Changes in older people’s care profiles during the last 2 years of life, 1996–1998 and 2011–2013: a retrospective nationwide study in Finland

1 Aal onenM, e al. BMJ Open 2017;7:e015130. doi:10.1136/bmjopen-2016-015130 Open Access Abs Ac Objec i es The ime o dea h is inc easingly pos poned o a e y high age. How his change a ec s he use o ca e se ices a he popula ion le el is unknown. This s udy analyses he ca e p o iles o olde people du ing hei las 2 yea s o li e, and in es iga es how hese p o iles di e o he s udy yea s 1996–1998 and 2011–2013. Design Re ospec i e c oss-sec ional na ionwide da a d awn om he Ca e Regis e o Heal h Ca e, he Ca e Regis e o Social Ca e and he Causes o Dea h Regis e . The da a included he use o hospi al and long- e m ca e se ices du ing he las 2 yea s o li e o all hose who died in 1998 and in 2013 a he age o ≥70 yea s in Finland. Me hods We cons uc ed ou ca e p o iles using wo c i e ia: (1) numbe o days in ound- he-clock ca e ( s a home) in he las 2 yea s o li e and (2) ca e ansi ions du ing he las 6 mon hs o li e (ie, end-o -li e ca e ansi ions). esul s Be ween he s udy pe iods, he a e age age a dea h and he numbe o diagnoses inc eased. Mos olde people (1998: 64.3%, 2013: 59.3%) li ed a home un il hei las mon hs o li e (p o ile 2) a e which hey mo ed in o hospi al o long- e m ca e acili ies. This p o ile became less common and he p o iles wi h a high use o ca e se ices became mo e common (p o iles 3 and 4 oge he in 1998: 25.0%, in 2013: 30.9%). People wi h demen ia, women and he oldes old we e o e - ep esen ed in he la e p o iles. In bo h s udy pe iods, ewe han one in en s ayed a home o he whole las 6 mon hs (p o ile 1). conclusions Pos ponemen o dea h o a e y old age may ansla e in o mo e se e e disabili y in he las mon hs o yea s o li e. Ca e sys ems mus be p epa ed o longe pe iods o long- e m ca e se ices needed a he end o li e. In ODuc IOn D i en mainly by longe a e age li espans, popula ion ageing c ea es signi ican new challenges o heal h and social se ice sys ems wo ldwide.1 2 Findings on heal h ends and he de elopmen o heal hy li e expec ancy a y, bu wha is cons an is ha , on a e age, physical and cogni i e capa- bili ies will decline wi h inc easing old age and app oaching dea h.3–7 The use o ca e se ices8–10 and ca e ansi ions, ha is, mo e- men s be ween ca e si es,11–13 a e hus na u- ally mo e equen a he end o li e (EoL) han ea lie . Declining old-age mo ali y has pos poned he EoL pe iod and dea h o olde ages, he eby p oducing new challenges o medical se ices and long- e m ca e. In esponse o he inc easing need o ca e and ising cos s induced by popula ion ageing and longe li e, se e al coun ies ha e ealloca ed esou ces om ins i u ional ca e o suppo o li ing a home wi h home help o in home-like ca e acili ies.14–17 Finland is no excep ion. Since he la e 1990s, he p o i- sion o adi ional ins i u ional long- e m ca e has dec eased, while he a ailabili y o se ice housing wi h 24 hou s assis ance, classi ied as communi y long- e m ca e, has inc eased.18 This change has coincided wi h an inc ease in he modal age o dea h om Changes in olde people’s ca e p o iles du ing he las 2 yea s o li e, 1996–1998 and 2011–2013: a e ospec i e na ionwide s udy in Finland Ma i Aal onen,1 Leena Fo ma,1 Ju a Pulkki,1 Jani Rai anen,1,2 Pekka Rissanen,1 Ma ja Jylha1 To ci e: Aal onenM, Fo maL, PulkkiJ, e al. Changes in olde people’s ca e p o iles du ing he las 2 yea s o li e, 1996–1998 and 2011–2013: a e ospec i e na ionwide s udy in Finland. BMJ Open 2017;7:e015130. doi:10.1136/ bmjopen-2016-015130 ►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- 015130). Recei ed 11 No embe 2016 Accep ed 30 Augus 2017 1Facul y o Social Sciences (Heal h Sciences) and Ge on ology Resea ch Cen e , Uni e si y o Tampe e, Tampe e, Finland 2The UKK Ins i u e o Heal h P omo ion Resea ch, Tampe e, Finland co espondence o D Ma i Aal onen; ma i. s. aal onen@ u a. i Resea ch s eng hs and limi a ions o his s udy ►The majo s eng h o his s udy is he a ailabili y o in o ma ion on all ound- he-clock ca e use and days li ed a home du ing he las 730 days o li e o e e yone who died a age 70 o olde in 1998 o in 2013 in Finland. ►The size and eliabili y o he s udy popula ion p o ides a unique s udy composi ion ha enables he iden i ica ion o he olde popula ions’ ca e p o iles a he na ional le el, and clea ly e eals how hose p o iles ha e changed be ween he s udy yea s. ►The da a lacked in o ma ion on unc ional abili y, a ailabili y o in o mal ca e and o mal home ca e, all o which a e impo an con ibu o s o ca e use. ►De ailed in o ma ion on which ca e si es people ac ually mo ed be ween could no be iden i ied. g oup.bmj.com on Decembe 14, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om 2Aal onenM, e al. BMJ Open 2017;7:e015130. doi:10.1136/bmjopen-2016-015130 Open Access Figu e 1 The ou ca e p o iles. End o li e (EoL) e e s o he las 6 mon hs o li e. 83 o 86 yea s and a doubling in he numbe o people aged 85 o olde .19 Al hough he impac s o highe age and he close- ness o dea h on ca e use a e well known, we s ill lack a good unde s anding o pa e ns o ca e use among olde people. We ha e i ually no in o ma ion abou how hese ca e pa e ns will be a ec ed by he ongoing e o m o ca e se ices. In his s udy, we ook ad an age o exis ing na ionwide ca e egis e s o in es iga e (1) ca e p o iles in he las 2 yea s o li e o he en i e popula ion who died a he age o 70 o olde in Finland in 1998 o in 2013 and (2) he changes in hese p o iles be ween 1998 and 2013, a pe iod du ing which long- e m ca e was e o med. Me hODs Da a The da a o he s udy we e d awn om na ional egis- e s and included he use o hospi al and long- e m ca e se ices du ing he las 2 yea s o li e o all people who died in 1998 (n=34 116) and in 2013 (n=38 087) a he age o 70 o olde in Finland. In o ma ion on he use o ca e se ices o he las 730 days o each indi idual was ob ained om he Ca e Regis e o Heal h Ca e and he Ca e Regis e o Social Ca e (Na ional Ins i u e o Heal h and Wel a e). These egis e s include in o ma- ion on all ca e use deli e ed in all ound- he-clock ca e acili ies. The in o ma ion in he egis e s is illed o e e yone, hus he e is p ac ically no missing in o ma ion. The decision o ocus on he las 2 yea s o li e was based on he knowledge ha he use o di e en ca e se ices inc eases mos sha ply du ing his pe iod.8 9 The Causes o Dea h Regis e (S a is ics Finland) p o ides he da e o dea h and all causes o dea h (immedia e, unde lying, in e media e and con ibu ing causes) as s a ed on each indi idual dea h ce i ica e. Da ase s we e linked using pe sonal iden i y codes (PICs) ha emain unchanged h oughou he pe son’s li e ime. The esea che s had no access o he PICs. Pe mission o access he egis e s was ob ained om each egis a . The esea ch plan was app o ed by he Pi kanmaa Hospi al Dis ic E hics Commi ee (Decision R08192). The diagnoses we e iden i ied in he ca e egis e s and he Causes o Dea h Regis e and we e ca ego ised acco ding o he In e na ional Classi ica ion o Diseases (ICD10): demen ia (F00–F03, G30), cance (C00–C97), diabe es (E10–E14), psychosis, dep essi e symp oms o o he men al heal h diso de s excluding demen ia (F04–F99), Pa kinson’s disease o o he neu ological diseases (G00–G99 excluding G30), ch onic as hma and ch onic obs uc i e pulmona y disease (COPD) o o he espi a o y diseases (J00–J99), hip ac u e (S72), s oke (I60–I69), ischaemic and o he hea diseases excluding heuma ic and alcoholic hea diseases (I20–I25, I30– I425, I427–I52), and o he diseases o he ci cula o y sys em (I00–I15, I26–I28, I70–I99). ca e se ices The ca e se ices used and he amoun o ime spen ou side ca e acili ies was ollowed o 730 days p io o each indi idual’s dea h. The ound- he-clock ca e se ices s udied he e included p ima y ca e hospi al (heal h cen e inpa ien wa ds o e ing sho - e m hospi al ca e and ins i u ional long- e m ca e); specialised ca e hospi al (uni e si y hospi als and cen al, dis ic and p i a e hospi als); nu sing homes (ins i u ional long- e m ca e) and se ice housing wi h 24 hou s assis ance (communi y long- e m ca e). Se ice housing wi h 24 hou s assis ance (also known as shel e ed housing wi h 24 hou s assis ance) ep esen s a ela i ely new o m o ound- he-clock long- e m ca e acili ies and di e om nu sing homes. Mos nu sing homes a e publicly owned, while mo e han hal o he se ice housing acili ies wi h 24 hou s assis ance acili ies a e owned by he p i a e sec o o non-go e nmen al o ganisa ions. Municipali- ies o e ed ca e in publicly owned nu sing homes p e i- ously, howe e , in ecen yea s, hey ha e inc easingly pu chased se ices om p i a ely owned se ice housing ha p o ides 24 hou s assis ance. The eligibili y o long- e m ca e placemen ei he in a nu sing home o in se ice g oup.bmj.com on Decembe 14, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om 3 Aal onenM, e al. BMJ Open 2017;7:e015130. doi:10.1136/bmjopen-2016-015130 Open Access Table 1 Desc ip ion o he s udy popula ion, ca e ansi ions and days in ound- he-clock ca e* Yea o dea h (n) 1998 (34 116) 2013 (38 087) Age a dea h, mean† 82.4 84.3** Age (%)‡ 70–79 36.4 27.2** 80–89 46.8 46.7** 90+ 16.8 26.1** Gende (%)‡ Men 40.8 43.5** Women 59.2 56.5** Diagnoses (%)*‡ Demen ia 21.7 33.9** Cance 23.1 25.8** Diabe es 13.8 16.1** Men al diso de s 7.1 6.4** Neu ological 10.1 11.6** Respi a o y 43.6 45.5** Hip ac u e 7.3 6.1** S oke 22.8 19.9** Ischaemic hea disease 55.3 53.2** O he ci cula o y disease 27.7 41.2** Numbe o diagnoses, mean § 2.3 2.6** Numbe o ansi ions, mean (median)¶ In las 24 mon hs 7.8 (5.0) 7.9 (5.0) In las 6 mon hs 3.2 (2.0) 3.3 (2.0)** Days in ca e, mean (median)¶ Las 24 mon hs in o al 220.3 (82.0) 257.4 (93.0)** P ima y ca e hospi al 103.1 (22.0) 65.0 (18.0)** Specialised ca e hospi al 20.2 (7.0) 14.0 (4)** Nu sing home 96.6 (0) 63.5 (0)** Se ice housing wi h 24 hou s assis ance na 113.2 (0)** Days a home, mean (median)¶ 509.7 (648.0) 472.6 (637.0)** All hose who died a age 70 