The relationship between adolescent depressive symptomology and substance abuse

The research aimed to investigate the phenomena of adolescent depressive symptomology, substance abuse and the relationship between the two phenomena in a South African context. The influence of moderator vari­ ables was also examined. Another objective was to de­ termine risk factors for the before mentioned. This was done by using a questionnaire with a sample of 1298 conveniently selected adolescents in a South African high school. It was determined that adolescents be­ come progressively unhappier from 13 to 17 years of age. Girls are more depressed than boys. This research also revealed that adolescent depressive symptomology is significantly and positively correlated with earlier age of onset of substance abuse as well as frequency of usage. There appear to be no gender differences in substance abuse but teenagers from different ethnic and language groups differ in their use of substances. Risk factors for depression and substance abuse included a conflict relationship with parents, the experience of ma­ jor stressful events, dissatisfaction with school grades and friends’ use of substances. Opsomming Die doel met die navorsing was om die vlak van adolessente depressiewe simptome, die gebruik van dwelmmiddels en die verband tussen die twee in ‘n SuidAfrikaanse konteks te ondersoek. Die invloed van mod­ erator veranderlikes is ook nagevors. ‘n Verdere doelwit was om risikofaktore ten opsigte van voorgenoemde aspekte te bepaal. Dit is gedoen deur middel van ‘n vraelys met ‘n gerieflikheidsteekproef van 1298 adolessente'van ‘n Suid-Afrikaanse hoërskool. Daar is gevind dat adolessente progressief ongelukkiger word vanaf 13tot 17-jarige ouderdom. Dogters is meer depressief as seuns. Hierdie navorsing het ook getoon dat die vlak van adolessente depressiewe simptome beduidend en positief korreleer met die vroeëre aanvangsgebruik van dwelmmiddels sowel as met die frekwensie van gebruik. Geen verskil kom tussen adolessente van verskillende geslagte voor in die gebruik van dwelmmiddels nie, maar jeugdiges van verskillende etniese en taalgroepe verskil in hul gebruik van dw elm m iddels. R esikofak tore behels ‘n konflikverhouding met ouers, die belewing van voorvalle van hoë stres, ontevredenheid met skoolprestasies en vriendskap met maats wat dwelmmiddels gebruik.


Introduction
Findings from epidemiological studies have indicated that the prevalence of a number of health-compromising behav iours such as substance abuse and psychiatric disorders such as major depressive disorder (MDD) increase dra matically during adolescence (Windle & Davies 1999:823).Both depression and substance abuse among teenagers are issues of concern because of the deleterious effect of both phenomena on the adolescent.Regarding depres sion, suicide attempts by adolescents in different parts of the country are reported in the media (Keppler 2003:1).In addition, various South African studies have documented increasing rates of substance abuse among the youth (Flisher, Parry, Evans, Lombard & Muller 1998:5;Oven 2001:1;Parry&Pluddemann 1999:2).Pienaar (2003:9) quotes a spokesperson from the Medical Research Council stating that the percentage of patients younger that 20 years of age who were treated for drugs from the start of 2000 until the end of 2002, has increased from 17% to 24% in Cape Town, 19% to 22% in Durban, 17% to 25% in Gauteng and 7% to 15% in Mpumalanga.Fourty six children in Gauteng were in rehabilitation clinics for dagga, 27 for inhalants, 20 for alcohol and two for dagga as well as Mandrax.Given the emergence and relative high rate of dysfunctional be haviour and psychiatric disorders in adolescence, this re search focusses on the relationship between substance abuse and adolescent depressive symptomology.

