Factors influencing the relationship between alcohol consumption and risky sexual behaviour among young people: A systematic review

Abstract Within the literature, there is a well-established relationship between alcohol use and risky sexual behaviour. Three different but related explanations have been identified to account for the relationship, namely social, biological and individual. Although these explanations and the associated factors have been well explored in the literature, there is currently no empirical initiative that has shown how these factors interact with each other within the alcohol use and risky sexual behaviour relationship. The aim of the systematic review is to review and synthesise existing literature on the relationship between alcohol use and risky sexual behaviour among young adults (18–24 years old). Seventy-one articles were included in the review subsequent to a systematic search of the literature. The review highlighted three thematic domains relating to personality influences, social determinants and interpersonal factors, contributing to a more comprehensive understanding of these factors. At a foundation level, more research is required to gain new insights, discover new ideas and/or increase knowledge of a phenomenon, i.e. factors influencing the relationship between alcohol use and risky sexual behaviour among young adults.


PUBLIC INTEREST STATEMENT
There is a substantial body of evidence documenting the relationship between alcohol use and risky sexual behaviour. However, there is a dearth of empirical evidence highlighting the nature and complexity of this relationship, especially as it relates to intensity and frequency of alcohol use and contextual variables such as socio-economic status as well as gender and age. Similarly, the links between motivations for alcohol use and sexual behaviours are not well understood. Furthermore, it is noteworthy that the alcohol use-risky sexual behaviour link has not been adequately explored in young adults. The current study reviewed the literature available since 2000 relating to factors contributing to why young adults engage in alcohol use and risky sexual behaviour. A systematic review of global research was conducted. Results indicated that personality influences, social determinants and interpersonal factors were related to alcohol use and risky sexual behaviour.

