Children’s Exposure to Parental Conflict after Father’s Treatment for Alcoholism

Abstract

Objective

This study investigated children of alcoholics’ (COAs’) exposure to inter-parental conflict before and after their fathers received alcohol treatment and compared exposure levels to a community comparison sample.

Method

This study included 67 couples with a treatment-seeking male alcoholic partner and children aged 4–16. The alcoholic fathers and their relationship partners provided data at baseline and at six and twelve months follow-up. A community comparison sample of 78 couples with children in the target age range completed similar longitudinal assessments. It was hypothesized that treatment of paternal alcoholism would be associated with a decrease in COAs’ exposure to conflict, and that among remitted patients exposure to conflict would decrease to the level found in the community sample.

Results

Prior to the father’s alcohol treatment, the children of the treatment sample were exposed to significantly more conflict between their parents than in the community comparison sample. After the fathers received alcohol treatment, COAs’ exposure to conflict significantly decreased at both the six and twelve month follow-ups compared to baseline. Children of remitted alcoholics did not differ significantly in levels of exposure to conflict at six months follow-up compared with the community sample as predicted. However, at twelve months remitted alcoholics reported significantly more exposure to conflict compared to the community sample.

Conclusions

Decreased child exposure to parental conflict is a benefit associated with the father’s treatment for alcoholism, and it may lead to improvements in COAs’ functioning after parental treatment.

1. Introduction

Alcoholism among parents is a serious problem,. Alcoholic couples have more conflicts and risk of partner violence (Leonard, 2005). Children of alcoholics (COA) experience significant maladjustment (Obot & Anthony, 2004). Treatment for alcohol dependence decreases partner violence and verbal aggression (e.g. Murphy & Ting, 2010).. Studies show that COAs whose fathers were stably remitted after alcohol treatment had comparable levels of emotional health compared to a community sample (Burdzovic Andreas, et al., 2006, Moos & Billings 1982). One year after treatment, children whose fathers remained mostly abstinent showed lower and decreasing adjustment problems (Burdzovic Andreas & O’Farrell, 2007) and fathers with more treatment involvement, AA attendance, and abstinence had children with fewer externalizing problems (Burdzovic Andreas & O’Farrell, 2009).

1.1. Purpose and Predictions of Study

A review by Klostermann and Kelley (2009) noted the lack of research on conflict exposure among COA’s. The purpose of the present study was to investigate whether treatment for paternal alcohol dependence is associated with decreases in COAs’ conflict exposure.

This study examined data from men in alcoholism treatment and their female partners with assessments at baseline, 6- and 12-month follow-ups. A community comparison sample was also assessed to provide a realistic normative baseline (Kendall, Marrs-Garcia, Nath, & Sheldrick, 1999). We predicted exposure to parental conflict in COAs would be greater than the comparison sample before fathers’ treatment and would decrease significantly after treatment. We further predicted that fathers’ recovery from alcoholism would reduce child conflict exposure to the level of the comparison sample.

2. Method

Harvard Medical School and VA Boston Institutional Review Boards approved the study.

2.1. Samples

A detailed description of eligibility and recruiting are available elsewhere (Taft et al., 2006, 2010).¹

2.1.1. Alcoholism treatment sample (COA sample)

Participants were from a sample of 181 men in treatment. Eighty-one couples (45%) had a child who met the target child criteria, of whom 67 (83%) completed all assessments.

2.1.2. Community comparison sample

The community comparison sample of 145 couples was recruited using random-digit dialing. The male partner could not have sought treatment in the prior year. Eighty-four couples (58%) had a child who met the target child criteria, of whom, 78 (93%) completed all assessments.

2.1.3. Target child selection algorithm

A target child was chosen for each couple to guard against violating statistical assumptions regarding independent observations. A step-wise algorithm selected children most closely related to the male partner. Of 67 target children, 41 were biological children of the alcohol dependent men with their current partner, 2 were biological children of the men from a previous partner, 23 were the partner’s biological children and 1 was adopted.

