used to report distribution of the variables in the study. Bivariate correlations were used to study the strength of the associations between scaled variables. The associations between several dichotomous variables and scale values are studied by one-way ANOVA analyses. Finally, a hierarchical regression analysis, was conducted to weight the significance of physical and psychological abuse in relation to PTSD.
3. Results
Descriptive statistics are in Table 1. The average age was 26 years. Parental drinking lasted for more than 13 years in average before child had reached the age of 18. Thirty-one participants (6%) reported experiencing physical abuse only during their childhood, 126 (26%) reported experiencing psychological abuse only, 185 (38%) reported experiencing both forms of abuse, and 140 (29%) reported neither having experienced physical nor psychological abuse. Fourteen participants did not answer the questions about type and length of abuse. Both types of abuse lasted for many years. On average physical abuse lasted for almost 8 years and psychological abuse lasted for almost 10 years. Almost 2/5 never spoke about the problems relating to parental drinking in the family to anyone.
Table 1. Descriptive statistics (N = 496).
|
Mean |
SD |
Age |
26.3 |
4.9 |
Years of parental drinking while aged 0 - 18 |
13.2 |
5.2 |
Years of childhood physical abuse while aged 0-18 (Only for clients who experienced physical abuse) |
7.9 |
5.2 |
Years of childhood psychological abuse while aged 0 - 18. (Only for clients who experienced psychological abuse) |
9.8 |
5.6 |
|
N |
% |
Childhood physical abuse |
218 |
44 |
Childhood psychological abuse |
311 |
63 |
Spoke to no one about parental drinking |
188 |
38 |
Only spoke to non-drinking parent or siblings |
101 |
20 |
Only spoke to persons outside the immediate family |
67 |
14 |
Spoke to persons both inside and outside the immediate family |
140 |
28 |
Nobody outside the immediate family knew about parental drinking |
100 |
20 |
Table 2 shows a number of important associations. Childhood experiences of abuse and parental drinking were positively correlated with levels of general psychological distress, post-traumatic stress, negatively with work and social adjustment and social support and were unrelated to self-esteem. The number of years of parental drinking had strong positive associations to the number of years of both physical and psychological abuse, and a negative association to social support but no associations were found in relation to self-esteem and PTSD scores. Physical and psychological abuse were strongly associated; they had similar associations to CORE and work and social adjustment but differed in that only psychological abuse was associated with PTSD scores. PTSD scores were positively associated with CORE, and work and social adjustment scores. This group of four symptom measures had negative associations to self-esteem and social support.
Table 2. Correlations among extent of parental alcohol abuse, physical, and emotional abuse and the scales of the study and scale descriptives.
|
Scale |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
|
|
|
|
|
1 |
Years of alcohol problems |
|
|
|
|
|
|
|
|
|
|
|
|
2 |
Extent of physical abuse |
0.25** |
|
|
|
|
|
|
|
|
|
|
|
3 |
Extent of emotional abuse |
0.41** |
0.63** |
|
|
|
|
|
Mean |
SD |
Alpha |
No. of Items |
Range |
4 |
PTSD-8 |
0.03 |
0.11 |
0.15** |
|
|
|
|
20.2 |
5.0 |
0.79 |
8 |
8 - 32 |
5 |
CORE |
−0.01 |
0.19** |
0.12* |
0.51** |
|
|
|
49.83 |
21.07 |
0.94 |
34 |
2 - 122 |
6 |
Crisis Support Scale |
−0.13** |
0.10 |
−0.11 |
−0.14** |
−0.23** |
|
|
3.3 |
1.2 |
0.82 |
7 |
2 - 42 |
7 |
Rosenberg Self-esteem |
0.03 |
−0.05 |
−0.08 |
−0.34** |
−0.68** |
0.26** |
|
24.1 |
5.8 |
0.77 |
10 |
1 - 40 |
8 |
WSAS |
0.09 |
0.16* |
0.13* |
0.39** |
0.58** |
−0.21** |
−0.52** |
16.5 |
7.7 |
0.90 |
5 |
1 - 35 |
Note: * p < 0.05, ** p < 0.01.
Talking to nobody was associated with lower self-esteem (F(1,495) = 7.36, p < 0.01) and talking to both immediate family members and persons outside the immediate family was associated with higher self-esteem (F(1,494) = 5.82, p < 0.02). No statistically significant associations between talking to others and the other scales were found. People outside the immediate family knowing about the family’s alcohol problem was associated with lower scores with regards to CORE (F(1,494) = 3.90; p = 0.05). There were no statistically significant associations between others knowing about the alcohol problems in the family and the other scales.
