Examining retail purchases of cigarettes and nicotine replacement therapy in Finland

Resea ch Pape Tobacco Induced Diseases 1 Examining e ail pu chases o ciga e es and nico ine eplacemen he apy in Finland Da id S. Timbe lake1,2, Johanna Joensuu1, Te hi Ku ko3, A ja H. Rimpelä1,4,5, Jaakko Ne alainen1 ABSTRACT INTRODUCTION Finland’s success in achie ing he goal o i s obacco endgame la gely depends on ec i ying de iciencies in he deli e y o smoking cessa ion se ices. One such weakness, which has no been documen ed wi h empi ical da a, is misuse o nico ine eplacemen he apy (NRT). This s udy’s objec i e was o examine pu chase pa e ns o NRT o es ima ing imp ope use o he medica ion. The s udy was based on he assump ion ha du a ion o a pu chase episode is indica i e o ei he p ope use o misuse o NRT. METHODS The pa icipan s (n=728), who pu chased a leas one NRT p oduc in 2016 (mos ly gum/lozenge), we e selec ed h ough en ollmen in a la ge cus ome loyal y p og am in Finland (LoCa d). Pa icipan s we e ca ego ized in o one o i e g oups acco ding o hei longes pu chase episode o NRT, de ined by pu chases made in consecu i e, 4-week in e als. RESULTS Mos pa icipan s, who did no adhe e o NRT guidelines, ei he pu chased he medica ion o oo sho (≤4 weeks, 63.5%) o oo long (>24 weeks, 13.2%) o a pu chase episode. Median pu chases o NRT in he i s mon h o use we e one and ou in he o me and la e , espec i ely. In con as o o he g oups, pe sis en use s (>24 weeks) did no cu ail pu chases o NRT ac oss se e al 4-week in e als, sugges ing po en ial o dependence on NRT. CONCLUSIONS The obse a ion ha mos pu chase episodes we e e mina ed p ema u ely is consis en wi h su eys epo ing widesp ead NRT misuse. Gi en unce ain y o g ea e egula ion o NRT sales h ough legisla ion, i would be p uden o Finnish e aile s o p omo e p ope use o he he apy. AFFILIATION 1 Facul y o Social Sciences, Tampe e Uni e si y, Tampe e, Finland 2 P og am in Public Heal h, Uni e si y o Cali o nia, I ine, Uni ed S a es 3 The Social Insu ance Ins i u e o Finland, Helsinki, Finland 4 PERLA - Tampe e Cen e o Childhood, You h and Family Resea ch, Tampe e Uni e si y, Tampe e, Finland 5 Depa men o Adolescen Psychia y, Tampe e Uni e si y Hospi al, Tampe e, Finland CORRESPONDENCE TO Da id S. Timbe lake. P og am in Public Heal h, Uni e si y o Cali o nia, An ea e Ins uc ion & Resea ch O ices, I ine, CA, Uni ed S a es E-mail: [email p o ec ed] KEYWORDS nico ine eplacemen he apy, ciga e es, obacco endgame, ime-se ies da a, e ail pu chases Recei ed: 16 No embe 2018 Re ised: 30 Ma ch 2019 Accep ed: 15 Ap il 2019 Published by Eu opean Publishing on behal o he In e na ional Socie y o he P e en ion o Tobacco Induced Diseases (ISPTID). © 2019 Timbe lake D.S. This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0 In e na ional License. (h ps://c ea i ecommons.o g/licenses/by/4.0/) Tob. Induc. Dis. 2019;17(May):39 h ps://doi.o g/10.18332/ id/108537 INTRODUCTION P o iding nico ine eplacemen he apy (NRT) o smoke s wi hou a doc o ’s p esc ip ion, ia an o e - he-coun e (OTC) pu chase, was a signi ican ini ia i e o inc ease accessibili y and use o he p o en he apy1. C i ics a gue ha easie access o he medica ion educes he in e ac ion be ween pa ien and p o ide , which is c i ical o a success ul qui a emp . O he s con end ha e en a e p esc ibing medica ion, physicians seldom p o ide p ope cessa ion counseling2,3. In coun ies such as Finland, pha macis s p o ided counseling because OTC sales o NRT we e ini ially es ic ed o pha macies. Bu , he Finnish Pa liamen app o ed a law ha de egula ed he NRT ma ke in 2006, allowing e ail sales in g oce y s o es, kiosks and gasoline s a ions4. P ices o NRT p oduc s, which declined by 15% a e de egula ion, we e he leas expensi e in hype ma ke s, ollowed by Resea ch Pape Tobacco Induced Diseases Tob. Induc. Dis. 2019;17(May):39 h ps://doi.o g/10.18332/ id/108537 2 supe ma ke s and pha macies5. Consequen ly, sales o NRT inc eased d ama ically in Finland6, bu no in pha macies7 whe e consul a ion occu s on he p ope use o he medica ion. The ini ial conce n abou libe alizing he NRT ma ke ocused on smoke s’ po en ial o p olonging use o he medica ion beyond he 3-mon h pe iod se by he US Food and D ug Adminis a ion (FDA). This conce n was la gely dispelled by s udies epo