2.1.4. Demographic and diagnostic characteristics

Table 1 shows demographic characteristics for the treatment sample (N=67) and community comparison sample (N=78). There were some significant differences between the samples. To determine whether the variables which were significantly different between the samples would affect subsequent analyses, correlation analyses were conducted. The variables that were consistently significantly correlated with conflict exposure were the male and female partner’s education. The repeated measure ANOVA of exposure to conflict and group membership was run controlling for these variables which did not affect the significance or interpretation of the comparisons between the samples, and so were not included as covariates.

Table 1.

Comparison of Potential Covariates between Treatment and Community Samples

Characteristics	Treatment	Community	Test Statistic	p-value
Child Characteristics
Age (M, SD)	8.55 (3.32)	9.12 (4.07)	t=.92	.34
Boy (n, %)	32 (47.7 %)	45 (57.7 %)	χ2=1.42	.23
Child’s Relationship with Male Partner			χ2=18.03	<.001**
Biological (n, %)	44(65.7%)	73(93.6%)
Not Biological (n, %)	23(34.3%)	5(6.4%)
Family Characteristics
Male Partner’s age (M, SD)	38.1(6.03)	40.87(7.01)	t=2.52	.013*
Female Partner’s age (M, SD)	35.1(5.83)	38.65(5.9)	t=3.63	<.001**
Male Partner’s Race (n, %)			χ2=3.67	.055
Non-Hispanic White (n, %)	56 (83.5%)	73 (93.6%)
Other (n, %)	11 (16.5%)	5 (6.4%)
Female Partner’s Race (n, %)			χ2=5.42	.020*
Non-Hispanic White (n, %)	57(85.1%)	75(96.2%)
Other (n, %)	10(14.9%)	3(3.8%)
Father’s yrs education (M, SD)	12.33(1.93)	14.73 (2.38)	t=6.59	<.001**
Mother’s yrs education (M, SD)	13.73(2.34)	14.93(2.16)	t=3.22	.002**
Years Living Together (M, SD)	9.83(5.83)	14.79(6.88)	t=4.72	<.001**
No. of children in household (M, SD)	2.4 (1.01)	2.2 (.95)	t=1.04	.299

Note. Treatment Sample N = 67, Community Comparison Sample N = 78

^*p < .05

^**p < .01

For the treatment sample, the current (past 6 months) substance use disorder (SUD) diagnoses on the SCID for the male patients were alcohol dependence (n=63; 94%) alcohol abuse (n = 2; 3%)² drug abuse or dependence (n = 34; 51%). Female partners in the treatment sample had low levels of current SUD diagnoses – alcohol dependence (n=5; 7%), alcohol abuse (n = 3; 4%), and drug abuse or dependence (n = 3; 4%). In the community sample, both partners had low levels of SUD diagnoses – alcohol dependence (n = 5; 6% for men, n = 2; 3% for women), alcohol abuse (n = 2; 3% for men, none for women), and drug abuse or dependence (none for men, n = 1; 1% for women).

2.2. Procedures

2.2.1. Alcoholism treatment received by fathers

Participants were recruited from four programs in Massachusetts in three levels of care: inpatient/residential (n = 37, 55%), intensive outpatient or day treatment (n = 16, 23%), and outpatient counseling (n = 14, 20%).

2.2.2. Baseline and follow-up data collection

Data on drinking and exposure to parental conflict were collected at baseline, 6- and 12-month follow-ups from both samples.

2.3. Measures

2.3.1. Measure of child exposure to parental conflict

Exposure to conflict was measured using the O’Leary-Porter Scale of Overt Hostility (Porter & O’Leary, 1980) with both partners perception of positive and negative interactions in the presence of the target child. The O’Leary Porter Scale (OPS) is a 10-item scale using a 5-point rating ranging from “never” to “very often,” with higher total scores reflecting lower exposure to conflict. The questions include conflicts over finances, discipline, verbal, and physical hostility as well as one question about affection. Reports were collected from both partners and combined by item, such that the more severe rating was used for each item. The OPS is a reliable measure of overt hostility with a Cronbach’s alpha of .86 and test-retest reliability of .96 over a two week period (Porter & O’Leary, 1980). The OPS has been found to correlate significantly with measures of marital adjustment (Emery & O’Leary, 1982; Emery & O’Leary, 1984) and conduct problems (Johnson & O’Leary, 1987; Porter & O’Leary, 1980).