Introduction
There is a substantial body of evidence documenting the relationship between alcohol use and risky sexual behaviour (see Adams et al., 2014;Campbell, Williams, & Gilgen, 2002;Fritz et al., 2002;Mbulaiteye et al., 2000;Weiser et al., 2006). While the literature indicates a relationship between alcohol use and risky sexual behaviour, there is little empirical evidence highlighting the nature and complexity of this relationship, especially as it relates to intensity and frequency of alcohol use and contextual variables such as socio-economic status as well as gender and age (Cooper & Orcutt, 2000;Leigh & Stall, 1993). Similarly, the links between motivations for alcohol use and sexual behaviours are not well understood (Patrick & Maggs, 2010). It is noteworthy that high-risk sexual behaviours and the resulting health effects disproportionately affect young people.
Evidence from a 2010 National Survey on Drug Use and Health conducted among young people, predominantly urban youth located in New York City, indicated that alcohol use was highest among young adults from 21 to 25 years of age with 45.5% of these individuals reporting binge drinking and 18% reporting heavy drinking (binge drinking on five or more days) in the past month (Griffin, Scheler, Acevedo, Grenard, & Botvin, 2012). Furthermore, in a study conducted in the US, approximately 50% of all new sexually transmitted infections (STIs) in the US occurred among 15-24 year olds, and almost 50% of all new HIV infections in the US occurred in young people aged 24 years or under (CDC, 2003, Griffin et al., 2012. Brown and Vanable (2007) also found that alcohol use prior to a sexual encounter was strongly correlated to unprotected sexual intercourse encounters involving casual partners. Important to note previous research has focused on adolescents or youth in formal educational settings (Pithey & Morojele, 2002), in other words those attending school, college and/or university. Additionally, it is important at this point to note the definition of both adolescents and young adults.
According to the APA (2002), the most commonly used chronologic definition of adolescence includes ages 10-18; nevertheless, it could include 9-26 years depending on the source. "The current lack of consensus of an operational definition of adolescent chronology can be attributed to a number of factors, including: the appreciated continuity of human development; a recognition of individual, cultural, gender and racial variability; the ascribed relative salience of specific developmental milestones, and a perpetually refined science of human development in a dynamically evolving society" (Curtis, 2015, p. 9). Curtis (2015) proposes an operational definition of adolescence based in developmental science that includes ages from 11 to 25 years. In this definition, "early adolescence" and "young adulthood" are substages of this critical transitional period. With early adolescence ranging from 11 to 13 years, adolescence from 14 to 17 years and young adulthood from 18 to 24 years old. For the purpose of this study, Curtis' (2015) operational definition will be taken into account when referring to young adults.
The relationship between alcohol use and risky sexual behaviour is complex and cannot be explained by a single mechanism. This relationship instead reflects multiple underlying causal and non-causal processes (Cooper, 2006). While there is a clear relationship between alcohol use and risky sexual behaviour, research identifies them as two separate variables interacting at a specific given point. The use of alcohol has globally been identified as the third leading risk factor for poor health (World Health Organisation (WHO), 2010). The deleterious effects of alcohol result in an estimated 2.5 million deaths every year, of which a significant proportion occur amongst young people (WHO, 2010). Previous research has shown that a range of alcohol-related problems, including poor class attendance, hangovers, trouble with authority, injuries and even fatalities, are commonly experienced by young adults engaging in heavy drinking (see Foster, Neighbors, & Young, 2013;Hingson, 2010;Hingson, Heeren, Winter, & Wechsler, 2005). Despite these findings, heavy alcohol consumption (AC) among young adults remains normative and is increasing in prevalence within university populations (Stewart & Devine, 2000). Stewart and Devine further state that studies have shown that about one-third of undergraduate students drink at a level that produces acute physical, psychological, social and academic problems (e.g. hangovers, lowered self-esteem, sexual misconduct, missing classes). Crawford and Novak (2007) reported that students who perceive heavy drinking as a common activity at school are more likely to increase their levels of AC in order to gain social acceptance and avoid negative peer evaluations. Despite having specific policies designed to reduce students' levels of consumption, binge drinking remains a frequent recurrence and is escalating (Crawford & Novak, 2007).
Previous research (see Baer, 1994;Critchlow, 1987;Darkes & Goldman, 1993;Evans & Dunn, 1995;Stacy, Widaman, & Marlatt, 1990) has identified a number of specific factors associated with heavy drinking, including demographic characteristics (gender and fraternity/sorority membership); descriptive and injunctive social norms; enhancement, social, coping and conformity drinking motives, expectancies and tendencies; and subjective evaluations of positive and negative alcohol effects. Surprisingly little research has evaluated the relative contribution of different factors in predicting AC and related problems as well as the direct impact of these constructs on binge-drinking consequences in a systematic manner (Neighbors, Lee, Lewis, Fossos, & Larimer, 2007;Turrisi, Wiersma, & Hughes, 2000). Hall, Holmqvist, Simon, and Sherry (2004) define risky sexual behaviour in terms of the behaviour itself, as well as the nature of the relationship between partners. Simply put, it can take several forms, ranging from acquiring a large number of sexual partners, to engaging in risky sexual activities and sexual intercourse under the influence of substances such as cocaine or alcohol (Hall et al., 2004). Risky sexual behaviour is identified as the second highest risk factor for harm in high mortality developing countries and constitutes 10.2% of the global burden of disease (Rehm & Room, 2003, as cited in Morojele et al., 2006). Globally there are an estimated 357 million new cases of STIs each year (WHO, 2016), with the highest rates among 20-24 year olds (Karl & Gabriele, 2005). Young adults are vulnerable for a number of reasons: their tendency to have multiple sexual partners (concurrent or sequential) as well as difficulty accessing effective STD prevention services (Charnigo et al., 2013); unprotected sexual intercourse (Gullete & Lyons, 2006;Khasakhala & Mturi, 2008); social, economic and contextual settings (Khasakhala & Mturi, 2008); sexual debut for most people occurs during their teenage years (Khasakhala & Mturi, 2008); the urge to have sex and curiosity play a role in risky behaviour but to name a few (Caron, Davis, Wynn, & Roberts, 1992;Keeling, 1995;Mickler, 1993;Okafor & Obi, 2005;Sells & Blum, 1996). Simultaneously, there has been a marked increase in the unwanted personal and social consequences associated with these behaviours such as higher prevalence of STDs, unintended pregnancies, school dropouts and heightened demands on the health and human service agencies (Langer, Warheit, & McDonald, 2001) As reported by Langer et al. (2001), these negative social and personal consequences associated with these changes in sexual attitudes and behaviours have attracted the interest of researchers and scholars from a variety of disciplines. Researchers collectively have examined the impact of biological and psychological predispositions as well as elements derived from our social and cultural environments (Langer et al., 2001). These include psychosocial variables (e.g. norms, attitudes and self-efficacy) as well as personality traits such as sensation seeking and impulsivity (Charnigo et al., 2013;Hoyle, Fejfar, & Miller, 2000;Noar, Zimmerman, Palmgreen, Lustria, & Horosewski, 2006).
As Cooper (2002) noted, "the relationship between Alcohol use and risky sexual behaviour appears to be both complex and highly circumscribed" (p. 115), varying with characteristics of the individual drinker and the sexual situation. Although some multiple-event studies have supported the hypothesis that AC increases the likelihood of engaging in risky sexual behaviour, the findings are mixed (Cooper, Peirce, & Huselid, 1994;Graves & Hines, 1997;Morrison et al., 2003;Testa & Collins, 1997). Some authors point to an interplay between individual and environmental factors reflecting a cyclical and mutual dynamic of the individual influencing or being influenced by their environment (Choquet, 2004;Clapp, Segars, & Voas, 2002DeJong et al., 2006;Harford, 1979;Jessor, 1998;Wechsler et al., 2002). Other authors have found stronger results when the outcome measure was number of casual sexual partners than when it was frequency of condom use (Cooper, 2002). Graves and Hines (1997) found that AC was more common in sexual events that involved partners known for a short period of time; however, results regarding the relationship between AC and condom use were inconsistent. The interrelationships between partner type, intoxication and condom use make it difficult to disentangle alcohol's role in unprotected sex (Abby, Parkhill, Buck, & Saenz, 2007). Although alcohol researchers focus on alcohol's role in risky sexual behaviour, many theories of health behaviour have been applied to sexual risk taking and STD and HIV prevention (Albarracin et al., 2005).
Prior literature focused on adolescents, while less is known about how associations between substance use and risky sexual behaviour may change across young adulthood. The transition to adulthood is marked by dramatic increases in freedoms and responsibilities that occur at the same time that an individual's ability to self-regulate is still emerging (King, Nguyen, Kosterman, Bailey, & Hawkins, 2012). However, how the associations between substance use disorders and risky sexual behaviours unfold across young adulthood remains unclear. (King et al., 2012). Despite the strong link between alcohol use and risky sexual behaviour, researchers have not adequately explored this relationship among young adults. The studies that have been conducted focused primarily on adolescents in formal educational settings (Cooper, 2002;Flisher et al., 1996cited in Pithey & Morojele, 2002, clinic-based populations receiving treatment for STIs and HIV (Kalichman, Simbayi, Cain, & Jooste, 2007). It is important to note that the majority of studies that have linked alcohol use and risky sexual behaviour have been carried out in the developed world and among high school or college students (Khasakhala & Mturi, 2008). In a study conducted by Griffin et al. (2012), they found that these behaviours peak during the early to mid-twenties as young people live more independently and autonomously from their family of origin, enjoy new freedoms such as legal drinking and the ability to enter bars and nightclubs, and have increased opportunities for sexual and romantic relationships. With statistics suggesting that the age group of 18-25 years is most vulnerable in terms of abusing alcohol (South African Community Epidemiology Network on Drug Use [SACENDU], 2011), it is imperative to explore the link between alcohol use and risky sexual behaviour in this cohort. This would contribute to a more comprehensive understanding of the relationship between alcohol use and risky sexual behaviour as suggested by Adams et al. (2014).

