Depression

156 Developmental Psychopathology

events, depressed youth are less able' to generate ef-fective solutions to interpersonal problems. Consis-tent with this hypothesis, depressed children havepoorer social skills (Bell-Dolan, Reaven, & Peterson,1993) and they are less often chosen as playmatesor workmates by other ~hildren (R~d~IRh, Hammen.& Burge, 1994).

Are depressed children the victims or the initia-tors of negative social relationshipf? An ingeniousstudy by Altmann and Gotlib (19881 investigated thesocial behavior of depressed school-age children byobserving them in a natural setting: at play duringrecess. The authors found that depressed childreninitiated play and made overtures for social contactet least as much as did nondepressed children, andwere approached by other children just as often. Yet,depressed children ended up spending most of theirtime alone ..By carefully observing the sequentialexchanges between children, the researchers dis-covered the reason for this. Depressed children weremore likely to respond to their peers with whatwas termed "negative/aggressive" behavior: hitting,name-calling, being verbally or physically abusive.

These observations fit well with the model de,veloped by Patterson and. Capaldi (1990), in whichpeer relations are posited to play the role of medi-ators of depression. According to this model, a neg-ative family environment leads children prone to de-pression to enter school with low self-esteem; poorinterpersonal skills, aggressiveness, and a negativecognitive style. They are less able to perceive con-structive solutions to social problems' andale morelikely to be rejected by peers because of the waythey behave. Peer rejection, in turn, increases theirnegative view of self and thus their depression.

In order to test this model, Capaldi (1991, 1992)differentiated four groups of boys depending onwhether they demonstrated aggression, depressedmood, both aggression and depression, or neither.Boys were followed over a two-year period. fromgrades 6 to . While depression and adjustmentproblems t ded to abate over time in the depressedgroup, no such improvement occurred in the. twoother' bed groups. While, in general; aggres-sive behavior was more stable than depressedmood, condu t problems increased the risk of sub-sequen having a depressive mood. In fact, ag-

gression in grade 6 predicted depressed mood ingrade 8, while earlier depression did not predictlater conduct problems.

Capaldi conceptualizes the process leading fromaggression to depression as follows. (See Figure7.3.) Aggression and noxious behavior. alienates par-ents, peers, and teachers, resulting in more inter-personal conflict and rejection. Further, a~agreSSiOnleads to oppositional behavior in the cl sroom,which leads to learning deficits and poor kill de-velopment. Both of these factors result in profoundfailure experiences in the social and academicrealms. Failure and rejection, in turn, produce lowself-esteer;z. The impact of peer rejection. low aca-demic skills, and low self-esteem is associated withincreasingly serious deficits in adolescence, ulti-mately resulting in depression.

The Organic ContextEvidence that organic factors play an important eti-ological role in depression has emerged in studies of

_adult populations, while research evidence in regardto children is slight. (Our presentation follows Ham-men .and Rudolph, 1996. unless otherwise noted.)

Familial concordance rates provide evidence foragenetic component in depression. Children, ado-lescents, and adults who have close relatives withdepression are at considerable risk for developingdepression themselves. In fact, having a depressedparent is the single best predictor of whether a child

. will become depressed.However, simply demonstrating a correlation be-

tween parent and child depression fails to disentan-gle the relative influence of heredity and environ-·ment. For this, twin and adoption studies are needed.One study of adolescents compared monozygotic anddizygotic twins, biological siblings, half-siblings,and biologically unrelated step-siblings and foundsignificant genetic influences at lower levels of de-pression but significant environmental influences athigh levels of depression (Rendeet 81., 1993). An-other study investigated a large sample of monozy-gotic and dizygotic twin pairs agedS to ·16 years(Thapar & McGuffin, 1996). While environmentalfactors seemed to best account for depression inchildhood, evidence for a genetic component wasstrong in adolescence. In sum, while data support th~

, -I

I!,I-I-

-.'

Chapter 7 Disorders in the Depressive Spectrum and Child and Adolescent Suicide 157

Figure 7.3 Mediators of the .Effects of Child and Family. Factors on Depression.

Soun:e: Adapted from Patterson and Capaldi. 1990

theoryof a genetic component to depression in adultsandadolescents, other factors also playa major rolein the etiology of depression in children.

Research with adults indicates a neuroendocrineimbalance as an etiological agent;' particularly hy-persecretion of the hormone cortisol. This is not sur-prising since hormone production regulates mood,appetite, and arousal. all of which are adverselyaf-fected by depression. However, little evidence ex-ists for the role of cortisol in child depression.

Depression is also associated with low levelsof the neurotransmitter serotonin. Antidepressantmedications that act to increase serotonin _avail-ability. including tricyclics such as imipramine andthe new generation -of selective serotonin reuptakeinhibitors (SSRIs). such as Prozac, have been proveneffective in combating depression in adults. Again.

however. evidence in support of this mechanismin children is mixed at best. In many controlledstudies. antidepressant medications have not beeneffective in combating child depression. Fisherand Fisher (1996) reviewed thirteen double-blindplacebo-controlled studies published between 1965and 1994 and found only two cases in whicli anti-depressants relieved depression better than place-bos. In fact, some studies seemed to shoW"that theplacebo was more effective! Similar null findingsare reported in an exhaustive review by Sommers-Flanagan and Sommers-Flanagan (1996).

