Is There Successful Aging for Nonagenarians? The Vitality 90+ Study

Hindawi Publishing Co po a ion Jou nal o Aging Resea ch Volume 2012, A icle ID 868797, 9pages doi:10.1155/2012/868797 Resea ch A icle Is The e Success ul Aging o Nonagena ians? The Vi ali y 90+ S udy Lily Nos a y,1Ty i Sa keala,2An i He onen,1and Ma ja Jylh¨ a1 1Ge on ology Resea ch Cen e and School o Heal h Sciences, 33014 Uni e si y o Tampe e, Tampe e, Finland 2Finnish Cance Regis y, Ins i u e o S a ical and Epidemiological Cance Resea ch, Pieni Roobe inka u 9, 00130 Helsinki, Finland Co espondence should be add essed o Lily Nos a y, lily.nos a [email p o ec ed] Recei ed 23 Ma ch 2012; Re ised 12 Sep embe 2012; Accep ed 12 Sep embe 2012 Academic Edi o : Lo e a DiPie o Copy igh © 2012 Lily Nos a y e al. This is an open access a icle dis ibu ed unde he C ea i e Commons A ibu ion License, which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed. Objec i es. This s udy was designed (1) o es ima e he p e alence o success ul aging among nonagena ians based on six diffe en models and (2) o in es iga e whe he success ul aging is associa ed wi h socio-demog aphic ac o s. Me hods.Amailedsu ey was conduc ed wi h people aged 90+ in Tampe e in 2010. Responses we e ecei ed om 1283 people. The p e alence o success ul aging was measu ed by six mul idimensional models including physical, social, and psychological componen s. Age, sex, ma i al s a us, le el o educa ion, and place o li ing we e s udied as ac o s associa ed wi h success ul aging. Resul s.Thep e alenceo success ul aging a ied om 1.6% o 18.3% depending on he model applied. Success ul aging was mo e p e alen in men, and also mo e p e alen among communi y-li ing people. In mos models, success ul aging was also associa ed wi h younge age, being ma ied, and a highe le el o educa ion. Discussion. Models which emphasize he absence o disease and ac i i y as c i e ia o success ul aging may no be he mos ele an and applicable in oldes old. Ins ead, p e e ence should be gi en o models ha ocus mo e on au onomy, adap a ion and sense o pu pose. Age-sensi i e app oaches would help us be e unde s and he po en ial o success ul aging among indi iduals who al eady ha e success in longe i y. 1. In oduc ion Inc easing longe i y is one o he g ea achie emen s o ou ci iliza ion, bu i has also gi en ise o discussion abou good and success ul aging. The concep o success ul aging has a ac ed much deba e, bu he e is s ill no uni e sally accep ed de ini ion o s anda d measu emen ool o i . The Encyclopedia o Aging de ines success ul aging as su i al (longe i y), heal h (lack o disabili ies), and li e sa is ac ion (happiness) [1]. I appea s ha he main sou ces o difficul y lay in he ambigui y o he meaning o “success,” in he complexi y o he aging p ocess, he apid changes aking place in socie y, and he changing cha ac e is ics o he olde popula ion. Discussions on success ul aging ha e aken wo main pe spec i es: one de ines success ul aging as a s a e o being, while he o he unde s ands i as a p ocess o adap a ion, desc ibed as doing he bes wi h wha one has [2]. S udies aking he adap a ion app oach ha e o en ound ha olde people hemsel es eel hey a e aging success ully, e en hough adi ional quan i a i e models say o he wise [3,4]. Success ul aging as a s a e o being, hen, is an objec i e measu able condi ion a a ce ain poin in ime, demons a ing he posi i e ex eme o no mal aging. The mos in luen ial model o success ul aging as a s a e o being was in oduced by Rowe and Kahn [5–8], who cha ac e ize “success” as absence o disease and disabili y, main ained physical and men al unc ioning, and ac i e engagemen wi h li e. Many s udies and de ini ions ake he iew ha success ul aging is possible only among indi iduals wi hou disease and impai men . Ob iously such ca ego iza ions a e likely o exclude mos olde people, ypically he oldes -old, om he possibili y o success ul aging [9]. Success ul aging is o cou se impossible in he absence o aging. S ill, acco ding o Bowling [3], longe i y is only a ely men ioned in lay o biomedical de ini ions. In s udies using quan i a i e measu es, younge age is