o olde in 1998 and 2013. *The Causes o Dea h Regis e , The Ca e Regis e o Heal hCa e, and The Ca e Regis e o Social Wel a e. †pValue e e s o independen samples - es be ween 1998 and 2013. ‡pValue e e s o χ2 es by c oss abs be ween 1998 and 2013. §Includes he abo e men ioned diagnoses. ¶pValue e e s o Mann-Whi ney U es be ween 1998 and 2013. **p<0.001. na,in o ma ion no a ailable. housing wi h 24 hou s assis ance is needs based. Se ice housing acili ies a e conside ed mo e home-like han nu sing homes, bu he o me a e expec ed o p o ide ca e o simila heal h condi ions and esiden g oups as he la e , namely, nu sing homes do. The munici- pali ies hold he main esponsibili y o he unding o nu sing home ca e, and in addi ion he esiden s pay an income-based mon hly ee o ca e and esidence. On he con a y, in se ice housing wi h 24 hou s assis ance he ees which he esiden s pay ou o pocke a y, hus he unding is no as egula ed as i is o nu sing homes.20 People we e classi ied as li ing a home when hey we e no in any o he abo e desc ibed ca e acili ies. Howe e , we ha e no da a on he use o in o mal ca e o o mal home ca e se ices. The 1998 s udy popula ion e e s o people who died in 1998: hei ca e use hus ook place du ing 1996–1998. The 2013 s udy popula ion e e s o people who died in 2013 and whose ca e use ook place du ing 2011–2013. Key concep s Use o ca e e e s o he numbe o days s ayed in di e en ound- he-clock ca e acili ies in he las 2 yea s o li e. Ca e ansi ion is de ined as mo ing om one’s home o a ca e acili y o om one ca e si e o ano he o a home discha ge o a leas one nigh . Ca e p o ile e e s o he whole ajec o y o days s ayed a di e en si es and he equency o ca e ansi ions in he las 24 mon hs. The concep EoL e e s he e o he inal 6 mon hs o li e. Iden i ica ion o p o iles Based on ea lie indings by ou esea ch g oup9 20 21 and o he s,6–8 22 we conside ed being in ound- he-clock ca e e sus li ing a home and he expe ience o ca e ansi ions as he mos impo an desc ip o s o pa e ns o ca e du ing he las yea s o li e. The e o e, we i s ca ego ised each indi idual in o a g oup by using wo c i e ia ( igu e 1): (1) whe he he deceden had li ed he majo i y o ime in he las 2 yea s o li e in ound- he-clock ca e (≥365 days) o a home (≥365 days) and (2) whe he o no he deceden had any ca e ansi ions du ing he las 6 mon hs o li e (ie, EoL ca e ansi ions). Based on hese c i e ia, we cons uc ed ou ca e p o iles. The i s p o ile consis ed o people who li ed mos o he ime in he las 2 yea s a home and had no ansi ions o ca e acili ies in he las 6 mon hs. The second p o ile included hose who li ed mos o hei las 2 yea s a home and had ca e ansi ions in he las 6 mon hs. The hi d p o ile comp ised people who li ed mos o he ime in ound- he-clock ca e and had EoL ca e ansi ions. The ou h p o ile included people who li ed mos o he ime in ound- he-clock ca e and s ayed he las 6 mon hs in he same ca e acili y. These ca ego- ies co e ed 97.8% and 98.6% o all indi iduals who died in 1998 o in 2013, espec i ely. Indi iduals who did no all in o any o hese ca ego ies (2.2% in 1998 and 1.4% in 2013) we e designa ed as ‘unclassi ied’ and hus excluded om u he analysis. s a is ical analysis The cha ac e is ics o he di e en p o iles we e desc ibed by age, gende , diagnos ic g oups, numbe o ca e ansi ions in he las 2 yea s and in he las 6 mon hs, and numbe o days in ca e du ing he las 2 yea s. Fi s , g oup.bmj.com on Decembe 14, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om 4Aal onenM, e al. BMJ Open 2017;7:e015130. doi:10.1136/bmjopen-2016-015130 Open Access Figu e 2 P o iles 1–4 in 2013. Righ scale: p opo ion (%) o people who had ansi ions 24mon hs o 1 mon h be o e dea h. Le scale: a e age numbe o days in di e en ca e acili ies o a home 24mon hs o 1 mon h be o e dea h.EoL, end o li e. he s udy popula ion was ca ego ised in o ou di e en p o iles and desc ibed among hose who died in 2013. Second, he p o iles we e iden i ied among hose who died in 1998. Thi d, he p o iles in 1998 and 2013 we e compa ed. χ2 es s we e pe o med o compa e he p opo ion o di e en age g oups, men, women and he p opo ion o people wi h di e en diagnoses in di e en p o iles o hese yea s. Independen sample - es s we e pe o med o compa e he mean age and mean numbe o diagnoses o he di e en p o iles be ween 1998 and 2013. Mann-Whi ney U es s we e conduc ed o compa e he numbe o ansi ions in he las 2 yea s and he las 6 mon hs, and he numbe o days in ca e o he di e en p o iles. The da a we e analysed using IBM SPSS S a is ics V.22. P o ile igu es we e cons uc ed using Excel (Mic o- so O ice 2013). esul s The numbe o people who died a age 70 o olde o alled 34 116 in 1998 and 38 087 in 2013. The a e age age a dea h was 82.4 yea s in 1998, and 84.3 yea s in 2013 ( able 1). The p opo ion o women in he sample was lowe in 2013 (56.5%) han in 1998 (59.2%). The p opo - ion o hose wi h demen ia, cance , diabe es, Pa kinson’s disease o o he neu ological diseases, ch onic as hma and COPD