Father-Child Similarities in Patterns and Onset of Polysubstance Use during Adolescence

Introduction: Early onset substance use predicts problems in adulthood, and may threaten the well-being of future offspring. Whereas substance-specific familial risks to adolescents are important (e.g., parent to child tobacco use), polysubstance use is common, particularly for substances other than alcohol and for youth showing early onset. Studies typically ignore or are unable to examine use of multiple substances. Presently we identify heterogeneity both in the developmental timing of adolescents’ first use and in the types of substances they use. We then test whether adolescents’ polysubstance use onset patterns predict similar patterns in their own children.

Methods: At-risk boys [generation two (G2) of a 3-generational sample] were assessed prospectively from ages 10 to 45 years. G2 ages at which each of four substances (alcohol, tobacco, marijuana & other drugs; ATMO) were first used (“age of onset”) were derived from annual reports across ages 10-19 years. G2 men’s G3 offspring participated in another study; G3 ages of ATMO onset were determined from four assessments across the same age span. Preliminary data are from 146 G3 (nested in 93 G2 fathers) who were old enough to have participated. By May 2019, sample sizes will be >200 for G3 and >105 for G2.

Results: On average G3 (run separately by gender) showed later onset for alcohol and tobacco than their fathers (p < .001). Discrete time survival mixture analyses identified three adolescent polysubstance use onset classes for G2 and two for G3. G2 classes were: Early-ATMO (14%; all substances by age 13-15), Later-AT (52%; primarily alcohol and tobacco, with onset by age 17-19), or Later-ATMO (34%; similar onset, but including marijuana and often other drugs). G3 classes were Middle-ATMO (13%; all substances by age 17-19 and all but other drugs by 15-17) or Later-ATM (87%; no other drug use, and onset on other substances across ages 15-19). G3 and G2 onset classes were significantly associated, controlling for G3 gender and G2 age when G3 was born. Specifically, G3 were more likely to be in the Middle-ATMO class if their father was in the G2 Early-ATMO than in the G2 Later-AT class. Increased sample size may enhance discernment of heterogeneity in G3, and whether it is the timing versus the pattern of their fathers’ adolescent polysubstance use that confers greatest risk.

Conclusions: The timing and nature of polysubstance use onset are not wholly confounded in adolescence, and such patterns may affect risk for polysubstance use in the next generation. Differential risk suggests further nuance to case identification and prevention may be possible