Research Paper (APA STYLE) as soon as possible

Familial Influences on Substance Use by Adolescents and Young Adults Author(s): Michael H. Boyle, Mark Sanford, Peter Szatmari, Kathleen Merikangas and David R. Offord Source: Canadian Journal of Public Health / Revue Canadienne de Santé Publique, Vol. 92, No. 3 (MAY / JUNE 2001), pp. 206-209 Published by: Canadian Public Health Association Stable URL: http://www.jstor.org/stable/41993309 Accessed: 03-08-2017 05:40 UTC

REFERENCES Linked references are available on JSTOR for this article: http://www.jstor.org/stable/41993309?seq=1&cid=pdf-reference#references_tab_contents You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide

range of content in a trusted digital archive. We use information technology and tools to increase productivity and

facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

http://about.jstor.org/terms

Canadian Public Health Association is collaborating with JSTOR to digitize, preserve and extend access to Canadian Journal of Public Health / Revue Canadienne de Santé Publique

This content downloaded from 208.95.50.183 on Thu, 03 Aug 2017 05:40:06 UTC All use subject to http://about.jstor.org/terms

Familial Influences on Substance Use

by Adolescents and Young Adults Michael H. Boyle , PhD,1 Mark Sanford, MB ChB,1 Peter Szatmari , MD,1

Kathleen Merikangas, PhD,2 David R. Offord, MD1

Substance use by adolescents and young adults is increasing after a decade-long decline in use during the 1980s.1"4 Efforts to understand factors leading to drug use in this age group have focused on genetic, individual and environmental influences

(especially family, peer and school environ-

mental factors).5 Evidence suggests that genetic factors account for a large share of observed familial aggregation of use and abuse of alcohol and illicit drugs, and that both general and drug-specific genetic influences are at play.6 However, the rela- tive importance of environmental and genetic factors is unclear. For example, a recent population-based twins study found little evidence of family environmental fac-

tors causing alcohol-related disorders in contrast to earlier clinic- and court-based

studies.7 In contrast, in a U.S. adoption cohort, substantial environmental effects

were reported for sibling aggregation (but not for parent-child aggregation) of alco- hol use and problem drinking (r = 0.24 in non-biological siblings, r = 0.45 in same sex/same age siblings vs. r = 0.01 (p>0.05) for opposite sex/dissimilar age siblings).8

A criticism of the literature in this area is

that very few studies have used representa-

tive community-based samples to evaluate these issues free from the referral biases

1. Department of Psychiatry and Behavioural Neuroscience and Centre for Studies of Children

at Risk, McMaster University, Hamilton, ON 2. Departments of Psychiatry and Epidemiology,

Yale University, New Haven, CT Correspondence and reprint requests: Michael H. Boyle, Centre for Studies of Children at Risk, McMaster University Faculty of Health Sciences and Hamilton Health Sciences Corporation, Patterson Building, Chedoke Division, Hamilton Health Sciences Corporation, Box 2000, Hamilton, ON, L8N 3Z5, Tel: 905-521-2100 ext 77365, E-mail: [email protected] Dr. Michael Boyle is supported by a Canadian Institutes of Health Research Scientist Award and

Dr. Peter Szatmari is supported by a Senior Fellowship Award from the Ontario Mental Health Foundation.

inherent in clinic samples.9 Two such stud-

ies report on sibling effects. In the U.S. National Survey on Drug Abuse, drug use by teenagers correlated with drug use by fathers (from r = 0.17 for current cigarette

use to r = 0.31 current marijuana use), by mothers (from r = 0.04 (p>0.05) for current alcohol use to r = 0.30 for lifetime mari-

juana use), and by older siblings (r = -0.07 (p>0.05) for lifetime cocaine use to r = 0.30 for lifetime marijuana use).10 In a second study, alcohol use was highly con- cordant among spouses (r = 0.67) but less so among parents and offspring and sib- lings (from r = 0.25 to r = 0.44). 11