Rationale
Within the literature, there is a well-established relationship between alcohol use and risky sexual behaviour (Adams et al., 2014;Campbell et al., 2002;Fritz et al., 2002;Mbulaiteye et al., 2000;Weiser et al., 2006). This relationship has emerged as a major health concern especially among young adults between the ages of 18-25 (Morojele et al., 2004;SACENDU, 2011). Freeman and Parry (2006) have identified three different but related explanations to account for the relationship, namely social, biological and individual. Although these explanations and the associated factors have been well explored in the literature, there is currently no empirical initiative that has shown how these factors holistically interact with each other within the alcohol use and risky sexual behaviour relationship. The current study hopes to contribute in this regard by systematically reviewing the literature with the ultimate aim of informing future studies. It is anticipated that this would contribute to a more comprehensive understanding of the factors influencing the relationship between alcohol use and risky sexual behaviour among young adults and better inform intervention strategies.

Aim
The aim of the systematic review is to review and synthesise existing literature on the factor that influence the relationship between alcohol use and risky sexual behaviour among young adults (18-24 years old).

Method
We performed a systematic review of published studies since 2000 to identify the factors influencing the relationship between alcohol use and risky sexual behaviour among young people (age 18-24 years). The findings of this review will provide evidence-based knowledge critical for addressing the aim and specific objectives.

Review question
• What factors influence the relationship between alcohol use and risky sexual behaviour among young adults?

Article search
Various resources, published in English from the year 2000 to present, were consulted for the review. Based on the nature of the topic, ScienceDirect and EBSCOhost were the main sources for the search. Databases within EBSCOhost included PsycARTICLES, SocINDEX, and Academic Search Complete. Once the articles had been appraised, manual searching of reference lists then took place. An initial review of relevant literature was conducted to identify key studies in the field. The following keywords formed part of the initial search within the above-mentioned databases: prevalence, alcohol consumption, risky sexual behaviour, alcohol use and risky sexual behaviour, participants, risk and protective factors, extraneous factors, at-risk behaviour and risk-taking behaviour. Within the searches that were initially conducted, it was found that the initial keywords did not yield the expected number of studies pertaining to the topic or no results were yielded. It was therefore decided to use the following terms: alcohol consumption, risky sexual behaviour, young people, young adults, risk and protective factors, contributing factors and influences.

