Social support as a mediator between problem behaviour and gambling: a cross-sectional study among 14-16-year-old Finnish adolescents

Social suppo as a media o be ween p oblem beha iou and gambling: a c oss-sec ional s udy among 14–16-yea -old Finnish adolescen s Tiina Räsänen, 1 Tomi Lin onen, 2 Asko Tol anen, 3 Anne Konu 1 To ci e: Räsänen T, Lin onen T, Tol anen A, e al. Social suppo as a media o be ween p oblem beha iou and gambling: a c oss- sec ional s udy among 14–16-yea -old Finnish adolescen s. BMJ Open 2016;6:e012468. doi:10.1136/bmjopen-2016- 012468 ▸P epublica ion his o y o his pape is a ailable online. To iew hese iles please isi he jou nal online (h p://dx.doi.o g/10.1136/ bmjopen-2016-012468). Recei ed 30 Ap il 2016 Re ised 3 No embe 2016 Accep ed 8 No embe 2016 Fo numbe ed a ilia ions see end o a icle. Co espondence o D Tiina Räsänen; asanen. iina.a@s uden .u a. i ABSTRACT Backg ound: Du ing he adolescen pe iod, isk- aking beha iou inc eases. These beha iou s can comp omise he success ul ansi ion om adolescence o adul hood. The pu pose o his s udy was o examine social suppo as a media o o he ela ion be ween p oblem beha iou and gambling equency among Finnish adolescen s. Me hods: Da a we e ob ained om he na ional School Heal h P omo ion S udy (SHPS) om he yea s 2010 and 2011 (N=102 545). Adolescen s we e classi ied in he mos homogeneous g oups based on hei p oblem beha iou ia la en class analysis. Resul s: Pa h analysis indica ed ha social suppo was nega i ely associa ed wi h p oblem beha iou , and p oblem beha iou and social suppo we e nega i ely ela ed (excep o social suppo om iends among boys) o gambling. Social suppo om pa en s and school media ed, albei weakly, he ela ions be ween p oblem beha iou and gambling among gi ls and boys. Conclusions: P oblem beha iou may a ec gambling h ough social suppo om school and pa en s. Thus p e en ion and in e en ion s a egies should ocus on s eng hening adolescen s’social suppo . In addi ion, because o he clus e ing o di e en p oblem beha iou s ins ead o concen a ing on a single o m o p oblem beha iou mul iple-beha iou in e en ions may ha e a much g ea e impac on public heal h. INTRODUCTION Du ing he adolescen pe iod, isk- aking beha iou inc eases. 1 Risk- aking includes beha iou s ha a e done unde one’sown oli ion, ha e unce ain ou comes and can comp omise he success ul ansi ion om adolescence o adul hood. 23 Risk- aking beha iou s also mani es in he o m o p oblem beha iou s, ac ions ha a e socially disapp o ed and esul in social sanc ions. These beha iou s can co-occu ; hus, in ol e- men in one p oblem beha iou inc eases he isk o in ol emen in o he p oblem beha iou s, which is known as p oblem beha iou synd ome. 45 Gambling may be a componen o he p oblem beha iou syn- d ome: p oblem gambling and egula gam- bling (a leas once a week) a e associa ed wi h conduc diso de , subs ance use and delinquency. 6–9 Howe e , i is no known i non-p oblem gambling and gambling ha is no necessa ily egula is ela ed o he p oblem beha iou synd ome. In Finland in 2010 gambling was allowed o adolescen s aged 15 yea s and o e (excep In e ne and casino gambling) and in 2011, 15-yea olds could legally gamble a slo machines (all o he gambling games had age limi o 18 yea s), which a e widely a ailable in Finland. Adolescen s ha e also had posi i e a i udes owa ds gambling in Finland. 10 Likewise, s udies ha e shown ha pa en s and eache s do no hink ha gambling is a g ea conce n among you h. 11 12 Acco ding o p oblem beha iou heo y, i is he balance be ween p o ec i e and isk S eng hs and limi a ions o his s udy ▪La ge scale, popula ion-based s udy explo ed he ela ionship be ween gambling, p oblem beha - iou and social suppo . ▪S udies ha e examined pa hways om social suppo o bo h gambling and o he p oblem beha iou s, bu esea ch ha ies o explain how adolescen s’access o social suppo accoun s o he associa ion be ween gambling equency and p oblem beha iou , is lacking. ▪Tempo al ela ionship be ween hese ac o s emains unclea , because o he c oss-sec ional na u e o he s udy. ▪I may be ha he inal sample did no include he adolescen s who we e engaging in he highes le els o gambling and he highes le els o p oblem beha iou . ▪Since his s udy was based solely on adolescen sel - epo s, he e is he possibili y o o e o unde - epo ing. Räsänen T, e al.BMJ Open 2016;6:e012468. doi:10.1136/bmjopen-2016-012468 1 Open Access Resea ch ac o s ha de e mines ones’in ol emen in p oblem beha iou . 5 Social suppo p o ided by indi iduals and ins i u ions can be desc ibed as in e pe sonal ela ion- ships ha influence an indi idual’s unc ioning. Social suppo is pa en s’closeness, moni o ing and ca ing, eache s’in e es in hei s uden s, and iends’suppo - i eness. 