one o he mos egula p edic o s o success ul aging [10,11], and he a e o “success” d ops d ama ically in e y old age. This may la gely be due o he usual ocus on physical de ici s. 2 Jou nal o Aging Resea ch Indeed, se e al esea che s ha e emphasized he need o use mul idimensional models and o adop diffe en concep ual app oaches o s udying diffe en age g oups [3,12]. Recen ly, Young e al. [13] sugges ed ha success ul aging may coexis wi h diseases and unc ional limi a ions i compensa o y psychological and social mechanisms a e used. Thei model conside s h ee impo an p inciples: he he e ogenei y o aging, mul iple pa hways o success ul aging, and indi idual compensa ion mechanisms o adjus o age- ela ed changes. The oldes -old g oup o nonagena ians mee s he key biomedical c i e ion o success ul aging ha is longe i y. They a e also a apidly g owing age g oup ha is he e oge- neous in e ms o heal h and unc ioning: a la ge majo i y ha e some heal h p oblems bu a e independen in basic e e yday ac i i ies [14]. In his s udy, we in es iga e success ul aging in an unselec ed popula ion o nonagena ians, applying se e al diffe en models ha include physical, social, and psycho- logical dimensions. The models diffe wi h espec o he h eshold o “success” on he physical, social, and psycho- logical dimensions. Ou aim is no o in oduce an ideal o uni e sal model, bu a he o demons a e he a ia ion in he p e alence o success ul aging by applying diffe en c i e ia. The i s objec i e o his s udy was o cons uc six diffe en models o success ul aging and o use hese models o es ima e he p e alence o success ul aging among nonagena ians. The second objec i e was o in es iga e whe he success ul aging in nonagena ians, de ined in se e al diffe en ways, is associa ed wi h sociodemog aphic ac o s. 2. Me hods 2.1. Da a. The Vi ali y 90+ s udy is a popula ion-based mul idisciplina y esea ch p og am on nonagena ians in he ci y o Tampe e, Finland. In he con ex o his p og am, mailed su eys we e conduc ed wi h all communi y-dwelling people in 1996 and 1998, and wi h bo h communi y-dwelling and ins i u ionalized people ou imes since 2001. This s udy used he da a om he mailed su ey in 2010. A ques ionnai e was sen o all indi iduals aged 90 o o e in Tampe e (N=1630). Responses we e ecei ed om 1283 people, gi ing a esponse a e o 79%. P oxy esponses we e ob ained om 22% o he subjec s who we e hemsel es unable o answe he ques ions. Fo addi ional 20%, he esponden chose he answe s bu someone else helped in eading he ques ions o w i ing down he answe s. The esea ch p o ocol was app o ed by he Ci y o Tampe e E hics Commi ee. In o med consen was ob ained om all esponden s o hei legal ep esen a i es. 2.2. Independen Va iables. We explo ed he associa ions o i e sociodemog aphic ac o s wi h success ul aging: age, sex, ma i al s a us, le el o educa ion, and place o li ing. Age was ca ego ized in o h ee g oups: 90-91, 92- 93, and 94–107. Ma i al s a us was classi ied as cu en ly ma ied and cu en ly unma ied, including ne e ma ied, di o ced, and widowed. Educa ion was ca ego ized in o ou g oups as low (no mo e han elemen a y schooling), middle (lowe seconda y school), high ( oca ional school, olk high school, o uppe seconda y school), and highes (college and academic educa ion). Place o li ing was dicho omized as communi y (p i a e and se ice housing) and ins i u ion ( esiden ial ca e, se ice housing wi h 24-hou assis ance, and hospi als). 2.3. Componen s o Success ul Aging. Ou dependen a iable was success ul aging. I was desc ibed by six diffe en models ha we e cons uc ed using psychical, social, and psychological indica o s. The physical componen included h ee elemen s: dis- eases, unc ioning, and senses. The pa icipan s we e asked whe he hey had been old by a doc o ha hey had (1) a hea p oblem, (2) s oke, (3) ci cula o y p oblems in he b ain, (4) diabe es, (5) a h i is, (6) Pa kinson’s diseases, (7) hip ac u e, o (8) demen ia o memo y p oblems. Fo he measu emen o unc ional abili y, he pa icipan s we e asked whe he hey we e able o pe o m independen ly (a) h ee mobili y ac i i ies: mo ing abou indoo s, walking 400 me e