The objective of this study is to examine

family influences on tobacco, alcohol, and marijuana use by 12-24 year olds in a rep- resentative general population sample liv- ing at home. The following questions are addressed: 1) what is the prevalence of sub-

stance use among family members?; 2) does sibship structure influence aggrega-

tion or clustering of substance use in fami-

lies?; 3) does gender or age influence within-family correlations of substance use?;

and 4) is substance use among younger sib- lings associated more strongly with parental substance use or substance use by older siblings? Answers to these questions will provide insight into the relative importance of parental and sibling effects and whether observed sibling aggregation is likely to be due to shared genetic or environmental effects.

METHODS

Survey design Information for this research comes

from the Ontario Health Survey (OHS: ref. 12). The sample was obtained using a stratified multi-stage cluster design: there were 35,479 dwellings selected and 87.5% of these yielded a person knowledgeable

206 REVUE CANADIENNE DE SANTÉ PUBLIQUE VOLUME 92, NO. 3

ABSTRACT

This study uses data from the Ontario Health Survey to examine within-family influences (sibship number, age and sex com- position; family structure and parental sub- stance use) on the use of tobacco, alcohol and marijuana in households (N=4,643) among offspring aged 12 to 24 years. Using a modification of the kappa statistic, concor- dance among siblings is modest generally and undifferentiated across substance type. Concordance is stronger among sibships that are either all male or older (19-24 years) and

is particularly strong for siblings < two years

apart in age. The dominant influence of sub- stance use behaviour appears to be from older siblings to younger siblings and not from parents to offspring. Sibling concor- dance for substance use suggests that the treatment and prevention of substance use (and abuse) among adolescents and young adults might be enhanced by including a family focus, especially where there are two or more siblings at home.

ABRÉGÉ

L'étude, fondée sur les données de l'Enquête

sur la santé en Ontario, porte sur les influences

intrafamiliales (nombre, âge et sexe des mem- bres de la fratrie, structure familiale et consom-

mation de drogues par les parents) sur la con-

sommation de tabac, d'alcool et de marijuana dans des familles (n = 4 643) d'au moins un

parent et un enfant de 12 à 24 ans. Une analyse

Kappa modifiée révèle une concordance frater-

nelle modeste dans l'ensemble, peu tranchée selon le type de drogue. La concordance est plus

prononcée lorsque la fratrie se compose unique-

ment de garçons ou comporte un membre plus

âgé (19 à 24 ans), particulièrement lorsque l'écart entre les membres est de deux ans ou

moins. L'influence dominante, en matière de

consommation de drogues, semble s'exercer des

aînés aux cadets et non des parents aux enfants. La concordance constatée entre les membres de

fratries porte à croire que l'on pourrait améliorer

le traitement et la prévention de la consomma-

tion (et de l'abus) de drogues chez les adoles- cents et les jeunes adultes en y intégrant une composante familiale, surtout lorsqu'il y a plus d'un enfant à la maison.

This content downloaded from 208.95.50.183 on Thu, 03 Aug 2017 05:40:06 UTC All use subject to http://about.jstor.org/terms

MAY - JUNE 2001 CANADIAN JOURNAL OF PUBLIC HEALTH 207

about the household who agreed to pro- vide information on sociodemographic and general health variables of all residents in a

personal interview. Self-administered ques-

tionnaires on more personal health issues (including substance use) were left for completion by all household members aged 12 years and older.

Sample for analysis The sample consisted of families with at least one offspring aged 12 to 24 years and at least one natural parent in the home. All

household members other than offspring and their natural parent(s) were excluded. Applying these criteria yielded a potential sample of 23,915 individuals living in 6,976 families. Failure of individuals to

return self-administered questionnaires reduced the sample by 24.5% (5,861/23,915). Missing data on one or more of the variables needed for the analy-

ses reduced the sample by another 14.9% (3,564/23,915). A comparison of respon- dents and nonrespondents based on inter- view data is presented in Table I. Most dif- ferences are less than 5% and none are

larger than 10%. The following groups (>5% differences) may be over-represented:

one-parent households; parents born in Canada who were younger and better edu- cated; and siblings who were older and female.