o o he espi a o y diseases, and o he diseases o he ci cula o y sys em, in las 2 yea s o li e, was highe in 2013 han in 1998. Hip ac u e, s oke, ischaemic and o he hea diseases, and men al diso de s dec eased be ween he s udy yea s. The a e age numbe o diagnoses in he las 2 yea s o li e and he numbe o ca e ansi ions in he las 6 mon hs o li e inc eased. The o al numbe o days in ca e pe capi a inc eased due o he inc ease in days in se ice housing wi h 24 hou s assis ance; he numbe o days in o he ca e acili ies dec eased be ween he s udy yea s, as he numbe o days li ed a home also dec eased. The ou ca e p o iles cons uc ed in 2013 we e: (1) communi y esiden s (8.4% o he whole s udy popula- ion); (2) communi y esiden s wi h EoL ca e ansi ions (59.3%); (3) ca e esiden s wi h EoL ansi ions (14.7%) and (4) pe manen ca e esiden s (16.2%) ( igu e 2). In all p o iles, he a e age age, he p opo ion o hose who died a age 90 o olde and he p opo ion o men we e highe in 2013 han in 1998. P o ile 1, communi y esiden s, was equally p e alen in bo h s udy yea s. Olde people in his g oup made li le use o ca e se ices, especially ca e p o ided in long- e m ca e acili ies ( able 2). All o he olde people in his g oup s ayed a home o he las 6-mon h pe iod. This p o ile was he younges o all he p o iles al hough he a e age age and he p opo ion o hose who died a he age o 80 o olde inc eased. P o ile 1 was he only one wi h a male majo i y. The numbe o diagnoses was lowe (a e age 1.8) han in he o he p o iles, al hough he numbe was highe in 2013 han in 1998. O e hal had ischaemic hea disease and one- hi d had o he ci cula- o y diseases. P o ile 2, communi y esiden s wi h EoL ca e ansi- ions, was he single bigges p o ile and accoun ed o o e hal he s udy popula ion, al hough he p o ile was less p e alen in 2013 han in 1998 ( able 2). People in g oup.bmj.com on Decembe 14, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om 5 Aal onenM, e al. BMJ Open 2017;7:e015130. doi:10.1136/bmjopen-2016-015130 Open Access Table 2 Desc ip ion o p o iles P o ile 1: Communi y esiden s P o ile 2: Communi y esiden s wi h EoL ca e ansi ions P o ile 3: Ca e esiden s wi h EoL ca e ansi ions P o ile 4: Pe manen ca e esiden s Yea o dea h 1998 2013 1998 2013 1998 2013 1998 2013 % o popula ion (n) 8.5(2889) 8.4 (3207) 64.3 (21 951) 59.3 (22568) 10.5 (3574) 14.7 (5610) 14.5(4958) 16.2(6176) Popula ion Age, mean* 79.4 81.1** 81.4 83.2** 85.2 86.8** 86.3 87.8** Age(%)† 70–79 55.2 44.9** 41.7 32.0** 19.5 15.0** 16.8 12.9** 80–89 37.0 41.2** 45.9 47.7** 54.6 48.2** 49.9 44.6** 90+ 7.8 13.9** 12.4 20.3** 25.9 36.8** 33.3 42.5** Gende † Men 51.4 57.6** 45.2 48.3** 31.1 34.5** 23.9 27.0** Women 48.6 42.4** 54.8 51.7** 68.9 65.5** 76.1 73.0** Numbe o diagnoses, mean* 1.5 1.8** 2.4 2.7** 2.7 3.0** 2.3 2.3 Diagnoses (%)† Demen ia 5.7 10.8** 12.0 20.4** 40.8 59.3** 56.5 70.4** Cance 7.0 8.9** 31.0 36.4** 12.9 13.3 6.5 8.1* Diabe es 8.9 12.1** 15.0 17.9** 16.0 17.8* 10.0 10.1 Men al diso de s 5.9 8.2** 6.3 5.8* 10.4 8.4* 8.5 5.6** Neu ological 3.9 5.6* 9.8 11.7** 15.8 14.8 11.9 11.3** Respi a o y 14.9 19.9** 43.8 48.0** 52.9 56.7** 51.3 39.9** Hip ac u e 1.2 1.5 6.7 6.1* 15.8 10.5** 7.4 4.7** S oke 12.0 10.3* 22.1 19.2** 28.6 25.9* 27.0 21.5** Ischaemic hea disease 74.2 66.5** 57.7 57.7 52.5 51.3 37.4 34.0** O he ci cula o y disease 25.8 35.8** 30.6 46.4** 25.7 41.2** 18.1 25.0** Desc ip ion o p o iles Numbe o ansi ions,‡ mean (median) In las 2 yea s 1.8 (0) 1.6 (0)** 9.7 (7) 9.9 (7) 8.3 (6) 9.1 (6)** 2.5 (0) 2.7 (0) In las 6 mon hs 0 0 4.5 (3) 4.7 (4)** 2.8 (2) 3.4 (3)** 0 0 Days in ca e,‡, mean (median) 14.8 (0) 10.4 (0)** 78.7 (53) 77.5 (52)** 610.9 (665) 645.0 (728)** 672.5(730) 688.2(730)** P ima y ca e hospi al 5.7 (0) 5.9 (0) 42.1 (20) 43.7 (24)** 128.0 (41) 84.1 (24)** 398.7(461) 148.9 (0)** Specialised ca e hospi al 3.5 (0) 2.4 (0)** 23.6 (0) 16.2 (9)** 23.2 (5) 14.5 (3)** 14.4 (0) 11.0 (0)** Nu sing home 5.5 (0) 0.6 (0)** 13.0 (0) 6.8 (0)** 450.1 (518) 122.7 (0)** 259.1 (0) 259.5 (0) Con inued g oup.bmj.com on Decembe 14, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om 6Aal onenM, e al. BMJ Open 2017;7:e015130. doi:10.1136/bmjopen-2016-015130 Open Access P o ile 1: Communi y esiden s P o ile 2: Communi y esiden s wi h EoL ca e ansi ions P o ile 3: Ca e esiden s wi h EoL ca e ansi ions P o ile 4: Pe manen ca e esiden s Yea o dea h 1998 2013 1998 2013 1998 2013 1998 2013 Se ice housing na 1.4 (0) na 10.7 (0) na 424.8 (521) na 266.9 (0) Home 715.2 (730) 719.6 (730)** 651.4 (677) 652.5 (680)** 119.1 (65) 84.6 (2)** 57.5 (0) 41.8 (0)** Compa ison be ween 1998 and 2013. EoL, end o li e, e e s o he las 6 mon hs o li e. Numbe o diagnoses: mean numbe o majo diagnos ic g oups in Causes o Dea h egis e , Ca e Regis e o Heal hCa e and Ca e Regis e o Social Wel a e. Diagnoses: p e alence by p o iles in 1998 and in 2013 in Causes o Dea h egis e , Ca e Regis e o Heal hCa e and Ca e Regis e o Social Wel a e. *0.001≤p<0.005, **p<0.001. *pValue e e s o independen samples - es be ween 1998 and 2013. †pValue e e s o χ2 es by c oss abs be ween 1998 and 2013. ‡pValue e e s o Mann-Whi ney U es be ween 1998 and 2013. na,in o ma ion no a ailable. Table 2 Con inued his p o ile li ed mainly a home, and mos o hei ound- he-clock ca e use was clus e ed in hei las ew mon hs o li e. The a e age numbe o days in ca e acili ies du ing he las 2 yea s was only 77.5 days, indica ing ha people mos ly spen less han 3 mon hs in ca e. The use o long- e m ca e se ices was in equen . The numbe o ca e ansi ions was highe han in o he p o iles and inc eased om 1998 o 2013. This was he second younges p o ile ye he p opo ion o people aged 90 o olde inc eased especially. P o ile 2 was almos equally common among men and women. The numbe o diag- noses was a he high and inc eased om 1998 o 2013. The mos common diagnosis was ca dio ascula disease, ollowed by espi a o y diseases. Cance was clea ly mo e equen in his p o ile han in he o he p o iles and inc eased o e ime: o e one- hi d o he people in his p o ile we e diagnosed wi h cance . The numbe o ca e esiden s wi h EoL ca e ansi ions (p o ile 3) inc eased om 1998 o 2013 ( able 2). Long- e m ca e use was equen , especially se ice housing wi h 24 hou s assis ance. People inc easingly mo ed om long- e m ca e o hospi al as dea h app oached. The ca e ansi ions we e a he high and inc eased om 1998, bu hey we e s ill lowe han in p o ile 2. P o ile 3 was domi- na ed by women and people aged 80 o o e . The a e age numbe o diagnoses inc eased om 1998 o 2013. The p e alence o bo h demen ia and espi a o y diseases was o e 50%, and was highe in 2013 han in 1998. The numbe o pe manen ca e esiden s (p o ile 4) inc eased somewha o e ime ( able 2). This was he oldes p o ile: he a e age age o dea h was 87.8 yea s, and 42.5% we e aged 90 o olde . The days li ed in ca e acili ies we e also he highes o all in his p o ile: pe ma- nen ca e esiden s spen no mo e han app oxima ely 1 mon h a home. The numbe o ca e ansi ions was a he low, and hey did no change be ween he s udy yea s. The a e age numbe o diagnoses also emained unchanged. The majo i y we e women, and h ee in ou o he pe manen ca e esiden s had demen ia. The p e - alence o demen ia inc eased no ably be ween he s udy yea s. DIscussIOn In his s udy o Finnish people who died a he age o 70 o olde in 1998 and 2013, we cons uc ed ou ca e p o iles in he las 2 yea s o li e. One o he main ind- ings was ha he majo i y o olde people li ed ou side ca e acili ies un il he e y las mon hs o li e. This, howe e , became less equen o e he s udy pe iod, and he ca e p o iles wi h high le els o ca e se ice use became mo e p e alen . The p opo ion o hose who li ed majo i y o he ime in ca e and had ca e ansi- ions in he las 6 mon hs o li e inc eased in pa icula . These changes in olde people’s ca e p o iles occu ed simul aneously o when he Finnish heal hca e policy has p omo ed de-ins i u ionalisa ion o ca e and empha- sised li ing a home as long as possible. The e o e, he g oup.bmj.com on Decembe 14, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om 7 Aal onenM, e al. BMJ Open 2017;7:e015130. doi:10.1136/bmjopen-2016-015130 Open Access ca e in las 2 yea s o li e has de eloped in he opposi e di ec ion o wha he ecommenda ions o ca e policy sugges ed. The numbe o people who died a a e y old age wi h demen ia, and wi h mul i-mo bidi y, inc eased be ween he s udy yea s. I is likely ha inc easing longe i y and mo bidi y con ibu ed o he inc ease in ca e p o iles ha showed high le els o ca e use.9 22–25 Lunney e al6 7 iden- i ied ou EoL unc ional decline ajec o ies: he ail y g oup, cance g oup, sudden dea h and o gan ailu e g oup. The pa e ns o unc ional decline illus a e he le el o independence and he need o ca e nea he ime o dea h; he sudden dea h and he cance g oup we e he mos independen , while he ail y g oup was he leas independen and used long- e m ca e se ices he mos . In ou s udy, p o iles 1 and 2 likely include he sudden dea h and cance g oups, espec i ely, ep e- sen ed in Lunney’s s udy, while p o iles 3 and 4 sha e cha ac e is ics ha a e in common wi h he ail y g oup ha is domina ed by e y old people and women. The changes in ca e se ice use a he EoL and in he s udy popula ion do imply ha he p opo ion o hose who belong o he ail y g oup a he EoL has inc eased be ween he s udy pe iods. The inc ease in he pe cen ages and he numbe s o people aged 90 o olde and hose wi h demen ia we e highes in he ca e p o iles 3 and 4. Howe e , hese numbe s inc eased in all he p o iles, e en among hose who li ed a home un il hei las mon hs. We do no know he unc ional s a us o hese indi iduals o how independen ly hey managed in hei e e yday li es, bu he e is no e idence ha he unc ional capaci y o people aged 90 o o e has imp o ed in he pas decade.26 I is possible hen ha inc easing numbe s o he e y old and ail a e li ing hei las mon hs o li e mainly a home. I his is ue, mo e esou ces will need o be made a ailable o home ca e se ices o ensu e he e