Adolescent depressive symptomology
Relatively current prevalence figures for Major Depressive Disorder (MDD) puts the figure for adolescent sufferers of depression at about 5-10% (Kaplan & Sadock 1998:1245), although Lewinsohn (1999:271) believes this figure could be as high as 20%.One study by McGee, Feehan, Williams andAnderson, published in 1992 (in Windle &Davies 1999:825), reported that prior to age 12 years, boys had higher rates of depression than girls; subsequent to age 12 years, girls had higher rates of depression than boys, with the ratio estimated at about 4 :1 .In their study with over 1 000 adolescents, Windle and Davies (1999:839) found the ratio of depression for girls relative to boys, 2.5 : 1. Ac cordingly, Lewinsohn (1999:272) determined depression in twice as many girls than boys, beginning at age 12 to 13.
The term depression is widely used to cover both transient dysphoric loss-related emotional states and clinical depres sion leading to profound disturbance of mental and social functioning (Gotlief & Hammen 1996:2).According to Kaplan and Sadock (1998:553), a depressed mood and a loss of interest are the key symptoms of depression.The depressive symptoms influence and interact with sets of other symptoms in the domains of cognition, behaviour and bodily functioning and interfere with the acquisition of im portant developm ental tasks and skills (G otlief & Hammen 1996:3, 66).Adolescent depression is closely linked to functional impairment and disease (Lewinsohn, Seeley, Hibbard, Rohde & Sack 1998:1120), subsequent adult depression (Harrington 2001:21) and increased sui cide risk (Kaplan & Sadock 1998:553).
Adolescent depression has an insidious onset; it is chronic rather than acute; it is episodic and there is a larger propor tion of co-morbidity than in adult depression.Biederman and Spencer (1999:276) as well asFaraone (1999:271) indi cate that adolescent depression has atypical features, par ticularly mood reactivity.However, the physical and cog nitive changes of the adolescent period make it difficult to establish the limits of depressive disorder.Depressed ado lescents tend to have multiple problems which include academic performance, substance abuse, antisocial behav iour, truancy, lost of interest, sadness, irritability, low self esteem, hopelessness, pessimism, self-criticalness, agita tion, school anxiety, loneliness, feelings of being devalued by parents, social withdrawal, somatic complaints, unusual attention to death as well as complaints of boredom, anger and aggression (Beck, Steer & Brown 1996:5;Gotlief & Hammen 1996:43;Kaplan & Sadock 1998:542,553;Lamarine 1995:391;Santrock 2000:463;Smith, Bern & Nolen-Hoeksema 2001:427).According to the literature, adoles cent depression may, amongst other factors, be increased by substance abuse (Kaplan & Sadock 1998:553).