Much of the research these investigators re-viewed was based on -the older type of antide-pressants rather than the new SSRIs. which maybe both more effective and safer. Recent studiesfocusing on the efficacy of SSRIs in child and

158 Developmental Psychopathology

~I~--~--~~-.-------

Cognitiverepresentations of

seff, others, Faroily'-'-~_

e,xp~rie~ -T"- -

" ,

l~erjiersonal"., ..

,.~C'C?",petence, :';..

Ufestress

Figure 7.4 Hammen's Multifactorial Transactional Model of Child and Adolescent Depression.

Source: Hammen and Rudolph. 1996

adolescent depression are somewhat more promis-ing (DeVane & Sallee, i996). However, effectsizes are ,still small, and in some studies resultsemerged only for ratings of global improvementand not for symptoms of depression (Emslie, Ken-nard, & Kowatch, i995).

In summary, research on child depression, whilelimited, suggests that organic theories of etiologyderived from studies of adults cannot be applied aseasily to children. Further, without prospective datademonstrating that biological indicators predate theonset of depression, there remains some question asto whether these are the cause or result of depres-sion, Ultimately, the picture is likely to be a com-plex and transactional one. Experiences and moodact on biology, and, in turn, biology reciprocally af-fect cognitions, emotions, and memory (Post &Weiss, 1997).

Integrative Developmental ModelA comprehensi e developmental psychopathologymodel of depression has been put forward by Ham-men (1991, i992; Hammen & Rudolph, 1996) andis presented ir; Figure 7.4. The case study that waspresemed in 'X J also illustrates the elements of

r

While acknowledging that there are many path-ways to depression, Hammen's model places dys-functional cognitions at the forefront. First, how- ,ever, the stage for the development of these negativecognitions is set by family factors, such as a de-pressed parent, insecure attachment, and insensitive.or.rejecting caregiving. Adverse interpersonal expe-riences contribute to the child's development of neg-ative schemata: of the self as unworthy, others asundependable and uncaring, and relationships ashurtful or unpredictable. The depressive cognitive

· style also involves the belief that others' judgmentsprovide the basis for one's self-worth, as well as a

, ,

tendency to selectively attend to only negativeevents and feedback about oneself.

Further, Hammen highlights the fact that the re-lationships among affect, cognition, and behaviorare dynamic and transactional. For example, nega-

·tive cognitive styles lead to problems in interper-sonal functioning, which act both as vulnerabilitiesto depression and as stressors-m-theirown right. Thenegative attributions of depressed children interferewith the development of adequate coping and socialskills, and they respond to interpersonal 'problems

·through ineffective strategies such as withdrawal oracquiescence. These strategies not only fail to re-solve interpersonal problems but even exacerbate

lII

Chapter 7 Disorders in the Depressive Spectrum and Child and Adolescent Suicide 159

J -:

them.:.increasing experiences of victimization, rejec-tion, and isolation. Therefore, the negative cognitivestyles and poor interpersonal problem-solving skillsassociated with depression further disrupt socialrelationships, .undermine ihe child's. competence,induce stress.rand confirm the child's negative be-liefs about the self ana the world.

As development proceeds, these cognitive and in-terpersonal vulnerabilities increase the ;elihOOd thatindividuals will respond with depressio when facedwith stress during development.Hamme's model de-scribes three aspects to the role of stress in depres-sion. First, as described above, individualsvulnerableto depression may actuallygeneratesome of their ownstressors. In this way they contribute to the aversive-ness of their social environments, as well as consoli-dating their negative perspectiveson the world.An il-lustration of this kind of process later in developmentis "assortative mating," the tendencyof individuals tochoose partners who mirror or act on their vulnera-bilities.For example, Hammen and colleagues findthat depressed women are more likely than others tomarry men with a diagnosablepsychopathology,and,in turn, to experience marital problems and divorcewhich contribute further to their depression.

Second, the association between stress and "de-pression is mediated by the individual's cognitivestyle and interpretation of the meaning of stressfulevents. While· life stress increases the likelihood ofpsychopathology in general, it is the tendency to in-.

. terpret negative events 'as disconfirmationsof one'sself-worth that leads to depression in particular:Third, certain groups of children are at high riskbecause they are exposed to the specific kinds. ofstressors that increase depression. These includemaltreated children, those whose parents are emo-tionally disturbed, those in families with high lev-els of interparental conflict, or those who live in sit-uations of chronic adversity that diminish the entirefamily'S morale and sense of well-being.

Biological factors can comeinto play at any pointin the cycle. For example, individual differences intemperament may contribute to problems' in chil-dren's relationships with parents and others: Biolog-ical factors can affect children's ability to cope withstressful circumstances, as well as increasing theirvulnerability to depression as a reaction to stress.