13 S udies ha e shown ha social suppo om pa en s, iends and school all play a c ucial ole in ado- lescen s’p oblem beha iou . 14–16 The e is also some e i- dence indica ing ha amily cha ac e is ics influence adolescen s’gambling. 17 Specifically, highe le els o pa en al a achmen , moni o ing and supe ision a e linked o dec eased gambling, while lowe le els o pa - en al us and communica ion a e associa ed wi h inc eased gambling. 18 Low le els o pa en al moni o ing o adolescen s is also associa ed wi h p oblem gambling du ing young adul hood. 19 Addi ionally, adolescen p oblem gamble s end o pe cei e lowe le els o amil- ial and pee suppo . 20 While p e ious s udies ha e examined sha ed p edic o s be ween gambling, p oblem beha iou s and pa en ing, he findings o hese s udies do no p o ide e idence ha pa en al supe ision o moni o ing a e ela ed o gambling as i is he case o o he p oblem beha iou s. 21 22 Mo eo e , iends’ app o al o gambling, associa ing wi h de ian pee s, associa ing wi h iends who gamble and ha ing iends who ha e gambling p oblems a e associa ed wi h adoles- cen gambling. 82324 Howe e , he only school cha ac e - is ics ha ha e been s udied in he con ex o gambling a e suspension a es, low school commi men and schools’ ewa ds o p osocial in ol emen . These we e ei he weakly o no a all associa ed wi h gambling. 25 26 S udies examining he associa ions be ween schools’ social suppo and gambling, a e lacking. Ea lie s udies ha e demons a ed ha he e a e gende di e ences in gambling as well as p oblem beha iou . Fo example, gambling is mo e common among men 27 and men also end o s a gambling a a younge age. 28 Addi ionally, men ha e a highe isk o hea y alcohol use and smoking. 29 30 Howe e , li le is known abou gende di e ences in ela ion o co-occu ence o gambling equency and p oblem beha iou s. I is also unclea i he e a e gende di e - ences in associa ions be ween gambling, p oblem beha - iou and social suppo . In s udies ha assess gambling, p oblem beha iou and isk/p o ec i e ac o s, a limi ed numbe o p oblem beha iou s (eg, gambling, subs ance use and delinquency) a e examined and he ocus is on pa en al moni o ing, indi idual ac o s (impulsi i y, mo al disen- gagemen ) and pee delinquency, 21 22 31 32 he eby excluding social suppo om impo an indi iduals and ins i u ions, like iends and school. P e ious s udies ha e examined pa hways om social suppo o bo h gambling and o he p oblem beha iou s, bu esea ch ha ies o explain how adolescen s’access o social suppo accoun s o he associa ion be ween gambling equency and p oblem beha iou , is lacking. Thus, he ole o social suppo in he associa ion be ween p oblem beha iou and gambling has no been in es i- ga ed. As ea lie s udies ha e shown ha gambling is associa ed wi h p oblem beha iou and he deg ee o social suppo ha an adolescen ecei es is ela ed o bo h gambling equency and p oblem beha iou ; i is easonable o expec ha social suppo plays an impo - an ole in he ex en o which p oblem beha iou s a e ela ed o gambling equency. The e o e, using a popula ion-based su ey o 8 h and 9 h g ade boys and gi ls, he aim o his s udy was o in es iga e he ela ions be ween p oblem beha iou and gambling equency, and social suppo as a po en ial media o o his associ- a ion (ie, se e as a p o ec i e ac o ). I is assumed ha highe social suppo would be nega i ely associa ed wi h p oblem beha iou and gambling. METHOD Se ing In Finland, compulso y basic comp ehensi e school s a s a age 7 (1s g ade) and ends a he age o 16 yea s (9 h g ade). Du ing he 8 h and 9 h g ade boys and gi ls a e be ween 14 and 16 yea s o age. The Finnish gambling sys em is based on a s a e mon- opoly; he p ofi s a e e u ned o socie y and used o common good. In addi ion, he games a e widely a ail- able in public places. The gambling age limi was se o 18 yea s in Oc obe 2010. The limi was also applied o slo machines in July 2011, and p io o ha he age limi was age 15 yea s. A he ime o su ey comple ion, he age limi o gambling was 18 yea s and 15 yea s o slo machines. Da a Da a we e ob ained om he na ional School Heal h P omo ion S udy (SHPS) in 2010 and 2011, which included ques ions on adolescen s’gambling, p oblem beha iou and social suppo . In 2010, he s udy was conduc ed in sou he n, eas e n and no he n Finland and in wes e n and cen al Finland in 2011. Thus, he 2010 and 2011 su eys co e all o Finland, and in each yea only a ew municipali ies e use o pa icipa e. 33 The municipali ies ha did no pa icipa e do no di e om he o he s and his pa o he d op-ou is no likely o pose a h ea o gene alisabili y. In 2010, 78% o Finnish 8 h and 9 h g ade s we e su eyed, and