Adolescent substance abuse
In the literature, a distinction is made between substance use and substance abuse.Adolescent substance use re fers to the non-pathological use of, experimentation with or the occasional irregular use of psychoactive substances.The National Drug Master Plan (South Africa (Republic) Department of Welfare 1999:1) states that the greatest number of drug experimenters is of school-going age and the age of first experimentation has dropped.Such experi mentation or occasional use is not catered for in the DSM-IV-R and is often referred to as 'recreational' use.Some researchers view adolescent use o f substances as a socialisation process (although not endorsing such behav iour) and believe it is mostly self-limiting (Bukstein 1995:26;Kaminer 1994:42).Moreover, Schedler and Block (in Kaminer 1994:42) found that youngsters who had experimented with psychoactive substances were psychologically healthier than either frequent users or abstainers.Some of these adolescents, however, will go on to develop drug depend ency to the detriment of their well-being.
Rates of drug abuse show considerable variation by age (Kaplan & Sadock 1998:383), with a positive correlation between age and level of substance abuse (Rocha-Silva 1998:18;Leggett 1999:9).It should be noted that the age of onset of a particular stage of substance use and the fre quency of use at an early stage are strong predictors of progression to later stages (Kandel, Yamaguchi & Chien 1992:447;Windle 1990:86).Clayton (1992:21-28) postulated the following stages in substance use: initiation, continua tion, maintenance and progression, progression across drug classes leading to polysubstance abuse and dependence, and finally regression, cessation and relapse cycles.A position at one stage does not necessarily imply the user will progress to the next stage and this is especially true of adolescents (Bukstein 1995:24;Clayton 1992:26;Newcomb 1992:255).Of significance for this study is the finding that those who abuse one substance and are depressed are more likely to progress to the next stage of substance abuse than those who are not depressed (Kaplan, Landa, Weinhold & Shenker 1984:600).
Both depression and substance abuse tend to be chronic conditions and both are associated with psychosocial mor bidity.When these disorders are co-morbid, functioning is further im paired (Rao, Ryan, Dahl, Birmacher, Rao, Williamson & Perel 1999:1114).The relationship between depression and substance abuse is highly complex.Some researchers found a cause-effect relationship between the severity of dependence and depression (Coelho, Rangel, Ramos, Martins, Prata & Barros 2000:103;Patton, Hibbert, Rosier, Carlin, Caust & Bowes 1996:225;Stefanis & Kokkevi 1986:126).One study found that depression on the one hand, and cigarette, marijuana, and harder drug use on the other hand among the suburban students in their sample were positively related (Way, Staber, Nakkula & London 1994:331-357).Feelings of meaninglessness and lack of purpose were important variables in the connection be tween adolescent depression and substance abuse (Kinnier et al 1994a: 101-111).Thus a strong mediating relationship between purpose in life with depression as precursor and substance abuse as consequence was determined.Unful filled needs (possibly including depression) may propel adolescents into destructive behaviour such as substance abuse (Mainous III, Martin, Oler, Richardson & Haney 1996:807).
Having depressive symptoms is associated with a higher risk of developing alcohol-related problems in adolescents (Chen, Anthony & Crum 1999:38; Windle & Davies 1999:823-844).Depression increases the probability of sub stance abuse but only if the peer group is tolerant of such behaviour, or the person is a member of a deviant subgroup (Simons, Whitbeck, Conger & Melby 1991:466).This indi cates the presence of risk factors associated with adoles cent depressive symptomology and substance abuse.

Risk factors associated with adolescent depressive symptomology and substance abuse
Henry et al (1993:469-480) found the strongest predictors of substance abuse were multiple drug use and concurrent conduct problems.Lewinsohn, Gotlib andSeely (1995:1227) found external behaviour problems more predictive of sub stance abuse and internal behaviour problems to be more predictive of MDD.By surveying clinical and epidemio logical retrospective studies, researchers have shown that there is about an even distribution of MDD prior to and following an onset of drug use (Stefanis & Stefanis 1999:30).Conduct disorder, MDD and Attention Deficit Hyperactiv ity Disorder (ADHD) may all be important concomitants of substance abuse in males, while in females, depression may be the prim ary variable related to substance abuse (W hitm ore, M ikulich, Thom pson, Riggs, A arons & Crowley 1997:87) .
Risk factors that occur in both MDD as well as Substance Abuse Disorder were identifed as current depression, in ternal behaviour problems, problems with coping skills, in terpersonal conflict with parents, dissatisfaction with grades and external behaviour problem s (Lew insohn et al 1995:1226).Of these, Aseltine (1998:549) found low family support to be the most distinctive factor associated with the co-occurrence of depression and substance abuse.Windle and Davies (1999:823-844) found support for unique risks associated with specific subgroups in their study of 1000 adolescents.Particularly significant and distinguish ing for their mixed group of heavy drinking and depression youths, were childhood external problems, stressful life events, high levels of current substance abuse and high levels of delinquency.By scoring the above mentioned risk factors for adolescents, clinicians may estimate the prob ability of their developing either disorder, or both.
Adolescence is the highest risk period for onset of both depression and substance abuse (Rao, Daley & Hammen 2000:215).Dually diagnosed adolescents are at a higher risk for lifelong dysfunction and even suicide (Rao et al 2000:221).Hence, the importance of further investigation of the issue.