Multiple regression analysis was conducted with PTSD total score as the dependent variable. Table 3 shows the final model. Female gender and lack of high school diploma were introduced as step 1, and they explained 4% of the PTSD variance. Emotional/Psychological abuse was added in step 2 together with experienced parental suicide threat. This step contributed with 4.6% of the variance. Number of lifetime traumas constituted step 3 which brought the adjusted R2 up to 11.4%. Step 4 was formed by negative social support and self-esteem and yielded together another 4.5% to the explained PTSD variance. The 5th and final step included general distress and work and social adjustment, and these two factors added another 5% to the variance bringing it up to 25% all together. Three factors contributed at lower steps: Education (steps 1 - 3), suicide threat (steps 2 - 3), and self-esteem (step 4), but they lost their significance in the final step.
Table 3. Hierarchical regression analysis for variables predicting ITQ-PTSD symptoms.
Step |
Variable |
β |
SE |
Sign. |
F* |
df |
Adj. R2 |
1 |
Constant |
|
|
|
10.585 |
2461 |
0.040 |
Gender |
0.174 |
0.571 |
<0.001 |
|
|
|
High School |
−0.140 |
0.486 |
0.002 |
|
|
|
2 |
Constant |
|
|
|
11.915 |
4459 |
0.086 |
Gender |
0.159 |
0.558 |
<0.001 |
|
|
|
High School |
−0.129 |
0.482 |
0.005 |
|
|
|
Emo. abuse |
0.126 |
0.458 |
0.007 |
|
|
|
Suicide threat |
0.157 |
0.299 |
0.001 |
|
|
|
3 |
Constant |
|
|
|
12.954 |
5458 |
0.114 |
Gender |
0.165 |
0.550 |
<0.001 |
|
|
|
High School |
−0.098 |
0.482 |
0.032 |
|
|
|
Emo. abuse |
0.122 |
0.451 |
0.008 |
|
|
|
Suicide threat |
0.101 |
0.308 |
0.036 |
|
|
|
No. traumas |
0.186 |
0.152 |
<0.001 |
|
|
|
4 |
Constant |
|
|
|
17.676 |
7456 |
0.201 |
Gender |
0.122 |
0.528 |
0.004 |
|
|
|
High School |
−0.059 |
0.461 |
0.174 |
|
|
|
Emo. abuse |
0.133 |
0.432 |
0.002 |
|
|
|
Suicide threat |
0.064 |
0.296 |
0.163 |
|
|
|
No. traumas |
0.153 |
0.145 |
<0.001 |
|
|
|
Neg. support |
0.148 |
0.139 |
<0.001 |
|
|
|
Self-esteem |
−0.241 |
0.037 |
<0.001 |
|
|
|
5 |
Constant |
|
|
|
18.210 |
9454 |
0.251 |
Gender |
0.146 |
0.514 |
<0.001 |
|
|
|
High School |
−0.071 |
0.447 |
0.094 |
|
|
|
Emo. abuse |
0.112 |
0.422 |
0.009 |
|
|
|
Suicide threat |
0.059 |
0.290 |
0.196 |
|
|
|
No. traumas |
0.116 |
0.144 |
0.009 |
|
|
|
Neg. support |
0.125 |
0.136 |
0.003 |
|
|
|
Self-esteem |
−0.076 |
0.043 |
0.141 |
|
|
|
Gen. distress |
0.099 |
0.087 |
0.031 |
|
|
|
Work adjust. |
0.250 |
0.031 |
<0.001 |
|
|
|
Note: *all models sign. p < 0.001.
4. Discussion
The amount of childhood abuse was a high as 70% and even 80% for girls. Emotional/psychological abuse was more prevalent and longer lasting than physical abuse, however they co-occur often. The abuse mingles with parental drinking that had been present through most of the participants’ childhood. In the final model, the hierarchical regression analysis revealed six factors: Female gender, emotional/psychological abuse, high number of life-time traumas, negative social support, general distress, and having many work and social adjustment problems contributed to the PTSD variance.