ing ha only 6% o pa icipan s pu chased and used NRT o an ex ended pe iod o a leas six mon hs8,9. I is belie ed ha mos smoke s p olong hei use o NRT in o de o qui o educe smoking a he han sa ia e an addic ion. Consequen ly, esea che s ad oca ed ha guidelines o du a ion o NRT use be ex ended beyond he 3-mon h maximum10,11. A mo e p essing issue ela es o he majo i y o smoke s who p ema u ely e mina e he use o NRT12. Using a longi udinal design, Zhang e al.12 epo ed ha smoke s who used NRT o less han ou weeks had a dec eased likelihood o achie ing abs inence ela i e o he non-NRT use s (OR=0.51, 95% CI: 0.38–0.67). The ine ec i eness o using NRT o such a b ie pe iod is conce ning because s udies, such as he ITC Fou -Coun y Su ey13, indica e ha OTC NRT is p ema u ely e mina ed among he majo i y o use s. Balm o d e al.13 epo ed ha 62.9% and 76.3% o use s o OTC NRT had only used he medica ion o ≤4 weeks and ≤8 weeks, espec i ely. Such s udies, howe e , a e limi ed by he use o sel - epo ed da a ha canno be e i ied. The al e na i e o analyzing ime-se ies pu chases is one means o ob ia ing he inaccu acies ha occu wi h su ey da a. The cu en s udy used e ail da a o es ima ing he pe cen age o NRT use s who pu chased he medica ion o oo sho o oo long o a pu chase episode. An examina ion o e ail pu chases is pa icula ly needed in ligh o Finland’s goal o end he use o obacco by 2030 (i.e. <5% p e alence). The Roadmap o a Tobacco–F ee Finland speci ies se e al s eps in ea ing obacco dependence14 as pa o he la ge goal o ec i ying de iciencies in smoking cessa ion se ices. As e idence o one such de iciency, Finland ecei ed he low sco e o 5/10 poin s o obacco ea men on he 2016 Tobacco Con ol Scale15. One no able ecommenda ion in he Roadmap is o eimbu se smoke s o medica ions and emo e Sec ion 54a om he Medicines Ac , he la e pe mi ing sales o OTC NRT in e ail ou le s (e.g. g oce y s o es). The absence o pu chase da a, which p eda es he de egula ion o Finland’s NRT ma ke in 2006, p ecludes assessmen o whe he pu chases om e ail ou le s con ibu e o g ea e misuse o NRT ela i e o pu chases om pha macies. The cu en s udy is nei he designed o es e ec i eness o egula ed e sus de egula ed NRT ma ke s, no is i designed o examine pa e ns o use and mo i es o pu chasing NRT. Ins ead, his s udy aims o examine he du a ion o NRT pu chase episodes o es ima ing he du a ion o episodes o NRT use. The goal is o e i y he ex en o NRT misuse13 h ough es ima ion o pu chase episodes ha a e e mina ed p ema u ely. The indings could se e as baseline es ima es in he case o u u e policy change, and p o ide e aile s in o ma ion ha could acili a e e o s o inc ease cus ome s’ adhe ence o NRT guidelines. METHODS Sample selec ion Pa icipan s o his s udy we e selec ed h ough hei en ollmen in a cus ome loyal y p og am headed by he S G oup, a comme cial en e p ise possessing mo e han 45% o he ma ke sha e in Finland16. The S G oup’s la ge ma ke sha e minimizes bu does no nega e he possibili y ha s udy pa icipan s could ha e pu chased ciga e es and NRT om o he sou ces. Using a da abase con aining cus ome s’ email add esses, he S G oup con ac ed membe s o he cus ome loyal y p og am in he HOK–Elan o e ail co-ope a i e in sou he n Finland. The email inqui y sough membe s’ consen o elease basic demog aphic in o ma ion (age, gende , pos al code) and pu chases made in he yea 2016. Release o he da a, which excluded pe sonal iden i ie s, was in ended o esea ch pu poses and app o ed by he e hics commi ee o he Uni e si y o Helsinki Re iew Boa d in he humani ies and social and beha io al sciences. App oxima ely 5% o membe s o he loyal y p og am consen ed o he da a elease, yielding a sample o 13274 cus ome s. Compa ed o he gene al popula ion in sou he n Finland, consen ing membe s o he loyal y p og am we e p edominan ly emale (abou 68%), middle-aged (abou 46 yea s old), and mo e educa ed, which can be a ibu ed o highe pa icipa ion a es16. This is e idenced by he s a is ic indica ing ha 56% e sus 33% o consen ing loyal y Resea ch Pape Tobacco Induced Diseases Tob. Induc. Dis. 2019;17(May):39 h ps://doi.o g/10.18332/ id/108537 3 membe s and he gene al popula ion, espec i ely, had a uni e si y deg ee17. Ou inal analy ic sample, which excluded S G oup pe sonnel, consis ed o cus ome s who pu