s, using s ai s and (b) wo ADL ac i i ies: ge ing in and ou o bed and d essing and und essing. The esponse op ions, (1) yes, wi hou difficul y; (2) yes, wi h difficul y; (3) only wi h help; (4) no a all, we e ca ego ized as independen (1 + 2) and dependen (3 + 4). The pa icipan s we e also asked whe he hey we e able o ead he newspape , wi h glasses i hey used glasses ( ision), and o hea wha ano he pe son was saying when hey we e alone wi h hem, wi h hea ing aid i hey used a hea ing aid (hea ing). The psychological componen was desc ibed by h ee a iables. The pa icipan s we e asked whe he hey suffe ed om dep ession o had dep essi e eelings (yes, no). P esen sel - a ed heal h was ca ego ized as a e age o good ( e y good, ai ly good, and a e age) and poo ( ai ly poo and poo ). Sel - a ed heal h was included in he psychological componen s because i is a subjec i e measu e wi h no p ede e mined c i e ia: i e lec s no only he mo e objec i e componen s o heal h, bu also and impo an ly he age- ela ed way in which he indi idual adjus s and adap s o diffe en heal h p oblems [15]. The pa icipan s we e also asked whe he hey hough i was good o people o li e o be 100 yea s (yes, no). The social componen was measu ed by wo ques ions: he equency o mee ings wi h child en (six ca ego ies om oday o yes e day o se e al yea s ago) and he equency o alking on he phone wi h amily membe s o iends (six ca ego ies om oday o yes e day o se e al yea s ago). One- i h (20.1%) o he esponden s had no child en. I hese pa icipan s had had elephone con ac s du ing he pas wo weeks, hey we e ca ego ized as ha ing had con ac wi h child en. The pe cen age o missing da a a ied be ween he diffe en a iables. The highes igu es we e eco ded o wo psychological a iables. Pa o he eason o his was ha hese ques ions we e no asked o p oxy esponden s. Mos o hese pa icipan s li ed in ins i u ions and had mul iple heal h p oblems. To a oid educing he numbe o pa icipan s in he analyses, we ca ego ized bo h p oxy esponses and o he missing alues in hese wo a iables a he nega i e ex eme (poo sel - a ed heal h and hinking Jou nal o Aging Resea ch 3 Success ul aging Physical componen Psychological componen Social Diseases Physical unc ioning Dep ession Li ing will Mee ing child en Telephone use Vision and componen hea ing Sel - a ed heal h Figu e 1: Th ee componen s o success ul aging. ha i is no good o li e o be 100). This impu a ion was done o a oid o e es ima ion o he p e alence o success ul aging, which would happen i he ailes pa icipan s we e los om he analyses. 2.4. Cons uc ing Six Models o Success ul Aging. Following Rowe and Kahn [6] and Young e al. [13], we de ined suc- cess ul aging as consis ing o h ee componen s as shown in Figu e 1. Six diffe en models we e cons uc ed wi h diffe en h esholds. The main diffe ences be ween he models a e in he physical componen , whe e we de ined ou al e na i e c i e ia o “success,” anging om mos o leas demanding as ollows: C i e ion 1: absence o disease + good ision and hea ing + independence in all i e ac i i ies. C i e ion 2: less han h ee diseases, no demen ia, good ision and hea ing, and independence in ADL and mo ing abou indoo s (independen in 3 easie ac i i ies). C i e ion 3: no demen ia, good ision and hea ing, and independence in all i e ac i i ies. C i e ion 4: good ision and hea ing, and indepen- dence in all i e ac i i ies. In he psychological componen , “success” was de ined as absence o dep essi eness, a e age o good sel - a ed heal h, and ag eemen wi h he iew ha i is good o li e o be 100. In he social componen , “success” was de ined as ha ing me one’s child en and ha ing alked on he phone wi h amily membe s o iends du ing he pas wo weeks. The six models o success ul aging we e cons uc ed as ollows: Model 1: Physical componen c i e ion 1 and psycho- logical componen & social componen . Model 2: Physical componen c i e ion 2 and psycho- logical componen & social componen . Model 3: Physical componen c i e ion 3 and psycho- logical componen & social componen . Model 4: Physical componen c i e ion 4 and psycho- logical componen & social componen . Model 5: Physical componen c i e ion 3 and psycho- logical componen . Model 6: Physical componen c i e ion 3 and social componen . 