Developmental influences also enter into the pic-ture in a number of ways. First, difficulties that oc-cur earlier in development may have particularlydeleterious effects, diverting children to a deviantpathway from which it is difficult to retrace theirsteps. Once on a deviant trajectory, children becomeincreasingly less able to make up for failures to de-

" velop early stage-salient competencies. Accumu-lated stress may also alter the biological processesunderlying depression, especially in young chil-dren, whose systems are not yet fully matured. Sec-ond, cognitive development can influence depres-sion. As we have seen, young children's thinkingtends to be undifferentiated and extreme, con-tributing to an "all pr nothing" kind of reasoning.A negative cognitive style formed at an early age,therefore, may be particularly difficult to changeonce consolidated. Third, the organizational viewof development argues that the connections amongcognition, affect, behavior, and contextual factorsstrengthen over time.Thus, over the life course, de-pressivepatterns are integrated into the self system,become increasingly stable, and require lower·thresholds for activation.

Hammen's model is relatively new, and it is inthe nature of research in developmental psy-chopathology that decades must pass before we havedata available that fully test a given model by trac-

. ing pathways of development from infancy to adult-hood. Therefore, it is too early. to say whether thisis an accurate account of the developmental psy-chopathology of depression. To date, parts of themodel have held up to empirical scrutiny. Rudolph,Hammen, and Burge (1994) demonstrated links be-tween child depression and negative cognitions _about self and other, negative representations offamily and peer relationships, biases in social in-formation processing, and poor interpersonal skills.However, in a study of adults, Hammen and eel-leagues (1995) found that, while attachment-relatednegative cognitions and life stress predicted depres-sion one year later, the results were not specific tosymptoms of depression. Therefore. it may be thatHammen's model actually represents a generalmodel for the development of psychopathology, one

. that can be appliedjo the understanding of depres-sion but is not specific to it.

160 Developmental Psychopathology

Intervention

PhaJ"acotherapyCAs noted above, while some studies indicate that ~he

new antidepressants (SSRIs) reduce depressivesymptoms in children, results are mixed. Undesir-'able side effects also occur, including restlessnessand irritability, insomnia, gastrointestinal discom-fort, mania, and psychoticreactiom'[ (De VanJ &Sallee, 1996). There are advocates fOr their contin-ued use, who cite the low rates. of serious side ef-fects and the devastating consequences of untreateddepression (Kye & Ryan, 1995). However, othersare strongly opposed, arguing that their use is un-ethical given that their effectiveness is not supportedby the existing research (Pellegrino, 1996).

Despite questions about the effectiveness of an-tidepressants with children, JPey are being pre-scribed at arr increasing rate:fin 1996, U.S. physi-cians wrote 735,000 prescriptions for SSRIs forchildren ages 6 to 18, a rise of 80 percent in onlytwo years (Clay, 1997). Prozac now comes in pep-permint flavor especially designed for children')

As with adult depression, fOJ:child depression theusual recommendation (if not the usual practice) isto use antidepressant medic'Jon only as an adjunctto other forms of treatment.[Many factors thatcon-tribute to depression-stressful life circumstances, .poor parent-child relationships, family conflict anddissolution, low self-esteem, arid negative cognitivebiases, for example=-cannot be' changed by psy-chopharmacologj)and can be better addressed by.psychotherapy with .the individual child or the fam-ily (Dujovne, Barnard, & Rapoff, 1995).

Psychodynamic PsychotherapyPsychodynamic treatments for depression focus.broadly on problems in underlying personality or-ganizarion, tracing these back to the negative child-hood experiences from which depression emerges ',

(The goals of therapy are to decrease.self-criticismand negative self-representations and to help thechild to develop more adaptive defense mechanismsin order to be able to continue on a healthy courseof emotional developmentjwith younger c~dr~n,the therapist may use playas a means of bringing

d>.

these issues into the therapy room, with the focusshifting to d{scussion as children become more cog-nitively mature (Speier et al., 1995).

Psychodynamic approaches rarely provide out-come studies beyond individual case reports. How-ever, Fonagy an~ Target (1996) investigated theeffectiveness of aldevelop~entally oriented psycho-analytic approach with children. (This is described inmore detail in Chapter 17.) Results showed that thetreatment was effective, particularly for in~malizingproblems such as depression asd anxiety. oungerchildren (i.e., under II years) responded st. How-ever, the treatment was no quick cure; the best resultswere found when treatment sessions took place 4 to5 times per week over a period of two years.

Cognitive Behavior,!l Therapy .An example of a cognitive-behavioral approach isthe Coping with Depression Course for Adolescence(Lewinsohn et al., 1996), a downward extension ofa treatment program originally designed for adults.This intervention includes role playing to teach in-terpersonal and problem-solving techniques, cogni-tive restructuring to decrease maladaptive cogni-tions such as "Nothing ever turns out right for me,"and self-reinforcement techniqueyStudieS of the ef-fectiveness. of this approach show that, for the 80percent of adolescents who improve, treatment gainsare lasting. Cognitive behavioral therapies for childdepression are the most extensively researched, and,o:,erail, findings concerning their effectiveness are

. very positive (see Marcotte, 1997, and Southam-Gerow et aI., 1997, for reviews).