in 2011 81% o Finnish 8 h and 9 h g ade s pa icipa ed in he s udy. Responden s who had no answe ed o e hal o he ques ions, o had no exp essed hei sex o g ade, we e emo ed. Also hose who we e no in school when su ey was conduc ed o could no answe independ- en ly did no pa icipa e. 33 SHPS is a c oss-sec ional su ey adminis e ed h oughou he school day in he 8 h and 9 h g ades o lowe seconda y schools. Answe ing was olun a y. Pa en al consen was no equi ed acco ding o e hical guidelines in Finland; gen- e ally, su eys conduc ed du ing school days do no 2Räsänen T, e al.BMJ Open 2016;6:e012468. doi:10.1136/bmjopen-2016-012468 Open Access need pa en al pe mission. In o al, 102 545 8 h (mean age=14.9, SD=0.4) and 9 h (mean age=15.9, SD=0.4) g ade s uden s pa icipa ed in he s udy. Fo y-nine pe cen o su ey esponden s we e boys. Measu es P oblem beha iou P oblem beha iou was measu ed h ough 11 di e en o ms o p oblem beha iou s conce ning adolescen s’ uancy, bullying, delinquency, smoking, using snu , alcohol d inking, d unkenness- ela ed d inking, using alcohol and medicine o in oxica ion, using medicine o in oxica ion, glue-sni fing and d ug use. Mo e de ailed in o ma ion abou ques ions ega ding p oblem beha iou can be ound elsewhe e. 34 Since p oblem beha iou indica o included measu es wi h mul iple beha iou ypes and esponses, which we e no necessa - ily iden ically scaled, la en classes we e es ima ed. Based on la en class analysis (LCA) adolescen s we e di ided in o ou g oups. F om hese g oups hose who pa ici- pa ed in p oblem beha iou he mos and he leas we e selec ed om he da a o s udy ex emi ies o he p oblem beha iou phenomenon. Social suppo om iends Social suppo om iends was eques ed by ques ion: Do you ha e a his momen any close iend whom you can alk wi h confiden ially almos e e y hing? Al e na i es we e I do no ha e any iends, one close iend, wo close iends and se e al close iends. Social suppo om school Social suppo om school based on h ee s a emen s and wo ques ions abou schools’and eache s’suppo . S a emen s we e: eache s encou age me o imp ess my own opinions a classes, eache s a e in e es ed abou how I am doing and eache s a e ea ing s uden s ai ly. The scale o hese was o ally ag ee, ag ee, disag ee and o ally disag ee. S uden s we e also asked: do you ha e oubles ge ing along wi h you eache s. Al e na i es we e no a all, ai ly li le, ai ly much and e y much. As well adolescen s we e eques ed i hey had oubles wi h school wo k, how o en hey go help om school. Al e na i es we e always when I need i , mos o he ime, a ely and almos ne e . All o he s a emen s and ques ions we e coded om one o ou , ou indica ing good social suppo . F om hese fi e a iables he sum a iable was c ea ed, which had one componen s uc- u e wi h C onbach’sα alue o 0.7. Social suppo om pa en s Social suppo was examined by ou ques ions. S uden s we e asked do hei pa en s know all o his/he iends. This ques ion had al e na i es bo h pa en s do (coded as h ee), only mo he / a he knows (coded as one) and nei he o hem know (coded as ze o). Second ques ion was, do you pa en s know whe e you spen you F iday and Sa u day e enings, which had al e na i es hey know always (coded as h ee), know some imes (coded as one) and mos o he ime hey do no know (coded as ze o). Adolescen s we e also asked can hey alk wi h hei pa en s: almos ne e (coded as ze o), some imes (coded as one), ai ly o en (coded as wo) and o en (coded as h ee). The las ques ion was i adolescen s had oubles wi h school wo k, how o en hey go help om home. Al e na i es we e always when I need i (coded as h ee), mos o he ime (coded as wo), a ely (coded as one) and almos ne e (coded as ze o). These ou ques ions measu ing he social suppo om pa en s had one componen s uc u e, which had C onbach’sα alue o 0.6. Gambling Gambling was eques ed by one ques ion: How o en do you gamble? Al e na i es we e on 6–7 days pe week, on 3–5 days pe week, on 1–2 days pe week, less han once a week, less han once a mon h, I ha e no gambled du ing p e ious yea . S a is ical analysis Basic s a is ical analyses we e conduc ed wi h IBM SPSS S a is ics V.22, and s uc u al equa ion modelling was conduc ed wi h Mplus V.7.0. 