chased a leas one NRT p oduc in he yea 2016 (n=728). Measu es The measu es we e he ime, da e, and expendi u es o ciga e es and NRT in he yea 2016. Ca ego ies o NRT consume s we e based on he pu chase o one o mo e p oduc s in consecu i e 4-week in e als. The ca ego ies, which we e based on clinical guidelines and empi ical indings12, co espond o: he ea ly e mina o s (≤4 weeks o use); hose who used NRT o he minimum pe iod (5–8 weeks); hose who used NRT o he s anda d ecommended pe iod (9–12 weeks); hose who ex ended hei use o NRT acco ding o e ised ecommenda ions (13–24 weeks)10; and hose who exceeded he maximum pe iod o NRT use (>24 weeks). The s udy pa icipan s we e ca ego ized in o one o he i e mu ually exclusi e g oups acco ding o hei longes pu chase episode, absen a pu chase lapse exceeding 28 days. Fo example, i a pa icipan had pu chased NRT on 5 Feb ua y, 21 Feb ua y and 16 Ma ch, and made a subsequen pu chase on 9 June, hen he pa icipan was classi ied as ha ing pu chased NRT o 5–8 weeks o use ( ange: 29–56 days). Fo hose pa icipan s who had mul iple bu equal pu chase episodes, he longes episode in days was selec ed. An in e up ion in he pu chase o NRT exceeding one mon h was used p e iously in signi ying a new episode o use9. Ou classi ica ion scheme did no accoun o hose who made bulk pu chases o NRT o use o e a leng hy ime pe iod. Fu he mo e, he classi ica ion scheme was no based on inciden pu chases, and, hus, does no necessa ily ep esen indi iduals who used NRT o he i s ime. Fo example, an ea ly e mina o who made a pu chase in Janua y 2016 may ha e pu chased NRT h oughou he la e pa o 2015. Based on sales da a om he S G oup, nico ine gum and lozenge cons i u ed 96.9% o he o al sales olume o all nico ine eplacemen p oduc s. Sales o he nico ine pa ch, which we e low and unchanging om 2006 (de egula ion) h ough 20136, we e sligh ly highe in pha macies compa ed o o he ou le s. Gi en i s low o e all olume, i is unlikely ha he nico ine pa ch was supplemen ed wi h copious pu chases o nico ine gum and lozenge. Da a eleased by he S G oup included nei he he p oduc ype (e.g. gum, lozenge, pa ch) no p oduc cha ac e is ics such as b and, package size, o dosage in millig ams. Expendi u es o NRT we e a ailable o analysis, bu , o e lapped oo much o di e en ia e p oduc ypes. Thus, we op ed no o ex apola e package size om expendi u e o es ima ing he supply o NRT o use o e a designa ed pe iod. Ins ead o ex apola ing package size, we used consecu i e pu chases o e 4-week in e als as a measu e o du a ion o NRT use, which does no imply ha he pu chase(s) p o ided a su icien supply o NRT o any gi en mon h. Al hough expendi u es o NRT we e no used in ex apola ing package sizes, hey we e summed o e in e als as a desc ip i e way o illus a ing pu chase di e ences among he i e g oups o NRT consume s. Da a analysis The da a analyses en ailed he use o pai ed and non–pai ed s a is ical es s. Non-pa ame ic me hods we e employed due o he skewed dis ibu ions o pu chases and expendi u es o NRT and ciga e es. The K uskal–Wallis es on anks, analogous o a one-way ANOVA, was used o compa ing pu chases and expendi u es ac oss he i e g oups o NRT consume s. One no able compa ison ac oss g oups was he numbe o NRT pu chases made in he i s ou weeks o a pu chase pe iod. The Wilcoxon signed ank es was used o es ing di e ences wi hin pai s ac oss ime pe iods, by NRT consume g oup. Pu chases in he i s ou weeks se ed as he baseline o compa isons o pu chases made in subsequen 4-week in e als. The wi hin-pai es s acili a ed assessmen o whe he NRT pu chases s ayed he same o diminished ollowing he i s ou weeks o a pu chase episode. RESULTS Cha ac e is ics o NRT consume s The 728 pa icipan s made a o al o 9300 NRT and 21601 ciga e e pu chases in 2016. Mos pu chases o NRT and ciga e es occu ed on F iday, making up 16.2% and 17.4% o o al sales, espec i ely. A highe pe cen o emales e sus males (66.9% s 57.0%) and young e sus he old (72.9% s 61.6%) p ema u ely e mina ed use o NRT. Tes s o independence Resea ch Pape Tobacco Induced Diseases Tob. Induc. Dis. 2019;17(May):39 h ps://doi.o g/10.18332/ id/108537 4 indica ed a s a is ically signi ican associa ion be ween NRT consume g oup and sex (χ2 (4)=10.8; p=0.03), bu no age g oup (χ2 (8)=10.5; p=0.3) (Table 1). The ea ly