Has the role of personal income in alcohol drinking among teenagers changed between 1983 and 2013: a series of nationally representative surveys in Finland

Has he ole o pe sonal income in alcohol d inking among eenage s changed be ween 1983 and 2013: a se ies o na ionally ep esen a i e su eys in Finland Tomi Lin onen, 1,2 Jaakko Ne alainen 2 To ci e: Lin onen T, Ne alainen J. Has he ole o pe sonal income in alcohol d inking among eenage s changed be ween 1983 and 2013: a se ies o na ionally ep esen a i e su eys in Finland. BMJ Open 2017;7: e013994. doi:10.1136/ bmjopen-2016-013994 ▸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-2016-013994). Recei ed 23 Augus 2016 Re ised 7 Feb ua y 2017 Accep ed 17 Feb ua y 2017 1 Finnish Founda ion o Alcohol S udies, Helsinki, Finland 2 Uni e si y o Tampe e, School o Heal h Sciences, Tampe e, Finland Co espondence o D Tomi Lin onen; omi.lin onen@ alkoholi u kimussaa io. i ABSTRACT Objec i e: A o dabili y is known o be a key de e minan o alcohol consump ion, possibly e en mo e impo an in adolescence. Rela ing adolescen d inking pa e n de elopmen s o e a pe iod o ime o ends in adolescen income can yield in o ma ion on he signi icance o pa en al con ol o adolescen income. Design: Biannual epea ed c oss-sec ional su eys. Se ing: Na ionally ep esen a i e samples. Pa icipan s: 14-yea -old adolescen s in he pe iod 1983–2013 in Finland (N=33 771). Resul s: Adolescen s’alcohol d inking pa e n was signi ican ly associa ed wi h hei disposable income. The OR o mon hly d unkenness e sus abs inence was 6.6 (95% CI 5.0 o 8.8) among gi ls and 9.0 (6.3 o 13.0) among boys in he highes income g oup compa ed wi h he lowes . Howe e , he associa ion be ween income and d inking pa e n weakened conside ably du ing he 30-yea pe iod. Conclusions: Disposable income has been a signi ican p edic o o adolescen alcohol d inking in he pas 30 yea s. Howe e , in he ecen yea s, he amoun o disposable money has dec eased in impo ance. INTRODUCTION A he popula ion le el, he p ice o alcohol is known o be a key de e minan o alcohol d inking in ensi y. 12 I can be assumed ha a o dabili y is e en mo e impo an in adolescence since hei pu chasing powe is significan ly lowe han in adul hood. In he ques o de e minan s o adolescen d in- king, 3 i may hus be use ul o look a he issue o alcohol a o dabili y among een- age s. In ac , changes o e ime in adoles- cen disposable income ha e been shown o be associa ed wi h adolescen d inking ends. 4 While popula ion income changes end o be ela i ely slow, changes in adolescen income ollow a di e en logic. 5 Fu he mo e, d inking la ge amoun s o alcohol un il d unk seems mo e a ec ed by so-called pe iod e ec s (eg, mac oeconomic changes, alcohol p ice changes) han d ink- ing small amoun s. 6 The p opo ion o Finnish 14-yea -olds epo ing ha hey do no d ink alcoholic be e ages dec eased om 60% in he ea ly 1980s o a ound 30% in 1997–1999, bu inc eased again o o e 60% in 2013. 7 A he o he end o he d inking pa e n spec um, he p opo ion o 14-yea -olds d inking un il d unk a leas once a week has s ayed be ween 0% and 2% du ing he 30-yea pe iod om 1983 o 2013. The mos no able changes ha e been obse ed in he p opo - ion o 14-yea -olds d inking un il d unk a leas once a mon h. Bo h in he ea ly 1980s and ea ly 2010s, a ound 5% o 14-yea -old gi ls and boys d ank un il d unk a leas mon hly. Howe e , in he peak yea o 1999, 16% o gi ls and 11% o boys epo ed d ink- ing un il d unk a leas once a mon h. 8 Simila findings on ecen dec eases in ado- lescen d inking ha e been obse ed in he UK, 9 he USA, 10 Sweden, 11 Aus alia 12 and many o he high-income coun ies. 13 S eng hs and limi a ions o his s udy ▪La ge na ionally ep esen a i e samples. ▪Thi y-yea ime se ies wi h he iming o he s udy, sampling and da a collec ion me hod simila h oughou he s udy pe iod. ▪Possible measu emen bias due o sel - epo ed indica o s. ▪The esponse a e declined o e he yea s om 83% in 1983 o 38% in 2013. ▪The use o eal income ha accoun s o he e ec s o in la ion. Lin onen T, Ne alainen J. BMJ Open 2017;7:e013994. doi:10.1136/bmjopen-2016-013994 1 Open Access Resea ch g oup.bmj.com on June 1, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om While se e al ac o s associa ed wi h adolescen alcohol ini ia ion ha e been s udied, 14 15 adolescen dis- posable income has ecei ed li le esea ch in e es , al hough i po en ially has significan implica ions o subs ance use. Recen ly, he Ins i u e o Alcohol S udies in he UK explo ed po en ial explana ions o declining you h d inking and wound up ocusing hea ily on a o dabili y. 16 E en hough he legal alcohol pu chase age is 18 yea s in Finland, adolescen s widely epo ha i is easy o pu chase alcoholic be e ages. 17 18 In add- i ion, he unde aged ha e he op ion o asking olde iends o pu chase he be e ages using hei money. Adolescen s’disposable income in wes e n coun ies is ypically ecei ed as allowances, gi s om pa en s and ela i es, o om household wo k and ex e nal pa - ime wo k. 51920 In he USA in 1997, a ound hal o hose aged 14–15 yea s we e employed a leas pa ime. 21 In he No dic coun ies, a su ey by Danske Bank in 2012 epo ed ha 19% o hose aged 15–17 yea s ea ned money om wo k ou side he amily. 22 Ve y ew s udies ha e analysed income among young people who s ill li e a home o he sociodemog aphic de e minan s a ec - ing income. 