 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Department of Psychiatry, New York University School of Medicine, New York, New York
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
METHODS. At time 1, structured interviews were administered, separately and in private, to male and female youth (
age = 16.3) and their fathers; adolescents were reinterviewed 
1 year later (at time 2). Structural equation modeling was used to examine the interrelationship of time 1 paternal tobacco and illicit drug use, father–child relations, adolescent psychological adjustment, and peer group factors and adolescent smoking at time 2. A supplementary analysis assessed the same model with control on the adolescent's age, gender, frequency of contact with the father, and the father's treatment status.
RESULTS. The structural equation model showed a mediational pathway linking paternal tobacco and drug use to a weak and conflictual father–child relationship, which was associated with greater adolescent maladjustment, which in turn was related to deviant peer affiliations, which predicted adolescent smoking at time 2. There was also a direct path from paternal tobacco and drug use to adolescent time 2 smoking. The supplementary analysis found no significant differences between the models with and without control.
CONCLUSIONS. Findings provide evidence of the mechanisms that underlie the association between paternal drug use characteristics and smoking in the adolescent child. Clinical implications suggest the importance of the father–child relationship to smoking prevention programs for at-risk youth.
Key Words: paternal drug use • at-risk youth • adolescent smoking • father-child relationship
Abbreviations: SEM—structural equation model • GFI—goodness-of-fit index
Increasing evidence suggests that the children of drug abusers may be at elevated risk for maladjustment and problem behaviors, including substance use.1–3 Surprisingly little research, however, has examined the impact of parental drug abuse on smoking in the adolescent child. The adolescent children of drug-abusing parents may be especially susceptible to tobacco use, which in turn has been linked, both concurrently and prospectively, with drug use and psychopathology.4, 5 An understanding of smoking in this at-risk population, therefore, may help to elucidate the mechanisms through which parental drug use has an adverse impact on the child. To our knowledge, the present study is the first longitudinal investigation of the interrelationship of several psychosocial constructs with respect to adolescent smoking in the children of drug-abusing fathers. Specifically, we examined the prospective association of paternal tobacco and illicit drug use, the father–child relationship, the adolescent's psychological adjustment, and aspects of the peer group, with respect to cigarette smoking in a cohort of at-risk youth.
Although few studies have examined the impact of parental drug use on their children's cigarette smoking, findings have suggested that the offspring of drug users have elevated rates of tobacco use.1, 6 Results of the more extensive literature on parental tobacco use and children's smoking have been somewhat inconsistent, but most studies have shown an increased risk for offspring smoking in both adolescence and adulthood.7–9 In general, previous research suggests that parental tobacco and drug use may have a considerable impact on the initiation and persistence of smoking in the adolescent child.10, 11 Li et al,11 for example, found that parents' tobacco use was significantly associated with the later use of cigarettes, marijuana, and alcohol by their adolescent children, with the strongest influence found for smoking. These researchers also showed that any parental substance use was significantly related to higher levels of adolescent smoking. In addition, Sutherland and Willner12 found an association between both the independent and combined effects of parental tobacco and drug use and increased rates of smoking in the adolescent child. Research suggests that aspects of the parent–child relationship (eg, smoking rules13), the adolescent's personality (eg, refusal self-efficacy14), and peer group influences (especially tobacco use15) may mediate the association of parental and adolescent smoking, although there is relatively little research in this area. Studies that examine the indirect effect of parental drug use on adolescent smoking are scarcer still, but findings also support the possibility of mediational effects (eg, Li et al11). In general, however, the association of the parent–child relationship and adolescent smoking has not been addressed in the literature in the context of paternal drug use.