e mina o s, who did no make consecu i e pu chases ac oss 4-week in e als, composed almos wo- hi ds o he en i e sample. O he NRT p oduc s, 15.2% and 11.6% we e pu chased by he ea ly e mina o s du ing Janua y 2016 and he o he ele en mon hs o he yea (χ2 (1)=7.8; p=0.005), espec i ely, sugges ing ha se e al ea ly e mina o s may ha e pu chased he medica ion as pa o a New Yea ’s esolu ion. Among he ea ly e mina o s (n=462), 237 had pu chased NRT only one ime wi h a median expendi u e o €6.6 ( ange: €2 – €30.1). These one- ime pu chase s o NRT equen ly pu chased ciga e es as e idenced by a high median numbe o ciga e e pu chases (31) and ciga e e expendi u es (€251.8). The pa icipan s who exceeded NRT ecommenda ions (>24 weeks, n=96), e e ed o as pe sis en use s, had he ewes ciga e e pu chases due in pa o he la ge pe cen (50%) o use s who did no make a single ciga e e pu chase in 2016. The pe sis en use s had a median pu chase gap o NRT ha was hal he alue obse ed in hose who bough NRT Table 1. Cha ac e izing consume s o nico ine eplacemen he apy (NRT) by hei demog aphics and pu chases o NRT and ciga e es in 2016 (n=728 ) All g oups Ca ego ies o NRT Consume s Tes s a is ic Ea ly e mina ion Minimum Recommended Ex ended Exceeded Sample size 728 462 66 43 61 96 Du a ion o NRT use de ining g oupa ≤4 weeks 5–8 weeks 9–12 weeks 13–24 weeks 25–52 weeks Demog aphics Sex (%) Female 64.8 68.4 51.5 58.1 55.7 65.6 χ2 (4)=10.8* Male 35.2 31.6 48.5 41.9 44.3 34.4 Age g oup (%) <30b 9.6 11.0 10.6 4.6 9.8 4.2 χ2 (8)=10.5 30–49 58.5 58.0 56.1 72.1 49.2 62.5 ≥50 yea s 31.9 31.0 33.3 23.3 41.0 33.3 NRT Pu chases Median pu chases/ yea 3 1 10 12 23 49 χ2 (4) =530*** Median pu chases/ pe iodc 2 1 4 6 13 43.5 χ2 (4) =578*** Pu chase gap in daysc,d 9 12 12.7 12 8 6 χ2 (4) =82.5*** NRT Expendi u es (€) Median/yea 32.3 13.6 90.1 155.7 233.3 633.0 χ2 (4) =466*** Median/pe iodc21.6 9.1 40.6 69.4 145.7 585.1 χ2 (4) =480*** Ciga e e Pu chases & Expendi u e % Pu chased in 2016 67.6 71.6 65.1 67.4 67.2 50.0 χ2 (4)=17.2** Median pu chases/ yea e 24 27 34 18 17 11 χ2 (4) =4.6 Median expendi u es (€) /yea e 180.4 211.9 224.0 105.5 105.2 91.4 χ2 (4) =7.2 a Ca ego ies a e de ined acco ding o pu chases made in consecu i e 4-week in e als (see Me hods), b Includes small numbe o adolescen s (n=18), c Co esponds o he pe iod de ining each g oup, d Median was calcula ed om he dis ibu ion o medians o indi idual pa icipan s, excluding hose who pu chased NRT only one ime, e Limi ed o hose who made a ciga e e pu chase in 2016 (n=492), K uskal-Wallis es o non-pa ame ic da a (one-way ANOVA on anks). *p<0.05, **p<0.01, ***p<0.0001. Resea ch Pape Tobacco Induced Diseases Tob. Induc. Dis. 2019;17(May):39 h ps://doi.o g/10.18332/ id/108537 5 o e he ecommended pe iod (6 s 12; o all g oups, χ2 (4)=82.5; p<0.0001). The pu chase gaps we e assessed wi hin pe iods o consecu i e NRT pu chases (gaps ≤28 days), hus nega ing he possibili y ha la ge gaps could be a ibu ed o a la e pu chase o NRT. Almos 5% o he en i e sample (n=35) had pu chased NRT o 13 consecu i e 4-week in e als in 2016. Pu chases o NRT o e ime Fo he i s mon h o use, pu chases o NRT a ied signi ican ly ac oss he i e g oups o consume s (χ2 (4)=416.9; p<0.0001), wi h he median pu chase anging om one o ou in he ea ly e mina o s and pe sis en use s, espec i ely (Table 2). Simila pu chase pa e ns we e obse ed ac oss he NRT consume g oups o he subsequen 4-week in e als. In examining wi hin-pai di e ences, we obse ed s a is ically signi ican declines in NRT pu chases om baseline (i.e. i s ou weeks o a pu chase episode) o he subsequen 4-week in e als in he g oups using NRT o e he minimum, ecommended, and ex ended ime pe iods. Fo example, hose who ollowed minimum ecommenda ions had a median o h ee pu chases in he i s mon h o use, bu only 1.5 pu chases in weeks i e h ough eigh (S=624.5; p<0.001). Declines ollowing he i s ou weeks o pu chasing NRT we e e iden o e e y g oup wi h excep ion hose who exceeded he guidelines. The la e had a median o ou pu chases in weeks 1–4, weeks 5–8, weeks 9–12, and weeks 13–16. By weeks 25–28, a signi ican decline in NRT pu chases om baseline had been obse ed (S=760.5; p<0.001). Ye , he pe sis en use s’ pu chases and expendi u es we e a he s able o e he whole pe iod. Table 2. A compa ison o he qua