The research question, aims and hypotheses
The question underpinning this research is: What is the relationship between adolescent depressive symptomology and substance abuse?The research aims to investigate em pirically the phenom ena of adolescent depressive symptomology, adolescent substance abuse and the rela tionship between the two phenomena in a South African context.Apart from aiming to determine the levels of de pressive symptomology and substance abuse in the sam ple, it was hypothesised that:

Research design
The research employed a survey design with a sample of all the learners from grade eight to grade 12 in an urban, former ' model C' school (approximately 1300 learners), with which one of the researchers was associated.Thus the sample was a convenience, non-probability sample.Although generalisations have to be made with caution, the school is similar to many former 'model C ' schools in urban areas and, to that extent, the results are relevant for that popula tion of adolescents.

Research instrument
A questionnaire, based on an in-depth literature review, was designed and used to collect the data.The question naire comprised 69 items in the following sections: Section A covered biographical information of the respond ents namely age, grade (and if a grade was ever repeated), gender, ethnic group and home language.It was envis aged that this information would be useful in contextualising the results.Section B comprised items relevant to the levels of adoles cent depressive symptomology.A five-point Likert scale was provided for this section.
Section C dealt with items that focussed on substance abuse and which included the following three parts: P a rt I determined if the respondent had ever used a par ticular substance (the options provided were "never used" and "used").P a rt II asked when a respondent had first used a substance (the options were "never used, before 13 years of age, 13-14 years, 15-16 years and 17 years and older").P a rt III determined how often a substance had been used in the previous 30 days (options included "never used, 2-5 times, 6-10 times and more than 10 times this last month").
Section D consisted of items to determine the possible role of common risk factors, namely stressful life events, inter personal conflict with parents, dissatisfaction with grades, absenteeism, lateness for school, delinquency, ADHD and suicide attempts.Respondents could choose between "Yes" and "No".

Validity and reliability of the questionnaire
Validity is concerned with whether a test measures what it is supposed to measure.Two measures of validity were considered when compiling the questionnaire, namely con tent validity and face validity (De Vos & Fouché 1998:82).
The content validity was determined by the literature study and by the judgment of experts in adolescent depressive symptomology and substance abuse.According to Leedy (1993:41) two important questions are: Is the instrument measuring what it is supposed to measure?Is the sample being measured adequate to be representative of the phe nomena?Positive confirmation in this regard was obtained by experts in the field.Face validity also relies on the judg ment of an expert.Only after it was confirmed that the questionnaire seemed to measure what it was supposed to measure, was the questionnaire finalised.
Reliability refers to the degree of consistency with which an instrument measures.In this study, the reliability of the scaled items of the questionnaire was calculated by means of the Cronbach alpha coefficient.It was measured at 0.83, which is good for this type of questionnaire.

Pilot study
A pilot study with learners at a school not included in the research was conducted.After the pilot study, a few items were reformulated before the questionnaire was edited and finalised.For example, one item requested the adolescents to indicate their use of inhalants.For clarity, 'glue' and 'thinners', were added as examples of inhalants.Likewise, the names 'speed' and 'ravers' were added as examples of stimulants.

Administration of the questionnaires
Ethical measures were undertaken in accordance with the guidelines given in Strydom (1998:24-35).Permission for administering the questionnaire was requested from the school's Governing Body.The parents were informed of the questionnaire by the school's newsletter and given the opportunity to withdraw their children.Learners were also informed about the purpose of the study, prior to the ad ministration of the questionnaire and they were asked to participate voluntarily.Thus informed consent was ob tained.Respondents and the school remained anonymous.The questionnaires were administered in class time.After capturing of the data, the Statistical Programme for the So cial Sciences (SPSS) was used to analyse the data.

Biographic information
The biographic information revealed the fol lowing about the sample composition.Respondents' responses regarding adoles cent depressive symptomology, substance abuse and the possible role of risk factors were determined and are as follows.