2.3.2. Frequency of substance use and abstinence by alcoholic fathers

Both partners completed the Timeline Follow-Back Interview (TLFB; Sobell & Sobell, 1996) to measure the number of drinking and heavy drinking days (i.e., ≥ 6 standard drinks), and other drug use.

2.3.3. Remission as treatment outcome

Based on their substance use the year after baseline, patients were categorized as remitted or relapsed (Moos, Finney & Cronkite, 1990). Patients were classified as remitted if in the year after baseline they were: (a) completely abstinent or drinking < 3 oz. of alcohol per day for no more than 10% of the interval; (b) free from illicit drug use except marijuana for no more than 10% of the interval; (c) no hospitalization for substance use; (d) no legal problems from substance use; and (e) no employment problems from substance use.

3. Results

As an omnibus analysis, a 2 Groups (alcoholic and community samples) by 3 Time periods (baseline, 6-, and 12-months follow-up) repeated measures ANOVA was run with OPS scores as the dependent variable. Results showed significant effects for Group (F (1, 143) = 27.96, p<.001) with higher conflict exposure in the treatment sample and for Time (F (2, 286) = 16.11, p<.001) with conflict exposure decreasing over time. The interaction showed a non-significant trend, (F (2, 286) =24.85, p=.074).

3.1. COAs’ Conflict Exposure Before Treatment Compared to Community Sample

As predicted, at baseline, COA (n=67) had significantly more conflict exposure (i.e., lower OPS scores) than community sample children (n=78) [M_{Alc Baseline} = 22.6 (6.18) vs. M_{Comm Baseline} = 28.3 (5.99); t (1, 143) = 5.58, p = <.001].

3.2. Improvements in COAs’ Conflict Exposure Following Treatment

As predicted, following treatment, COA’s exposure to conflict decreased significantly (see Table 2). Paired sample t-tests showed a significant decrease in COAs’ exposure to conflict between baseline and 6- month follow-up [M_Alc-Baseline= 22.6 (6.18) vs. M_Alc-6mo= 24.6 (6.27); t(66)=3.02, p=.004] and between baseline and 12-month follow-up [M_Alc-Baseline= 22.6 (6.18) vs. M_{Alc-12 mo}= 25.3 (6.3); t(66)=3.57, p=.001]. There was no significant difference for the treatment sample between six and twelve months.

Table 2.

O’Leary Porter Scale Mean (S.D.) Scores in Treatment and Community Samples

Sample	Baseline	6-months	12-months
Community (n=78)	28.3 (SD=6)	28.6 (SD=5.6)	29.7 (SD=5.3)
Treatment (n=67)	22.6 (SD=6.2)a	24.6 (SD=6.3)b	25.3 (SD=6.3)b
Treatment Subset who Remitted (n=20)	24.0 (SD=6.8)	26.9 (SD=5.4)	26.4 (SD=6.4)
Treatment subset Non-Remitted (n=43)	22.0 (SD=5.8)	23.7 (SD=5.6)	24.2 (SD=6.0)

Note. A lower score on the O’Leary Porter Scale reflects greater conflict exposure.

^aThe treatment sample had significantly greater conflict exposure than the community sample at the baseline assessment.

^bThis score in the treatment sample showed significant decrease in conflict exposure from the baseline assessment.

3.3 Comparisons of Conflict Exposure between Remitted and Relapsed Families

We ran analyses of covariance (ANCOVA) with outcome status (remitted, relapsed) as the grouping variable, controlling for baseline conflict exposure, predicting conflict exposure at 6- and 12-month follow-ups. Included were 20 remitted, 43 relapsed, and 4 patients missing data. There were no significant differences at the 6-month[M_Remitted-6mo = 26.9 (5.4) vs. M_{Relapsed-6mo =} 23.7(5.6); F(1,63) = 2.81, p = .099] and 12-month follow-ups [M_{Remitted-12mo} = 26.4 (6.4) vs. M_{Relapsed-12mo =} 24.2(5.9); F(1,63) = .684, p = .412].