Family TherapyA comprehensive review by Dujovne, Barnard, &Rapoff (1995) examines the relative effectivenessof a number of different treatments for childhooddepression. They conclude that rc~IY-fOcuSed

.. treatments (family therapies) .warrant pnmary con-sideration, given the roles of the family sitwition.parent-child relationships, and parent depression in

. the developm~nt of.de?ress~ve sp~c~m disorders.Consistent With this, \Lewmsohn s group (1996)found that the effectiveness of their cognitive-behavioral intervention for depressed children wasenhanced by the addition of interventions with the

)

I -I -

Ii

Chapter 7 Disorders in tDe Depressive Spectrum and Child and Adolescent Suioce 161

';

parents. Group sessions are held in which parentsare given the opportunity to discuss issues relatedto depression and to .learn the same infrpersonalcommunication and conflict resolution skills beingtaught to their children.] .

Arr effective multifaceted approach to treatingchild depression in the family c9ntext is describedby Stark and associates (l996).l{nterventions withthe child include the use of individual and grouptherapy in order to increase positive mood and ex-pectations, restructure maladaptive schemata, andenhance social skills. Interventions aimed at thelarger system include parent training and familytherapy in order to reduce the environmental stresses.that contribute to the development of depression.Further, consultation with teachers is provided topromote and reinforce children's use of adaptivecoping strategies during the school daY;)

. Prevention"~ Efforts to prevent the development of childhood de-

pression have focused on those most at risk-namely, children of depressed mothers. For exam-ple, Gelfand and colleagues (1996) developed an in-tervention program for depressed mothers and theirinfants. Home visits were made by trained nurseswhose goals were to increase depressed mothers'parenting efficacy as well as,to foster more positivemother-infant interactions)Mothers who partici-pated improved in reported depression and per- .ceived stress, and both their own arid their infants'overall adjustment improved. Children of motherswho participated were also less avoidant in their at-tachment than other children; however, they werealsp more resistant. .

(An intriguing study by Malphurs and colleagues(1996) targeted another at-risk sample, depressedteenage mothers. Mothers were observed interact-ing with their infants and were differentiated interms of whether they demonstrated a withdrawn oran intrusive parenting style. Specific types of inter-ventions were designed t-;;tieIpcounter these prob-lematic patterns) For example, intrusive motherswere coached ~ imitate their children's behavior,thus giving children more opportunities to initiateand influence the flow of the interaction. In con-trast, withdrawn mothers were coached to keep theirr .

infant'S attention, thus increasing the level of mu-tual interest and engagement. Results suggest thateach specific coaching strategy improved the inter-actional behavior of the type of depressed motherfor whom .it ~.as de~elopey

_ J ,

Child and Adolescent Suicide

Definitions and pretalencecbs we begin our discussion of suicide, we must im- -mediatelydistinguish among three categories: suici-dalthoughts, suicide attempts, and completed suicide.

_JOur review' follo~ ~~land and Zigler,1993, ~x~ _.cept where noted.) SUlczdal thoughts,~nce consid-ered-to be rare in childhood, are in fact disconcert-ingly prevalent. Studies of U.S. high school studentshave found that 63 percent experienced suicidalthoughts, while 54 percent of college students hadc~nsidered suicide at least once in their Iives)(!)Suicide atte.0lZEGpicaily involve using a slow-acting method under circumstances in which dis-covery is possible. The act is most often in reac-tion to an interpersonal conflict or significant-stressor. Although the attempt 'is unsuccessfu9 it:-'may nevertheless be serialS, serving as "practice"for a futur.elethal attempt~pproxirnately, 7.per~ent . _of U,S. high school students. attempt SUICIdeIn agiven year (Centers for Disease:Control, 1995), andthere are reports of repeated and apparently serious~'~'attempts at suicide ~ng preschopler~Rosenthal& Rosenthal, I984).(further, while as many as 10percent of college students report having made asuicide attempt, only 2 percent of those had soughtmedical or psychological help)Therefore, our sta-tistics on the prevalence of suicide attempts may beunderestimates.

(j)Coppleted suicide, while, rare,is a significantproblem among adole~uicide is the thirdleading cause of death among 15- to 19-year~oldadolescents in the United States, in line behind ac-cidents and hornicidesl (Garland & Zigler, 1993).

(Further, suicide among the young is increasing at an. 'IV' cQ.Salarming rate, with rates rising much more dramat-ically than in the general populationjwhile suicidein the general population has increased 17 percent

d.

162 [)eolejelplTlenta sychopa ology .

201816

5uidde rate per 14100,000 in pop~ . 12.