35 Basic s a is ical analyses included χ 2 - es ( o ca ego ical da a) and Mann-Whi ney U es ( o con inuous da a), which a e used o de ec s a is ically significan di e ence be ween wo o mo e g oups. LCA was used o iden i y g oups o adolescen s based on hei in ol emen in di e en p oblem beha iou s. The pu pose was o iden i y he mos homogeneous g oups ha include ex eme classes o p oblem beha iou . LCA p o ides fi s a is ics and sig- nificance es s o assess he numbe o classes ha bes fi he da a. 36 To selec he numbe o classes o he model, he analysis was fi s un wi h one class, hen wi h wo and h ee classes and so on. This p ocedu e was ollowed un il o he highes possible numbe o classes we e un (which was in his case 11). Model solu- ions we e assessed by examining he en opy alues (nea 1.0), he Lo-Mendell-Rubin es (a p alue ha p o ides in o ma ion abou s a is ically significan imp o emen in fi o he inclusion o an addi ional class) as well as he in e p e a ion o he esul s. 37 Fo u he analysis, all indi iduals we e classified in o classes whe e class membe ship was based on he highes pos- e io p obabili ies. Based on he LCA esul s, pa h analysis was conduc ed o hose who pa icipa ed in he mos and he leas p oblem beha iou s. The analysis was conduc ed o examine i social suppo media es he associa ion be ween p oblem beha iou and gambling. In his model, gambling equency was eg essed on he h ee o ms o social suppo , and o ms o social suppo we e hen eg essed on p oblem beha iou (figu e 1). Analyses we e un sepa a ely o 8 h and 9 h g ade gi ls and boys gi en he possible di e ences in he ela ion be ween gambling and p oblem beha iou . All he Räsänen T, e al.BMJ Open 2016;6:e012468. doi:10.1136/bmjopen-2016-012468 3 Open Access models we e sa u a ed, so he model fi s we e necessa ily pe ec . RESULTS La en class analysis Adolescen s who did no gamble du ing he p e ious yea (37.3%) o had missing alues on he main ou come a iable (1.3%) we e excluded. This was ca ied ou because we we e only in e es ed in adoles- cen s who we e gambling. Fo he emaining 62 956 indi iduals, LCA was conduc ed sepa a ely o 8 h and 9 h g ade gi ls and boys o iden i y g oups o adolescen s based on hei p oblem beha iou s. Fo all 8 h and 9 h g ade gi ls and boys, he op imal model had ou classes wi h en opy alues anging om 0.88 o 0.83. The ea- sonably high en opies and he Lo-Mendell-Rubin es p alues sugges ed a good model fi o he ou class model. Howe e , among 8 h g ade gi ls and 9 h g ade boys, he p alues we e lowe han 0.05, indica ing ha he fi e o mo e class model was a be e fi . Howe e , a e aking in o accoun he en opies and he in e - p e a ion o he esul s, he ou class model was chosen o all pa icipan s ( able 1). Class 1 included adolescen s who pa icipa ed in all o ms o p oblem beha iou s a he highes equency (see e . 33). Fo example, adolescen in class 1 smoked a leas daily, d ank alcohol on a weekly basis and had ied d ugs a leas once. You h in class 2 also pa ici- pa ed in p oblem beha iou , bu hei pa icipa ion in p oblem beha iou s was mo e expe imen al. In addi ion, mos o he you h in class 2 did no use alcohol wi h medicine o in oxica ion, use medicine o in oxica ion and did no pa icipa e in glue sni fing. In class 3, ado- lescen s smoked obacco less o en han once a week and consumed alcohol only a couple o imes a mon h. The you hs in class 4 did no pa icipa e in any o m o p oblem beha iou . F om hese g oups, class 1 and class 4 we e selec ed, including adolescen s pa icipa ing egula ly in all examined p oblem beha iou s and hose no pa icipa ing in any, which allowed o he s udy o he ex emes o p oblem beha iou (N=29 870). Gambling, p oblem beha iou and social suppo The nex se o analyses compa ed he adolescen s who pa icipa ed in he mos p oblem beha iou and he adolescen s ha pa icipa ed in he leas p oblem beha - iou . Boys gambled mo e han gi ls did (χ 2 (4) =3280.182, p<0.001); howe e , pa icipa ion in p oblem beha iou was mo e common among gi ls (χ 2 (1) =1313.636, p<0.001) ( able 2). Gi ls expe ienced mo e social suppo om iends (χ 2 (4)=553.19, p<0.001), and boys ecei ed mo e suppo om school (U=88080721, p<0.001) and om pa en s (U=84026360, p<0.001) ( able 2). 9 h g ade s gambled (boys: χ 2 (4) =11 395 p=0.022), (gi ls: χ 2 (4)=25.416, p<0.001) and pa icipa ed in p oblem beha iou (boys: χ 2 (1) =126.024, p<0.001), (gi ls: χ 2 (1)=367.030, p<0.001) mo e o en han 8 h g ade s. Eigh h g ade boys also expe ienced mo e social suppo om iends (χ 2 (3) =25.933), p<0.001), school (U=52467017.0, p<0.01), and pa en s (51780278.5, p<0.001) han 9 h g ade boys. Eigh h g ade gi ls expe ienced mo e social suppo om school (U=9696512.5, p<0.001) and pa en s (U=9817557.0, p<0.01), bu he e we e no s a is ical di - e ences in social suppo om iends ( able 2). Among gi ls who gambled on weekly basis, 50% o 8 h g ade s and 74% o 9 h g ade s also pa icipa ed in p oblem beha iou . Among he boys, he pe cen age was 14% o 8 h g ade s and 22% o 9 h g ade s. Pa h analysis The s anda dised pa ame e es ima es (β) shown in able 3 indica ed ha p oblem beha iou was signifi- can ly and posi i ely ela ed o adolescen s’gambling equency (pa h c). The βindices also showed