Variables and measures

The substance use variables were mea-

sured independently for respondents based on their answers to self-administered ques-

tionnaires. Use of cigarettes refers to daily consumption; use of alcohol refers to aver-

age weekly consumption of 2 or more drinks in the previous 12 months; use of marijuana refers to lifetime use of marijua- na, cannabis or hash at least 5 times.

Sociodemographic variables were derived from responses provided by the knowl- edgeable household member to the interviewer-administered questionnaire.

Analysis Prevalences are estimated by the number

of users per 100 for each category of sub- stance use. A modification of the kappa statistic proposed by Fleiss13 is used to esti-

mate familial aggregation. In this method,

TABLE I

Distribution of Selected Characteristics of Respondents and Nonrespondents

Characteristics Respondents Nonrespondents Households (6976) (4643) (2333) Urban residence 74.5 70.2 <6 members in household 93.4 91.8 Income below the poverty line 1 4.4 11 .0 One parent in home 1 8.3 1 3.0

Parents (12,919) (7749) (5170) Female 54.1 49.7 Age in years <50 78.4 71 .2 Birthplace Canada 75.5 66.6 Secondary school complete 64.1 58.2 Working 79.1 79.1

Siblings (1 0,996) (6741 ) (4255) Female 49.2 42.2 Age in years 1 8-24 61 .1 54.6 In school 83.0 78.2

TABLE II

Substance Use Prevalence per 100 Among Parents and Siblings

Tobacco Alcohol Marijuana (current) (one year) (lifetime)

Parents Mothers (4194) 25.9 18.9 7.5 Fathers (3555) 27.8 42.3 9.5

Siblings Sex Females (3423) 11.6 7.8 10.6 Males (3318) 16.3 16.8 15.7

Age in years 12-14(2073) 2.1 1.6 1.3 15-17(2048) 12.2 4.9 8.9 18-20(1580) 20.8 22.1 19.3 21-24(1040) 25.0 29.3 28.0

TABLE III

Family Composition and Aggregation of Substance Use Among Siblings

Feature

Category (n) Tobacco Alcohol Marijuana

Overall (1 727) 0.1 9 (0.02) 0.1 5 (0.02) 0.20 (0.02) Number of siblings 2 (1 405) 0.20 (0.03) 0.1 4 (0.03) 0.20 (0.03) >3 (322) 0.1 5 (0.03) 0.1 8 (0.03) 0.1 8 (0.03)

Sex composition All female (368) 0.09 (0.05) 0.09 (0.05) 0.1 7 (0.05) All male (41 5) 0.24 (0.04) 0.1 7 (0.04) 0.28 (0.04) Mixed (994) 0.20 (0.03) 0.1 4 (0.03) 0.1 7 (0.03)

Age composition All 1 2-1 8 (971 ) 0.20 (0.03) 0.1 6 (0.03) 0.1 6 (0.03) All 1 9-24 (1 82) 0.30 (0.06) 0.25 (0.06) 0.29 (0.06) Mixed (574) 0.1 1 (0.03) 0.03 (0.03)* 0.1 3 (0.03)

* not significant at p<0.05

an analysis of variance approach is taken to adjust for variations in the number of sib- lings per family. Absolute differences between kappas in excess of 0.10 are defined arbitrarily as 'important' effects. Considered in this analysis are the effects of sibship structure (sibship size, sex com- position and age composition) on familial aggregation. Kappa is used also to examine the effects of gender and age on within-

family correlations by estimating levels of concordance on substance use between

pairs of family members, including spous- es, spouses and siblings and siblings sub- divided by sex and age. Households with four or more siblings [43/4,643 (0.9%)] are excluded from this analysis to prevent large sibships from exerting undue weight on the analysis. Finally, logistic regression analysis is used to estimate the strength of