108642O~-r----r----.----r----.----~---.----.----.----'--

-e- White males

1960--62 63-65 66-68 69-71 72-74 75--77 71h'10. 8,.,83 84--86. 87-M

-0- NonWhite males

___ WhIte females -{}- Nonwhite females

Figure 7.5 U.S. Youth Suicide Rates for 1S- to 19-Year-Olds by Race and Gender.

Source: Garland and Zigler. 1993

since the 1960s, among adolescents it has increased200 percent, to 11.3 per 100,000. (See Figure 7.5.)Rates for younger children are lower: In 1991, thesuicide rate for .children aged 5 to 14 was 0.5 per100,000. A protective factor for younger childrenmay be that they have.more, ..difficulty. accessinglethal means; consequently, there are 14.4 attemptsfor every completed suicide in'lO- to ll-year-olds(Pfeffer et al., 1994). Most suicides (70 percent) OC-cur in the horne. Firearms are the most frequentmethod used b~ both males and females (59 per-.cent), followed by hanging for males and drug in-gestion for females.tIn all age groups, females are more likely than

males to attempt suicide, while males are more likelyto succeed. Females attempt suicide at least threetimes as often as males do, whereas males ~omp1etesuicide about four times as often as females) The ex- .planation for this appears to lie in the choice ofmethod. In contrast to male suicides, two-thirds ofwhom die by self-inflicted gunshot wounds;the typ-ical young female attempter ingests drugs at home.The latter case is called low-lethality behavior be-cause of the length of time needed for the method totake effect and the likelihood that someone will findthe attempter before it is too late to resuscitate. Itshould not be assumed, however, that young womenrare less serious about wanting to die. Females are

more likely to have an aversion to violent methods,and sometimes young people's understanding ofhow deadly a drug can be is simply inaccurate. Fur-ther, these statistics often come from mental healthclinics and ignore one very important group incar-

.~ .males.·lf W~ illcl.I.Ided...waks In.juvaUkdetention facilities in these statistics, the genderdifferences in suicide attempts might not be so great.

Etiology

The Intrapetsonal ContextPsychological characteristics distinguish some ado-lescent suicides, the majority of whom have a di-agnosable psychopathology (Beautrais, Joyce, &Mulder, 1996). For example,(83 percent of youthswith suicidal ideation show signs of depression. The.relationship between sllifide and depression is a sig-nificant but complex onf. Most depressed youths are

.:not suicidal. Further, while Harrington 'and co1-leagues' (1994) longitudinal research s~s that

_childhood -depression is a strong predictor. of at-tempted suicide in adulthood, the key seems to bethe association between childhood depression andadult depression. In other words, depressed childrenwho grow up to be nondepressed adults are not atrisk for suicide.

I I

I

!i -II -!

_ J ,

Chapter 7 Disorders-in the Depressive Spectrum and Child and Adolescent Suicide 163

There are other significant predictors of youthsuicide besides depression. Clues to these can befound in a study of 3,000 suicidal youths attendinga free medical clinic (Ado links, 1987). Feelingstates preceding theirsuicide attempts were angerfirst, followed by loneliness, worry about the future,remorse orsharne, and hopelessness. The reasonsyouths gave for attempting suicide were; ill order,relief from an intolerable state of mind or escape )from. an impossible situation, making people un-derstand how desperate they feel, making someonesorry or getting back at someone, trying to influ-ence someone or change someone's mind, showinghow much they loved someone or finding outwhether someone really loved them, and seekinghelp. Many had been preoccupied with thoughts ofdeath for an extended period of time, but onlyaround half of the adolescents said they actuallywanted the attempt to succeed. Typically, despite thelong, gestation period of suicidal thoughts, the ac-tual attempt was made with little-premeditation.

These data have two important implications.First, impulsivity is implicated in suicide. Impul-sivity may be seen in many ways, including lowfrustration tolerance and lack of planning, poorself-control, disciplinary problems, poor academicperformance, and risk-taking behavior. Substance

• abuse is found in 15 to 33 percent of suicide com-pIeters, with suicidal thoughts increasing after theonset of substance use. Substance abuse may play"a: role in increasing impulsivity, clouding judgment,

vand disinhibiting self-destructive behavior. Otherdisorders of impulse control. including eating dis-orders, are also related to an increased risk for sui-cide (Berman & Jobes, 1991).

Secondly, ~nger and aggression emerge as animportant part of the suicide constellation) About

.70 percent of suicidal youth exhibit conduct.disor-.rler and antisocial behavior (Berman & Jobes,

1991). Childhood conduct disorder also has been'shown to predict adult suicidality independently ofdepression (Harrington et al., 1994). Achenbachand colleagues' (1995b) six-year longitudinal studyalso shows that suicidal ideation is predicted, notby depression, but by earlier signs of externalizingdisorders: for boys, intbe.form of aggressiveness,and for girls, in the form. of delinquent behavior.

The link between conduct disorder and suicide mayalso be strongest for boys, with the combination ofdepression and conduct problems particularly toxic.