iles o NRT expendi u es and qua iles o numbe o NRT pu chases (in pa en hesis) ac oss g oups o consume s by ou -week pe iods o NRT use 4-Week Pe iod o NRT Use Pe cen ile Ca ego ies o NRT Consume s Ea ly e mina ion ≤4 weeks (n=462 ) Minimum 5–8 weeks (n=66 ) Recommended 9–12 weeks (n=43 ) Ex ended 13–24 weeks (n=61 ) Exceeded 25–52 weeks (n=96 ) Weeks 1–4*** 25% 5.2€ (1)a15.2€ (2) 17.7€ (2) 21.7€ (2) 29.8€ (3) 50% 9.1€ (1) 23.3€ (3) 35.4€ (3) 39.2€ (3) 56.0€ (4) 75% 20.1€ (2) 44.7€ (4) 52.0€ (4) 55.3€ (5) 89.7€ (6) Weeks 5–8*** 25% NAb6.6€ (1) 10.0€ (1) 16.8€ (2) 32.6€ (3) 50% NA 15.8€ (1.5) 19.9€ (2) 30.2€ (3) 55.5€ (4) 75% NA 34.0€ (2) 43.1€ (3) 43.7€ (4) 83.8€ (6.5) Signed ank es S=624.5** S=206.5** S=157* S=77.5 Weeks 9–12*** 25% NA NA 10.0€ (1) 15.0€ (2) 30.0€ (3) 50% NA NA 18.5€ (1) 29.4€ (2) 50.4€ (4) 75% NA NA 30.0€ (2) 43.2€ (4) 85.8€ (6) Signed ank es S=267** S=349** S=321.5 Weeksc 13–16*** 25% NA NA NA 14.9€ (2) 26.0€ (3) 50% NA NA NA 29.9€ (2) 53.7€ (4) 75% NA NA NA 38.6€ (3) 88.5€ (6) Signed ank es S=316.5** S=175 Weeksc 25–28 25% NA NA NA NA 20.8€ (2) 50% NA NA NA NA 47.4€ (3) 75% NA NA NA NA 72.7€ (5) Signed ank es S=760.5** a Each cell ep esen s he qua ile 1s (25%), 2nd (50%), o 3 d (75%) o NRT expendi u es in eu os and qua ile o numbe o NRT pu chases (in pa en hesis), b No applicable, c The 4-week pe iod was selec ed because e e yone wi hin he espec i e NRT g oup had pu chased NRT du ing he speci ied pe iod. ***p<0.0001 co esponds o a K uskal–Wallis es on anks o NRT pu chases ac oss g oups o consume s; **p<0.001 and *p<0.05 co espond o he Wilcoxon signed ank es o NRT pu chases ac oss 4-week pe iods, using he i s ou weeks o a pu chase episode as he baseline compa ison. Resea ch Pape Tobacco Induced Diseases Tob. Induc. Dis. 2019;17(May):39 h ps://doi.o g/10.18332/ id/108537 6 Concu en pu chases o ciga e es and NRT The pe cen age o concu en pu chases o ciga e es and NRT in he i s and las mon h o a pu chase episode, shown in Figu e 1, excludes he pa icipan s who ne e pu chased ciga e es in 2016. Concu en pu chases we e coded in e ms o pai s o consecu i e pu chases o he wo p oduc s. Fo example, i a cus ome pu chased a ciga e e pack on day 5, ano he ciga e e pack on day 17, and NRT on day 20, hen a single concu en pu chase would be coded as ha ing occu ed wi hin he same week. In his way, he coding scheme selec ed he sho es pe iod be ween a ciga e e and NRT pu chase. In hei single mon h o NRT use, he ea ly e mina o s had a lowe pe cen age o concu en pu chases ela i e o o he g oups, which occu ed despi e equen pu chases o ciga e es h oughou he yea . Concu en pu chases on he same day o same week, which occu ed equen ly o he o he g oups, changed sligh ly among hose who pu chased NRT o e minimum and ex ended pe iods. In con as , he decline in he pe cen age o concu en pu chases on he same day/week was mo e e iden o hose who ollowed ecommenda ions (75.7% o 56.8%) and hose who exceeded ecommenda ions (75.2% o 63.7%). Figu e 1. Concu en pu chases o ciga e es and NRT in he i s mon h (n=844 pu chases) and las mon h (n=286 pu chases) o a pu chase episode DISCUSSION The s udy’s indings sugges ha mos consume s in Finland a e no pu chasing NRT o e a su icien pe iod o ime. Ou es ima e o hose who pu chased NRT o ou weeks o less (63.5%) is ema kably simila o use o he medica ion o e he same pe iod epo ed om he ITC Fou –Coun y Su ey (62.9%)13. The cu en s udy also e ealed ha he amoun o NRT pu chased in he i s ou weeks o a pu chase episode was e y small in hose who e mina ed ea ly ela i e o hose who p oceeded o pu chase NRT o e leng hie pe iods. The co ela ion be ween amoun and du a ion o NRT pu chases could be indica i e o success o ailu e o a cessa ion a emp 18. As epo ed in o he s udies13,19, he wo mos commonly ci ed explana ions o p ema u ely e mina ing NRT use, elapse o smoking and medica ion side e ec s, a e legi ima e easons acco ding o manu ac u e s’ guidelines. A hi d explana ion may be aced o smoke s’ conce ns abou becoming dependen on NRT. Some smoke s om Finland, who pa icipa ed in an online o um (i.e. STUMPPI qui line)20, ad oca ed he s ic and Insu icien (≤ 4 weeks) 0% 40% 20% 60% 90% 10% 50% 80% 30% 70% 100% Fi s Mon h (n=430) Fi s Mon h (n=111) Fi s Mon h (n=74) Fi s Mon h (n=108) Fi s Mon h (n=121) 1 Mon h only