2.5. Analysis. The p e alence o success ul aging in diffe en sociodemog aphic ca ego ies was compa ed by c oss abu- la ion using he Chi-squa e es . Logis ic eg ession models we e used o assess he independen associa ions o diffe en models o success ul aging wi h sociodemog aphic ac o s. Odds a ios (ORs) and 95% con idence in e als (95% CI) we e calcula ed. These analyses we e pe o med using he SPSS Package 16. 3. Resul s Mos o he pa icipan s (85.9%) we e unde 95 yea s o age, and mo e han 80% we e women. These igu es well e lec he dis ibu ions in he gene al popula ion. Only 12.1% we e s ill ma ied and 37.5% li ed in an ins i u ion. The majo i y had no mo e han elemen a y schooling (Table 1). Hea p oblems, a h i is, and demen ia we e he mos equen diseases, and only 14.7% o men and 10.2% o women did no ha e any o he eigh condi ions lis ed in he ques ionnai e. Fou in en esponden s we e independen in all i e ac i i ies, and se en in en we e independen in ADL and mo ing abou indoo s. Acco ding o diffe en c i e ia, 5.3 o 25.2% we e aging success ully i only he physical 4 Jou nal o Aging Resea ch Table 1: Popula ion cha ac e is ics. Cha ac e is ic F equency % Age (N=1283) 90–91 44.5 92–93 25.5 94+ 30.0 Gende (N=1283) Women 81.2 Men 18.8 Ma i al s a us (N=1267, missing 16) Unma ied 87.9 Ma ied 12.1 Educa ion (N=1234, missing 49) Low 56.4 Middle 9.9 High 22.7 Highe 11.0 Place o li ing (N=1278, missing 5) Communi y 62.5 Ins i u ion 37.5 componen was conside ed. In he psychological componen , he p e alence o success ul aging was 20%, in he social componen he igu e was ma kedly highe a 75%. Men had be e sco es han women in bo h he physical componen (mos c i e ia) and he psychological componen (Table 2). The p e alence o success ul aging a ied be ween he six models (Table 3). I was lowes (1.6%) o Model 1, which equi ed absence o all diseases, independence in all i e ac i i ies, and good ision and hea ing, in addi ion o he psychological and social componen s, and highes (18.3%) o Model 6, which diffe ed om Model 1 in ha diseases o he han demen ia we e allowed, and he psychological componen was no included. Success ul aging was signi ican ly mo e p e alen in men han women and among communi y-li ing han ins i u ionalized indi iduals, ega dless o he model. Acco ding o mos models, success- ul aging was mo e equen among hose aged 90–93 han hose aged 94+, among ma ied people, and among hose wi h a highe educa ion. Finally, logis ic eg ession models we e calcula ed o examine he independen associa ion o diffe en sociode- mog aphic indica o s wi h he six models o success ul aging (Table 4). In ou models, highe age was independen ly associa ed wi h less success ul aging. Gende was ano he p edic i e a iable, and in all models excep model 6, men we e signi ican ly mo e success ul in aging han women. Highe educa ion was a signi ican p edic o in wo models, and in Model 6 bo h hose wi h a high and he highes educa ional le el diffe ed signi ican ly om hose wi h he lowes le el o educa ion. Ma i al s a us did no play an independen ole, bu place o li ing was a signi ican de e minan o success ul aging in all bu Model 1. 4. Discussion This pape examined one he mos p ominen concep s in aging esea ch, success ul aging, by cons uc ing six diffe en models o measu e i among nonagena ians. The models we e based on wo k by Rowe and Kahn, Rowe, and Young e al. [6–8,13,16], al hough no he exac same indica o s we e used. Acco ding o Young e al. [13] and Rowe and Khan [6], success ul aging is ypically unde s ood as comp ising h ee main domains: physical (in Young e al.: physiological), psychological, and social (in Young e al.: sociological). The esul s showed ha he p e alence o success ul aging a ies ma kedly om one model o ano he , s anding a 1.6% o Model 1 ha equi ed he absence o any disease, independence in unc ioning, and he abili y o hea and ead, as well as mee ing he psychological and social c i e ia, and a 18.3% o Model 6, which equi ed he absence o demen ia, independence in unc ioning, he abili