(f Capaldi (l9?2) found that, among boys whoshowed a combination of depression and aggres-sion, school failure, poor relationships with parentsand peers, and low self-esteem resulted in suicidalideation two years later)

However, it is important to recognize that suici-dal adolescents are a diverse group. Some may ex-hibit no apparent problems or disorders. They mayappear to be "model" youth who keep their anxiety,perfectionism, and feelings of failure to themselves.

The Interpersonal Context:Family InfluencesThe family context is also important, although a sig-nificant weakness of many family studies is that theyare retrospective rather than prospective, (Here wefollow the review by Wagner, 1997. unless other-wise noted.) Assessing suicide only after it is at-tempted does not provide convincing evidence that .family-factors lead up to suicide.rA number of studies have confirmed the exis- .:tence of a significant degree of family dysfunctionand adverse childhood experiences among suicideattempters) (Beautrais, Joyce, & Mulder, 1996).Prospective studies show that suicidal ideation andsuicide attempts are predicted by low levels of par-ent warmth, communicativeness, support, and emo-tional responsiveness. and high levels of violence,disapproval, harsh discipline; abuse, and generalfamily conflict. Retrospective studies show that at-tempters and their parents describe the family ashaving lower cohesion, less support. and pooreradaptability to change. Suicide attempters also aremore likely to report feeling that they are unwanted ..or burdensome to their families. There is a signifi-cantly high level of psychopathology 'among familymembers, particularly suicide and depression.

Perceived lack of support from parents also hasbeen implicated as a significant predictor of ado-lescent suicidal thinking (Harter & Marold, 1994).Further, Harter and Whitesell (1996) found thatthe depressed youths exhibiting the least suicidal .ideation were those who perceived themselves tohave more positive relationships" with parents and

164 Developmental Psychopathology

more parent support. Thus supportive parent-childrelationships IDf-yprovide a buffer against suicidal-ity in at-risk children. .

Finally, exposure to the suicidal behavior ofanother person in the family or immediate socialnetwork is more 'Common in suicidal adolescentsthan in controls. This has been referred to as a con-tagion effect: Children who are exposed to suicidalbehavior, especially in family members or peers,are more likely to attempt it themselves. Such ex-posure should be regarded as accelerating the riskfactors already present rather than being a sufficientcause of suicide.

The Interpersonal Context:Peer RelationsPerceived (Jackof peer supportJ(HfI1er & Marold,1994) and~or social adjustmenj (Pfeffer et al.,1994) have been identified as risk factors. Suicidalyouth are more likely than others to feel ignoredand rejected by peers. They also report havingfewer friends and are concerned that their friend-ships are contingent-that they must behave a cer-tain way in order to be accepted by agemates. Per-ceived social failures, rejection, humiliation.' andromantic disappointments are common precipi-tants of youth suicide.

The Superordinate ContextSocioeconomic disadvantage has also been associ-ated with suicide (Beautrais, Joyce, & Mulder, 1996),with children growing up in poverty being at greaterrisk for suicidal fu.oqghts, attemp~, and completi~ns.

~~"( ..s~U~ "'.t)"",,-~ .s~~~ '> ~""'1 trt ~~ ~ ~~\

Developmental CourseThe question often arises as to whether youn/chil-dren who attempt suicide are really trying to. killthemselves, and therefore whether their attemptswarrant serious concern or presage future suicidal- .ifY:Doubt about whether they really intend to die issupported by cognitive-developmental research onchildren's limited understanding of the concept ofdeath, as weU as studies showing that suicidal chil-dren have a .limited understanding of the perma-nency of death (Cuddy-Casey & Orvaschel, 1997).However, ~ngitudinal research is consistent in

c.' .. '

showing that childhood suicide attempts are a strongpredictor of spbsequent attempts and completions.For example, \';ix to eight years after their first at-tempt, suicidal children were six times more likelythan other children to have made another suicide

_attempY:Pfeffer et,al:, 1994). Most subse~~~nt at-tempts occurred Within two years of the initial a~- .tempt, and over half of those who continued to be4uicidal made multiple attempts. Therefore, suicideattempts in children should not be dismissed as mereattention-getting behavior, since those who engagein them are at risk for more serious attempts andpossible completions in the future.

Among(adolescent suicide atternpters, Adolinks(1987) found that the majority improved within onemonth. However, about one-third subsequently ex-perienced major difficulties in the form of increasedpsychological and physical disorders, interpersonalproblems, and increased criminal behavior. One inten repeated the attempt, with boys succeeding more • ;often than girls, The risk for future disturbances wasparticularly strong ill teenage males)

Integrative Developmental ModelsA classic reconstructive account is provided by Ja-cobs (1971), who investigated fifty 14~ to 16-year-olds who attempted suicide: A control sample ofthirty-one subjects, matched for age, race, sex, andlevel of mother's education was obtained from a 10- 'cal high schooL Through an intensive, multi-tech-nique investigation, Jacobs was able to reconstructa five-step model of the factors leading up to suici-dal attempts:

1. Long-standing history of problems from earlychildhood. Such problems included parental di- .vorce, death of a family member, serious illness,parental alcoholism. and school failure, Subse-quent research has shown that it is ahigb level

- of intrafamilial conflict along with a lack of sup-port for the child that is the riskfactor, not a par-

, . ticular family consteUation such as divorce orsingle parenthood (Weiner, 1992).