Adolescent depressive symptomology
Table I illustrates the percentage of respond ents in the sample who experienced different kinds of depressive symptomology.(Lewinsohn 1999).
From the literature review, a number of null-hypotheses were formulated which stated that no statistically signifi cant differences existed between various groups regarding depressive symptomology.These were tested by means of t-tests and analysis of variance.Results indicated that different ethnic or home language groups did not differ significantly with regard to depressive symptomology.However, significant differences were found between the following groups: • different ages (F-value= 10.066, p<0.01) • various grades (F-value= 10.462, p<0.01) • those that had or had not failed at school (t-value=0.01,p<0.05) • different genders (t-value=2.095,p<0.05).
The happiest group (least depressive symptomology) was the 13-14 year olds (mean=3.5583)who were significantly happier than the 17 years and older group (mean=3.5029).Accordingly, the happiest group were the grade eights (mean=3.7465),and the unhappiest, the grade 11 's.(3.4848).
Those who had repeated a grade reported significantly more depressive symptomology (mean=3.5112)than those who had never repeated a grade (mean=3.6109).Regarding gen der, the adolescent boys in the sample were significantly happier (mean=3.6280)than the girls (mean=3.5652).This confirms previous findings (Lewinsohn 1999) that girls are more prone to depression than boys.
Table II illustrates the percentage of respondents who indi cated that they had never used substances as well as those who had used them.In addition, the table illustrates how old the respondents were at the time of first use.

Table II: Age of first use of substances dependency between current use of substances and age
Significant differences (p<0.01) were found between differ ent age groups for all substances, except for inhalants, stimulants and heroin.The age groups differed in their use of cigarettes and tobacco (X2=l 19.882); alcohol (X2=133.5);dagga (X2=48.613);cocaine (X2= 3 1.502); hallucinogens (X2=22.064)and ecstasy (X2=92.888).The pattern revealed that the 17 years and older group used the substances mentioned significantly more than the 15 to 16 years old or the 13 to 14 years old adolescents.For example, 73% of the 17 year old group, 57.4% of the 15 to 16 year olds, 36.2% of the 13 to 14 year olds and 16.7% of those younger than 13 years had used alcohol the previous 30 days.This  Regarding current substance use, the pattern was similar to that revealed by Table II, except that Table III indicates significantly lower current use of substances.For exam ple, 41% (100-59%) in comparison to 79% of the adoles cents had used cigarettes and tobacco; 54% (100-46%) in comparison to 65.3% had used alcohol; and 15% (100-85%) in comparison to 34.4% had used dagga in the previous 30 days.Four percent and less of the adolescents indicated current use of stimulants, cocaine, heroin, hallucinogens and ecstasy.This confirms reports in the literature that some adolescents experiment with drugs but do not con tinue using them on a regular basis.
Null-hypotheses that stated that there was no statistically significant dependency between membership of various groups and current use of substances were formulated and tested.CW-square analyses revealed the following: that adolescents use more substances as they grow older.This confirms previous research, for example that of Rocha-Silva (1998) and Legget (1999), as mentioned.

dependency between current use of substances and grade
Significant dependencies were found between grade level and learners' use of cigarettes and tobacco (X2'147.118,p<0.01), alcohol (X2=170, p<0.01), dagga (X2=74.095,p<0.01) and ecstasy (X2 '26.108, p<0.05).As learners progress to higher grades, they increase their use of these substances significantly, in accordance with the above findings for age.Alcohol shows an increase by grade until grade 11 where it peaks with a marginal drop in use in grade 12.

dependency between current use of substances and gender
In contrast to what was expected from the literature that more males abuse substances, no significant gender differ ences were found in current use of substances.

dependency between current use of substances and ethnic group
Significant differences between ethnic groups were found in the current use of cigarettes and tobacco (X2 = 36.434,p<0.01) and alcohol (X2 = 47.385,p<0.01) in the following way: 53.6% of the Coloureds; 42.8% of the Whites; 33.8% of the Africans; 28.9% of the Indians and only 15% of the Asians had used cigarettes and tobacco in the 30 days preceding the questionnaire completion.Thus, there is a significant dependency between ethnic group and use of tobacco.Regarding alcohol, current use revealed a similar pattern: 64.3% of the Coloureds; 58.5% of the Whites; 41.7% Africans; 37.3% Indians and 23.1% Asians had used alcohol in the 30 days preceding questionnaire completion.Hence tobacco and alcohol are used, from most to least, by Coloureds, Whites, Africans, Indians and lastly Asians.