y o hea and ead, and mee ing he social c i e ia. Howe e , he main socioeconomic p edic o s emained la gely he same ac oss he models. I is ob ious ha he absence o disease is he mos demanding c i e ion o measu ing success ul aging. Dis- ease and a leas some unc ional de e io a ion a e almos ine i able in e y old age. Only 11% o he nonagena ians in ou s udy had no majo disease, and only 5.3% we e bo h ee o disease, able o hea and see, and independen in i e daily ac i i ies (physical dimension c i e ion 1). Ve y ew ea lie s udies ha e a emp ed o es ima e he p e alence o success- ul aging in nonagena ians o in e y old age in gene al. on Fabe e al. [2] classi ied only 10% o communi y-dwelling and 1.9% o ins i u ionalized pa icipan s aged 85 o o e as success ul age s. In he NonaSan eliu s udy by Fo miga e al. [17], he igu e was 12% wi h communi y-dwelling nonagena ians. These s udies emphasized he ole o heal h and physical unc ioning, bu also included some social o quali y-o -li e measu es. I is clea ha especially when he ocus is on he physical dimension, success ul aging will be e y a e among people expe iencing longe i y. Rowe and Kahn [6] included p oduc i e ac i i ies in hei model o success ul aging bu hese can ha dly be expec ed om nonagena ians. Ho gas e al. [18] showed ha he daily ac i i ies o indi iduals aged 90 o o e diffe ed om o he age g oups, and in all ca ego ies his age g oup was engaged in signi ican ly less ac i i y han o he s. This implies ha he social dimension o success ul aging among he oldes old should be measu ed using diffe en c i e ia and agains diffe en ac i i ies han in he case o he younge old and should be seen in ela ion o he si ua ion o he bes pe o me s in he same age g oup. In c oss-sec ional analysis, we limi ed ou examina ion o socioeconomic p edic o s ha a leas po en ially ha e played a ole in he li es o he indi iduals o a longe ime, and, wi h he excep ion o place o li ing, a e no supposed o be in luenced by ac o s ha we e hough o be componen s o success ul aging. In mos s udies age has eme ged as one o he s onges p edic o s o success ul aging [18]. In ou s udy, pe sons aged 94 o o e we e less likely o mee he success ul aging c i e ia han he younge age g oups. The diffe ence Jou nal o Aging Resea ch 5 Table 2: F equency (%) o he a iables composing h ee componen s o success ul aging in men and women. Va iables Men (N=226–238) % Women (N=1006–1038) % P alue All (N=1227–1283) % Physical componen No hea p oblem 42.4 47.1 0.192 46.2 No s oke 96.2 94.4 0.250 94.7 No ci cula o y p oblems in b ain 78.6 79.3 0.798 79.2 No diabe es 85.7 88.8 0.178 88.2 No a h i is 69.7 54.3 <0.001 57.2 No Pa kinson’s disease 99.6 98.3 0.148 98.6 No hip ac u e 89.2 81.3 0.003 82.8 No demen ia 66.4 59.2 0.033 60.6 No disease 14.7 10.2 0.045 11.0 2 o less diseases wi h no demen ia 39.5 29.8 0.004 31.6 Able o see 72.8 59.9 <0.001 62.3 Able o hea 68.1 71.5 0.299 70.9 Able o see and hea 53.2 48.2 0.170 49.2 Independen in i e ac i i ies 58.3 34.9 <0.001 39.2 Independen in 3 easie ac i i ies 83.0 72.5 <0.001 74.5 C i e ion 1∗7.7 4.8 <0.001 5.3 C i e ion 2∗28.3 26.1 0.500 26.5 C i e ion 3∗24.8 19.1 0.050 20.2 C i e ion 4∗33.9 23.2 <0.001 25.2 Psychological No dep essi eness 87.8 79.6 0.004 81.2 Sel - a ed heal h a e age o good 72.3 61.4 0.002 63.4 Willing o li e up o 100 yea s 42.4 24.8 <0.001 28.1 Psychological componen 34.0 16.7 <0.001 20.0 Social engagemen Me child en du ing pas 2 weeks 92.7 93.5 0.644 93.4 Phone con ac s 84.6 79.4 0.070 80.4 Social componen 78.8 74.4 0.166 75.2 ∗C i e ions. C i e ion 1: No disease, and able o hea and ead, and independen in all i e ac i i ies. C i e ion 2: Less han 3 diseases, no demen ia, able o hea and ead, and independen in h ee easie ac i i ies. C i e ion 3: No demen ia, able o hea and ead, and independen in 5 ac i i ies. C i e ion 4: Able o hea and ead, independen in 5 ac i i ies. 6 Jou nal o Aging Resea ch Table 3: P e alence o success ul aging (%) acco ding o he six models in diffe en socioeconomic ca ego ies. Models o success ul aging∗ 123456 To al p e alence 1.6 6.3 5.7 6.8 6.3 18.3 Age 90–91 1.9 7.9 6.5 7.5 7.5 23.8 92–93 2.5 6.8 