Inclusion/exclusion criteria
Included studies were original English-language research articles published in peer-reviewed literature which reported on both quantitative and qualitative studies that focused on the factors impacting on the relationship between alcohol use and risky sexual behaviour among young people. The review did not attempt to categorise studies in terms of these two methods of research, but instead to provide a comprehensive picture of studies exploring the factors contributing to alcohol use and risky sexual behaviour. The review included both males and females between the ages of 18-24 years. Any racial, ethnic, cultural or religious groups were eligible for inclusion, regardless of geographic region. Articles pertaining to sexual violence, sexual coercion, intimate partner violence, HIV/AIDS and alcohol use disorders were not included in the review as there has been a paucity of research conducted to adequately addressed the extraneous factors contributing to the relationship between alcohol use and risky sexual behaviour (Morojele et al., 2006;Scott-Sheldon et al., 2012). Furthermore, studies focusing on age groups other than young adults and studies focusing on substance use in general were also excluded from the review.

Quality assessment
Study quality assessment is relevant to every step of a review. The assessment is crucial in evaluating the strengths, weaknesses and benefits of the assumptions and conclusions made in the study, as well as exploring heterogeneity and informing decisions regarding suitability of meta-analysis. In addition, they help assess the strength of inferences and make recommendations for future research. The selection of studies to be included in the review was thus assessed by utilising an adapted version of the Evaluation Tool for Qualitative Studies and The Evaluation Tool for Quantitative Research Studies (Long, Godfrey, Randall, Brettle, & Grant, 2002;Long & Godfrey, 2004).

Data extraction
Using the template in Table 1 (Appendix), data was extracted at various stages, namely assessment of eligibility, assessment of quality, assessment of study characteristics and extraction of study findings. The table utilised for data extraction has been formatted to extract data specifically relevant to the research question, which includes the study authors, aim, sample size, participant characteristics, research design, outcomes, themes and selfconcept domain.

Data synthesis
Data was summarised by means of tabulation of study characteristics, quality and effects as well as statistically if they are sufficiently similar and if they are of adequate quality to explore differences between studies and combining their effects (meta-analysis). The Textual Narrative Analysis approach as proposed by Lucas, Arai, Baird, and Roberts (2007) was utilised. This method synthesises the studies used for review by study characteristics, context, quality and findings that are reported on. Structured summaries were then developed, expanding on and illuminating the context of the extracted data. This method was utilised as this approach typically groups studies into more homogenous groups, synthesising different types of research evidence and making explicit the diversity in study designs and contexts (Lucas et al., 2007).

Results
The aim of the systematic review is to review and synthesise existing literature on the factors that impact on the relationship between alcohol use and risky sexual behaviour among young adults (18-24 years old). The results of the systematic review are structured as follows: type of research, scope and focus of research, context of the studies, quality (content and method of the studies), age cohort, theory and findings, as proposed by Lucas et al. (2007) in the Textual Narrative Analysis. A detailed summary of the final search procedure will be outlined below followed by a comprehensive explanation of the findings.

Article search procedure
The keyword search yielded a total of 4,149 articles with 196 titles identified as relevant and were included for the abstract appraisal. The 147 articles that were excluded did not meet the inclusion criteria as they focused on other substances, HIV/AIDS-related cases, intimate partner violence and other age groups. Of the 196 abstracts, 29 were included for full-text appraisal. After a fulltext appraisal, only 23 articles were included in the review. Based on reference list mining of the included articles, an additional 28 articles were identified as relevant for full-text appraisal. This process is outlined in Figure 1.

Type of research
The 51 articles included in the review can broadly be categorised into two types, that is, empirical work and review articles on alcohol use and risky sexual behaviour. Forty-five articles were classified as empirical articles utilising both qualitative and quantitative methods, as well as longitudinal studies with various age cohorts of young people ranging from children in school, adolescents in high school, college and university students, employed and unemployed youth, drinkers and non-drinkers. Additionally, six of the articles were review articles, consisting of literature reviews and systematic reviews focusing on various aspects of the relationship between alcohol use and risky sexual behaviour, such as the nature of the relationship, the motivations for engaging, and risk and protective factors.

Scope and focus of research
The research focus of the 45 studies in the review can generally be divided into three thematic domains, namely (1) Personality influences on alcohol use and risky sexual behaviour.
(2) Social determinants of alcohol use and risky sexual behaviour.
(3) Interpersonal factors related to alcohol use and risky sexual behaviour (see Table 2).
These domains will be discussed in greater detail below.