Studies of non–drug-using parents suggest that less parental nurturance and involvement increase the risk for adolescent smoking.16 Harakeh et al,17 for example, showed that a positive parent–child relationship was linked with less adolescent smoking both concurrently and prospectively. One construct that may link the father–child relationship with adolescent smoking is the youth's psychological adjustment. According to family interactional theory18 as well as previous investigations, adolescents with weak parental bonds and greater parent–child conflict are at elevated risk for psychological maladjustment and maladaptive behaviors, such as greater depression, anxiety, delinquency, and tobacco and drug use.16, 19, 20
A father who abuses drugs may engender a less cohesive and more stressful family environment,7, 21 and his ability to form a caring and supportive relationship with his child may be seriously compromised, eg, as a result of emotional dysregulation or unavailability, lack of financial resources, comorbid psychopathology, and physical health problems.22, 23 The adolescent children of fathers who use drugs may have weak and conflictual paternal bonds, which place them at risk for externalizing and internalizing symptoms as well as problem behaviors.24, 25
Numerous aspects of maladjustment have been linked with adolescent tobacco use in both cross-sectional and longitudinal investigations.5, 26 According to the theory of problem behavior27 and the findings of other research,28 there is a strong concordance between adolescent smoking and problem behaviors such as rebelliousness, depression, poor school achievement, emotional undercontrol, and drug use. Moreover, several investigators have found that adolescents with poorer psychological adjustment and greater problem behaviors, including smoking, are more likely to affiliate with deviant peers.15, 29 There also may be a reciprocal effect between adolescents and their peer groups wherein maladjusted youth are more likely to select deviance-prone peers, who in turn influence the adolescents' problem behaviors.30
Research has consistently demonstrated the importance of the peer group for adolescent smoking.31, 32 Simons-Morton,33 for instance, found that an increase in the number of friends who smoke was predictive of smoking initiation in adolescents. Bryant and Zimmerman34 also showed that adolescents who perceive that their friends do well in school are less likely to smoke cigarettes. Furthermore, numerous studies have shown that adolescent smoking is associated with several problem behaviors within the peer group (eg, deviance, poor grades, substance use) in addition to peer smoking.35–37 Although findings in the literature have been inconsistent regarding the relative impact of parents and peers on adolescent smoking, a growing body of evidence suggests that their influences may be interrelated, eg, a weak parent–child bond may increase the influence of deviant peers on the likelihood of adolescent smoking.37 However, the exact nature of the relationship between parenting behaviors and peer influences is not yet understood with regard to adolescent smoking.33
The objective of the present study was to examine the interrelations of risk factors that are associated with adolescent smoking among children of drug-abusing fathers. Specifically, our conceptual model hypothesized that (1) greater paternal tobacco and illicit drug use would be associated with a weak and conflictual father–child relationship, which (2) would be linked with the adolescent's psychological maladjustment, which (3) would increase the likelihood of adolescent smoking 1 year later (time 2), and also have a reciprocal relationship with a deviance-prone peer group, and (4) problem behaviors in the peer group would also predict adolescent smoking at time 2. We also expected that greater paternal tobacco and illicit drug use would have a direct path to adolescent time 2 smoking, as well as be mediated by the father–child relationship. In addition, we used a study design that incorporated responses from both father and child. An advantage of this procedure is that it helps to reduce single-respondent biases in the study of father–adolescent relations.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
At time 1, each father and adolescent pair was administered individual structured interviews by trained interviewers, separately and in private. For the present analysis, adolescents were reinterviewed (by the same interviewer, whenever possible) at time 2, 1 year later. A 12-month interval between time waves was selected because of the rapid transitions in smoking status during adolescence, as well as the difficulty in retaining a highly transient cohort. Interviews took 2 hours, and all participants were paid for each interview to compensate for their time and effort. In keeping with institutional and federal guidelines for the protection of human subjects, written informed consent was obtained from all participants. In addition, fathers or legal guardians provided written informed consent for their child's participation when he or she was younger than 18 years at time 1. Institutional Review Board approval was also obtained, as was a Certificate of Confidentiality from the National Institute on Drug Abuse, prohibiting access to interview information.
Attrition in this study (39%) was attributed to the inability to relocate adolescent participants, despite repeated attempts to contact them or their fathers and mothers or the adjunct contact people furnished at time 1. Compared with youth who completed both waves of interviews, adolescents whom we could not relocate at time 2 reported greater paternal drug use (t = 2.98; P 
.01) and more paternal cigarette smoking (t = 2.35; P .01) as well as less peer achievement (t = 2.15; P .05), at time 1.