19 20 Mac o-le el de e minan s include changes in he na ional economy and in cul u al no ms. In an economic down u n, eenage s find i ha de o ge pa - ime jobs because o he highe le el o gene al unemploymen . 23 24 Mic o-le el de e minan s include age, gende , amily ype, place o esidence, income and he socioeconomic s a us o he amily. 25 In e es ingly, adolescen s om disad an aged backg ounds may ha e mo e money a hei own disposal han young people om be e -o amilies. 20 26 In Finland, adolescen s om single-pa en amilies seem o ha e significan ly highe income han hose om nuclea amilies when adjus ing o o he amily socioeconomic indica o s. 5 In Finland, he disposable income o 14-yea -olds has gene ally inc eased om a ound €6 in 1983 o €12 pe week in 2013, wi h a sligh dec ease in he ea ly 1990s associa ed wi h an economic dep ession 25 (figu es we e adjus ed o money alue changes). In he 1980s, boys aged 14 yea s had oughly 5% mo e money a hei dis- posal han same-aged gi ls, bu his gende gap widened o a ound 10% in he 2010s. In ac , s udies ha e sug- ges ed ha gende is mo e significan in explaining eenage s’incomes han socioeconomic indica o s like amily backg ound o place o esidence. 527 Family backg ound may play an impo an ole in alcohol d inking and disposable income among adoles- cen s. Mo e d inking among adolescen s om amilies wi h less educa ed pa en s has been iden ified; 428 howe e , an opposing obse a ion has been epo ed. 29 No di e ences in disposable income among Finnish 14-yea -olds by pa en al educa ion o p o essional s a us we e obse ed in a ecen s udy. 25 All in all, i seems ha when s udying he d inking and income o 14-yea -olds, amily backg ound mus be aken in o accoun . Since 14-yea -olds ecei e hei income mainly om hei pa en s, income may se e as a way o con ol hei o sp ings’d inking. Mo e insigh in o how disposable income a ec s adolescen d inking can be ob ained by looking a adolescen d inking pa e n de elopmen s o e a pe iod o ime, and ela ing he ends o changes in adolescen income. The cu en pape examines asso- cia ions be ween alcohol d inking and income among Finns aged 14 yea s in a 30-yea pe iod om 1983 o 2013. Since p e ious esea ch 30 31 has shown ha , in addi ion o d inking alcoholic be e ages pe se, d inking un il d unk is a significan de e minan o p oblems among adolescen s, he whole spec um o d inking om abs inence o weekly d unkenness is analysed he e. METHODS The s udy da a we e de i ed om a na ionwide moni o - ing sys em o adolescen heal h beha iou s in Finland, he Adolescen Heal h and Li es yle Su ey (AHLS). 78 Sel -adminis e ed ques ionnai es we e mailed o mu u- ally independen , na ionally ep esen a i e samples o eenage s; o he cu en s udy, we selec ed hose aged 14 yea s, as p e ious esea ch has indica ed ha changes in d inking pa e ns ha e been la ges in his age g oup. 6 Two o , ecen ly, h ee einqui ies we e sen o non- esponden s. In addi ion o he pape ques ion- nai e, a digi al e sion was a ailable on he in e ne in 2009, 2011 and 2013. Samples we e ob ained om he Na ional Popula ion Regis e Cen e and we e based on pa icula da es o bi h, so ha all Finns bo n on he sample days we e included. All p ocedu es pe o med in he s udy we e in acco d- ance wi h he e hical s anda ds o he ins i u ional esea ch commi ees and wi h he 1964 Helsinki decla - a ion and i s la e amendmen s o compa able e hical s anda ds. The su ey se ies o igina ed om 1977 and ex ends o he yea 2013. Each da a collec ion yea , he s udy plan and da a collec ion p ocedu e we e e iewed by e hics commi ees. In he ea ly yea s, he E hics Commi ees o he Depa men o Public Heal h a he Uni e si y o Helsinki and he Pi kanmaa Hospi al Dis ic , Finland, app o ed he s udy p o ocol. Recen ly, he E hics Commi ee o he Tampe e egion (Finland) app o ed he s udy p o ocol: filling in he ques ionnai e was conside ed as adolescen s’consen o pa icipa e. No pa en al consen was eques ed. In case he espon- den ’s pa en s wished o inspec he ques ionnai e, he esponden was ins uc ed o p esen he ques ionnai e o his/he pa en s be o e answe ing he ques ions. The mean age o esponden s was 14.6 yea s. Da a we e collec ed om Feb ua y o Ap il e e y o he yea . The iming o he s udy, sampling and da a collec ion me hod we e simila h oughou he s udy pe iod. Sample sizes a ied; he numbe s o esponden s anged om 3105 in 1989 o 8390 in 1997, while esponse a es a ied om 83% in 1983 o 47% in 2011 and 38% in 2013. 7 The esponse a es we e highe among gi ls. Al oge he , 15 761 boys and 18 010 gi ls esponded o he su eys. 2Lin onen T, Ne alainen J. BMJ Open 2017;7:e013994. doi:10.1136/bmjopen-2016-013994 Open Access g oup.bmj.com on June 1, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om The ques ionnai e consis ed o 12 pages each yea . The majo i y o he ques ions emained unchanged h oughou he yea s. The equency o alcohol use was in es iga ed wi h he ques ion ‘How o en do you use alcohol? T y o include also hose imes you consumed only small amoun s o alcohol’. The al e na i es we e ‘daily’,‘a ew imes a week’,‘once a week’,‘a ew imes a mon h’,‘abou once a mon h’,‘abou once in 2 mon hs’,‘3–4 imes a yea ’,‘once a yea o less e- quen ly’,‘I do no use alcohol’. D unkenness was mea- su ed wi h he ques ion ‘How o en do you use alcohol un il you a e eally d unk?’The al e na i es we e ‘once a week o mo e o en’,‘once o wice