This content downloaded from 208.95.50.183 on Thu, 03 Aug 2017 05:40:06 UTC All use subject to http://about.jstor.org/terms

FAMILIAL INFLUENCES

TABLE IV

Kappa Estimates of Agreement Between Family Members on Substance Use

Members (n) Tobacco Alcohol Marijuana Spouses (3106) 0.30(0.02) 0.32(0.02) 0.44(0.03) Father-Son (2728) 0.14(0.02) 0.09(0.02) 0.05(0.02) Father-Daughter (2492) 0.04(0.02) 0.06(0.01) 0.05(0.02) Mother-Son (3106) 0.06(0.02) 0.07(0.02) 0.01(0.02)* Mother-Daughter (301 7) 0.16(0.02) 0.12(0.02) 0.08(0.02) Brothers < 2 yrs (290) 0.45 (0.08) 0.44 (0.07) 0.33 (0.08) Brothers > 2 yrs (333) 0.02 (0.05)* 0.07 (0.04)* 0.06 (0.07)* Sisters < 2 yrs (240) 0.27(0.09) 0.34(0.11) 0.23(0.09) Sisters > 2 yrs (284) 0.06(0.07)* 0.05(0.06)* 0.14(0.07) Brothers-Sisters <2 yrs (51 6) 0.23(0.06) 0.22(0.06) 0.28(0.06) Brothers-Sisters > 2 yrs (579) 0.1 1 (0.04) 0.1 1 (0.04) 0.1 0 (0.04)

* not significant at p<0.05

TABLE V

Strength of Association Estimated by Relative Odds and 95% Confidence Intervals in Logistic Regression Between Substance Use in Youngest Sibling

Characteristic Tobacco Alcohol Marijuana

Youngest sibling Male gender 1 .46 (0.99-2.14) 2.4 7t (1 .53-3.97) 1 .42 (0.93-2.1 7) Age in years 1 .37t (1 .28-1 .47) 1 .59t (1 .46-1 .75) 1 .46f (1 .35-1 .59)

Parents

Lone parent 1 .36 (0.91 -2.04) 0.91 (0.53-1 .56) 1 .24 (0.79-1 .96) Substance used 1 .37 (0.93-2.03) 1 .09 (0.68-1 .74) 1 .91 (0.96-3.80)

Siblings >2 in home 0.63 (0.36-1 .11) 0.65 (0.32-1 .30) 0.78 (0.43-1 .42) Substance used 4.08t (2.75-6.05) 3.77f (2.37-5.99) 5.34t (3.48-8.1 8)

t p<0.001

association between substance use by the youngest sibling in the family (dependent variable) and use of the same substance by parent(s) and older sibling(s).

RESULTS

Prevalences and sibship structure Prevalences for all substances used are

higher among fathers than among mothers and among male siblings than among female siblings (Table II). Among siblings, there is a gradient of increasing substance use by age that begins to taper off among 21-24 year olds.

Table III shows that overall levels of

aggregation are modest for use of tobacco, alcohol and marijuana (k=0.19, 0.15, 0.20). Estimates of aggregation exhibit small variation by sibship size. Sibships composed entirely of males versus females exhibit higher levels of aggregation for use of tobacco (k=0.2 4 versus 0.09) and mari-

juana (k=0.28 versus 0.17). Age composi- tion exerts the most potent effect on esti- mates of aggregation: sibships composed

entirely of young adults aged 19 to 24 years show the highest levels of aggregation

for all categories of substance use.