dependency between current use of substances and language group
In addition to the above, significant dependencies were determined between language group and substance use for cigarettes and tobacco (X2=23.389,p<0.05) and for alco hol (X2=26.427,p<0.01) as follows.In the 30 days preced ing questionnaire completion, significantly more Afrikaans (43,4%) and English (42,8%) speaking learners had smoked than learners speaking an African language (32,2%), or a foreign language (23,6%).Significantly more English (57,1%) and Afrikaans (56,5%) speaking learners had also used alcohol in the previous 30 days than learners speak ing a fo reig n lan g u ag e (43,7%) or an African lan guage (40,9%).This indi cates that the English and Afrikaans-speaking learners sm oked and drank m ore than the other language groups in the sample.

The relationship between adolescent depressive symptomology and substance abuse
Two null-hypotheses were stated, focussing on the corre lation between adolescent depressive symptomology and age of first use or current use of substances.When these were tested, the correlations were significant (p<0.01) and positive, although low (0.22 and 0.27 respectively).This implies that as learners become more depressed, they tend to be (a) younger when starting with drugs and (b) have a higher incidence of current drug use.This confirms previ ous findings as indicated.

Risk factors for depression and substance abuse
In the literature review, various risk factors were identified for depression as well as substance abuse.The extent to which the subjects in the sample experienced these risk factors, is illustrated in Table IV.
Table IV illustrates that the risk factors with the highest frequencies were dissatisfaction with marks (58%), having experienced major stressful events (52%), friends' use of substances (49%) and fighting with parents (30%).Ac cording to Table I, suicide ideation was the symptom of depression with the highest frequency in the sample.Ac cording to Table II and III, alcohol was the substance most often used by the subjects.To determine the dependency between the abovementioned four risk factors and (a) sui cide ideation and (b) alcohol use, null-hypotheses were stated.Chi-square analysis was used to test the hypoth eses.

Conclusions
Adolescents experience many depressive symptoms.The level in this study is significantly higher than those reported by previous studies using adolescents in other contexts.Feel ings of aggression, anger, worthlessness and guilt were experienced by more than half of the sample, while suicide ideation was reported by 75%.In addition, adolescents become progressively sadder from 13 to 17 years of age.
Learners who fail grades are unhappier than the others and females are more depressed than males.
The prevalence rates of substance abuse found in this study are similar to those reported in the literature.Some adolescents experiment with substances, but do not con tinue using them.However, current use as found in this study is of concern, especially of tobacco, alcohol and dagga.
Learners increase their consumption of cigarettes and to bacco, alcohol, dagga and ecstasy as they grow older and move up the grades.Learners who fail at school are par ticularly at risk.Alcohol use levelled off in grade 12 and this is possibly the result of maturation.More Coloured and White, and more English and Afrikaans learners smoke and drink than other ethnic and language groups.How ever, no gender differences were found for current use of substances.If more females than males are depressed, this has not translated into substance abuse and this implies that depression has different consequences according to gender.
As expected from the literature study, there was a signifi cant and positive correlation between the number of de pressive symptoms experienced by the adolescents and substance use.As the learners have more and more symp toms of depression, they tend to be younger when starting with drugs and to have more current drug use.
Of the risk factors identified in the literature, the following seems to be most significant for depression (suicide idea tion) and alcohol use or abuse: a conflict relationship with parents, experiencing major stressful events and most im portantly, dissatisfaction with school grades and friends' use of substances.
Further research needs to be done on the rise of depressive symptomology in late adolescence, the longitudinal paths of adolescent substance users and abusers and on the in teractions between depression and substance abuse.Quick, accessible, screening instruments for depression in learn ers should be developed by researchers for South African high school learners so that those learners at risk, can be identified.These learners could then be referred to profes sionals and be monitored within the school system.