ha he e was a significan nega i e associa ion be ween p oblem beha iou and he di e en ypes o social suppo ; howe e , among 9 h g ade gi ls, social suppo om iends was no s a is ically significan (pa hs d– ). Social suppo om pa en s and school we e nega i ely ela ed wi h gambling (pa hs g and i). Among boys, social suppo om iends was significan and posi i ely associa ed wi h gambling; howe e , among gi ls social suppo om iends was significan ly and nega i ely associa ed wi h gambling (pa h h) ( able 3; also see figu e 1). The s a is ically significan albei e y small indi ec e ec s indica ed ha social suppo media ed he ela- ionship be ween p oblem beha iou and gambling e- quency, bu social suppo om iends did no media e his associa ion. The o al e ec o p oblem beha iou Figu e 1 Pa h model o he associa ions be ween he social suppo om iends (SSF), social suppo om school (SSS), social suppo om pa en s (SSP), p oblem beha iou and gambling equency among 8 h and 9 h g ade boys and gi ls. 4Räsänen T, e al.BMJ Open 2016;6:e012468. doi:10.1136/bmjopen-2016-012468 Open Access on p oblem gambling equency consis ed o a di ec e ec and an indi ec e ec h ough social suppo ( able 3). Fo y- wo pe cen and 49% o he a iance (R 2 ) was explained by he associa ion be ween gambling and social suppo o 8 h g ade and 9 h g ade boys, espec i ely. Fo 8 h g ade gi ls, 63% o he a iance was explained and 69% o he a iance was explained o 9 h g ade s. DISCUSSION This la ge scale, popula ion-based s udy explo ed he ela ionship be ween p oblem beha iou , gambling and social suppo among 14–16-yea -old adolescen s. The main pu pose o he s udy was o examine whe he social suppo would media e his associa ion. Social suppo om pa en s and school media ed weakly he ela ions be ween p oblem beha iou and gambling among gi ls and boys. Social suppo om iends did no media e his associa ion. Thus, he da a sugges ha he ela ion be ween p oblem beha iou and gambling can be explained only by he deg ee o social suppo ecei ed om school and pa en s; he e may be o he issues o phenomena ha a e mo e impo an in media - ing his associa ion. In acco dance wi h he Mun e al 38 esea ch, his s udy indica ed ha p oblem beha iou did clus e a he bo h ends o he li es yle spec um (ie, ‘p oblem’o ‘non- p oblem’beha iou pa e ns among adolescen s). Wi hin hese wo classes, he e we e adolescen s who pa icipa ed egula ly in all o ms o p oblem beha iou s and hose who did no pa icipa e in any p oblem beha- iou s. Addi ionally, his s udy ound ha p oblem beha iou and gambling we e associa ed, wi h p oblem beha iou inc easing as he equency o gambling inc eased. This is simila o p e ious s udies. 72232 These esul s may indica e ha in ol emen in one p oblem beha iou also inc eases he isk o in ol emen in o he p oblem beha iou s, as demons a ed by p oblem beha - iou heo y. 413 To he bes o ou knowledge, his is he fi s s udy o examine whe he social suppo om school is asso- cia ed wi h gambling equency. Ea lie s udies ha e assessed school cha ac e is ics, such as suspension a es, school commi men and schools’ ewa ds o p osocial in ol emen . 25 26 In he cu en s udy, social suppo om school was significan ly ela ed o gambling e- quency, wi h lowe suppo ela ed o highe gambling pa icipa ion. In addi ion, low social suppo om pa en s and iends was ela ed o inc eased gambling, which is consis en wi h he findings o Ha doon e al. 20 In hei s udy, hey epo ed significan di e ences be ween gambling se e i y g oups in hei le els o amily and pee suppo . Specifically, hei esul s indi- ca ed ha non-gamble s and social gamble s expe i- enced mo e social suppo han a - isk and pa hological gamble s. Howe e , in he cu en s udy, his pa e n was ound only o emale adolescen gamble s. Fo boys, Table 1 Resul s o he LCA 8 h g ade boys 9 h g ade boys 8 h g ade gi ls 9 h g ade gi ls n o classes en opy Lo-Mendell- Rubin n o classes en opy Lo-Mendell- Rubin n o classes en opy Lo-Mendell- Rubin n o classes en opy Lo-Mendell- Rubin 1 class model 19 471 NA NA 21 711 NA NA 9432 NA NA 12 342 NA NA 2 class model c1=6805 c2=12 666 0.86 0.00 c1=25 768 c2=37 188 0.85 0.00 c1=3591 c2=5841 0.86 0.00 c1=5445 c2=6897 0.82 0.00 3 class model c1=1975 c2=6813 c3=10 683 0.90 0.00 c1=9482 c2=26 304 c3=27 170 0.89 0.00 c1=1734 c2=3706 c3=3992 0.88 0.00 c1=2645 c2=3707 c3=5990 0.88 0.00 4 class model c1=586 c2=5377 c3=3065 c4=10 443 0.88 0.04 c1=911 c2=4734 c3=7233 c4=8833 0.86 0.00 c1=513 c2=2123 c3=3240 c4=3556 0.86 0.00 c1=1473 c2=4145 c3=3169 c4=3555 0.83 0.00 5 class model c1=567 c2=2829 c3=4543 c4=2956 c5=8576 0.81 0.70 c1=503 c2=2290 c3=6086 c4=4142 c5=8690 0.85 0.00 c1=138 c2=1165 c3=1641 c4=3013 c5=3475 0.86 0.00 c1=159 c2=1910 c3=2865 c4=3543 