Personality influences on alcohol use and risky sexual behaviour
Studies in this domain focused on how personality traits influence their motivation to engage in risky alcohol use, risky sexual behaviour and/or the combination of the two. More specifically the articles focused on concepts like, self-efficacy, self-regulation, impulsivity, selfesteem, self-awareness, self-consciousness, sensation seeking but to name a few. This domain included 14 studies. Using Textual Narrative Analysis as a guide, it is discussed specifically in terms of (1) content, (2) age cohort, (3) context, (4) method, and (5) theory.
3.2.1.1. Content. The content and specific focus of the research studies in this domain ranged from elucidating the joint contribution of sensation seeking and impulsivity to decision-making to risk behaviours (Charnigo et al., 2013), evaluating self-consciousness, self-awareness, drinking identity, sensation seeking, self-esteem, as a moderator or contributor to risky alcohol use and subsequent risky sexual behaviour (Gullete & Lyons, 2006;Miller et al., 2003). Studies (Quinn & Fromme, 2010) also focused on self-regulation as buffering risk associated with alcohol use and as a protective factor to heavy drinking and unprotected sex. Other studies focused specifically on gender differences (Morojele et al., 2006), individual competencies (Stuewig et al., 2015), distinct facets of impulsivity and its contribution to alcohol use outcomes (Shin, Hong, & Jeon, 2012), examining restraint and temptation (Rinker & Neighbors, 2013), and drinking refusal self-efficacy (DRSE) (Oei & Morawska, 2004).
3.2.1.2. Age cohort. The review of literature showed that studies in this research area focused predominantly on young adults between the ages of 17 and 26 years. Two longitudinal studies focused on a specific cohort at two points in their lives, one focused on a mean age of 14.6 years and the second point was at a mean age of 22.8 (Griffin et al., 2012), the second study's first point was at age 10 − 12 years and the second point was at 18 − 21 years (Stuewig et al., 2015).
3.1.2.3. Context. The context of the research studies which focused on alcohol use and risky sexual behaviour among young adults were conducted primarily in various states in the US, specifically various states in the US. Studies were predominantly conducted amongst university students in both the private and public sectors. Studies were also conducted in both low-income communities and middle-income communities.
3.1.2.4. Method. The majority of the studies utilised quantitative methods (n = 9 of 10) with most utilising self-reported questionnaires or computer or online-based surveys. Three studies were longitudinal studies that each had two contact points with individuals that reported on how behaviours at point one could be possible predictors for behaviour at point two (Griffin et al., 2012;Quinn & Fromme, 2010;Stuewig et al., 2015). All studies (n = 9) were generally descriptive studies that utilised correlation analysis and one utilising structural equation model (Murry, Simons, Simons, & Gibbons, 2013). The one review study focused on key constructs of alcohol expectancies (AEs) and DRSE to explain the acquisition and maintenance of binge drinking.
3.1.2.5. Theory. There were only 5 of the 10 studies which used a theoretical framework to guide their study. Four of the eight studies utilised the Five-Factor Model of Personality (Charnigo et al., 2013;Miller et al., 2003;Shin et al., 2012;Stewart & Devine, 2000) to synthesise and interpret the findings of their study. Other studies utilised Transtheoretcal Model of Behavioral Change (Gullete & Lyons, 2006) and Alcohol Expectancy Theory (Oei & Morawska, 2004). One study utilised the theory behind self-regulation and sensation seeking to synthesise and explain the findings (Table  3) (Quinn & Fromme, 2010).

Social determinants of alcohol use and risky sexual behaviour
The studies in this domain explored the social, contextual and environmental factors impacting on the relationship between alcohol use and risky sexual behaviour. More specifically the articles focusing on social aspects including social norms, motives assumptions and beliefs in the social context. Contextual aspects focused on the context in which alcohol use and risky sexual behaviour takes place. Lastly environmental aspects focus on the historical construction of communities.
3.2.2.1. Content. The content and specific focus of the research studies in this domain consisted of an array of focal areas including assumptions of environmental approaches to alcohol use (Clapp et al., 2002), normative perceptions (Lewis, Patrick, Mittmann, & Kaysen, 2014), social norms, demographics, drinking motives and AEs in predicting AC and related problems (Neighbors et al., 2007), the nature of the relation among drinking beliefs, drinking tendencies, perceived sexual control (Walsh et al., 2013), behavioural consequences (Turrisi et al., 2000) and effects of alcohol use on condom use (Abbey, Saenz, & Buck, 2005). Further studies explore alcohol use and risky sexual behaviour (Abby et al., 2007;Orchowski & Barnett, 2012;Seth, Wingood, DiClemente, & Robinson, 2011). Contextual aspects include the extent to which Spring Break drinking and sexual behaviours are related (Patrick, 2013), rates of risky sexual behaviour among women bar drinkers (Parks, Hsieh, Collins, Levonyan-Radloff, & King, 2008) and reducing sexual risk behaviours among university students in particular (Connor, Psutka, Cousins, Gray, & Kypri, 2013). Further studies focus on identifying patterns of alcohol use behaviours and AEs among women (Stappenbeck et al., 2013), the relationship between binge drinking, "reflection impulsivity" alcohol-related expectancies and unplanned sexual behaviour in a sample of young social drinkers (Townsend et al., 2011). Environmental aspects include studies focusing on the extent of the current alcohol problem as well as a historical perspective (Vicary & Karshin, 2002), how alcohol misuse increases the occurrence of sexual risk behaviour in South African communities (Pithey & Morojele, 2002) and describing an example of the use of latent variable modelling to create measures of complex phenotypes and environments that illustrate the utility of the general versus specific conceptualisation (Bailey, Hill, Meacham, Young, & Hawkins, 2011).