a mon h’,‘less e- quen ly’,‘ne e ’. Using hese wo ques ions, an indica o o d inking pa e ns was cons uc ed: non-d inke , d inks only small amoun s, d inks un il d unk occasion- ally, d inks un il d unk a leas once a mon h, d inks un il d unk a leas once a week. The measu es o d unkenness ha e been ound o be easonably eliable and alid. 32 33 The ques ion on weekly disposable income, ‘How much money do you ha e, on a e age, a you disposal pe week (no including housing, ood and clo hing expenses)?’, was kep exac ly he same h oughou he yea s. The esponse ca ego ies, howe e , we e changed o eflec changes in mone a y alue and he adop ion o he eu o in 2002. The esponses we e ans o med in o 2013 eu o using he cos -o -li ing index p o ided by S a is ics Finland. The yea ly dis ibu ions o all he a iables a e shown in able 1. The p opo ion o i ems missing da a in he ou come a iable (d inking s yle) was 1.7% while 2.8% o he esponses we e missing o he main explana o y a iable (disposable income). Be ween 1% and 3% o he in o ma ion was missing in he o he explana o y a iables, bu anged om 4% o 11% by s udy yea in he case o he a he ’s educa ion. In he eg ession models, he ORs o he g oup wi h missing da a on a he ’s educa ion mos esembled he g oup wi h lowes educa ion. Indi iduals wi h missing da a on any o he indica o s we e excluded om he models. The e ec i e sample sizes we e 13 543 o boys and 15 731 o gi ls, 87% o he su ey esponse. Sepa a e mul inomial eg ession models we e con- s uc ed o he gende s o p edic he fi e-ca ego y d inking pa e n indica o ; abs inence was used as he e e ence ca ego y. S udy yea and quad a ic s udy yea e ms we e used as con inuous explana o y a iables accoun ing o ime ends. The income a iable con- ained six ca ego ies and he lowes ca ego y (€2 pe week o less) se ed as he e e ence ca ego y. In e ac ion e ms be ween income and ime a iables we e also en e ed in o he models o s udy possible di - e en ial changes in ime by income ca ego ies. In he nex s ep, amily backg ound a iables we e en e ed in o he models. In o de o s udy he in e ac ion be ween income and ime u he , he s eng h o he associa ion was s udied by c oss- abula ing income wi h he d inking pa e n a iable sepa a ely o each s udy yea . A γ coe ficien 34 wi h 95% CIs is shown o hese associa ions o each yea . Nex , a simple eg ession o he γcoe fi- cien on su ey yea was pe o med. The γcoe ficien is an o e all measu e o associa ion be ween wo o dinal a iables, and can be in e p e ed much like he o dina y linea co ela ion coe ficien . A se o po en ial con ounde s was included in eg es- sion models. The amily s uc u e is an indica o o whe he o no he esponden cu en ly li es wi h his o he pa en s (nuclea amily s non-nuclea amily). The place o esidence is a combined indica o b inging oge he sample-based (popula ion egis y) and sel - epo ed in o ma ion (me opoli an a eas/ci ies/ owns/ illages/ u al a eas). Sel - epo ed da a on he a he ’s o gua dian’s educa ion was classified in o basic (9 yea s o less), seconda y (9–12 yea s) and e ia y (o e 12 yea s). The a he ’s o gua dian’s p o essional s a us was coded om open-ended esponses: 35 uppe whi e colla , lowe whi e colla , ag icul u e and o es y, blue colla . Finally, we conduc ed wo sepa a e sensi i i y analyses: one o in es iga ing he sensi i i y o he γcoe ficien o he changes in he income dis ibu ion and he second o assessing he sensi i i y o he model o alcohol p ice changes. In he fi s analysis, we agg ega ed he lowes and highes income ca ego ies so ha , in each s udy yea , he lowes and highes income ca ego y included as close o 20% o he o al numbe o esponden s as pos- sible. In he second analysis, he yea ly eal p ice indices o alcoholic be e ages, which we e a ailable o he yea 2000 onwa ds, we e added as co a ia es in o he models. When es ing o he ole o alcohol p ice changes, we fi s an gende -specific models wi h yea s 2001–2013 wi hou he p ice indices, and hen compa ed he co e- sponding models wi h he indices added. IBM SPSS S a is ics so wa e was used o he analyses. RESULTS Alcohol d inking and d unkenness inc eased among bo h gi ls and boys in all income g oups om 1983 o he la e 1990s, a e which i dec eased (figu e 1). Weekly d unkenness in he lowes income g oup (€2 pe week o less) was non-exis en h oughou he pe iod 1983–2013 and mon hly d unkenness was a e. In he highes income g oup (€17 o mo e pe week), a ound 40% o gi ls and boys d ank un il d unk a leas occa- sionally in he s udy pe iod o 1983 un il he la e 1990s, a e which d unkenness dec eased conside ably. The p opo ion o 14-yea -olds who d ank alcoholic be - e ages bu ne e un il d unk s ayed la gely unchanged h oughou he s udy pe iod. Bo h ime and quad a ic ime we e s a is ically signifi- can p edic o s in he model when p edic ing mon hly d unkenness, occasional d unkenness and d inking o small amoun s o alcoholic be e ages; non-d inke s o med he e e ence g oup ( able 2). Weekly disposable income p edic ed d unkenness and d inking only small Lin onen T, Ne alainen J. BMJ Open 2017;7:e013994. doi:10.1136/bmjopen-2016-013994 3 Open Access g oup.bmj.com on June 1, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om Table 1 Dis ibu ions o all a iables used in he s udy by he yea o da a collec ion 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 Gende Boys % 48 48 48 47 46 46 46 45 47 45 