2. Acceleration of problems in adolescence. Farmore important than earlier childhood problemswas the frequency of distressing events occur;ring within the last five years for the suicidal

Chapter 7 Disorders in the Depressive Spectrum and Child and Adolescent Suicide 165

youths; for example, 45 percent had dealt with -divorce in the previous five years as comparedto only 6 percent of the control group. Terrni-\nation of a serious romance was also much

- higher among the suicidal group, as were arrestsand jail sentences.

3. Progressive failure to cope and isolation frommeaningful social relationships. The suicidal andcontrol groups were equally rebellious in termsof becoming disobedient, sassy, and defiant.However, the coping strategies of suicidal ado-lescents were characterized much more by with-drawal behavior, such as avoiding others and en-gaging in long periods of silence (see also Spiritoet al., 1996). The isolation in regard to parentswas particularly striking. For example, while 70percent of all suicide attempts took place in thehome, only 20 percent of those who reported theattempt had informed their parents about it. Inone instance an adolescent telephoned a friendwho lived miles away, and he, in turn, telephonedthe parents who were in the next room.

4. Dissolution of social relationships. In the daysand weeks preceding ~e attempt, suicidal ado-lescents experienced th~rea1cing off of social re-lationships, leading to the feeling of hopelessness.

5. Justification of the suicidal act, giving the ado-lescent permission to make the q,ttempt.This jus-tification was reconstructed from 112 suicidenotes of adolescents and adults attempting andcompleting suicide. The notes contain certain re-curring themes; for example, the problems areseen as long-standing and unsolvable, so deathseems like the only solution. The authors of suchnotes also state that they know-what they are do-ing, are sorry for their act, and beg indulgence.The motif of isolation and subsequent hopeless-ness is prevalent.

r _

Another comprehensive account of the develop-ment of suicidal ideation is offered b1Earter (Har-ter, Marold, & Whitesell, 1992; Harteree Marold,1991, 1994), who integrates her own research withthat of others. Her model reconstructs the succes-sive steps that ultimately eventuate in suicidalideation in a nonnative sample of 12- to IS-year-olds. (See Figure 7.6.)

Immediately preceding and highly related to sui-cidal ideation is what Harter calls the depressioncomposite, which is made up of three interrelatedvariables: low global self-worth, negative affect, andhopelessness. The first two are highly correlated-the lower the perceived self-worth, the greater thefeelings of negative mocd, _ J ,

Moreover, the depressive composite is rootedboth in the adolescents' feelings of incompetenceand in their lack of support froln family andfriends. These two variables of. competence andsupport are, in turn, related in a special way. Inregard to competence, physical appearance, peerlikability, and athletic ability are related to peersupport, while scholarly achievement and behav-ioral conduct are related to parental support. Fi-nally, adolescents identify more strongly withpeer-related competencies, with the others beingregarded as more important to parents than tothemselves.

Analyses of the data revealed that peer-relatedcompetencies and support were more stronglyrelated to the depressive composite than wereparental-related competencies and support, per-haps because the former are more closely con-nected with the adolescents' own self-concept.However, parental support was important in dif-ferentiating the adolescents who were only de-pressed from those who were depressed and hadsuicidal ideation. Further, the quality of supportwas crucial, Regardless of the level, if adolescentsperceived they were acting only to please parentsor peers, their self-esteem decreased and depres-sion and. hopelessness increased. On the other.hand, uncoriditional support helped adolescentsminimize the depressive composite.

In regard to the question of which came first, low-ered self-worth or depression, the data indicate thatcausationcan go in either direction. Some adolescentsbecome depressed when they experience .loweredself-worth, while others become depressed over otheroccurrences such as rejection or conflict, which inturn lower self-worth.

To answer the question, Why adolescence? Har-ter and colleagues (1997) marshal a number of find-ings concerning this period. IIi adolescence, self-awareness, self-consciousness, intiospection, and

166 Developmental Psychopathology

PIiYSl~~P~~RAN(_E.· " c'PEER LlKfJ'.BIUTY· 1 _ •

ATHLETIC CdMPETENCE :.