c5=3865 0.84 0.76 Bold cha ac e s ep esen hose pa icipa ing in p oblem beha iou he mos and he leas . LCA, la en class analysis; NA, no applicable. Räsänen T, e al.BMJ Open 2016;6:e012468. doi:10.1136/bmjopen-2016-012468 5 Open Access social suppo om iends was posi i ely ela ed o gam- bling equency. I migh be ha gambling is social ac i - i y o boys, bu o gi ls gambling may be an ac i i y ha subs i u es no mal social con ac s. P oblem gam- bling may occu in he absence o iends, whe eas non- p oblema ic gambling may be pa o a social pas ime. 20 Gambling was mo e common among boys, which is con- sis en wi h p e ious s udies; 623 howe e , o e 50% o 8 h g ade and 74% o 9 h g ade gi ls who gambled on a weekly basis pa icipa ed in p oblem beha iou . This indica es ha ac i e gambling among gi ls is s ongly linked o p oblem beha iou and hus a phenomenon ha should be s udied mo e ex ensi ely. Al hough he findings o his s udy indica ed ha social suppo media ed he associa ion be ween p oblem beha iou and gambling equency, he Table 2 Adolescen s’in ol emen in gambling and p oblem beha iou , and expe iencing SSF, SSS and SSP based on hei sex and g ade 8 h g ade boys 9 h g ade boys 8 h g ade gi ls 9 h g ade gi ls Boys s gi ls 8 s9 g ade Pe cen Gambling p<0.001 e a:0.278 p=0.022 e a:0.01 Less han once a mon h 38.4 34.1 73.7 69.9 Less han once a week 28.8 30.3 15.7 18.1 1–2 days pe week 18.7 19.6 5.0 6.2 3–5 days pe week 8.1 9.3 2.2 3.0 6–7 days pe week 6.1 6.8 3.4 2.8 Pa icipa ing in p oblem beha iou 5.3 9.3 12.6 29.3 p<0.001 e a: 0.210 p<0.001 e a: 0,14 SSF p<0.001 e a: 0.101 p=0.029 e a:0.015 Do no ha e any iends 13.7 15.3 7.1 6.7 One close iend 24.1 26.0 20.7 21.6 Two close iends 20.5 19.1 27.5 27.2 Se e al close iends 40.9 39.0 44.5 44.3 Missing alues 0.8 0.7 0.2 0.3 x (SD) SSS 2.8 (0.5) 2.7 (0.6) 2.7 (0.5) 2.7 (0.6) p<0.001 e a: 0.003 p<0.001 e a: 0.002 SSP 2.2 (0.6) 2.2 (1.2) 2.1 (0.7) 2.0 (0.7) p<0.001 e a: 0.005 p<0.001 e a: 0.001 n 11 029 9744 4069 5028 29 870 29 870 E a alues indica e he s eng h o associa ions. ss , social suppo om iends; ssp, social suppo om pa en s; sss, social suppo om school. Table 3 Pa h model o he di ec and indi ec e ec s be ween he ss , sss, ssp, pb and gambling equency (gb) among 8 h and 9 h g ade boys and gi ls 8 h g ade boys 9 h g ade boys 8 h g ade gi ls 9 h g ade gi ls ββββ Di ec e ec c (pb o gb) 0.35*** 0.36*** 0.34*** 0.36*** Pa h om pb o sss (pa h e) -0.30*** -0.36*** -0.43*** -0.46*** Pa h om pb o ss (pa h d) -0.09*** -0.07*** -0.08*** -0.01 Pa h om pb o ssp (pa h ) -0.37*** -0.39*** -0.19*** -0.50*** Pa h om sss o gb (pa h h) -0.11*** -0.13*** -0.06** -0.11*** Pa h om ss o gb (pa h g) 0.07*** 0.04*** -0.08*** -0.09*** Pa h om ssp o gb (pa h i) -0.09*** -0.08*** -0.10*** -0.06*** To al e ec c' (pb o gb) 0.41*** 0.43*** 0.42*** 0.44*** To al indi ec e ec c-c' 0.06*** 0.07*** 0.08*** 0.08*** Pa h om pb o sss o gb 0.03*** 0.05*** 0.03** 0.05*** Pa h om pb o ss o gb -0.01*** -0.00*** 0.01*** 0.00 Pa h om pb o ssp o gb 0.03*** 0.03*** 0.05*** 0.03*** ** p<0.01, ***p<0.001 6Räsänen T, e al.BMJ Open 2016;6:e012468. doi:10.1136/bmjopen-2016-012468 Open Access empo al ela ionship be ween hese ac o s emains unclea . Thus, addi ional esea ch is needed using longi- udinal da a. Addi ionally, i is impo an o no e ha while social suppo om iends, pa en s and school we e significan ly associa ed wi h gambling, he pa h coe ficien s we e low among boys and gi ls and may due o la ge sample size. Fu he mo e, pa h coe ficien s o he indi ec e ec s we e low. This may indica e ha social suppo does no play a la ge in he associa ion be ween p oblem beha iou and gambling. The e may be addi ional ac o s ha play a mo e impo an ole in his associa ion, such as ha ing iends o pa en s who: gamble, ha e a gambling p oblem o pa icipa e in o he p oblem beha iou s o indi idual ac o s such as low sel -con ol o conduc diso de . These ypes o ac o s ha e been examined in p e ious s udies. 7 8 18 21 23 24 Fu he mo e, gambling equency was assessed by a single i em, and he e was no alida ed p oblem gam- bling sc een used he ein. Howe e , he goal o he s udy was o s udy gambling in ol emen and i s associa ions wi h p oblem beha iou s and social suppo , no o iden- i y p oblem gamble s. The su ey did no collec add- i ional in o ma ion on gambling beha iou (eg, how much money is spen on gambling), no was in o ma- ion abou gambling ypes. Addi ionally, he defini ion o gambling was no gi en in he ques ionnai e, so ado- lescen s could in e p e gambling in many ways. 34 39 Also we did no use a alida ed measu e o social suppo . The in e nal consis ency o he scale o ‘social suppo om pa en s’was no op imal as indica ed by he low alue o