45 42 41 42 40 45 Gi ls % 52 52 52 53 54 54 54 55 53 55 55 58 59 58 60 55 Alcohol use D unk weekly % 0 0 1 1 1 1 1 2 2 1 1 1 1 0 0 0 D unk mon hly % 4 3 5 7 8 10 8 11 11 9 6 5 4 5 3 2 D unk occasionally % 17 13 19 20 20 22 23 25 23 21 18 16 13 15 14 10 D inks bu no un il d unk % 28 25 27 26 30 28 31 28 31 29 30 26 22 27 21 21 Does no d ink % 50 58 46 45 39 38 36 33 32 35 43 50 60 53 61 66 Missing % 1 1 2 1 1 2 1 1 2 5 3 3 0 1 1 0 Weekly disposable income €17 o mo e % 2 4 7 4 5 4 5 6 8 7 11 9 13 16 21 20 €10–€16 % 5 6 12 9 13 11 12 13 13 14 19 17 20 25 27 27 €7–€9 % 11 17 18 16 19 18 17 18 19 18 14 12 12 12 13 11 €5–€6 % 24 30 30 22 24 22 21 23 21 19 26 28 26 22 19 19 €3–€4 % 34 26 20 26 21 24 24 22 22 20 17 18 15 12 11 8 €2 o less ( e e ence) % 23 16 10 20 16 19 18 17 16 15 12 15 13 9 7 10 Missing % 1 2 4 2 2 2 2 1 2 7 2 2 2 4 4 6 Place o esidence Me opoli an a ea % 10 10 9 11 10 9 11 11 10 14 16 16 16 15 16 14 Ci ies % 17 16 17 16 15 15 15 16 16 16 15 15 15 15 14 18 Towns %27273329333235353331323331333432 Villages % 27 28 26 26 26 26 22 23 24 20 21 20 22 20 19 20 Ru ala eas %19171416151616151515151516161615 Missing % 0 2 1 1 2 2 1 1 2 4 2 2 1 1 1 2 Fa he ’s/gua dian’s educa ion Basic % 78 72 70 65 63 61 56 57 59 54 57 51 54 45 52 48 Seconda y % 12 13 12 13 15 17 18 15 13 13 12 14 12 11 12 11 Te ia y % 10 11 13 16 17 18 20 23 21 23 25 28 33 41 37 40 Missing % 0 4 5 6 5 5 6 5 6 11 5 7 0 2 0 1 Fa he ’s/gua dian’s p o essional s a us Uppe whi e colla % 15 16 18 21 22 22 24 24 34 36 33 34 36 39 40 43 Lowe whi e colla % 25 25 27 27 27 29 29 29 22 20 24 25 23 23 16 17 Ag icul u e and o es y % 13 11 10 9 9 7 6 6 7 6 6 6 6 4 6 6 Blue colla % 46 46 43 42 39 40 37 38 35 35 34 32 32 31 36 31 Missing % 1 2 2 2 4 2 3 2 3 3 2 2 3 3 2 4 Family s uc u e Nuclea amily % 80 79 78 79 77 77 76 75 74 73 75 74 76 76 79 78 Non-nuclea amily % 19 20 21 20 23 22 24 24 25 24 24 25 23 24 21 21 Missing % 1 0 1 0 0 1 0 1 1 3 1 1 0 1 0 1 To al N 3721 3353 8113 3105 7634 8189 8382 8390 8219 7292 6761 6503 5840 5516 4566 3535 4Lin onen T, Ne alainen J. BMJ Open 2017;7:e013994. doi:10.1136/bmjopen-2016-013994 Open Access g oup.bmj.com on June 1, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om amoun s in a consis en manne : he highe he income, he highe he odds we e o d inking. The in e ac ions be ween money and ime, and money and quad a ic ime we e bo h s a is ically significan . This eflec s he p esence o non-mono onic ime ends. Model-based es ima ed cumula i e p opo ions (p obabili ies) o d inking pa e ns o e ime a e displayed in figu e 1. Owing o he appa en in e ac ion, γcoe ficien s be ween d inking pa e ns and income a e seen o change in ime (figu e 2). In he 1980s, he associa ion was weake in 1985 and 1987. Since 1989, he γcoe fi- cien s gene ally dec eased. Howe e , he associa ion may ha e been somewha s onge in 2011 and 2013 han in he p eceding h ee measu emen s. In eg ession models o he γcoe ficien on su ey yea , he es ima ed eg ession coe ficien is −0.049 pe 10-yea in e al, sug- ges ing a declining end (p=0.00025). This is in ag ee- men wi h he in e ac ion es s epo ed abo e. Finally, po en ial con ounde s (place o esidence, a he ’s o o he gua dian’s educa ion, a he ’s o o he gua dian’s labou ma ke posi ion, amily ype) we e included in he eg ession models. The place o esi- dence was gene ally no a s a is ically significan p e- dic o o d inking pa e ns ei he among gi ls o boys; gi ls’mon hly d unkenness was mo e p obable in places o he han u al a eas. Low a he ’s o o he gua dian’s educa ion and low a he ’s o o he gua dian’s labou ma ke posi ion we e associa ed wi h sligh ly mo e se e e d inking pa e ns. Li ing in a amily o he han one con- sis ing o wo pa en s was associa ed wi h significan ly mo e se e e d inking, especially among gi ls. Sensi i i y analysis showed ha he income dis ibu ion change o e he yea s seems no o ha e a ec ed he ela ionship be ween income and d inking significan ly as he γcoe ficien s esul ing om his we e e y close o he o iginal ones p esen ed in figu e 2. In he second sensi i i y analysis, alcoholic be e age p ice indices we e added o he models. Among bo h gende s, he es ima es o pe sonal disposable income changed only ma ginally and no in a uni o m di ec ion. The es ima es o he p ice indices we e no s a is ically significan p edic o s in he models excep among boys o d inking un il d unk occasionally (yea s 2009 s 2003: p=0.045) and among gi ls o d inking bu no un il d unk (yea s 2007 s 2003: p=0.047). These wo a e likely o be chance findings ou o 56 s a is ical es s. DISCUSSION On he basis o a compa able se ies o na ionally ep e- sen a i e su eys in Finland since 1983, we conclude ha 14-yea -old adolescen s’alcohol d inking became mo e Figu e 1 Es ima ed alcohol d inking (weekly d unkenness, mon hly d unkenness, occasional d unkenness, d inking bu no un il d unk, does no d ink) p obabili ies among 14-yea -olds in he pe iod 1983–2013 in he lowes income g oup (€0–€2/week), he mos common income g oup (€5–€6/week) and he highes income g oup (€17 o mo e/week). Lin onen T, Ne alainen J. BMJ Open 2017;7:e013994. doi:10.1136/bmjopen-2016-013994 5 Open Access g oup.bmj.com on June 1, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om Table 2 Mul inomial logis ic eg ession models o 14-yea -old boys’and gi ls’d inking pa e ns. Boys D unk weekly D unk mon hly D unk