7.1 8.6 7.7 21.2 94+ 0.5 3.4 3.2 4.2 3.4 7.8 P alue 0.104 0.021 0.042 0.048 0.022 <0.001 Gende Men 4.7 12.4 11.1 13.2 12.4 22.3 Women 1 5 4.5 5.5 5.1 17.5 P alue <0.001 <0.001 <0.001 <0.001 <0.001 0.095 Ma i al s a us Ma ied 3.3 11.8 11.8 14.5 12.5 24.2 No ma ied 1.4 5.6 4.9 5.9 5.6 17.6 P alue 0.095 0.003 0.001 <0.001 0.001 0.03 Place o li ing Communi y 2.3 8.4 7.6 9.1 8.2 25.9 Ins i u ion 0.6 2.9 2.5 3.1 2.5 6.1 P alue 0.026 <0.001 <0.001 <0.001 <0.001 <0.001 Educa ion Low 1.3 4.7 4.6 6 5.8 14.4 Middle 1.6 7.4 4.9 4.9 5.7 20.5 High 2.1 7.5 7.1 8.2 7.2 22.6 Highe 3 12.5 9.6 11.9 9.6 29.1 P alue 0.51 0.005 0.093 0.058 0.377 <0.001 ∗Models o success ul aging. Model 1. Heal h and unc ioning c i e ion 1 + psychological + social. Model 2. Heal h and unc ioning c i e ion 2 + psychological + social. Model 3. Heal h and unc ioning c i e ion 3 + psychological + social. Model 4. Heal h and unc ioning c i e ion 4 + psychological + social. Model 5. Heal h and unc ioning c i e ion 3 + psychological. Model 6. Heal h and unc ioning c i e ion 3 + social. be ween he age g oups was signi ican o all excep Model 1, and i was g ea es in Model 6 whe e he o e all p e alence o success ul aging was highes . A e adjus ing o o he sociodemog aphic a iables, a signi ican age diffe ence s ill pe sis ed in ou models. In ou s udy, he p e alence o success ul aging was consis en ly highe o men, and in all excep he las model he diffe ences we e also signi ican a e he adjus men s. Ea lie s udies [10] show no consis en pa e ns o gende diffe ences, bu he esul s seem o be dependen on he model used. McLaughlin e al. [11] ound no gende diffe ence in p e alence, bu highe odds o success ul aging in women a e con olling o sociodemog aphic a iables. Ou indings among nonagena ians a e only pa ly explained by he high p e alence o disabili ies and disease in women, as men had clea ly be e sco es in he psychological componen as well. These dispa i ies a e likely o e lec diffe en ial su i al, li elong diffe ences in biological, heal h, and social condi ions. Ma i al s a us was associa ed wi h success ul aging in unadjus ed analysis bu no in he adjus ed models, whe e he une en age and gende dis ibu ion o he a iable was con olled o . Educa ion is known o ha e an impac on heal h and li e s yle, and i e lec s socioeconomic s a us; he e o e, i can also be conside ed a po en ial p edic o o success ul aging. Mos o he s udies e iewed by Depp and Jes e [10] ound no diffe ences acco ding o educa ional le el, bu he analysis by McLaughlin e al. [11] in he Heal h and Re i emen S udy showed ha he p e alence o success ul aging was highe in g oups wi h a highe educa ion and household income. The s udy o P uchno e al. [19] e ealed ha a highe le el o o mal educa ion is associa ed wi h success ul aging. Ou indings wi h an olde g oup han in hese s udies showed a g aded inc ease in he p e alence o success ul aging wi h highe educa ion, al hough he diffe ence was no signi ican o all models. The disc epancy be ween he indings may a leas pa ly be due o sampling bias. In se e al s udies ins i u ionalized people and hose o lowe social posi ion we e less likely o pa icipa e [10], while ou s udy ep esen s he whole age g oup in he egion. Place o li ing is no usually conside ed a p edic o o success ul aging and in many (bu no all, see e.g., on Fabe e al. [2]) s udies samples only include communi y- dwelling indi iduals. In ou s udy, we wan ed o ake accoun o he possibili y o success ul aging e en in an ins i u ion. Jou nal o Aging Resea ch 7 Table 4: Associa ions o success ul aging, acco ding o he six models, wi h socioeconomic cha ac e is ics. A mul i a ia e logis ic eg ession model, all he p edic o s included in he model simul aneously. Odds a ios (OR) and 95% con idence in e als (CI). Models o success ul aging 123456 Age 90–91 2.74 (0.59–12.76) 1.93 (0.99–3.78) 1.68 (0.83–3.4) 1.43 (0.76–2.66) 1.82 (0.93–3.57) 2.85 (1.81–4.49) 92–93 3.93 (0.82–18.89) 2.15 (1.02–4.53) 1.76 (0.91–3.42) 1.93 (1.0–3.73) 2.14 (1.05–4.40) 2.90 (1.79–4.73) 94+ 1 1 1 1 1 1 Gende Men111111 Women 0.20 (0.07–0.54) 0.46 (0.26–0.82) 0.53 (0.29–0.96) 0.54 (0.31–0.94) 