Age cohort.
The review of literature showed that studies in this research area focused predominantly on young adults between the ages of 17 and 35 years. One longitudinal study focused on fifth graders at its first point of contact and adults were later retained at 24 years old (Bailey et al., 2011).

3.2.2.3.
Context. The context of the research studies focused young adults in relation to social, contextual and environmental aspects of alcohol use and risky sexual behaviour primarily in countries outside of South Africa, specifically various states in the US. Studies were predominantly conducted in schools, colleges and universities in both the private and public sectors. Studies were also conducted in both low-income communities and middle-income communities. Two studies were conducted in South Africa, one in an impoverished community on the Cape Flats (Abby et al., 2007) and the other in broader South African communities (Pithey & Morojele, 2002).

Method. It was found that most studies in this domain utilised quantitative methods (n = 16).
Various designs and data analysis techniques were used in these studies. The research designs used included correlational designs, cross-sectional and longitudinal designs. Many of these quantitative studies used independent and correlational techniques such as multiple, hierarchical and logistic regression (Abbey et al., 2005;Abby et al., 2007;Bogg & Finn, 2009;Connor et al., 2013;Lewis et al., 2014;Neighbors et al., 2007;Orchowski & Barnett, 2012;Parks et al., 2008;Seth et al., 2011;Turrisi et al., 2000). Two of the studies made use of latent analysis (Stappenbeck et al., 2013;Townsend et al., 2011), and one study made use of path modelling procedures (Walsh et al., 2013). Three of the articles in this domain were review articles (Pithey & Morojele, 2002;Vicary & Karshin, 2002).

Theory.
There were only three studies which used a theoretical framework to guide their study. These studies all utilised different theories to guide their study. Theories considered problem behaviour theory (Murry et al., 2013) and traumagenic dynamics theory (Walsh et al., 2013).  Another study utilised a conceptual model, more specifically family and economic stress theories and problem behaviour theory (Murry et al., 2013). Lastly, Oei and Jardim (2007) utilised the theory behind AEs and drink refusal self-efficacy to shape and guide their study (Table 4).

Interpersonal factors related to alcohol use and risky sexual behaviour
The studies in this domain focused on the family, peers, risk and protective factors, and personal attributes. The articles that focused broadly on the family and, in particular, parental and family influence and parental and family beliefs. Peer aspects focused on the relationship between young adults and their peers. These factors included perceived care of friends, role modelling, and perceived alcohol use and sexual behaviour of friends. Personal attributes refer to attitudes, race, gender, genetics and religiosity. Lastly, risk and protective factors also emerged in this domain. Using Textual Narrative Analysis as a guide, it is discussed specifically in terms of content, age cohort, context and method, and theory.
3.2.3.1. Content. The content and specific focus of the research studies in this domain broadly considered the relationship between alcohol use and risky sexual behaviour to more individual, family and peer aspects. More specifically, exploring the relationship between alcohol use and risky sexual behaviour as a phenomenon (Adams et al., 2014;Cooper, 2002Cooper, , 2006, associations across motivations for engaging in alcohol use and risky sexual behaviour (Patrick & Maggs, 2010), to illuminating our understanding of sexual behaviour (Marston & King, 2006), risk and protective factors (Kogan et al., 2010;Langer et al., 2001;Voisin, Hotton, Tan, & DiClemete, 2013), and protective behavioural strategies (LaBrie, Lac, Kenney, & Mirza, 2011). Individual aspects such as gender and psychological risk and protective factors (Park & Grant, 2005), cognitive and affective attitudes, personal normative beliefs, social determinants and expectations (Sonmez et al., 2006) and the impact of partner type (Brown & Vanable, 2007). Family relationships focused on parental mediation and critical thinking (Radanielina-Hita, 2015). Peer relationships, with specific reference social relationships (Townshend, Kambouropoulos, Griffin, Hunt, & Milani, 2014), perceived awareness and caring, or know or care about student's behaviour (Wetherill, Neal, & Fromme, 2010).