occasionally D inks bu no un il d unk OR 95% CI OR 95% CI OR 95% CI OR 95% CI Weekly disposable money €17 o mo e 14.58 (6.82 o 31.17) 9.01 (6.25 o 12.97) 3.47 (2.80 o 4.30) 1.95 (1.64 o 2.33) €10–€16 10.84 (5.22 o 22.49) 7.79 (5.54 o 10.94) 3.62 (3.01 o 4.37) 1.90 (1.63 o 2.22) €7–€94.75 (2.21 o 10.18) 5.45 (3.88 o 7.66) 3.10 (2.58 o 3.71) 1.99 (1.72 o 2.30) €5–€6 1.96 (0.88 o 4.37) 3.19 (2.28 o 4.47) 2.22 (1.87 o 2.64) 1.56 (1.37 o 1.78) €3–€4 0.60 (0.21 o 1.68) 1.90 (1.33 o 2.73) 1.65 (1.38 o 1.97) 1.37 (1.19 o 1.57) €2 o less ( e e ence) 1 1 1 1 Place o esidence Me opoli an a ea 1.27 (0.63 o 2.59) 1.09 (0.78 o 1.52) 1.13 (0.92 o 1.37) 1.00 (0.85 o 1.18) Ci ies 0.71 (0.35 o 1.43) 1.00 (0.74 o 1.35) 0.84 (0.69 o 1.01) 0.84 (0.72 o 0.97) Towns 1.00 (0.54 o 1.85) 1.19 (0.91 o 1.55) 0.97 (0.83 o 1.14) 0.85 (0.75 o 0.97) Villages 1.38 0.74 o 2.59 1.13 (0.85 o 1.50) 1.06 (0.90 o 1.25) 0.97 (0.85 o 1.12) Ru al a ea ( e e ence) 1 1 1 1 Fa he ’s/gua dian’s educa ion Basic 0.75 (0.45 o 1.26) 1.38 (0.97 o 1.96) 1.14 (0.99 o 1.32) 0.94 (0.83 o 1.06) Seconda y 0.51 (0.26 o 1.00) 1.28 (1.01 o 1.62) 0.87 (0.73 o 1.04) 0.96 (0.83 o 1.10) Te ia y ( e e ence) 1 1 1 1 Fa he ’s/gua dian’s p o essional s a us Uppe whi e colla 0.44 (0.25 o 0.79) 0.94 (0.74 o 1.18) 0.92 (0.80 o 1.06) 1.04 (0.92 o 1.18) Lowe whi e colla 1.02 (0.67 o 1.56) 1.03 (0.84 o 1.25) 1.03 (0.91 o 1.17) 1.05 (0.94 o 1.17) Ag icul u e and o es y 0.51 (0.19 o 1.35) 0.90 (0.62 o 1.30) 0.96 (0.77 o 1.19) 1.07 (0.89 o 1.27) Blue colla ( e e ence) 1 1 1 1 Family s uc u e In a nuclea amily 0.38 (0.26 o 0.56) 0.47 (0.39 o 0.56) 0.63 (0.55 o 0.70) 0.86 (0.77 o 0.95) No in a nuclea amily ( e e ence) 1 1 1 1 Gi ls D unk weekly D unk mon hly D unk occasionally D inks bu no un il d unk OR 95% CI OR 95% CI OR 95% CI OR 95% CI Weekly disposable money €17 o mo e 14.17 (7.06 o 28.48) 6.63 (5.01 o 8.78) 4.13 (3.38 o 5.05) 2.37 (1.99 o 2.84) €10–€16 9.71 (5.12 o 18.42) 5.10 (4.02 o 6.47) 2.96 (2.50 o 3.50) 1.68 (1.46 o 1.94) €7–€96.35 (3.32 o 12.16) 3.88 (3.06 o 4.91) 2.88 (2.45 o 3.39) 1.77 (1.54 o 2.04) €5–€62.21 (1.10 o 4.44) 2.53 (2.01 o 3.18) 2.27 (1.95 o 2.64 1.60 (1.41 o 1.81) €3–€4 1.67 (0.82 o 3.41) 1.29 (1.00 o 1.65) 1.63 (1.40 o 1.90 1.22 (1.08 o 1.39) €2 o less ( e e ence) 1 1 1 1 Place o esidence Me opoli an a ea 1.42 (0.73 o 2.77) 1.35 (1.02 o 1.78) 1.05 (0.88 o 1.26) 1.17 (1.00 o 1.37) Ci ies 1.12 (0.58 o 2.14) 1.26 (0.97 o 1.63) 0.91 (0.77 o 1.08) 0.91 (0.79 o 1.06) Towns 1.72 (0.98 o 3.04) 1.65 (1.31 o 2.08) 1.04 (0.90 o 1.21) 1.03 (0.90 o 1.17) Villages 1.04 (0.56 o 1.94) 1.57 (1.24 o 2.00) 1.13 (0.97 o 1.31) 0.99 (0.87 o 1.14) Ru al a ea ( e e ence) 1 1 1 1 Fa he ’s/gua dian’s educa ion Basic 1.55 (0.96 o 2.50) 1.52 (1.26 o 1.84) 1.28 (1.12 o 1.45) 1.06 (0.95 o 1.19) Seconda y 0.96 (0.53 o 1.74) 1.21 0.96 o 1.51) 1.05 (0.90 o 1.22) 1.03 (0.90 o 1.18) Te ia y ( e e ence) 1 1 1 1 Fa he ’s/gua dian’s p o essional s a us Uppe whi e colla 0.51 (0.31 o 0.84) 0.77 (0.64 o 0.92) 0.95 (0.84 o 1.08) 1.07 (0.95 o 1.20) Lowe whi e colla 0.95 (0.66 o 1.36) 0.93 (0.79 o 1.09) 0.98 (0.87 o 1.10) 1.01 (0.91 o 1.12) Ag icul u e and o es y 0.65 (0.28 o 1.49) 0.73 (0.53 o 0.99) 0.73 (0.60 o 0.90) 0.99 (0.84 o 1.18) Blue colla ( e e ence) 1 1 1 1 Family s uc u e In a nuclea amily 0.28 (0.21 o 0.39) 0.42 (0.36 o 0.48) 0.53 (0.47 o 0.59) 0.71 (0.64 o 0.78) No in a nuclea amily ( e e ence) 1 1 1 1 The ORs a e based on main e ec s eg ession models. Income has been adjus ed o money alue changes. Models include linea and quad a ic ime a iables. S a is ically signi ican coe icien s a e higligh ed in bold. 6Lin onen T, Ne alainen J. BMJ Open 2017;7:e013994. doi:10.1136/bmjopen-2016-013994 Open Access g oup.bmj.com on June 1, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om se e e h ough he whole d inking pa e n spec um om 1983 un il he la e 1990s, a e which d inking became less p e alen . Disposable income among his age g oup inc eased wo old du ing he 30-yea pe iod. The alcohol d inking pa e n was significan ly associa ed wi h disposable income a ailable o he adolescen . The OR o mon hly d unkenness e sus abs inence was 6.6 (95% CI 5.0 o 8.8) among gi ls and 9.0 (6.3 o 13.0) among boys in he highes income g oup, ecei ing €17 pe week o mo e when con as ed wi h he lowes income g oup ha had €2 o less. Income seems o ha e been s ongly associa ed wi h hea y d inking, which equi es mo e money. The associa ion was conside ably mo e mode a e o d inking small amoun s; he ORs we e unde 3.0 o all income ca ego ies when com- pa ed wi h he lowes income g oup. D inking la ge amoun s o alcohol un il d unk seemed mo e a ec ed by pe iod e ec s han d inking small amoun s. Fu he mo e, he associa ion be ween income and d inking pa e n weakened conside ably du ing he 30-yea pe iod. Conside ing simila obse a ions on declining adolescen d inking in nume ous o he coun- ies 13 and heo e ical delibe a ions, 16 i is likely ha he findings on dec easing impo ance o disposable income a ailable o adolescen s o hei d inking a e gene alis- able o high-income coun ies. The changes in ime in he ela ionship be ween income and d inking alcoholic be e ages a e also a ec ed by changes in be e age p ices. The de elop- men o alcohol p ices cap u ed by he eal p ice index indica ed inc easing a e age p ice o alcohol om he yea 2000 un il 2003, when i was 7% mo e expensi e han a he u n o he millennium. 