Las Mon h (n=79) Las Mon h (n=44) Las Mon h (n=72) Las Mon h (n=91) Pu chases exceeding 28 days Pu chases wi hin 28 days, bu no same week Pu chases wi hin 7 days, bu no same day Same day pu chases o ciga e es and NRT Recomm. (9–12 weeks)Minimum (5–8 weeks) Ex ended (13–24 weeks) Exceeded (>24 weeks) Resea ch Pape Tobacco Induced Diseases Tob. Induc. Dis. 2019;17(May):39 h ps://doi.o g/10.18332/ id/108537 7 limi ed use o NRT in an e o o a oid nico ine dependence. Ye , he same smoke s acknowledged ha such p ac ice could lead o an unsuccess ul qui a emp . In ou s udy, i is likely ha many ea ly e mina o s ailed o qui smoking as a unc ion o hei equen ciga e e pu chases and dependence on nico ine. A ou h explana ion o ea ly e mina ion could be aced o imp ope NRT use a ising om lack o ins uc ion and guidance. One s udy epo ed ha among US smoke s who used nico ine gum, only 58% ead any o he p oduc inse s21. Ou es ima e o he NRT pu chases exceeding 24 weeks (13.2%) was highe han expec ed. The mos likely explana ion is ha he US s udy, which epo ed 6% o pa icipan s who pu chased NRT o a minimum o six mon hs9, es ima ed incidence o pe sis en use, whe eas, ou s udy es ima ed p e alence o pe sis en use. Some pe sis en use s iden i ied in ou s udy in 2016 p obably used NRT h oughou 2015, leading o a highe es ima e han ha epo ed by Shi man e al.9. Ano he explana ion o ou highe es ima e is he Eu opean guideline ecommending he ex ended use o NRT22. In con as , he s udy conduc ed by Shi man e al.9 was conduc ed p io o he US FDA’s ecommenda ion o ex end du a ion o NRT use10. I is no able ha 71.6% e sus 50% o he ea ly e mina o s and pe sis en use s, espec i ely, had pu chased ciga e es in 2016. While we do no know he smoking s a us o hose who did no pu chase ciga e es, i is likely ha mos we e o me smoke s who eplaced hei habi wi h pha maceu ical nico ine. Cu en knowledge sugges s ha pe sis en use o NRT is less ha m ul han smoking ciga e es. Unlike o he g oups in ou s udy, he pe sis en use s did no cu ail hei use o NRT o se e al weeks ollowing ini ia ion o a pu chase episode. While his is a conce n, dependence on o al NRT is an uncommon occu ence23. In lieu o an amendmen o he Medicines Ac , e aile s could ake a numbe o ac ions o inc ease cus ome s’ adhe ence o NRT guidelines. They could, o example, inc ease cus ome s’ exposu e o he guidelines h ough he dissemina ion o mo e ins uc ional b ochu es. Ano he op ion, which is cu en ly being discussed by he S G oup, is an online da abase o pu chases o cus ome s en olled in he loyal y p og am. A he momen , conside a ion is being gi en o acking g oce y pu chases o imp o ing he heal h and nu i ion o cus ome s. I such a da abase we e o be de eloped, hen NRT pu chases could be eadily added along wi h ecommenda ions o inc easing cus ome s’ p ope use o he medica ion. An online ea u e would be pa icula ly use ul because o he high pe cen o young cus ome s (<30 yea s old) who use he In e ne bu do no p ope ly use NRT. While any one o hese e aile ini ia i es could modi y consume beha io , hey a e no p o en as e ec i e in e en ions. Limi a ions The ime-se ies analysis o e ail pu chases o ciga e es and NRT minimized he measu emen e o ha equen ly occu s in su ey da a. Ye , he e a e limi a ions ha need o be acknowledged wi h ega d o using pu chase da a. Fi s , he ciga e e and NRT pu chases could ha e been used by someone o he han he pu chase , such as a amily membe . Da a indica ing mo e han one smoke pe smoking household9 sugges he possibili y o mul iple qui a emp s wi hin he same household. Consequen ly, single pu chase s o NRT wi hin such households could ha e in la ed ou es ima e o NRT pu chases pe cus ome . The second limi a ion is ha a loyal y membe may ha e pu chased NRT in an S G oup ma ke , bu pu chased ciga e es elsewhe e (o ice e sa); consequen ly, nei he smoking s a us no p e alence o NRT use could be de e mined. Thi d, ou ca ego iza ion o NRT consume g oups was based on pu chase episodes in 2016, no pu chase inciden s. This limi a ion p ecluded us om es ima ing an indi idual’s isk o becoming a pe sis en use o NRT. We could no de e mine i he longes pu chase episode o NRT occu ed in 2016, o some pe iod p io o 2016. Fu he , we could no de e mine i an indi idual who pu chased NRT in he la e pa o 2016 