Problem gambling patterns among Australian young adults: Associations with prospective risk and protective factors and adult adjustment outcomes

Highlights

• New incidence problem gambling was the most common gambling pattern (3.6%).

• 2.6% of the sample were desistors, and 2.1% were persistent problem gamblers.

• Antisocial peers and alcohol use increased the risk of persistent problem gambling.

• Persistent problem gamblers had the most adjustment problems in adulthood.

Abstract

There is instability in the developmental course of problem gambling [PG] over time; however, studies that examine PG at an aggregate level obscure these variations. The current study employed data from a longitudinal study of Australian young adults to investigate: 1) PG patterns (i.e., resistance, persistence, desistence, and new incidence); 2) prospective risk and protective factors for these patterns; and 3) behavioural outcomes associated with these patterns. A sample of 2261 young adults (55.73% female) from Victoria, Australia, who were part of the International Youth Development Study completed a survey in 2010 (T1, age 21) and 2012 (T2, age 23) measuring PG (two items based on established measures), risk and protective factors, and behavioural outcomes. The majority of the sample (91.69%) were resistors (no PG at T1 and T2), 3.62% were new incidence PG cases, 2.63% were desistors (PG at T1 but not T2), and 2.07% reported persistent PG at T1 and T2. Individual civic activism was protective of new incidence PG, while affiliation with antisocial peers and frequent alcohol use increased the risk of persistence. Persistent problem gamblers also experienced the greatest number of poor behavioural outcomes at T2. New incidence was associated with internalising symptoms at T2, while desistance was not associated with any behavioural outcomes. In conclusion, each PG pattern was associated with different predictors and outcomes, highlighting the need to consider variation in the course of young adult PG in order to provide efficacious prevention and intervention approaches, and to protect against relapse.

Introduction

There is growing recognition of the varied course of problem or disordered gambling over time (LaPlante, Nelson, LaBrie, et al., 2008). Contrary to previously held beliefs that disordered gambling is a chronic and progressive condition, recent longitudinal studies have shown that there are patterns of change in individual gambling behaviour across time (e.g. LaBrie, R. A., et al., 2003, LaPlante, D. A., et al., 2008), despite overall rates of problem gambling [PG] remaining stable. These findings emphasize the importance of examining the gambling behaviour trajectories (longitudinal patterns) of individuals over time (Bray, Lee, Liu, et al., 2014). While there are a growing number of prospective longitudinal studies investigating predictors of PG, these studies largely investigate PG at an aggregate level (variable centred approach) without regard for the documented variability in individual PG patterns over time (e.g. Billi, R., et al., 2014, Hayatbakhsh, M. R., et al., 2006, Scholes-Balog, K. E., et al., 2014, Scholes-Balog, K. E., et al., 2015, Slutske, W. S., et al., 2005). Further, there is significant diversity in the types of potential predictors investigated, with these studies largely focused on individual level predictors of PG selected based on cross-sectional correlates of PG and other problem behaviours. Accordingly, these studies have yielded diverse findings. Studies such as these that investigate PG trends and predictors at an aggregate level may obscure sub-group differences in PG patterns which are identifiable with person-centred approaches (Welte, J. W., et al., 2004, Winters, K. C., et al., 2005), and which may have potentially important implications for intervention and prevention of relapse (LaPlante et al., 2008).

Different sub-groups of problem gamblers have been identified based on the course of the disorder over time (LaPlante et al., 2008). For example, Winters et al. (2005) studied a non-representative community sample of 305 adolescents in the United States (U.S.) at ages 16, 17 and 23 years. They found that the most prevalent pattern was resistance (no PG at any time-point; 60%), followed by new incidence (developed PG during the study; 21%), desistance (moved away from PG the study; 13%), and persistence (consistent PG throughout the study; 4%). A large representative study of Swedish adolescents and young adults found much smaller proportions for non-resistant pathways: 2.64% were desistors; 2.26% were new incidence; and 0.67% were persistors (LaBrie et al., 2003). Cultural, sample size and composition, and measurement differences may have resulted in discrepancies in the prevalence of the sub-groups in these two studies.

Nonetheless, while there are a growing number of studies investigating these different patterns of PG over time, there remains a lack of research investigating their behavioural, social, and contextual determinants (LaPlante et al., 2008). Understanding the factors that influence PG progression, and in particular, those associated with persistence of, or desistance from PG, will serve as a foundation for the development of efficacious prevention and intervention efforts (LaPlante et al., 2008). For example, understanding what makes an individual more or less likely (i.e., risk and protective factors) to show a pattern of desistance would also allow health care providers to better tailor treatment plans to help bring about positive changes and prevent relapse (LaPlante et al., 2008).

Similarly, to our knowledge there have been no published studies investigating outcomes associated with the different patterns of PG across time. For example, it is unclear whether desistance from PG is associated with lasting poor social and/or behavioural outcomes, such as engagement in other problem behaviours, lack of employment, and mental health problems, despite an absence of gambling problems. Such findings would have important ramifications for prevention of relapse and provision of appropriate support services following recovery from PG.

This exploratory study analysed data from a large prospective study of Australian youth to investigate temporal changes in young adult gambling behaviour and examine: (1) the prevalence of sub-groups with similar patterns of PG behaviour over time (defined as resistance, persistence, desistance, and new incidence; Winters et al., 2005); (2) social developmental and behavioural risk and protective factors for these PG patterns; and (3) young adult behavioural outcomes associated with these PG patterns. The social developmental risk and protective constructs were selected based on their established associations with PG (Scholes-Balog et al., 2014) and other problem behaviours (e.g., Hemphill, Heerde, Herrenkohl, et al., 2011), and were drawn from the Communities That Care (CTC) youth survey. The broad CTC framework is designed to comprehensively assess risk and protective factors across the social environmental domains that are considered to be influential in the development of youth behaviour (community, family, school, peer/individual) (Catalano & Hawkins, 1996) so as to facilitate community needs assessment for prevention and/or intervention (Arthur, Briney, Hawkins, et al., 2007).

It was hypothesised that the new incidence problem gambling sub-group would be the most prevalent problem gambling group identified, and the persistent problem gambling sub-group would be the least prevalent. Further, persistent problem gamblers were predicted to experience the greatest number of poor behavioural outcomes in young adulthood.