Within family correlations Concordance for the same categories of

substance use are much higher between parents (k=0.30 to 0.44) than between par- ents and their children (k=0.05 to 0.16) (Table IV). There appears to be a small gender-specific effect for tobacco use. Concordance between father-son and

mother-daughter are 0.14 and 0.16, respec- tively; between father-daughter and mother-

son, they are 0.04 and 0.06, respectively. Concordance on substance use between

siblings are influenced by both closeness in

age and sex. Siblings more than two years apart in age exhibit low concordance on substance use for all pairings (k=0.02 to 0.14). Siblings within two years of each other exhibit higher levels of concordance: these estimates are higher for brother (k=0.33 to 0.45) than for sister (k=0.23 to

0.34) and brother-sister pairings (k=0.23 to 0.28) (Table IV).

Parent and older sibling substance use Table V presents relative odds (RO)

between substance use reported by parents and older siblings and substance use reported by the youngest member of the sibship. Covariates are included to control for family status (one vs. two parents); sib-

ship size (3 or more vs. 2); and the age and gender of the youngest sibling. The strength of association between parental substance use and use of the same sub-

stance by the youngest sibling is small and nonsignificant (R0=1.09 to 1.91). These contrast with substance use by older sib- lings which is associated strongly and sig- nificantly with substance use by younger siblings (RO=3.77 to 5.34).

DISCUSSION

This study provides evidence for the aggregation of substance use among sib- lings in families. Overall, the effect is mod-

est, appears to be undifferentiated across substance type and is stronger among sib- ships that are either all male or older (19- 24 years). Within-family concordance is particularly strong for spouses and for sib-

lings < two years apart in age. Furthermore, the dominant influence of

substance use behaviour appears to be from

older siblings to younger siblings and not from parents to offspring.

There are a number of possible explana- tions for the strong between-sibling associ- ations in substance use observed in this

study. First of all, access to tobacco, alco- hol and drugs such as marijuana is deter- mined very much by age. The presence of older siblings already engaging in sub- stance use provides the means for younger siblings to obtain these substances. Second, older siblings constitute important role models, and in some instances the initia-

tion of younger siblings into substance use may arise from imitative behaviour. Third,

closeness in age will have important influ- ences on socialization processes. Siblings close in age can be expected to spend more time together through childhood and ado- lescence. This will lead to more shared

experiences in the family, neighbourhood, at school and with adolescent peer groups. An older sibling engaging in substance use will likely have peers engaged in the same

208 REVUE CANADIENNE DE SANTÉ PUBLIQUE VOLUME 92, NO. 3

This content downloaded from 208.95.50.183 on Thu, 03 Aug 2017 05:40:06 UTC All use subject to http://about.jstor.org/terms

activity who, in turn, could serve as impor-

tant role models for engaging in substance use and provide additional opportunities to access tobacco, alcohol and marijuana. The relatively weak associations between parental substance use and substance use by younger siblings should not be inter- preted to mean that parental behaviour has

no impact on substance use in the family. In addition to substance use behaviour,

there are other within-family processes such as afFectional ties, management prac- tices and the resolution of conflict that

may have important implications for sub- stance use among siblings.14

Study limitations There are a number of study limitations

that the reader should bear in mind when

considering the findings. One, all informa-

tion was collected by self-administered questionnaire completed at home, without opportunity to corroborate responses. Although studies exist in support of the reliability and validity of this method,15 it

is arguable that random error and under- reporting have led to an attenuation of associations. Two, the statistic used to measure concordance on substance use -

kappa - is compressed downward by low base rates.16 Three, although respondents and nonrespondents are comparable on sociodemographic characteristics, net data loss at 39.4% is high, and could include sample loss selective for substance use. Four, it is not possible to establish the chronology of substance use within fami- lies because the study design is cross- sectional. Five, data are not available on

intensity of contact between siblings or peer substance use, variables that would

help to clarify the modifying effects of age

on between-sibling agreement in substance use. And six, the study is restricted to the effects of family composition on substance use. There is no attempt to determine whether or not the socioeconomic circum-

stances of families act as effect moderators

or to model the impact of other important influences on substance use such as family processes and peers.