0.47 (0.27–0.83) 0.98 (0.06–1.50) Ma i al s a us Unma ied 1 1 1 1 1 1 Ma ied 0.85 (0.26–2.73) 1.17 (0.61–2.26) 1.60 (0.82–3.12) 1.71 (0.92– 3.16) 1.42 (0.74–2.7) 1.11 (0.69–1.81) Educa ion Low111111 Middle 1.64 (0.34–7.88) 1.77 (0.81–3.85) 1.16 (0.47–2.87) 0.89 (0.37–2.18) 1.10 (0.47–2.54) 1.57 (0.94–2.63) High 1.14 (0.39–3.33) 1.28 (0.72–2.3) 1.23 (0.68–2.22) 1.10 (0.64–1.91) 0.95 (0.54–1.69) 1.45 (1.00–2.11) Highe 1.35 (0.38–4.6) 2.03 (1.06–3.89) 1.48 (0.73–2.99) 1.45 (0.76–2.76) 1.14 (0.57–2.28) 2.00 (1.26–3.17) Place o li ing Communi y 3.11 (0.89–10.8) 2.48 (1.36–4.53) 2.66 (1.39–5.05) 2.64 (1.48–4.72) 3.18 (1.68–6.0) 4.30 (2.83–6.53) Ins i u ion 1 1 1 1 1 1 8 Jou nal o Aging Resea ch Howe e , he esul s showed ha he p e alence o success ul aging was clea ly lowe o hose li ing in ins i u ions, and his was also ue o he adjus ed models. Ou ea lie analyses (no shown he e) indica ed ha disease, disabili y, and p oblems wi h hea ing and seeing a e mo e p e alen in ins i u ions, as is sel - a ed heal h, which pa ly explains his inding. 4.1. S eng hs and Limi a ions. The majo s eng h o his s udy is ha i co e s he whole popula ion aged 90 o o e in he a ea conce ned, including ins i u ionalized people as well as p oxy esponses. The esponse a e was high. Ou ea lie and ongoing analyses sugges ha he in o ma ion on heal h and unc ioning collec ed by mailed ques ionnai es among nonagena ians is sufficien ly alid and eliable [20, 21]; pa icula ly as o a majo i y o hose suffe ing om demen ia, he answe s we e gi en by a p oxy esponden . In o de o gain a b oad and ho ough unde s anding o success ul aging, we included bo h physical, psychological, and social componen s in ou analyses. Unlike mos o he s udies, we also included he abili y o see and hea as an impo an con ibu ing ac o o independence and quali y o li e. The main limi a ions o ou s udy ha e o do wi h he measu es used o assess he social and psychological componen s. Ou only in o ma ion abou mee ing wi h o he people conce ned mee ings wi h child en; no da a we e a ailable abou o he amily membe s o iends. One i h o he esponden s had no child en, and we decided o gi e hem a posi i e sco e o social con ac s i hey had made o ecei ed any elephone calls du ing he pas wo weeks. One- i h o ou esponses we e om p oxies, who we e no asked abou sel - a ed heal h o li ing o be 100. The e o e, we had a high pe cen age o missing o p oxy answe s o wo ques ions ega ding he psychological dimension o success ul aging. In o de no o o e es ima e he p e alence o success ul aging, we sco ed his missing da a and p oxy answe s as nega i e. These kinds o p oblems a e una oidable in unselec ed samples o e y old people, bu hey none heless add some unce ain y o ou indings. Ano he ob ious limi a ion o ou s udy is ha we had no di ec ques ions designed o cap u e ou esponden s’ sel - e alua ions o hei li e. 4.2. Implica ions. Ou s udy in a nonselec ed popula ion o pe sons aged 90 o o e suppo s ea lie indings ha he p e alence o success ul aging is highly dependen on he model applied, bu in e e y case success ul aging is associa ed wi h age, gende , and socioeconomic s a us. Howe e , i is appa en ha wi h any model ha de ines success ul aging as a s a e o being and ha uses c i e ia commonly used o younge age g oups, success ul aging emains a a e si ua ion among he oldes old. An inc eased likelihood o heal h and unc ional p oblems, o en ollowed by educed oppo uni ies o ac i e social engagemen , is no ma i e consequences o biological aging and ypical o ex eme longe i y. The e o e, in e y old age, a he han models emphasizing he absence o disease and ac i i y, emphasis should be gi en o app oaches ocusing on au onomy, adap a ion, and sense o pu pose [3,22,23]. These age- sensi i e app oaches would help us be e unde s and he po en ial o success ul aging among hose indi iduals who ha e al eady had success in longe i y. Acknowledgmen s Ge on ology Resea ch Cen e is a join effo be ween he Uni e si ies o Jy ¨ askyl¨ a and Tampe e. This s udy was suppo ed by G an s om he CIMO Fund o L. Nos a y and om he Compe i i e Resea ch Funding o Tampe e Uni e si y Hospi al (9M025) o M. Jylh¨ a. The au ho s wish o hank Anna-Maija Koi is o, M.S. dg ee, o he s a is ical help and ad ice. Re e ences [1] E. B. Palmo e, “Success ul ageing,” in The Encyclopedia o Aging,G.L.Maddox,S.W.Sussman,R.C.A chleye al.,Eds., pp. 914–915, Sp inge , New Yo k, NY, USA, 1995. [2] M. on Fabe , A. Boo sma- an de Wiel, E. an Exel e al., “Success ul aging in he oldes old: who can be cha ac e ized as success ully aged?” A chi es o In e nal Medicine, ol. 161, no. 22, pp. 2694–2700, 2001. [3] A. Bowling, “Aspi a ions o olde age in he 21s cen u y: wha is success ul aging?” In e na ional Jou nal o Aging and Human De elopmen , ol. 64, no. 3, pp. 263–297, 2007. [4] A. Bowling, “E a um: Lay pe cep ions o success ul ageing: Findings om a na ional su ey o middle aged and olde adul s in B i ain (Eu opean Jou nal Ageing (2006) ol. 3 (123-136) 10.1007/s10433-006-0032-2),” Eu opean Jou nal o Ageing, ol. 4, no. 1, pp. 57–58, 2007. [5] J. W. Rowe and R. L. Kahn, “Human aging: usual and success ul,” Science, ol. 237, no. 4811, pp. 143–149, 1987. [6] J. W. Rowe and R. L. Kahn, “Success ul aging,” Ge on ologis , ol. 37, no. 4, pp. 433–440, 1997. [7] J. W. Rowe and R. L. Kahn, “Success ul aging,” Aging—Clinical and Expe imen al Resea ch, ol. 10, no. 2, a icle 142, 1998. [8] J. W. Rowe, Success ul Ageing, Dell, New Yo k, NY, USA, 1999. [9] M. Minkle and P. Fadem, “Success ul aging,” Jou nal o Disabili y Policy S udies, ol. 12, no. 4, pp. 229–235, 2002. [10] C. A. Depp and D. V. Jes e, “De ini ions and p edic o s o success ul aging: a comp ehensi e e iew o la ge quan i a i e s udies,” Ame ican Jou nal o Ge ia ic Psychia y, ol. 14, no. 1, pp. 6–20, 2006. [11] S. J. McLaughlin, C. M. Connell, S. G. Hee inga, L. W. Li, and J. S. Robe s, “Success ul aging in he Uni ed S a es: p e alence es ima es om a na ional sample o olde adul s,” Jou nals o Ge on ology—Se ies B, ol. 65, no. 2, pp. 216–226, 2010. [12] R. A. P uchno, M. Wilson-Gende son, and F. Ca w igh , “A wo- ac o model o success ul aging,” Jou nals o Ge on ology—Se ies B, ol. 65, no. 6, pp. 671–679, 2010. [13] Y. Young, K. D. F ick, and E. A. Phelan, “Can success ul aging and ch onic illness coexis in he same indi idual? A mul idimensional concep o success ul aging,” Jou nalo he Ame ican Medical Di ec o s Associa ion, ol. 10, no. 2, pp. 87– 92, 2009. [14] T. Sa keala, T. Nummi, M. Vuo isalmi, A. He onen, and M. Jylh¨ a, “Disabili y ends among nonagena ians in 2001–2007: i ali y 90+ s udy,” Eu opean Jou nal o Ageing, ol.8,no.2, pp. 87–94, 2011. Jou nal o Aging Resea ch 9 [15] M. Jylh¨ a, “Wha is sel - a ed heal h and why does i p edic mo ali y? Towa ds a uni ied concep ual model,” Social Science and Medicine, ol. 69, no. 3, pp. 307–316, 2009. [16] Y. Young, M. Y. Fan, J. M. Pa ish, and K. D. F ick, “Valida ion o a no el success ul aging cons uc ,” Jou nalo heAme ican Medical Di ec o s Associa ion, ol. 10, no. 5, pp. 314–322, 2009. [17] F. Fo miga, A. Fe e , D. Chi i e, M. Rubio-Ri as, S. Cue po, and R. Pujol, “P edic o s o long- e m su i al in nonagena - ians: he NonaSan eliu s udy,” Age and Ageing, ol.40,no.1, A icle ID a q127, pp. 111–116, 2011. [18] A. L. Ho gas, H. U. Wilms, and M. M. Bal es, “Daily li e in e y old age: e e yday ac i i ies as exp ession o success ul li ing,” Ge on ologis , ol. 38, no. 5, pp. 556–568, 1998. [19] R. A. P uchno, M. Wilson-Gende son, M. Rose, and F. Ca w igh , “Success ul aging: ea ly in luences and con empo- a y cha ac e is ics,” Ge on ologis , ol. 50, no. 6, pp. 821–833, 2010. [20] S. Goebele , M. Jylh¨ a, and A. He onen, “Sel - epo ed med- ical his o y and sel - a ed heal h a age 90. Ag eemen wi h medical eco ds,” Aging—Clinical and Expe imen al Resea ch, ol. 19, no. 3, pp. 213–219, 2007. [21] M. Vuo isalmi, Examining Sel -Ra ed Heal h in Old Age Me hodological S udy o Su ey Ques ions,Tampe eUni e si y P ess, Tampe e, Finland, 2007. [22] P. B. Bal es and M. M. Bal es, Success ul Aging: Pe spec i es om he Beha io al Sciences, Camb idge Uni e si y P ess, Camb idge, UK, 1990. [23] M. M. Bal es and L. L. Ca s ensen, “The p ocess o success ul ageing,” Ageing and Socie y, ol. 16, no. 4, pp. 397–422, 1996.