36 Owing o ax changes, p ices hen ell o 4% lowe han in 2000; his again changed, bo h due o he global financial c isis and na ional ax inc eases, so ha in 2013 alcohol was, on a e age, 28% mo e expensi e han in he yea 2000. Bee , he be e age a ou ed by he unde aged who d ink hea ily, 37 was 35% mo e expensi e in eal e ms han in he yea 2000. This figu e is, in ac , e y close o he co esponding inc ease in income among 14-yea -olds in he same pe iod. 25 Thus, he a o dabili y o alcohol among 14-yea -olds did no change du ing he pe iod. Money has been p esen ed as a key media o be ween pa en s and adolescen s. 19 38 In he No dic coun ies, 14-yea -olds do no wo k in la ge numbe s and hei wo k income is ela i ely low. 22 Thus, pa en s ha e had he means o con ol adolescen consump ion, including alcohol d inking. O he pa en ing ac o s, 15 howe e , complica e he pic u e. In Finland, a epo published in 2007 obse ed ha pa en al con ol o e hei o - sp ings’alcohol use had no igh ened. 17 Subsequen pape s om Finland wi nessed a g adual igh ening o pa en al a i udes owa ds adolescen d inking, 14 18 alcohol- ela ed pa en ing p ac ices 39 and con ol o selling alcoholic be e ages o mino s. 18 40 Simila obse - a ions om o he coun ies include igh ening o alcohol-specific pa en ing p ac ices be ween he yea s 2007 and 2011 in he Ne he lands, 41 which was shown o be associa ed wi h dec eased d inking among hose unde 16 yea s o age, and pa en al supe ision being associa ed wi h lesse d inking among 11-yea -olds in he UK. 42 Alcohol d inking is also a ec ed by pe iod- ela ed o coho - ela ed ac o s. 43 This pape makes an a emp o un eil he coho e ec s: he ac ha changes in he su ounding socie y a ec people o di e en ages in di - e en ways, c ea ing commonali ies o indi iduals. Coho e ec s we e iden ified using his same da a se in Finland: 6 con ex ual ac o s con ibu ed mo e o he de elopmen o d unkenness han d inking small amoun s. One possible explana ion o ou esul s can be ha ecen coho s a e mo e in e es ed in o he ways Figu e 2 The γcoe icien s and hei 95% CIs o he associa ions be ween alcohol d inking and disposable income among 14-yea -olds. Lin onen T, Ne alainen J. BMJ Open 2017;7:e013994. doi:10.1136/bmjopen-2016-013994 7 Open Access g oup.bmj.com on June 1, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om o spending ime and money, o example, using digi al echnology, han in d inking, especially o d unkenness. On he basis o su ey da a om 2006 in Finland, alcoholic be e ages do no seem o be a p io i y on he lis o possible ways o spend money. 44 Boys aged 11–17 yea s epo ed ha hey p e e o spend hei income on swee s, digi al games, mobile phones and hobbies, while same-aged gi ls p io i ised swee s, clo h- ing, mobile phones and going ou o ca es, o example. Alcohol and obacco we e he eigh h i em on bo h gende s’lis s. Un o una ely, no ime end in o ma ion exis s on adolescen consump ion p e e ences. Since he u n o he millennium, spending o income on alco- holic be e ages has me wi h new compe i ion om, o example, digi al gaming and o he amusemen and pas ime ac i i ies, including spo s, 45 46 which a e no eadily compa ible wi h d inking alcohol. We can only specula e ha he obse ed weakening connec ion be ween income and alcohol d inking in he ecen decade may eflec a change in he impo ance o alcohol in eenage s’consump ion—i may ha e been pa ly eplaced by spending on digi al echnology. Pa en al con ol o income does no seem an e ficien way o con olling alcohol d inking; emphasis may need o be di ec ed owa ds ensu ing ha eenaged adoles- cen s spend hei ime and ene gy on mo e cons uc i e ac i i ies han simply hanging ou . The indica o s ha e been de i ed om mailed sel - epo su eys and a e subjec o measu emen biases. Howe e , he bes possible ca e was aken o ensu e com- pa abili y ac oss he s udy yea s; o example, ega ding he iming, ques ions and sampling emained unchanged h oughou he yea s. The esponse a e was good un il ecen yea s; howe e , i was down o 38% in 2013. In o de o inc ease he esponse, he op ion o pa icipa e using he in e ne was in oduced in 2009; a ound 50% o he esponses ha e since been ecei ed ia he in e - ne . While e en in 2013 he a e emained 46% among gi ls, he esponse a e o only 30% among boys in 2013 means ha he esul s conce ning boys in ecen yea s mus be judged wi h cau ion. In analyses based on esponse iming and es – e es subsamples, i has been concluded ha while i is likely ha he p e alence o alcohol use is equal among esponden s and non- esponden s, d inking un il d unk is mo e common among non- esponden s. 43 Simila conclusions we e d awn om compa isons wi h he School Heal h P omo ions Su ey (SHPS), which is a na ionally ep e- sen a i e class oom su ey wi h a ound a 90% esponse a e. 47 The a es o abs en ion o 14–16 yea s old, o example, we e 27% (AHLS) and o 15–16 yea s old 25% (SHPS) in he yea 2001, and 52% (AHLS) and 50% (SHPS) in he yea 2013. 