Identification of risk factors associated

with substance use by adolescents and young adults is important from both pre- ventive and therapeutic viewpoints. Sibling concordance for substance use points to the importance of environmental effects and suggests that the treatment and pre- vention of substance use (and abuse) among adolescents and young adults might be enhanced by including a family focus, especially where there are two or more sib-

lings at home. Empirical support exists for family-based treatments for substance abuse and family-based preventive inter- ventions are under development.17 Evaluation of the effectiveness of these pre-

ventive programs should be a priority for research on interventions.

REFERENCES

1. Adlaf EM, Ivis EJ, Smart RG, Walsh GW. The Ontario Student Drug Use Survey: 1977-1995. Toronto: Addiction Research Foundation of Ontario, 1995.

2. Adlaf EM, Ivis FJ, Smart RG, Walsh GW. Enduring resurgence or statistical blip? Recent trends from the Ontario student drug use survey. Can J Public Health 1996;87:189-92.

3. Johnston LD, O Malley PM, Bachman JG. National Survey Results on Drug Use from Monitoring the Future Study, 1975-1995, Vol. 2, Secondary School Students [NIH Publication No. 96-4139]. Washington, DC: US Government Printing Office, 1996.

4. Poulin C, Elliott D. Alcohol, tobacco and cannabis use among Nova Scotia residents: Implications for prevention and harm reduction. Can Med Assoc J 1 997; 156:1 387-93.

5. Weinberg NZ, Rahdert E, Colliver JD, Glantz MD. Adolescent substance abuse: A review of the

past 10 years. J Am Acad Child Adolesc Psychiatry 1998;37:252-61.

6. Merikangas KR, Stolar M, Stevens DE, et al. Familial transmission of substance use disorders.

Arch Gen Psychiatry 1998;55:973-79. 7. Prescott CA, Kendler KS. Genetic and environ-

mental contributions to alcohol abuse and depen- dence in a population-based sample of male twins. Am J Psychiatry 1999;156:34-40.

8. McGue M, Sharma A, Benson P. Parent and sib- ling influences on adolescent alcohol use and misuse: Evidence from a U.S. adoption cohort. J Stud Alcohol 1 996;57:8- 1 8.

9. Light JM, Irvine KM, Kjerulf L. Estimating genetic and environmental effects of alcohol use and dependence from a national survey: A "quasi- adoption" study. J Stud Alcohol 1996;57:507-20.

10. Gfoerer J. Correlation between drug use by teenagers and drug use by older family members. Am J Drug Alcohol Abuse 1987;13:95-108.

11. Tambs K, Vaglum P. Alcohol consumption in parents and offspring: A study of the family cor- relation in a general population. Acta Psychiatr Scand 1990;82:145-51.

12. Ontario Ministry of Health. Ontario Health Survey 1990: User's Guide Vol. I Documentation. Toronto, Ontario Ministry of Health, 1992.

13. Fleiss JL. Statistical Methods for Rates and Proportions 2nd ed. New York: John Wiley & Sons, 1981.

14. Hawkins JD, Catalano RF, Miller JY. Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: Implications for substance use prevention. Psychol Bull 1 992; 1 1 2:64- 105.

15. Needle R, McCubbin H, Lorence J, Hochhauser M. Reliability and validity of adolescent self- reported drug use in a family-based study: A methodological report. Int J Addictions 1983;18:901-12.

16. Feinstein AR, Cicchetti DV. High agreement but low kappa: I. The problems of two paradoxes. J Clin Epidemiol 1990;43:543-49.

17. Hogue A, Liddle HA. Family-based preventive intervention: An approach to preventing sub- stance use and antisocial behaviour. Am J Orthopsychiatry 1999;69:278-93.

Received: August 22, 2000 Accepted: January 18, 2001

MAY - JUNE 2001 CANADIAN JOURNAL OF PUBLIC HEALTH 209

This content downloaded from 208.95.50.183 on Thu, 03 Aug 2017 05:40:06 UTC All use subject to http://about.jstor.org/terms