C onbach’sα. Howe e , alidi y is sup- po ed by he ac ha he p incipal componen analysis indica ed one componen s uc u e. In addi ion, da a we e collec ed om 78% o Finnish 8 h and 9 h g ade s in 2010 and om 81% o Finnish 8 h and 9 h g ade s in 2011. Howe e , i may be ha he final sample did no include he adolescen s who we e engaging in he highes le els o gambling and he highes le els o p oblem beha iou . Mo eo e , since his s udy was based solely on adolescen sel - epo s, he e is he possibili y o o e o unde - epo ing. Despi e he limi a ions, media ion e ec s we e de ec ed, which ha e heo e ical impo ance as well as significan implica ions o in e en ion. Specifically, p e- en ion and in e en ion s a egies should ocus on s eng hening adolescen s’social suppo om pa en s and school. In addi ion, because o he clus e ing o di - e en p oblem beha iou s and he ac ha p oblem beha iou was linked o gambling, ins ead o concen a - ing on a single o m o p oblem beha iou mul iple- beha iou in e en ions may ha e a much g ea e impac on public heal h. 40 Au ho a ilia ions 1 School o Heal h Sciences, Uni e si y o Tampe e, Tampe e, Finland 2 Finnish Founda ion o Alcohol S udies, Helsinki, Finland 3 Depa men o Psychology, Uni e si y o Jy äskylä, Jy äskylä, Finland Con ibu o s TR, AT, TL and AK we e in ol ed in s udy concep ion and design. TR, AT, TL and AK analysed and in e p e ed he da a. TR d a ed he manusc ip . AT, TL and AK c i ically e ised he manusc ip . Funding This s udy was inancially suppo ed by he Finnish Founda ion o Alcohol S udies. The Finnish Founda ion o Alcohol S udies had no in ol emen in he s udy design, da a handling, w i ing o submission o he manusc ip . The g an s awa ded by he Finnish Founda ion o Alcohol S udies o gambling esea ch a e based on a go e nmen con ac o s udying gambling- ela ed ha m. Compe ing in e es s None decla ed. E hics app o al The e hical commi ee o Tampe e Uni e si y Hospi al. P o enance and pee e iew No commissioned; ex e nally pee e iewed. Da a sha ing s a emen No addi ional da a a e a ailable. Open Access This is an Open Access a icle dis ibu ed in acco dance wi h he C ea i e Commons A ibu ion Non Comme cial (CC BY-NC 4.0) license, which pe mi s o he s o dis ibu e, emix, adap , build upon his wo k non- comme cially, and license hei de i a i e wo ks on di e en e ms, p o ided he o iginal wo k is p ope ly ci ed and he use is non-comme cial. See: h p:// c ea i ecommons.o g/licenses/by-nc/4.0/ REFERENCES 1. Michael K, Ben-Zu H. Risk- aking among adolescen s: associa ions wi h social and a ec i e ac o s. J Adolesc 2007;30:17–31. 2. Ig a V, I win CE. Theo ies o adolescen isk- aking beha io . In: DiClemen e RJ, Hansen WB, Pon on LE, eds. Handbook o adolescen heal h isk beha iou . New Yo k: Plenum P ess, 1996:35–51. 3. Lea he NC. Risk- aking beha iou in adolescence: a li e a u e e iew. J Child Heal hca e 2009;13:295–304. 4. Jesso R, Jesso SL. P oblem beha io and psychosocial de elopmen : a longi udinal s udy o you h. New Yo k: Academic P ess, 1977. 5. Jesso R, Tu bin MS. Pa sing p o ec ion and isk o p oblem beha io e sus p o-social beha io among US and Chinese adolescen s. J You h Adolesc 2014;43:1037–51. 6. Ba nes GM, Wel e JW, Ho man JH, e al. Gambling, alcohol, and o he subs ance use among you h in he Uni ed S a es. J S ud Alcohol D ugs 2009;70:134–42. 7. Wel e JW, Ba nes GM, Tidwell MC, e al. Associa ion be ween p oblem gambling and conduc diso de in a na ional su ey o adolescen s and young adul s in he Uni ed S a es. J Adolesc Heal h 2009;45:396–401. 8. Go i M, Po en e R, Pi ino A, e al. Rela ionship be ween gambling se e i y and a i udes in adolescen s: indings om a popula ion-based s udy. J Gambl S ud 2015;31:717–40. 9. Yip SW, Desai RA, S einbe g MA, e al. Heal h/ unc ioning cha ac e is ics, gambling beha io s, and gambling- ela ed mo i a ions in adolescen s s a i ied by gambling p oblem se e i y: indings om a high school su ey. Am J Addic 2011;20:495–508. 10. Salonen AH, Cas én S, Raisamo S, e al. A i udes owa ds gambling in Finland: a c oss-sec ional popula ion s udy. BMC Public Heal h 2014;14:982. 11. Shead WN, De ewensky JL, Mee kampe E. You mo he should know: a compa ison o ma e nal and pa e nal a i udes and beha io s ela ed o gambling among hei adolescen child en. In J Men Heal h Addic ion 2011;9:264–75. 12. De e ensky JL, S -Pie e RA, Temche CE, e al. Teache awa eness and a i udes ega ding adolescen isky beha iou s: is adolescen gambling pe cei ed o be a p oblem? J Gambl S ud 2014;30: 435–51. 13. Jesso R, Tu bin MS, Cos a FM, e al. Adolescen p oblem beha io in China and he Uni ed S a es: a c oss-na ional s udy o psychosocial p o ec i e ac o s. J Res Adolesc 2003;13:329–60. 