is a su icien supply o p o essionals ained in he ca e o olde people wi h demen ia and complex ch onic condi ions. Fu he , mo e esou ces and g ea e e o will p ob- ably ha e o be in es ed in in o mal ca e, which is usually p o ided by amily membe s o iends. Pa e ns o ca e use e lec no only he changes occu - ing in he popula ion s uc u e, bu also he ongoing e o m o he se ice sys em. Be ween he wo s udy yea s, he emphasis o long- e m ca e in Finland shi ed om ins i u ional ca e o communi y long- e m ca e o e ed in se ice housing wi h 24 hou s assis ance. Ou inding o an inc ease in he numbe o olde people ca ego ised in p o ile 3, including EoL ca e ansi ions, sugges s ha inc easing use o se ice housing may ha e led o a highe numbe o ca e ansi ions nea he ime o dea h, a esul suppo ed by ou p e ious s udy.20 Ou esul s also showed a decline in hospi al days and an inc ease in he numbe o diagnoses. In p e ious s udies, bo h he sho e pe iods o hospi al ca e27 and he inc eased numbe o diag- noses28 we e iden i ied as unde lying causes o he mo e equen ca e ansi ions. The e a e ew s udies on wha causes ansi ions om se ice housing wi h 24 hou s assis ance o hospi als.29 30 Some o he easons include symp oms ha a e oo di icul o be ea ed ou side hospi als, such as se e e alls o hip ac u es. T ansi ions o hese kinds o cases o hospi als a e likely o indica e adequa e ca e. Ye , a highe numbe o ca e ansi ions may also be indica i e o poo coo dina ion o ca e o hose su e ing om complex ch onic condi ions.1 The g owing numbe o ansi ions, pa icula ly close o dea h, may also jeopa dise he quali y o ca e and quali y o li e, especially in he mos ulne able pa ien g oups, such as hose wi h cogni i e impai men .31 Be e coo dina ion o ca e and imp o ed managemen o hospi al admissions and discha ges a e essen ial condi ions o a well unc- ioning and ageing iendly se ice sys em.1 When heal h and social ca e se ices a e eo ganised o cope wi h an inc easing numbe o olde clien s, i is impo an o gi e ca e ul hough o he unin ended consequences o hose changes. Ra he han ying o iden i y causal ela ionships be ween indi idual cha ac e is ics and ca e use, his s udy p o ides g ea e insigh in o ongoing demog aphic changes and how hese changes a e e lec ed in ac ual ca e use a he popula ion le el. The g ea es s eng h o his esea ch is in i s use o ex ensi e and eliable egis e da a ha enabled a e ospec i e ollow-up on all olde people who died in 1998 and in 2013. Finnish ca e egis e s a e conside ed highly eliable32 o p o iding accu a e in o ma ion on ca e se ice use and indi idual cha ac e is ics. The in o ma ion on ca e use, ca e an- si ions and pe sonal cha ac e is ics, such as age, gende and diagnoses, a e a ailable o e e yone; hus he e a e p ac ically no missing da a in hese egis ies. A ough classi ica ion, based on p e ious esea ch indings, was made on he basis o days s ayed in ca e and EoL ca e ansi ions. I e ealed clea di e ences be ween he ou p o iles in age, gende and equency o di e en diag- noses. Re ospec i e EoL ca e s udies a e no wi hou c i ique due o he obse a ion ime decided in ad ance, which does no a y be ween he s udy subjec s, and he p oblems in d awing associa ions be ween he subjec cha ac e is ics and he obse ed ca e.33 Ye , he pu pose o ou s udy was no o es ablish any p ospec i e ca e use ajec o ies o di e en diagnoses, o o e alua e clinical ca e pa hways ypical o speci ic diseases. The main goal o his pape was o p esen a ealis ic iew o he ex en o ca e se ice use in he las 2 yea s o li e in he olde popula ion du ing he ime when dea h is inc easingly pos poned o highe ages, and majo ca e e o ms a e conduc ed o educe ins i u ional ca e and inc ease li ing a home. Such a na ion-wide, e ospec i e s udy p o ides ex ensi e in o ma ion abou he ex en and he speci ic se ices ha a e being used in his popula ion du ing he las yea s o li e, ha is, du ing he ime pe iod when ca e use is usually a i s highes ,8–10 as well as o wha ex en se ice use has changed o e a ce ain ime pe iod. One o he weaknesses o his s udy is ha ou da a p o ided no in o ma ion on unc ional abili y o he g oup.bmj.com on Decembe 14, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om 8Aal onenM, e al. BMJ Open 2017;7:e015130. doi:10.1136/bmjopen-2016-015130 Open Access a ailabili y o in o mal ca e o o mal home ca e se ices, which a ec he need o and use o ound- he-clock ca e. In some cases, in o mal ca e may subs i u e o o mal ound- he-clock ca e se ices, and ice e sa.34 Based on ou esul s, we can only commen on ound- he-clock ca e se ice use occu ing ou side p i a e households. Mo e de ailed in o ma ion on which ca e si es people mo e be ween would gi e bene icial in o ma ion on he cu en ca e a angemen s nea he ime o dea h; howe e , hese single ca e ansi ions could no be iden- i ied in his s udy. The esul s o his s udy a e based on compa isons made be ween wo ime pe iods, and he e- o e, any assump ions o linea ends in de elopmen be ween 1998 and 2013 should be