48 Owing o he declining esponse and esponden selec ion based on d inking habi s, i is likely ha he figu es p esen ed he e conce n- ing d unkenness among 14-yea -olds a e sligh ly unde es- ima ed, especially a e he yea 2009. Since s udy d op-ou was mo e p onounced among hea ie d inke s, he dec ease in he associa ion be ween income and d inking pa e ns may be a ec ed such ha he dec ease in he s eng h o associa ion is o e es ima ed. Howe e , he selec ion bias is mos likely no enough o influence he gene al dec easing end such ha i would be an a e ac . Un o una ely, non- esponse analysis o dispos- able income is no a ailable in his s udy, and no 14-yea -olds’income ends ha e been published om o he adolescen da a se , ei he . School-based su ey da a wi h an a e age yea ly esponse a e o a ound 90% show a end o inc easing disposable income o 15–16 yea s old be ween he yea s 1995 and 2011, 39 in good ag eemen wi h he figu es p esen ed in his pape . The wo sensi i i y analyses, one es ing o sensi i i y o income dis ibu ion changes and he o he o sensi- i i y o alcohol p ice changes, showed ha he esul s we e obus o hese wo possible explana ions/con oun- de s. I is likely ha p ice is, howe e , a significan issue a he ime o pu chase, and he p ices 14-yea -olds a e o ced o pay o acqui ing alcohol illegally may be highe han he a e age e ail p ice. Un o una ely, in o ma ion on his phenomenon was no a ailable o he analyses. In addi ion, while alcohol a ailabili y is ce - ainly a significan ac o in unde aged d inking, alid indica o s o he expe ienced a ailabili y a he indi id- ual le el we e no a ailable o analyses. The dec eases in adolescen d inking simila o he ones epo ed he e o Finland ha e aised esea ch in e es in many o he high-income coun ies 13 and a decline in 11–15 yea s old d inking be ween he yea s 2002 and 2010 has been obse ed in 20 o he pa icipa - ing 28 coun ies. A ange o explana ions o his phe- nomenon ha e been specula ed on, including income changes, een cul u e changes, s ic e legal en o ce- men , changes in adul d inking and adul a i udes owa ds adolescen d inking, as well as policy and p e- en ion aimed a dec easing d inking. 12 13 16 49 Empi ical e idence on he dec ease in 14–17 yea s old d inking obse ed ecen ly in Aus alia shows he phe- nomenon o be dis ibu ed ac oss all socioeconomic sub- sec ions o he popula ion examined. 12 O he hypo heses and e idence e iewed ecen ly in he UK, he declining alcohol a o dabili y and be e pa en ing we e seen as he mos c edible explana ions o declin- ing you h d inking. 16 CONCLUSIONS I is jus as impo an o iden i y ac o s behind ecen dec eases in adolescen d inking as i is o analyse pas inc eases in d inking. The ac o s causing he down u n can be used in u u e alcohol policies and in e en- ions, 3 as well as in de eloping public policies and ac ion in gene al. In con as wi h he clea e ec s o alcohol policy changes in consump ion in he adul popula ion, 1 policy measu es including ax changes do no seem o ha e had a no iceable influence on alcohol d inking o e en d unkenness among he unde aged in Finland. 8 8Lin onen T, Ne alainen J. BMJ Open 2017;7:e013994. doi:10.1136/bmjopen-2016-013994 Open Access g oup.bmj.com on June 1, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om The cu en analyses u he complica e he pic u e: a p e iously iden ified p edic o o adolescen alcohol use, disposable income, has dec eased in impo ance, and he influence o pocke money, which has been an impo an ins umen o he pa en s o influence hei child en’s alcohol d inking, seems o ha e weakened o e he yea s. Con ibu o s Bo h au ho s join ly concei ed o he s udy idea and design, as well as pa icipa ed in he analysis o he da a and in e p e a ions o he esul s. TL w o e he i s d a , which was hen c i ically e ised by bo h au ho s. The inal e sion o he manusc ip was app o ed by bo h au ho s. Funding The da a collec ion was unded by he Minis y o Social A ai s and Heal h (Finland). Compe ing in e es s None decla ed. E hics app o al E hics Commi ee o he Tampe e egion (Finland). P o enance and pee e iew No commissioned; ex e nally pee e iewed. Da a sha ing s a emen Da a se a ailable on eques om he Uni e si y o Tampe e, School o Heal h Sciences, Tampe e, Finland. 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. Babo TF, Cae ano R, Casswell S, e al.Alcohol: no o dina y commodi y: esea ch and public policy. Ox o d: Ox o d Uni e si y P ess, 2010. 2. Wagenaa AC, Salois MJ, Kom o KA. E ec s o be e age alcohol p ice and ax le els on d inking: a me a-analysis o 1003 es ima es om 112 s udies. Addic ion 2009;104:179–90. 3. Pennay A, Li ings on M, MacLean S. Young people a e d inking less: i is ime o ind ou why. D ug Alcohol Re 2015;34:115–18. 4. Lin onen T, Rimpelä M, Vika A, e al. The e ec o socie al changes on d unkenness ends in ea ly adolescence. Heal h Educ Res 2000;15:261–9. 5. Lin onen T, Wilska T, Koi usil a L, e al. T ends in disposable income among eenage boys and gi ls in Finland om 1977 o 2003. In J Consum S ud 2007;31:340–8. 6. Lin onen T, Hä könen J, Rai asalo K, e al.Dec easing adolescen d inking: is he e e idence o con inua ion on o u u e adul coho s? APC-analysis o adolescen d inking in Finland, 1983–2013. Scand J Public Heal h 2016. doi:10.1177/ 1403494816665505. [Epub ahead o p in 26 Aug 2016]. 7. Kinnunen JM, Lind o s P, Pe e L, e al.The Adolescen Heal h and Li es yle Su ey 2013. Adolescen smoking, alcohol and subs ance use in 1977–2013. Helsinki: Minis y o social a ai s and heal h, 2013. 8. Lin onen T, Ka lsson T, Ne alainen J, e al. Alcohol policy changes and ends in adolescen d inking in Finland om 1981 o 2011. Alcohol Alcohol 2013;48:620–6. 9. Heal h and Social Ca e In o ma ion Cen e. S a is ics on alcohol: England. London: O ice o Na ional S a is ics, 2008. 