3.4. Conflict Exposure in COAs Whose Fathers Remitted Following Treatment

We predicted following fathers’ recovery, COA’s exposure to conflict would decrease to the level found in the community sample. As predicted, at the 6-month follow-up, COAs with remitted fathers (n=20) did not differ significantly from the community sample (n=78) [M_{Alc Remitted 6mo} = 26.9 (5.4) vs. M_{Comm 6mo} = 28.6 (5.6); t(96)=1.29, p=.199]. However at the 12-month follow-up COAs whose fathers remitted had significantly higher conflict exposure [M_{Alc Remitted 12mo} = 26.4 (6.4) vs. M_{Comm 12mo} = 29.7 (5.3); t(96)=2.38, p=.019]. The difference between samples at the 12-month follow-up may have been due to the significant decrease in conflict exposure in the community sample between 6- and 12-months [M_Comm-6mo = 28.6 (5.6) vs. M_Comm-12mo = 29.7 (5.3); t(77)=2.88, p=.005]. Conflict exposure did not change significantly between 6- and 12-month follow-ups in the remitted subsample [M_{Alc Remitted-6mo} = 26.9 (5.4) vs. M_{Alc Remitted-12mo} = 26.4 (6.4); t(19)=.46, p=.65].

4. Discussion

Study results supported the predicted association of fathers’ alcoholism treatment with decreases in their children’s exposure to parental conflict as an indirect benefit of treatment. Prior to treatment, children of the treatment sample were exposed to more conflict than the community sample. After treatment, the exposure to conflict significantly decreased at both the six and twelve month follow-ups for the entire treatment sample.

The results partly supported the prediction that following successful treatment, the levels of conflict exposure would decrease to levels found in the community sample. Six months after baseline the children of the remitted alcoholics did not significantly differ in levels of exposure to conflict compared with the community sample, although at the twelve month follow-up there was a significant difference. However, this difference at twelve months may be due to a significant decrease in exposure to conflict in the community sample, rather than a significant increase in the remitted sample. The results of the present study parallel prior research (Burdzovic Andreas, et al., 2006; Moos & Billings, 1982) showing that fathers’ successful treatment for alcohol dependence is associated with clinically significant decreases in COAs’ emotional maladjustment, comparable to levels found in the community.

This study provides evidence that an indirect benefit of alcohol dependence treatment for COAs is that they are exposed to significantly less parental conflict, which does not appear to return to pre-treatment levels up to a year after baseline. Improvement following treatment appeared to be present for the entire treatment sample, rather than limited only to treatment responders who achieve full-scale remission. This may indicate that even treatment attempts that do not lead to paternal remission may have some stabilizing effects. Decreased exposure to conflict, and the emotional turbulence associated with it, may lead to improvements in COA functioning. Clinical implications from this research indicate clinicians should target conflict in front of children, and discuss the harmful results.

The study’s strengths include incorporating reports of conflict exposure from both spouses to avoid bias, and, increasing external validity by comparing the COAs to a community sample. This study’s limitations include that it was a naturalistic rather than controlled treatment study, so treatment varied, which was not examined in the present study. The study was limited to families where only the male partner reported alcohol problems, which may not generalize to families with different configurations. Finally, age and gender of the child were not examined due to the study’s limited sample size. This study contributes to the COA literature by indicating that reduced conflict exposure appears to be a secondary benefit of paternal alcohol treatment.

Highlights.

• We compared conflict exposure in children of alcoholics (COA) to a community sample

• Before treatment, COA were exposed to more conflict than the community sample.

• COAs’ conflict exposure was decreased at 6- and 12-months follow-up.

• Conflict exposure was similar to the community sample at 6 months follow-up.

• Decreased conflict exposure after parental treatment may improve COA functioning.