made wi h cau ion. I is possible ha changes in diagnos ic p ac ices ha e con ibu ed o he inc eases seen in some o he diag- nos ic g oups s udied. Fo example he diagnos ics o demen ing diseases has imp o ed du ing he pas 20 yea s and i canno be excluded ha he egis e s a e he e o e likely mo e accu a e epo ing hese diseases a he end o ou s udy pe iod han in he beginning. Howe e , we a e no able o in es iga e he de elopmen o such diag- nos ic p ac ices in his pa icula s udy. Despi e hese sho comings, his a icle p o ides no el and eliable in o ma ion on ongoing ends in he pa e ns o ca e o an old popula ion li ing hei las yea s o li e. The demog aphic change and he ca e policies ha now p e e home ca e o e ins i u ional ca e a e no jus a Finnish phenomenon; hus, he esul s o his s udy should be aken se iously in he ca e planning o o he socie ies as well. As o inc easing longe i y, he key ques ion is whe he his change is accompanied by ei he a comp ession o an expansion o mo bidi y.1 25 In ei he case, he EoL pe iod is ine i ably cha ac e ised by mo bidi y and disabili y. E en i mo bidi y and he onse o ch onic condi ions, such as demen ia, a e u he pos - poned,35 he inc easingly high age a which his pe iod is li ed and expe ienced means he e will p obably be a con inued g owing need o ca e o a longe pe iod o ime be o e dea h. In o he wo ds, i is possible ha people will be able o enjoy an inc easing numbe o heal hy yea s, while a he same ime ha ing hei las yea s o li e be cha ac e ised by a longe pe iod o se ious disabili y han e e be o e. cOnclusIOns Mos olde people con inue o li e in hei homes un il he las mon hs o hei li es. This pa e n o ca e is in line wi h cu en ca e ecommenda ions, which emphasise he bene i s o li ing in a p i a e home, and ca e p o ided in non-ins i u ional se ings. Ye , li ing he mos o he las 2 yea s o li e in long- e m ca e se ings, inc eased om 1996–1998 o 2011–2013. The easons behind his de el- opmen should be s udied u he . Inc easing age a he ime o dea h, and an inc ease in mul i-mo bidi y, espe- cially demen ia, migh ha e con ibu ed o he inc ease in ca e use nea he ime o dea h. I is, he e o e, impo an o bea in mind ha li ing a home un il close o dea h may p esen new p oblems and challenges o he olde people hemsel es as well as o hei ca e p o ide s, espe- cially when hose las yea s and mon hs o li e a e inc eas- ingly li ed a a e y old age, and wi h mul i-mo bidi y. The ca e sys ems mus be p epa ed o he challenges o mo e se e e disabili y as well as o longe pe iods o long- e m ca e as needed a he end o li e. Acknowledgemen s This s udy was conduc ed as pa o he p ojec New Dynamic o Longe i y and he Changing Needs o Se ices (COCTEL). con ibu o s MA designed and conduc ed he analysis in collabo a ion wi h MJ, JR, LF, JP and PR. JR was esponsible o da a cons uc ion. MA p epa ed he i s d a o he manusc ip and MJ, LF, JP, JR and PR ga e c i ical commen s o he s udy composi ions, e ised he ex d a and app o ed he inal e sion. Funding The esea ch was suppo ed by he Academy o Finland (g an numbe 287372). compe ing in e es s None decla ed. e hics app o al The esea ch plan was app o ed by he Pi kanmaa Hospi al Dis ic E hics Commi ee, Finland. P o enance and pee e iew No commissioned; ex e nally pee e iewed. Da a sha ing s a emen No addi ional da a a e a ailable. Open Access This is an Open Access a icle dis ibu ed in acco dance wi h he C ea i e Commons A ibu ion Non Comme cial (CC BY-NC 4.0) license, which pe mi s o he s o dis ibu e, emix, adap , build upon his wo k non-comme cially, and license hei de i a i e wo ks on di e en e ms, p o ided he o iginal wo k is p ope ly ci ed and he use is non-comme cial. See: h p:// c ea i ecommons. o g/ licenses/ by- nc/ 4. 0/ © A icle au ho (s) (o hei employe (s) unless o he wise s a ed in he ex o he a icle) 2017. All igh s ese ed. No comme cial use is pe mi ed unless o he wise exp essly g an ed. eFe ences 1. Rechel B, G undy E, Robine JM, e al. Ageing in he Eu opean union. Lance 2013;381:1312–22. 2. Ch is ensen K, Doblhamme G, Rau R, e al. Ageing popula ions: he challenges ahead. Lance 2009;374:1196–208. 3. Len zne HR, Pamuk ER, Rhodenhise EP, e al. The quali y o li e in he yea be o e dea h. Am J Public Heal h 1992;82:1093–8. 4. Gu alnik JM, LaC oix AZ, B anch LG, e al. Mo bidi y and disabili y in olde pe sons in he yea s p io o dea h. Am J Public Heal h 1991;81:443–7. 5. Smi h AK, Wal e LC, Miao Y, e al. Disabili y du ing he las wo yea s o li e. JAMA In e n Med 2013;173:1506–13. 6. Lunney JR, Lynn J, Foley DJ, e al. Pa e ns o unc ional decline a he end o li e. JAMA 2003;289:2387–92. 7. Lunney JR, Lynn J, Hogan C. P o iles o olde medica e deceden s. J Am Ge ia Soc 2002;50:1108–12. 8. Ma ikainen P, Mu phy M, Me sä-Simola N, e al. Se en-yea hospi al and nu sing home ca e use acco ding o age and p oximi y o dea h: a ia ions by cause o dea h and socio-demog aphic posi ion. J Epidemiol Communi y Heal h 2012;66:1152–8. 