con inued o pu chase NRT h oughou 2017. An addi ional limi a ion was he inabili y o assess whe he he s udy’s indings can be gene alized om he sample o loyal y membe s who pu chased NRT o he popula ion o NRT consume s in sou he n Finland. One su oga e measu e o pu chasing NRT is smoking s a us. Ye , he absence o da a on smoking s a us in he sample and popula ion p ecluded us om weigh ing ou es ima es acco ding o popula ion p opo ions o smoke s and non-smoke s. Since he highly educa ed we e o e sampled in ou s udy, i is likely ha any such weigh ing scheme would Resea ch Pape Tobacco Induced Diseases Tob. Induc. Dis. 2019;17(May):39 h ps://doi.o g/10.18332/ id/108537 8 ha e yielded a p opo ion o ea ly e mina o s o NRT ha exceeded he p opo ion epo ed in he cu en s udy. Ye , in he absence o a weigh ing scheme, he es ima e o ea ly e mina ion o he apy was al eady in la ed by he highe pe cen age o emale cus ome s (who we e o e sampled) e sus male cus ome s who e mina ed pu chases o NRT p ema u ely. In con as , Balm o d e al.13 epo ed ha emale smoke s had g ea e odds, hough non- signi ican , han male smoke s o comple ing a cou se o ea men (OR=1.24, 95% CI: 0.95–1.61). The absence o p oduc ype (e.g. gum, lozenge) and p oduc cha ac e is ics (e.g. package size) p ecluded an accu a e assessmen o whe he a consume pu chased a su icien amoun o NRT o co e a 4-week pe iod. The al e na i e o using mo e de ailed da a om Nielsen Company’s Homescan Consume Panel was no an op ion because o he insu icien numbe o pa icipa ing households in Finland (n=5000; pe s. comm. E han Ma ko i z, Manage , Clien Solu ions, Nielsen). E en i mo e de ailed da a we e a ailable, such in o ma ion would no ha e accoun ed o he a iabili y in daily use o o al NRT9. CONCLUSIONS This s udy suppo s indings om su ey da a indica ing ha mos use s o NRT e mina e he he apy p ema u ely. In he la es p oposal o end obacco use in Finland by he yea 2030, a wo king g oup commissioned by he Minis y o Social A ai s and Heal h ecommended amending he Medicines Ac o g ea e moni o ing o he dis ibu ion channels o OTC NRT24. Ye , gi en he unce ain y o any amendmen o he Ac , i would be p uden o e aile s o ake s eps o inc ease cus ome s’ knowledge o p ope NRT use and awa eness o NRT pu chases. REFERENCES 1. Shi man S, Sweeney CT. Ten yea s a e he Rx- o- OTC swi ch o nico ine eplacemen he apy: wha ha e we lea ned abou he bene i s and isks o non- p esc ip ion a ailabili y? Heal h Policy. 2008;86(1):17- 26. doi:10.1016/j.heal hpol.2007.08.006 2. Shi man S, Fe guson SG, Hellebusch SJ. Physicians’ counseling o pa ien s when p esc ibing nico ine eplacemen he apy. Addic Beha . 2007;32(4):728-739. doi:10.1016/j.addbeh.2006.06.021 3. Solbe g LI, Asche SE, Boyle RG, Bouche JL, P onk NP. F equency o physician-di ec ed assis ance o smoking cessa ion in pa ien s ecei ing cessa ion medica ions. A ch In e n Med. 2005;165(6):656-660. doi:10.1001/a chin e.165.6.656 4. Ku ko T, Linden K, Vasama M, Pie ila K, Ai aksinen M. Nico ine eplacemen he apy p ac ices in Finland one yea a e de egula ion o he p oduc sales--has any hing changed om he communi y pha macy pe spec i e? Heal h Policy. 2009;91(3):277-285. doi:10.1016/j.heal hpol.2008.12.013 5. Aal o-Se ala, V and A Ala an a. E ec o de egula ion on he p ices o nico ine eplacemen he apy p oduc s in Finland. Heal h Policy. 2008. 86(2-3): p. 355-62. doi:10.1016/j.heal hpol.2007.11.013 6. Ku ko T. De egula ion o Nico ine Replacemen The apy P oduc s in Finland: Reason o Pha maceu ical Policy Changes and Re lec ions on Smoking Cessa ion P ac ices. Finland: Uni e si y o Helsinki; 2005. 7. Na ional Agency o Medicines and Social Insu ance Ins i u ion. Sales o nico ine eplacemen he apy p oduc s. Helsinki, Finland: F.S.o. Medicines; 2007. 8. Shi man S, Hughes JR, Di Ma ino ME, Sweeney CT. Pa e ns o o e - he-coun e nico ine gum use: pe sis en use and concu en smoking. Addic ion. 2003;98(12):1747- 1753. doi:10.1111/j.1360-0443.2003.00575.x 9. Shi man S, Hughes JR, Pilli e i JL, Bu on SL. Pe sis en use o nico ine eplacemen he apy: an analysis o ac ual pu chase pa e ns in a popula ion based sample. Tob Con ol. 2003;12(3):310-316. doi:10.1136/ c.12.3.310 10. Fuci o LM, Ba s MP, Fo ay A, Rojewski AM, Shi man S, Selby P, Wes R, Foulds J, Toll BA, W i ing Commi ee o he SRNT Policy and T ea men Ne wo ks. Add essing he e idence o FDA nico ine eplacemen he apy label changes: a policy s a emen o he Associa ion o he T ea men o Tobacco use and Dependence and he Socie y o Resea ch on Nico ine and Tobacco. Nico ine Tob Res. 2014;16(7):909-914. doi:10.1093/n /n u087 11. Zapawa LM, Hughes JR, Benowi z NL, Rigo i NA, Shi man S. Cau ions and wa nings on he US OTC label o nico ine eplacemen : wha ’s a doc o o do? Addic Beha . 