14. Deko icM. Risk and p o ec i e ac o s in he de elopmen o p oblem beha io du ing adolescence. J You h Adolesc 1999;28:667–85. 15. Fleming CB, Ca alano RF, Hagge y KP, e al. Rela ionships be ween le el and change in amily, school, and pee ac o s du ing wo pe iods o adolescence and p oblem beha io a age 19. J You h Adolesc 2010;39:670–82. 16. Ga ne ski N, Dieks a RF. Pe cei ed social suppo om amily, school, and pee s: ela ionship wi h emo ional and beha io al p oblems among adolescen s. J Am Acad Child Adolesc Psychia y 1996;35:1657–64. Räsänen T, e al.BMJ Open 2016;6:e012468. doi:10.1136/bmjopen-2016-012468 7 Open Access 17. McComb JL, Sabis on CM. Family in luences on adolescen gambling beha io : a e iew o he li e a u e. J Gambl S ud 2010;26:503–20. 18. Magoon ME, Inge soll GM. Pa en al modeling, a achmen , and supe ision as mode a o s o adolescen gambling. J Gambl S ud 2006;22:1–22. 19. Lee GP, S ua EA, Ialongo NS, e al. Pa en al moni o ing ajec o ies and gambling among a longi udinal coho o u ban you h. Addic ion 2014;109:977–85. 20. Ha doon KK, Gup a R, De e ensky JL. Psychosocial a iables associa ed wi h adolescen gambling. Psychol Addic Beha 2004;18:170–9. 21. Ba nes GM, Wel e JW, Ho man JH, e al. Sha ed p edic o s o you h ul gambling, subs ance use, and delinquency. Psychol Addic Beha 2005;19:165–74. 22. Wanne B, Vi a o F, Ca bonneau R, e al. C oss-lagged links among gambling, subs ance use, and delinquency om midadolescence o young adul hood: addi i e and mode a ing e ec s o common isk ac o s. Psychol Addic Beha 2009;23:91–104. 23. Cas én S, G ainge M, Lah i T, e al. A - isk and p oblem gambling among adolescen s: a con enience sample o i s -yea junio high school s uden s in Finland. Subs Abuse T ea P e Policy 2015;10:9. 24. Hanss D, Men zoni RA, Blaszczynski A, e al. P e alence and co ela es o p oblem gambling in a ep esen a i e sample o No wegian 17-yea -olds. J Gambl S ud 2015;31:659–78. 25. Lee GP, Ma ins SS, Pas ET, e al. Examining po en ial school con ex ual in luences on gambling among high school you h. Am J Addic 2014;23:510–17. 26. Scholes-Balog KE, Hemphill SA, Dowling NA, e al. A p ospec i e s udy o adolescen isk and p o ec i e ac o s o p oblem gambling among young adul s. J Adolesc 2014;37:215–24. 27. K is iansen S, Jensen SM. P e alence o gambling p oblems among adolescen s in he No dic coun ies: an o e iew o na ional gambling su eys 1997–2009. In J Soc Wel a e 2011;20:75–86. 28. Bu ge AN, Pie zak RH, Molina CA, e al. Age o gambling ini ia ion and se e i y o gambling and heal h p oblems among olde adul p oblem gamble s. Psychia Se 2004;55:1437–9. 29. Schul e MT, Ramo D, B own SA. Gende di e ences in ac o s in luencing alcohol use and d inking p og ession among adolescen s. Clin Psychol Re 2009;29:535–47. 30. Helldán A, Helako pi S, Vi anen S, e al. Heal h beha iou and heal h among he Finnish adul popula ion. Na ional Ins i u e o Heal h and Wel a e (THL), Repo 15/2013, 2013. h p://u n. i/URN: ISBN:978-952-245-931-2 31. Ba nes GM, Wel e JW, Ho man JH, e al. The co-occu ence o gambling wi h subs ance use and conduc diso de among you h in he Uni ed S a es. Am J Addic 2011;20:166–73. 32. Vi a o F, B endgen M, Ladouceu R, e al. Gambling, delinquency, and d ug use du ing adolescence: mu ual in luences and common isk ac o s. J Gambl S ud 2001;17:171–90. 33. Luopa P, Ki imäki H, Ma ikka A, e al. Nuo en hy in oin i Suomessa 2000–2013—Koulu e eyskyselyn ulokse . [Wellbeing o adolescen s in Finland 2000–2013.The Resul s o he School Heal h P omo ion s udy]. Na ional Ins i u e o Heal h and Wel a e (THL). Repo 25/2014. 34. Räsänen T, Lin onen T, Konu A. Gambling and p oblem beha io among 14- o 16-yea -old boys and gi ls in Finland. J Gambl Issues 2015;31:1–23. 35. Mu hén LK, Mu hén BO (1998–2015). Mplus use ’s guide. 7 h edn. Los Angeles, CA: 2015. 36. Nylund KL, Aspa ouho T, Mu hén BO. Deciding on he numbe o classes in la en class analysis and g ow h mix u e modeling: a Mon e Ca lo simula ion s udy. S uc Equ Model 2007;14:535–69. 37. Jung T, Wick ama KAS. An in oduc ion o la en class g ow h analysis and g ow h mix u e modeling. Soc Pe sonal Psychol Compass 2008;2:302–17. 38. Mun EY, Windle M, Schainke LM. A model-based clus e analysis app oach o adolescen p oblem beha io s and young adul ou comes. De Psychopa hol 2008;20:291–318. 39. Räsänen T, Lin onen T, Jo onen K, e al. Gi ls and boys gambling wi h heal h and well-being in Finland. J Sch Heal h 2015;85:214–22. 40. Nigg CR, Alleg an e JP, O y M. Theo y-compa ison and mul iple-beha io esea ch: common hemes ad ancing heal h beha io esea ch. Heal h Educ Res 2002;17:670–9. 8Räsänen T, e al.BMJ Open 2016;6:e012468. doi:10.1136/bmjopen-2016-012468 Open Access