Female gender and level of education are two factors that are well studied in trauma research. The prevalence of PTSD for women is twice that for men, a finding that is not explained by differential exposure to various trauma types [43]. High level of education is generally recognised as a protective factor associated with more resources, social support, and better coping skills. These two factors function as we expect, the same does emotional abuse, suicidal threats, and number of trauma events, all three of them being well-known risk factors for the endorsement of trauma symptoms and PTSD. Emotional abuse is associated with many different trauma reactions and PTSD [44]. While there are many studies on the effect of suicide and suicide attempts on offspring [45], there is a lack of studies that investigate the effects of parental suicide threats on the children but clinical experience suggests that there can be a considerable effect. Number of traumas is a factor whose importance is connected to the cumulative effects of Adverse Childhood Experiences (ACE) [46]. In contrast to the other factors, self-esteem and education appeared to function as buffers protecting against PTSD. However, negative social support and low self-esteem weakened the effect of education in the final models.
Besides the parental drinking and the abuse, the number of other traumas experienced was also a strong predictor for trauma symptoms in the participants. Our findings regarding significant associations between experienced physical and psychological abuse and increased levels of general distress, PTSD, self-esteem, and work and social adjustment correspond with other research regarding associations between abuse and AOPDs’ symptom levels [34] [47] [48]. Similarly, the findings that levels of social support correlated negatively with general distress, PTSD, work and social adjustment problems, and positively with self-esteem, matched Werner & Johnson’s [32] findings on the importance of social support for AOPDs outcomes.
We had expected that whom the AOPDs talked to during the childhood would have an impact on trauma symptoms. It turned out that it only had an impact on self-esteem that increased when talking to the non-drinking parent, to someone outside the family or both. Self-esteem was low when the child did not talk to anyone about the parental drinking, which was the case for almost 40%. When people outside the immediate family knew about the alcohol problems and the child acknowledged that, the amount of general distress was lower.
O’Leary [49] suggested that psychological abuse can be a precursor to physical abuse and that the effects of psychological abuse can be similar to the effects of physical violence. The association between psychological abuse and PTSD was well documented in a recent review by Dokkedahl and colleagues [50]. In a Swedish child population study, Hagborg and colleagues [51] found an interaction effects between gender and levels of emotional maltreatment with girls reported decreased mental health and mental well-being at lower degrees of emotional maltreatment compared to boys. Also, girls reported larger decreases in mental health in response to exposure of emotional maltreatment. This aligns well with the present findings in the regression analysis.
Recurrent suicide threats are a very effective component of psychological abuse. In the hierarchial model, it lost its predictive value when negative social support and low self-esteem was introduced in the model, reflecting a social breakdown in the family when the child feels let down and its core identity is under pressure. Self-esteem is a strong buffer against trauma symptoms but when the the general distress and the social adjustment problems increase, the positive effect of self-esteem is also weakened.
The findings in this study add to our understanding of the complexity of the field as certain specific features of abuse are highlighted in greater detail than in previous studies. The findings also support and add to previous conclusions suggesting that 1) AOPDS are heterogeneous population that should be studied as such [15] [16] and 2) that adverse childhood experiences can be viewed as a complex set of highly interrelated experiences [2] [6] [11] that affect offspring symptom levels in different ways.
Even though the findings in this study cannot directly be transferred to the general population as the sample is distressed and help-seeking, the findings about childhood abuse and severe parental problems can be tested in other community samples. The addition of a control group without alcohol abusing parents would add more value to the study. The childhood experiences of these clients entering a specialist service for AOPDs may differ from the childhood experiences of AOPDs entering other forms of primary care or at a later age. The study is also limited by its retrospective nature and thereby its roots in the biases of clients’ autobiographical memories. Responding to the survey challenged the denial experienced in many families with alcohol problems. The study was therefore suited to adults or counseling clients who were in the process of facing the impact of their childhood experiences. We would therefore not recommend this type of study with non-counseling samples or persons under 18.
5. Conclusion
While physical and psychological abuse is clearly important for levels of general AOPD distress, post-traumatic stress, self-esteem, and work and social adjustment, they differed as predictors of trauma and distress symptoms. Psychological abuse had a greater general impact than physical abuse. Girls were more psychologically abused and had higher levels of trauma symptoms. Talking to others about one’s parents’ alcohol problems during childhood was positively associated with self-esteem. Clinicians should be advised to pay at least as much attention to the psychological/emotional abuse as to the physical abuse and the neglect due to the parental drinking behavior.
Acknowledgements
Thanks to the clients of TUBA for their participation.