10. Johns on L, O’Malley PM, Bachman JG, e al.Moni o ing he u u e —na ional su ey esul s on d ug use, 1976–2012, Vol. 1, seconda y school s uden s. Ann A bo : Ins i u e o Social Resea ch, Uni e si y o Michigan, 2013. 11. Raninen J, Li ings on M, Lei man H. Declining ends in alcohol consump ion among Swedish you h—does he heo y o collec i i y o d inking cul u es apply? Alcohol Alcohol 2014;49:681–6. 12. Li ings on M. T ends in non-d inking among Aus alian adolescen s. Addic ion 2014;109:922–9. 13. Looze MD, Raaijmake s Q, Bog TT, e al. Dec eases in adolescen weekly alcohol use in Eu ope and No h Ame ica: e idence om 28 coun ies om 2002 o 2010. Eu J Public Heal h 2015;25(Suppl 2): 69–72. 14. Rai asalo K, Holmila M. Alkoholinkäy ö apojen pe iy yminen ja anhempien kas a uskäy ännö ”[He i abili y o alcohol use pa e ns and he pa en al educa ional p ac ices]. In: Lammi-Taskula, J, Ka onen S. eds. Lapsipe heiden hy in oin i 2014 [The well-being o amilies wi h child en 2014]. Helsinki: Na ional Ins i u e o Heal h and Wel a e, 2014:196–204. 15. Ryan SM, Jo m AF, Lubman DI. Pa en ing ac o s associa ed wi h educed adolescen alcohol use: a sys ema ic e iew o longi udinal s udies. Aus N Z J Psychia y 2010;44:774–7. 16. Bha acha ya A. You h ul abandon—Why a e young people d inking less? London: Ins i u e o Alcohol S udies, 2016. 17. Ka lsson T, Rai asalo K, Holmila M. Alaikäis en alkoholihankin ojen sie ämä ön helppous [The unbea able ease o access o alcohol among adolescen s]. In: Tige s ed C. ed. Nuo e ja alkoholi [You h and alcohol]. Helsinki: The Finnish Socie y o Alcohol and D ug Resea ch & Finnish You h Resea ch Socie y, 2007:149–160. 18. Rai asalo K, Huh anen P, Ahls öm S. Nuo en päih eiden käy ö sekä käsi ykse niihin lii y is ä iskeis ä ja saa a uudes a Suomen ESPAD-aineis on uloksia 1995–2011 [The use o in oxican s and pe cep ions o isk and a ailabili y based on he ESPAD Da a in Finland 1995–2011]. Yh eiskun apoli iikka 2012;77:72–80. 19. B usdal R, Be g L. A e pa en s gende neu al when inancing hei child en’s consump ion? In J Consum S ud 2010;34:3–10. 20. Wes P, Swee ing H, Young R, e al. A ma e ial pa adox: socioeconomic s a us, young people’s disposable income and consume cul u e. J You h S ud 2006;9:437–62. 21. Besen-Cassino Y. The cos o being a gi l: gende ea ning di e en ials in ea ly labo ma ke s. NWSA J 2008;20:146–60. 22. Danske Bank. Financial li e acy among 15–17 yea -old eenage s— based on a ma ke esea ch su ey among eenage s 15–17 yea s old in Denma k, No way, Sweden and Finland. Copenhagen: Danske Bank co po a e epo , 2013. 23. de Lange M, Ges huizen M, Wolbe s MHJ. “You h labou ma ke in eg a ion ac oss Eu ope”.Eu Soc 2014;16:194–212. 24. Mo ime JT. The e olu ion, con ibu ions, and p ospec s o he you h de elopmen s udy: an in es iga ion in li e cou se social psychology. Soc Psycholy Q 2003;75:5–27. 25. Wilska T, Lin onen T. The gende gap in eenage s’incomes. A 30-yea end in Finland 1983–2013. J You h S ud 2015;19: 421–37. 26. Koo eman P. Time, money, pee s, and pa en s; some da a and heo ies on eenage beha io . J Popul Econ 2007;20:9–33. 27. Raijas A, Wilska T. Huolenpi oa ja jakamis a— ahan ja ajan jakau uminen suomalaisissa lapsipe heissä [Ca ing and sha ing— he alloca ion o ime and money in households]. Discussion and wo king pape s 104/2007. Helsinki: Na ional Consume Resea ch Cen e , 2007. 28. Melo i R, He on J, Hickman M, e al. Adolescen alcohol and obacco use and ea ly socioeconomic posi ion: he ALSPAS bi h coho . Pedia ics 2011;127:e948–955. 29. Rich e M, Leppin A, Gabhainn S. The ela ionship be ween pa en al socio-economic s a us and episodes o d unkenness among adolescen s: indings om a c oss-na ional su ey. BMC Public Heal h 2006;6:289. 30. Kun sche E, Rossow I, Simons-Mo on B, e al. No ea ly d inking bu ea ly d unkenness is a isk ac o o p oblem beha io s among adolescen s om 38 Eu opean and No h Ame ican Coun ies. Alcohol Clin Exp Res 2012;37:308–14. 31. Samposalo H. Rela ionship be ween d inking pa e n and alcohol- ela ed ha ms and isk beha io s among 14- o 16-yea -old adolescen s in Finland. Tampe e: Tampe e Uni e si y P ess, 2013. 32. Lin onen T, Ahls öm S, Me so L. The eliabili y o d inking sel - epo s in adolescence. Alcohol Alcohol 2004;39:362–8. 33. Lin onen T, Rimpelä M. The alidi y o he concep o ‘sel -pe cei ed d unkenness’in adolescen heal h su eys. J Subs Use 2001;6:145–50. 34. Ag es i A. Ca ego ical da a analysis, hi d ed. New Yo k: Wiley, 2012. 35. In e na ional Labou O ganisa ion. In e na ional S anda d Classi ica ion o Occupa ions. Gene a: In e na ional Labou O ice, 1990. 36. O icial S a is ics o Finland. Yea book o alcohol and d ug s a is ics 2015. Helsinki: Na ional Ins i u e o Heal h and Wel a e, 2015. 37. Lin onen T, Konu A. D unkenness- ela ed alcoholic be e age choices among adolescen s. J Subs Use 2001;6:16–21. 38. Scho JB. Bo n o buy: he comme cialized child and he new consume cul u e. New Yo k: Sc ibne , 2004. 39. Rai asalo K, Holmila M. P ac ices in alcohol educa ion among Finnish pa en s: Ha e he e been changes be ween 2006 and 2012? D ugs Educ P e Pol, 2016. doi: 10.1080/ 09687637.2016.1183587. [Epub ahead o p in 1 Jun 2016]. Lin onen T, Ne alainen J. BMJ Open 2017;7:e013994. doi:10.1136/bmjopen-2016-013994 9 Open Access g oup.bmj.com on June 1, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om