Measures
Four latent variables were hypothesized on the basis of time 1 data: (1) paternal tobacco and illicit drug use, (2) the father–child relationship, (3) the adolescent's psychological adjustment, and (4) peer group factors. The dependent variable was adolescent cigarette smoking at time 2. Paternal tobacco and illicit drug use (time 1) consisted of a single-item scale that tapped the youth's perception of the greatest quantity of paternal cigarette smoking18 as well as an index of the youth's perception of the greatest frequency of paternal drug use18 (3 items that assessed the father's perceived use of marijuana, other illicit drugs, and nonprescription tranquilizers). The father–child relationship latent construct (time 1) included 6 scales, 2 of which comprised single items (father's smoking rules18 and father's educational expectations for his child39) and 4 of which were multi-item scales (father's child-centeredness [4 items, 
= .71],40 father's satisfaction with his child [6 items, = .86],18 father–child conflict [4 items, = .79],40 and the youth's identification with his or her father [10 items, = .91]).18 The father–child relationship scales were based on the father's report, with the exception of 2 scales (youth identification with the father and father–child conflict). The adolescent's psychological adjustment latent variable (time 1) was based on the adolescent's self-report of the following measures: rebelliousness (10 items, = .69),41 impulsivity (3 items, = .58),42 ego integration (5 items, = .60),18 responsibility (4 items, = .37),42 school achievement (5 items, = .63),18 externalizing problem behaviors (16 items that combined aggression [ = .76],43 delinquency [ = .75],44 and deviance [ = .71]44), and psychopathology (15 items that combined depression [ = .75],45 anxiety [ = .70],45 and interpersonal problems scales [ = .78]45). Peer group factors (time 1) were hypothesized on the basis of the youth's report of friends' smoking18 (single item), deviance (5 items, = .75),44 and academic achievement (3 items, = .61).18 These independent measures were shown to be valid and reliable in our own previous research and that of other investigators and were also found to predict adolescent tobacco use, drug use, and psychopathology.18, 46 The groupings for the measures were based on our conceptual framework as well as that of other investigators.18, 27 Smoking at time 2 was a self-report measure of the adolescent's cigarette smoking during the past year.18 The response range of the adolescent smoking scale was as follows: 1, none; 2, some but less than daily; 3, 1 to 5 cigarettes a day; 4, approximately half a pack a day; 5, approximately a pack a day; and 6, 1.5 packs a day or more. The mean frequency of adolescent past year smoking was 1.9 (SD: 1.4) at time 2 and 1.8 (SD: 1.4) at time 1. Although these respective means were not significantly different (t = –1.12; P = .26; degrees of freedom = 590), 17% of the youth increased smoking and 12% decreased smoking between times 1 and 2. Time 2 was selected as the outcome measure to allow for a temporal ordering of the latent constructs with respect to adolescent smoking. At time 2, 27% of the adolescents smoked cigarettes on a daily basis and 62.5% were abstainers. All Cronbach's for the scales were satisfactory, with one exception (adolescent responsibility), and the mean Cronbach's for all of the scales was .71.
Data Analysis
Latent variable structural equation models (SEMs) were used to examine the empirical validity of the processes hypothesized to be linked with adolescent time 2 smoking. SEM is a multivariate statistical method that evaluates both the measurement quality of a set of variables used to assess a latent construct (the measurement model) and the relationships among the latent constructs (the structural model). According to the guidelines set forth by Marsh et al,47 our sample size was sufficient for the use of SEM, which is best suited to a minimum of 200 observations. We used maximum likelihood methods to estimate the models by using LISREL VIII.48 The fit of the models was assessed with multiple indices: the goodness-of-fit index (GFI), the standardized root mean square residual, and the comparative GFI. For the nested models, we used 
2 tests to determine the improvement of the model at each incremental step. Supplementary analyses used SEM to test the same hypothesized model with control on the adolescent's age, gender, frequency of contact with the father, and the father's treatment status.
Fewer than 2% of the data were missing for some of the independent variables. Missing data were imputed using the cross-sectional (time 1) mean score for the item. There were no missing data for the measurement of adolescent tobacco use at time 2.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The adolescents' smoking at time 2 was not significantly related to the fathers' self-reported frequency of counseling sessions at time 1 (r = – 0.09; P = .16) or to the youth's time 1 report of the frequency of father–child contact (r = –0.10; P = .09). However, there was a trend toward significance for less frequent contact and greater adolescent time 2 smoking.