Age cohort.
The review of literature showed that studies in this research area focused predominantly on young adults between the ages of 15 and 35 years. One mixed-method study focused on 18-62 year olds (Townshend et al., 2014).

3.2.3.3.
Context. The context of the research studies focused on young adults in relation to family, peers and individual aspects relating to alcohol use and risky sexual behaviour primarily in countries outside of South Africa. Studies were predominantly conducted in colleges and universities in both the private and public sectors. Studies were also conducted in both low-income communities and middle-income communities. Two studies were conducted in South Africa, one in an impoverished community on the Cape Flats (Adams et al., 2014) and the other in one of the poorest suburbs in South Africa, average household income is less than R1500 pm and unemployment is high (Townshend et al., 2014). Four reviews were conducted within this domain focusing on published articles worldwide (Cooper, 2002(Cooper, , 2006Marston & King, 2006).

Discussion
While the relationship between alcohol use and risky sexual behaviour is well-established in the literature (see Adams et al., 2014;Campbell et al., 2002;Fritz et al., 2002;Mbulaiteye et al., 2000;Weiser et al., 2006), there are few empirical studies that demonstrate and unpack the nature of this relationship (see Cook & Clark, 2005;Morojele et al., 2006;Muchimba, Haberstick, Corley, & McQueen, 2013;Scott-Sheldon et al., 2012). Research shows that young adults engage in heavy drinking and experience a range of alcohol-related problems, including poor class attendance, hangovers and trouble with authorities, injuries and even fatalities Hingson, 2010;Hingson et al., 2005;Wechsler, Davenport, Dowdall, & Moeykens, 1994;Wechsler, Lee, Kuo, & Lee, 2000). The aim of the systematic review is to review and synthesise previous literature to discuss factors identified in the literature that influence the relationship between alcohol use and risky sexual behaviour. Thus, the articles in the current systematic review focused on literature highlighting these factors. The key factors influencing the relationship between alcohol use and risky sexual behaviour that emerged from the review were personality incluences, social determinants and interpersonal factors. Yet, determining alcohol's precise role in sexual risk taking has proven to be difficult. Past research has produced mixed results, depending on characteristics of individuals, their partners and the situation, as well as how the link between alcohol use and sexual behaviour was assessed (Abby et al., 2007). Patrick and Maggs (2010) suggested that the links between motivations for alcohol use and sexual behaviours are not well understood. However, within the literature the contributing factors for each overlap significantly, strengthening the relationship between the two concepts. While personality influences such as self-efficacy, self-regulation, self-awareness, sensation-seeking and so forth are important to understanding and determining drinking patterns, it has been suggested that dosage is an important aspect of alcohol use which is largely under the individuals' control. However, there is evidence that frequency of drinking occasions may be greatly influenced by social factors (Vogel-Sprott, 1974). For example, university or college students with high levels of sensation seeking may be at especially high risk to begin or escalate heavy drinking. However, articles reviewed to a large extent employed a cross-sectional research design, thus making it impossible to examine changes over time and, therefore, difficult to draw solid conclusions. The major implication for the studies that utilised longitudinal designs is the age periods at which data was collected. Collecting data before adolescence and then after does not make allowance for the period in between. Future longitudinal examination of the aetiology of alcohol use will allow for better understanding of the window of influence for specific risk and protective factors (Rutledge & Sher, 2001), and using multilevel modelling approaches would allow researchers to evaluate several predictors at once in order to determine their individual and combined effects (Kraemer et al., 2001).
Studies exploring the social, contextual and environmental factors conducted research on very similar populations, namely college students or university students predominantly from First World countries (the US, New Zealand and Australia). While only 4 out of 20 studies under this theme focused on developing countries (South Africa and Kenya). Therefore, the above results may be limited to the specific context of North American, New Zealand and Australian young adults' patterns of risky behaviour. Literature reviewed found that social norms, social context and social beliefs, as well as the broader environment, contribute to alcohol use and risky sexual behaviour. Thus, studies from other countries are needed in order to acquire knowledge on the significance of the cultural influences of drinking motives.
Literature reviewed focusing on factors relating interpersonal aspects concentrated on how the individual interacts with various other parts of their lives such as the personal attributes (age, race, gender, biological make up and religiosity), family, peers, romantic relationships, and broader risk and protective factors. The studies additionally highlight the impact of one's sociopolitical and historical situation shape decisions or lifestyles today (Cooper, 2006). Cooper (2006) further states that the belief that alcohol causally disinhibits sexual behaviour is firmly ingrained in our culture. Most people believe that drinking increases the likelihood of sexual activity, enhances sexual experience, and promotes riskier sexual behaviour (Cooper, 2006). Countries with an oppressive history have been shown to have a negative impact on an individual's healthy self-development, and more broadly on the development of the environment in which the individual resides. This is evident in studies conducted in South Africa (Adams et al., 2014;Townshend et al., 2014) and the US (amongst various ethnicities such as African-Americans, Hispanic and Asians) (Brown & Vanable, 2007;Kogan et al., 2010;Langer et al., 2001;Park & Grant, 2005;Voisin et al., 2013). This is consistent with previous research that states that while there is evidence that there is an inverse relationship between risky behaviours and age (Ajayi, Marangu, Miller, & Paxman, 1991;Akwara, Madise, & Hinde, 2003;Chassin, Pitts, & Prost, 2002;Johnston, O'Malley, Bachman, & Schulenberg, 2004;Kiragu & Zabin, 1995;O'Neill, Parra, & Sher, 2001;Ochollo-Ayayo & Schwarz, 1991;Schulenberg, O'Malley, Bachman, Wadsworth, & Johnston, 1996), it is an unlikely outcome in communities characterised by poverty and lack of resources and access to educational and employment opportunities. These factors, with a historical and political genesis, are mutually influencing, resulting in a recurring process of risky behaviour and impoverishment. Along with these external factors, individual factors such as perceptions of self-identity, social identity, selfefficacy, sense of belonging and hope for the future play a significant role in contributing towards risky behaviours (Adams et al., 2014).
Despite the large body of theory and research that supports alcohol's role in risky sexual behaviour, understanding the nature of this relationship has been more challenging than originally anticipated (for reviews, see Cooper, 2002;Halpern-Felsher, Millstein, & Ellen, 1996;Weinhardt & Carey, 2000). A possible contributing factor is that there were very few studies which utilised theoretical frameworks to structure the studies. The theoretical frameworks utilised attempted to summarise and make sense of individual and social factors contributing to alcohol use and risky sexual behaviour. These frameworks were used to understand the behaviour and help account for underlying mechanisms of the specific cohort's drinking and may ultimately explain points for more comprehensive intervention development and enhancement. Essentially, there is no one theory to explain the relationship between alcohol use and risky sexual behaviour. However, for an intervention to be successful a theory is needed to explain the relationship, understand the contributing factors and the patterns in the relationship and how it changes. Additionally, the use of theoretically based questionnaires with well-defined items is particularly important since the diversity of content of research in this field restricts the comparability of findings and makes conclusions difficult (Kuntsche, Knibbe, Gmel, & Engels, 2005).
According to Kuntsche et al. (2005), an explanation for these inconsistent results could be that motives affect drinking only in the onset phase and not the continuation to drink and indulge in risky sexual behaviour. From the research reviewed, there is consensus that internally caused motives such as coping, sense of belonging and self-efficacy are strongly related with personality traits are more consistently related to alcohol use across drinking situations. However, since there is a lack of longitudinal evidence it is nearly impossible to determine the long-term effects of different drinking motives on different alcohol-relatedoutcomes in different age groups within different contexts (Kuntsche et al., 2005).