2011;36(4):327-332. doi:10.1016/j.addbeh.2010.12.003 12. Zhang B, Cohen JE, Bondy SJ, Selby P. Du a ion o nico ine eplacemen he apy use and smoking cessa ion: a popula ion-based longi udinal s udy. Am J Epidemiol. 2015;181(7):513-520. doi:10.1093/aje/kwu292 13. Balm o d J, Bo land R, Hammond D, KM Cummings. Adhe ence o and easons o p ema u e discon inua ion om s op-smoking medica ions: da a om he ITC Fou - Coun y Su ey. Nico ine Tob Res. 2011;13(2):94-102. doi:10.1093/n /n q215 14. Minis y o Social A ai s and Heal h. Roadmap o a Tobacco- F ee Finland: Ac ion Plan on Tobacco Con ol. Helsinki, Finland: Minis y o Social A ai s and Heal h; 2014. Resea ch Pape Tobacco Induced Diseases Tob. Induc. Dis. 2019;17(May):39 h ps://doi.o g/10.18332/ id/108537 9 15. Joossens L, Raw M. The Tobacco Con ol Scale 2016 in Eu ope. Belgium, B ussels: A.o.E.C. Leagues; 2017. 16. Ne alainen J, E kkola M, Saa ija i H, Nappila T, Fogelholm M. La ge-scale loyal y ca d da a in heal h esea ch. Digi Heal h. 2018;4. doi:10.1177/2055207618816898 17. P elimina y and unpublished da a om he LoCa d da a collec ion o 2016-2018, 2019. 18. Raupach T, B own J, He bec A, B ose L, Wes R. A sys ema ic e iew o s udies assessing he associa ion be ween adhe ence o smoking cessa ion medica ion and ea men success. Addic ion. 2014;109(1):35-43. doi:10.1111/add.12319 19. Bu ns EK, Le inson AH. Discon inua ion o nico ine eplacemen he apy among smoking-cessa ion a emp e s. Am J P e Med. 2008;34(3):212-215. doi:10.1016/j.amep e.2007.11.010 20. Ku ko T, Linden K, Kols ela M, Pie ila K, Ai aksinen M. Is nico ine eplacemen he apy o e alued in smoking cessa ion? Analysis o smoke s’ and qui e s’ communica ion in social media. Heal h Expec . 2015;18(6):2962-2977. doi:10.1111/hex.12280 21. Bansal MA, Cummings KM, Hyland A, Gio ino GA. S op-smoking medica ions: who uses hem, who misuses hem, and who is misin o med abou hem? Nico ine Tob Res. 2004;6(Suppl 3):S303-310. doi:10.1080/14622200412331320707 22. Royal College o Physicians. Ha m educ ion in nico ine addic ion: Helping people who can’ qui . London, England: Royal College o Physicians; 2007. 23. Hughes JR, Pilli e i JL, Callas PW, Callahan R, Kenny M. Misuse o and dependence on o e - he-coun e nico ine gum in a olun ee sample. Nico ine Tob Res. 2004;6(1):79-84. doi:10.1080/14622200310001656894 24. Minis y o Social A ai s and Heal h: Smoke- ee Finland h ough be e obacco and nico ine policy. h ps://s m. i/a ikkeli/-/asse _publishe / yo yhma- upakka-ja- niko iinipoli iikkaa-kehi amalla-suomi-sa u omaksi. Accessed May 31, 2018. ACKNOWLEDGEMENTS The au ho s hank he S G oup o he use o e ail pu chases om membe s o he cus ome loyal y p og am. The p ima y au ho hanks he Fulb igh Finland Founda ion o p o iding sala y and o he esou ces which made his p ojec possible. We hank h ee anonymous e iewe s o hei help ul sugges ions. CONFLICTS OF INTEREST The au ho s decla e ha hey ha e no compe ing in e es s, inancial o o he wise, ela ed o he cu en wo k. T. Ku ko is a membe o he wo king g oup appoin ed by he Finnish Medical Socie y Duodecim and he Finnish Associa ion o Gene al P ac ice o he Tobacco and Nico ine Dependency, P e en ion and T ea men Cu en Ca e Guideline. The es o he au ho s ha e also comple ed and submi ed an ICMJE o m o disclosu e o po en ial con lic s o in e es . FUNDING The s udy was suppo ed by he Finland Fulb igh Founda ion. AUTHORS’ CONTRIBUTIONS D.S.T., J.J. and J.N. o mula ed he hypo hesis and analy ical me hodology. D.S.T. and J.J. coded he da a and conduc ed he s a is ical analyses. T.K. and A.H.R. p o ided policy insigh , guidance on clinical ecommenda ions and pa e ns o N.R.T. use ypically obse ed in p ac ice. D.S.T. w o e mos o he manusc ip and J.J., T.K., A.H.R. and J.N. p o ided c i ical e iew and edi ing. PROVENANCE AND PEER REVIEW No commissioned; ex e nally pee e iewed.