Structural Equation Modeling
We tested the measurement model as well as our conceptual model. The hypothesized and obtained models were similar, with the exception of 2 hypothesized paths that were found to be statistically nonsignificant and, therefore, omitted from the obtained model (ie, adolescent maladjustment did not have a reciprocal path to peer group factors or a direct path to smoking at time 2). All factor loadings were significant (P < .001), and all were >.20. These findings show that the indicator variables were a good measure of the latent constructs. The correlations among the variables that were derived from the covariance matrices and the information about the factor loadings from the measurement model are available from the authors on request.
Figure 1 presents the obtained structural model. The empirical results essentially confirmed our hypothesis. The following fit indices were obtained: GFI = 0.91, standardized root mean square residual = 0.059, and comparative GFI = 0.95. These results reflect a satisfactory model fit. As noted in the obtained model (Fig 1), paternal tobacco and illicit drug use at time 1 had a statistically significant path to a less nurturing and involved father–child relationship at time 1 (t = 4.33); a weak father–child relationship at time 1 had a significant path to the adolescent's psychological maladjustment (time 1; t = 5.93); and the adolescent's maladjustment had a significant path to peer smoking and deviance at time 1 (t = 6.02), which in turn had a significant direct path to adolescent smoking at time 2 (t = 3.05). There was also a significant direct path from paternal tobacco and illicit drug use at time 1 to adolescent time 2 smoking (t = 3.36). (All t values are significant at P < .05 and are based on a 2-tailed test.)
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Consistent with our findings regarding paternal tobacco and illicit drug use, most previous investigations have reported higher prevalence rates of smoking among children of cigarette smokers as compared with children of nonsmoking parents.9, 14 In addition, the relatively few studies that have specifically examined tobacco use in the children of drug-using fathers also have found elevated rates of smoking.1, 6 According to social learning theory, children may initiate smoking by the imitation of parental behavior.49 Furthermore, our finding of a direct effect of paternal tobacco and illicit drug use on adolescent smoking at time 2 may suggest that some aspects of the father's role modeling are not substance specific50; that is, the father's illicit drug use may serve as a model of the emotional regulatory effects of drugs, which is one of the reasons that adolescents smoke (eg, Chabrol et al51). It also is possible that the direct effect of paternal tobacco and drug use on adolescent smoking involves a common genetic diathesis, shared by father and child, which also may entail a vulnerability to multiple substances, eg, tobacco and opiates.52 Nonetheless, most studies on the genetics of smoking have shown a substantial environmental (ie, psychosocial) effect.53, 54 In this vein, our model also demonstrated an indirect path from paternal tobacco and illicit drug use to adolescent smoking at time 2, which was mediated by the parent–child relationship.
Our results suggest that paternal tobacco and illicit drug use may have an adverse impact on the father's childrearing practices and the establishment of a predominantly warm and nonconflictual relationship with his adolescent child. Fathers who smoke cigarettes may be less likely to have or to enforce antismoking rules for their child,55 and the adolescent may be more likely to smoke in the absence of such rules.13 In addition, the father may be more focused on procuring and taking illegal drugs than on his child's education and welfare. The adolescent child may also have less identification (defined in our study as admiration and emulation) with a father who is abusing drugs. There is currently scant empirical research on the father–child relationship in the context of paternal drug use,56 and even fewer studies, to our knowledge, have specifically looked at adolescent smoking as the outcome of interest. In a related vein, however, Rotheram-Borus et al57 examined the parent–child relationship in HIV-positive mothers and fathers, most of whom were drug users, and found that current parental drug use was highly associated with parent–child conflict. Brook et al58 also found that parental illicit drug use was inversely related to positive aspects of the parent–child bond in a community sample of young mothers and fathers and their children. In addition, our results are consistent with the few investigations of the mediational role of the parent–child relationship in the adjustment of the children of drug-using parents. Fals-Stewart et al,2 for instance, showed that interparental conflict and poor parenting skills linked paternal substance use with internalizing and externalizing symptoms in their preadolescent children. Similarly, Su et al21 found that negative life events and decreased family cohesion mediated the longitudinal association between maternal and/or paternal substance use disorders and both depression in and the use of alcohol and illicit drugs by their early adolescent child. (Of note is that these researchers also found a direct link between parental and offspring substance use, similar to our own finding.) Taken together with the present study, these preliminary results on the children of drug users point to the impact of paternal drug use on the father–child relationship, as well as the importance of the father–child bond to the adolescent's psychological adjustment.