Conclusions and Recommendations
Whilst there has been an increase in research attempting to ascertain and determine the key factors that make the relationship between the alcohol use and risky sexual behaviour significant, very few studies focused directly on the factors mediating and moderating this relationship. In spite of this, the studies and reviews which form part of this systematic review provide key insights into factors impacting on alcohol use and risky sexual behaviour. Further systematic enquiry and qualitative exploration at a primary level will allow for an increased comprehension as to how these factors contribute specifically to the relationship between alcohol use and risky sexual behaviour as some young adults may engage in risky sexual behaviour but not necessarily engage in alcohol use, or vice versa. Many of these studies reviewed were also quantitative studies; thus, an in-depth understanding, across time, will allow for a greater understanding of these factors. As Cooper (2006) expressed, the relationship between alcohol use and risky sexual behaviour is complex and cannot be explained by a single mechanism, but instead reflects multiple underlying causal and non-causal processes. Moreover, even the causal portion of this relationship is not apparent as a main effect but as an interaction. Therefore, there is a need to qualitatively explore young people's understandings, perceptions and motivations to engaging in alcohol use and risky sexual behaviour. While these findings will be context-specific, this will be the foundation for studies that aim to quantitatively understand this relationship. Furthermore, there remains a need for research among young adults as a whole and not exclusively among young adults in school, colleges, university or clinic/treatment setting. Other gaps in the literature point to the need for research across diverse socio-economic group.