Competence/Adequacy plus~opefrn,!ess (If Important) fon

Levelof plusHopelessness about;

P~ER SUP.PORT

-Level cif plus' .:'~~I.e~sness about

PARENT SUPP.ORT

-t:ompe&li£elMequ~cY plus-Ho~ele'mie~ (If Important) for:

~~~~i'&;~p~c~-liEHAVIOlto\L.CONDUq, .~.~- .•...~~;; :.:.;:' ~ -,: ~

D.epresslon Composite:."7' ~ . :." t

SEu:--WORTH·AFFECT,GENERAL HOPELESSNES.S

SUIOOAl- . loiAffON -

Risk Factors for Adolescent Suicidal Ideation.Figure 7.6

Source: Harter. Marold. and Whitesell, 1992

preoccupation with self-image increase dramatically,while self-esteem becomes more vulnerable. Peersupport becomes significantly more salient, althoughadolescents still struggle to remain connected withparents. For the first time, the adolescent can graspthe full cognitive meaning of hopelessness, while af-fectively there is an increase in depressive sympto-matology. Suicidal ideation is viewed as an effort tocope with or escape from the painful cognitions andaffects of the depressive. composite.

Interventionfrhe vast majority of suicidal adolescents provide

clues as to their imminent behavior; one study foundthat 83 percent of completers told others of their sui-cidal intentions in the week prior to their death. (Ourpresentation follows Berman and Jobes, 1991.)Most of the time such threats are made to familymembers or friends, who do not take them seriously,

try to deny them, or do not understand their irnppr-tance. Friends, for example, might regard reportingthe threats as a betrayal of trust. Thus, not only do'adolescents themselves not seek professional help,but those in whom they confide tend to delay or re-

. sist getting help. Consequently, an important goalof prevention is to educate parents and peers con-cerning risk signs.

Once an adolescent comes for professional help,'the immediate therapeutic-task is to protect the youthfrom self-harm through crisis intervention. Thismight involve restricting access to the means of com-mitring suicide, such as removing a gun from thehouse or pills from the medicine cabinet; a "no sui- -.cide contract" in which the adolescent agrees not tohurt himself or herself for an explicit time-limitedperiod; decreasing isolation by hav[;gsympatheticfamily members or friends with the adolescent at alltimeszgiving medication to reduce agitation or de- .pression; or, in extreme cases, hospitalization.

.' .

Chapter 7 Disorders in the Depressive Spectrum and Child and Adolescent Suicide 167

•

Suicide PreventionTurning again to Garland and Zigler's (1993) re-view, we find that two of the most commonly usedsuicide prevention efforts-suicide hot lines andmedia campaigns=-are only minimally effective.Communities with suicide hot lines have slightly re-

- duced suicide rates; however, hot lines tend to beutilized by only one segment of the population, Cau- .casian females. Even less helpful, well-meaning ef-torts to call media attention to the problem of sui-cide among teenagers may have the reverse effect.Several studies have shown increased suicide ratesfollowing television or newspaper coverage of sui-cide, particularly among teenagers.

School-based suicide prevention programs areextremely popular, with the number of schools im-plementing them increasing 200 percent in recent -,years. Goals of these programs are to raise aware-ness of the problem of adolescent suicide, train par-ticipants to identify those at risk, and educate youthabout community resources .available to thern.:However, a number of problems have been identi-fied with school-based suicide prevention efforts.For one thing, they may never reach the popula-tions most at risk because incarcerated youths, run-aways, and school dropouts will never attend theclasses. Even when students do attend the pro-

.grains, there are questions as to their-benefits. Theprograms tend to exaggerate the prevalence ofteenage suicide, while. at the same time de-ernpha-sizing the fact that most adolescents who attemptsuicide are emotionally disturbed. Thus, they ignoreevidence for th"e contagion effect and encourageyouth to identify with the case studies presented.By trying not to stigmatize suicide. these programsmay inadvertently normalize suicidal behavior andreduce social taboos against it.

•

••

Large-scale, well-controlled studies provide somebasis for these concerns. For example. one study of300 teenagers showed that attending a suicide pre-vention program slightly increased knowledgeabout suicide but was not effective in changing at-titudes about it. Boys in particular tended to change .. --in the undesirable-direction: more of them reportedincreased hopelessness and maladaptive coping af-ter expc\sure to the suicide program (Overholser etal., 198;). Another study of 1,000 youths found nopositive effects on.attitudes toward suicide. In fact,participation in the program was associated with asmall number of students responding that they now .thought suicide was a plausible solution to theirproblems. The students most at risk for suicide tobegin with (those who had made previous attempts)were the most likely to find the program distress-ing (Shaffer et al., 1991).

If suicide prevention programs are not the solution,what mightbe? While suicide is rare, Garland andZigler (1993) point out, the stressors and life prob-lems that may lead some youth to it are not. There-fore, successful prevention programs might be aimedtowardsuch risk factorsfor suicide as substance abuse,impulsivebehavior,depression, lack of social support,family discord, poor interpersonal problem-solvingskills, social isolation, and low self-esteem.

Recall that in Chapter 5 we dealt with the issue ofcontrol in the toddler-preschool period--control ofexcessive negativism and control of the bodily func-tions of eating and urination. We will now return tothis theme of control, exploring its manifestation in:the middle childhood period. We will examine twoextremes: excessive control, which is an important el-ement in anxiety disorders, and inadequate self-con-trol, which lies at the heart of conduct disorders.

, ..