Our model showed that an uninvolved and less nurturing father–child relationship increased the likelihood of adolescent affective and behavioral problems, such as internalizing (eg, depression, anxiety) and externalizing (eg, deviance, aggression) symptoms, as well as less ego integration and poor school achievement. This finding is consistent with family interactional theory18 and with research on community samples of non–drug-using parents and their children, including those from racial and ethnic minorities (eg, Harris et al59 and Velez and Ungemack60). In the present study, adverse aspects of the father's relationship with his child, such as noninvolvement (fewer smoking rules), less supportiveness (lack of child-centeredness), and rejection (dissatisfaction with the child), were linked with greater symptoms of anxiety and depression and more aggressive behaviors in the adolescent child.61, 62 Our findings are consistent, in part, with a recent longitudinal study by Buist et al,63 which showed reciprocal and stable inverse effects between less parent–child attachment and both adolescent internalizing and adolescent externalizing behaviors. Moreover, in accordance with both family interactional theory and social control theory, aspects of a weak paternal–child bond, including low identification with the father, may loosen the ties to conventional authority that the father represents64, 65 and incline the adolescent toward rebelliousness and deviance.66, 67
Our findings with respect to the peer group are supported by numerous previous investigations.31, 32, 68 Specifically, our model showed that adolescents who were deviance prone and had greater psychological maladjustment were more likely to select friends with greater deviance, more smoking, and poorer school achievement.36 Furthermore, our results demonstrated that affiliation with deviant peers was the most proximal link to adolescent smoking at time 2.32, 69 These results are in accordance with a longitudinal study by Moss et al,70 which showed that both paternal drug dependence and adolescent externalizing psychopathology increased the risk for adolescent affiliation with a deviant peer group. Patton et al71 also found that adolescents who had mostly smoking friends were 36 times more likely to smoke at follow-up than adolescents whose peers did not smoke. Although the mechanisms underlying the relationship between a deviance-prone peer group and adolescent smoking have not been established clearly, one possible explanation is that peer behaviors result in the adolescent's perception that smoking is a normative behavior.72 Blanton et al64 further posited that adolescents with friends who smoke may develop a positive image of a smoker, which in turn predicts the adolescent's smoking initiation. Peer influences on adolescent smoking also may include a social learning process in which members of the peer group model and reinforce smoking behaviors.68 Conformity to the peer group may be especially pronounced in adolescents with greater psychological maladjustment (eg, low ego integration36), weak paternal attachment,65 and less conventional behaviors (eg, low school achievement). Cigarette smoking, therefore, may represent an important social bond for at-risk adolescents.72
Limitations
Although the present study helps to elucidate the mechanisms that link paternal drug use and adolescent smoking, there are several limitations to the findings. First, although our longitudinal model examined the interrelation of several latent constructs and adolescent smoking, some factors were not assessed. In particular, the inclusion of maternal characteristics and aspects of the mother–child relationship in future research would add to a more complete understanding of both parents' respective and combined contributions to cigarette smoking in at-risk youth.
Second, our study did not address specific subgroups of smokers, such as adolescents at different stages of smoking. Research that examines smoking trajectories may clarify further the risk factors that are associated with each stage of smoking.
Third, because the time 1 constructs all were assessed concurrently, it is possible that their temporal ordering could be different from hypothesized in our model. However, our own previous research and that of other investigations35, 64, 70 support the temporal ordering of the constructs presented in our model. Similarly, our model can only present the temporal relations between the time 1 and outcome variables but cannot document causality. However, the longitudinal analysis of these constructs using structural equation modeling suggests the possibility of a causal ordering.
Fourth, we must be cautious in generalizing our findings because of biases that may stem from the community- and clinic-based composition of our sample, as well as the fact that the treatment programs from which we recruited may not be representative of other health care systems throughout the United States.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
We thank Linda Capobianco for invaluable assistance with the preparation of the manuscript.
|
FOOTNOTES |
|---|
Address correspondence to David W. Brook, MD, New York University School of Medicine, 215 Lexington Ave, 15th Floor, New York, NY 10016. E-mail: david.brook{at}med.nyu.edu
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
, Meeus W, van Aken MA. The reciprocal relationship between early adolescent attachment and internalizing and externalizing problem behaviour.
J Adolesc. 2004;27
:251
–266[CrossRef][ISI][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2006 American Academy of Pediatrics. All rights reserved. |
||
|
||