9. Fo ma L, Rissanen P, Aal onen M, e al. Age and closeness o dea h as de e minan s o heal h and social ca e u iliza ion: a case-con ol s udy. Eu J Public Heal h 2009;19:313–8. 10. Po AM, Po ai F, Visse G, e al. U iliza ion o acu e and long- e m ca e in he las yea o li e: compa ison wi h su i o s in a popula ion- based s udy. BMC Heal h Se Res 2009;9:139. 11. Menec VH, Nowicki S, Bland o d A, e al. Hospi aliza ions a he end o li e among long- e m ca e esiden s. J Ge on ol A Biol Sci Med Sci 2009;64A:395–402. 12. Lambe g JL, Pe son CJ, Kiely DK, e al. Decisions o hospi alize nu sing home esiden s dying wi h ad anced demen ia. J Am Ge ia Soc 2005;53:1396–401. 13. Menec VH, Nowicki S, S John PD. In- egion e sus ou -o - egion hospi aliza ions a he end o li e among olde u al esiden s: he ela ionship be ween pe sonal and sys em- ela ed ac o s. J Ge on ol A Biol Sci Med Sci 2011;66:1328–35. g oup.bmj.com on Decembe 14, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om 9 Aal onenM, e al. BMJ Open 2017;7:e015130. doi:10.1136/bmjopen-2016-015130 Open Access 14. Schon P, Lage g en M, Ka ehol I. Rapid dec ease in leng h o s ay in ins i u ional ca e o olde people in sweden be ween 2006 and 2012: esul s om a popula ion-based s udy. Heal h Soc Ca e Communi y 2016;24:631–8. 15. Connolly S, Gillespie P, O'Shea E. Es ima ing he economic and social cos s o demen ia in I eland. Demen ia 2014;13:5–22. 16. He nandez M. Assis ed li ing and esiden ial ca e in o egon: wo decades o s a e policy, supply, and medicaid pa icipa ion ends. Ge on ologis 2007;47:118–24. 17. de Meije C, Bakx P, an Doo slae E, e al. Explaining declining a es o ins i u ional LTC use in he Ne he lands: a decomposi ion app oach. Heal h Econ 2015;24 Suppl 1:18–31. 18. Na ional ins i u e o heal h and wel a e. S a is ical yea book o social wel a e and heal h ca e. h p://www. julka i. i/ bi s eam/ handle/ 10024/ 125386/ URN_ ISBN_ 978- 952- 302- 286- 7. pd ? sequence=1 (accessed 18 Ma 2016). 19 O icial S a is ics o Finland (OSF) (2016) Dea hs by age and gende 1980-2014. h p:// pxne 2. s a . i/ PXWeb/ pxweb/ i/ S a Fin/ S a Fin__ m__ kuol/ 020_ kuol_ au_ 102. px/ able/ ableViewLayou 1/? xid= 443 0829- a747- 4dc9- a0d - 2c351d1e02d7 (accessed 18 Ma 2016) 20. Aal onen M, Rai anen J, Fo ma L, e al. Bu densome ansi ions a he end o li e among long- e m ca e esiden s wi h demen ia. J Am Med Di Assoc 2014;15:643–8. 21. Aal onen M, Rissanen P, Fo ma L, e al. The impac o demen ia on ca e ansi ions du ing he las wo yea s o li e. Age Ageing 2012;41:52–7. 22. Nih ilaEK, Ma ikainen PT, Koskinen SV, e al. Ch onic condi ions and he isk o long- e m ins i u ionaliza ion among olde people. Eu J Public Heal h 2008;18:77–84. 23. Chen JH, Chan DC, Kiely DK, e al. Te minal ajec o ies o unc ional decline in he long- e m ca e se ing. J Ge on ol A Biol Sci Med Sci 2007;62:531–6. 24. Banaszak-Holl J, Fend ick AM, Fos e NL, e al. P edic ing nu sing home admission: es ima es om a 7-yea ollow-up o a na ionally ep esen a i e sample o olde Ame icans. Alzheime Dis Assoc Diso d 2004;18:83–9. 25. Cha e ji S, Byles J, Cu le D, e al. Heal h, unc ioning, and disabili y in olde adul s-p esen s a us and u u e implica ions. Lance 2015;385:563–75. 26. JylhaM, En o h L, Luukkaala T. T ends o unc ioning and heal h in nonagena ians: he i ali y 90+ s udy. Annual Re iew o Ge on ology and Ge ia ics 2013;33:313–32. 27. Un uh MA, T i edi AN, G abowski DC, e al. Does educing leng h o s ay inc ease ehospi aliza ion o medica e ee- o -se ice bene icia ies discha ged o skilled nu sing acili ies? J Am Ge ia Soc 2013;61:1443–8. 28. Aba shi E, Ech eld M, Van den Block L, e al. T ansi ions be ween ca e se ings a he end o li e in he Ne he lands: esul s om a na ionwide s udy. Pallia Med 2010;24:166–74. 29. Van aja K, SeinelaL, Val anne J. Elaman loppu aiheen sai aalasii o ja hoidon suunni elu ehos e ussa pal eluasumisessa ampe eella 2011. A ailable only in Finnish. Ge on ologia 2015;29:61–74. 30. Leh o V, SeinelaL, Aal onen M. Tehos e un pal eluasumisen asukkaiden sii o - aus a ekija ja sii ojen syy . A ailable only in Finnish. Ge on ologia 2015;29:12–23. 31. Gozalo P, Teno JM, Mi chell SL, e al. End-o -li e ansi ions among nu sing home esiden s wi h cogni i e issues. N Engl J Med 2011;365:1212–21. 32. Sund R. Quali y o he innish hospi al discha ge egis e : a sys ema ic e iew. Scand J Public Heal h 2012;40:505–15. 33. Bach PB, Sch ag D, Begg CB. Resu ec ing ea men his o ies o dead pa ien s. A s udy desing ha should be laid o es . JAMA 2004;22:2765–70. 34. Picka d L. Subs i u ion be ween o mal and in o mal ca e: a ‘na u al expe imen ’ in social policy in B i ain be ween 1985 and 2000. Ageing Soc 2012;32:1147–75. 35. Ngandu T, Leh isalo J, Solomon A, e al. A 2 yea mul idomain in e en ion o die , exe cise, cogni i e aining, and ascula isk moni o ing e sus con ol o p e en cogni i e decline in a - isk elde ly people (FINGER): a andomised con olled ial. Lance 2015;385:2255–63. g oup.bmj.com on Decembe 14, 2017 - Published by h p://bmjopen.bmj.com/Downloaded om