Trends in the use and costs of round-the-clock long-term care in the last two years of life among old people between 2002 and 2013 in Finland

RESEARCH ARTICLE Open Access T ends in he use and cos s o ound- he- clock long- e m ca e in he las wo yea s o li e among old people be ween 2002 and 2013 in Finland Leena Fo ma 1* , Ma ja Jylhä 1 , Ju a Pulkki 1 , Ma i Aal onen 1,2 , Jani Rai anen 1,3 and Pekka Rissanen 1 Abs ac Backg ound: The s uc u e o long- e m ca e (LTC) o old people has changed: ca e has been shi ed om ins i u ions o he communi y, and dea h is being pos poned o inc easingly old age. The aim o he s udy was o analyze how he use and cos s o LTC in he las wo yea s o li e among old people changed be ween 2002 and 2013. Me hods: Da a we e de i ed om na ional egis e s. The s udy popula ion con ains all hose who died a he age o 70 yea s o olde in 2002–2013 in Finland (N= 427,078). The cos s we e calcula ed using na ional uni cos in o ma ion. Bina y logis ic eg ession and Cox p opo ional haza d models we e used o s udy he associa ion o yea o dea h wi h use and cos s o LTC. Resul s: The p opo ion o hose who used LTC and he sum o days in LTC in he las wo yea s o li e inc eased be ween 2002 and 2013. The mean numbe o days in ins i u ional LTC dec eased, while ha o shel e ed housing inc eased. The cos s o LTC pe use dec eased. Conclusions: Use o LTC in he las wo yea s o li e inc eased, which was explained by he pos ponemen o dea h o inc easingly old age. Cos s o LTC dec eased as shel e ed housing eplaced ins i u ional LTC. Howe e , an accu a e compa ison o cos s o di e en ypes o LTC is di icul , and he socie al cos s o shel e ed housing a e no well known. Keywo ds: Long- e m ca e, Use and cos s, Las yea s o li e, Time end Backg ound The use o long- e m ca e (LTC) among old people is concen a ed in o he las yea s and mon hs o li e, and he use o LTC a he end o li e is highe in olde old han younge old age [1–4]. Dea hs a e inc easingly pos poned o la e old age: in Finland in 1994 ca. 32,500 dea hs occu ed a he age o 70+ and 4100 (12.6%) a he age o 90+, while in 2014 he co esponding igu es we e 38,800 and 9500 (24.5%) espec i ely (S a is ics Finland 2016). Thus he las yea s o li e a e being li ed a an olde age han be o e, and consequen ly he use o LTC nea he end o li e will p obably g ow [5]. A numbe o poli ical p og ams in Finland ha e a emp ed o change highly ins i u ionalized LTC p ac ices and ha e ecommended communi y-based se ices in- s ead [6, 7], as has also happened in many o he coun ies [8, 9]. In p ac ice, he p opo ion o old people using LTC has emained close o 10% since he 1990s, bu he use o ins i u ional ca e has dec eased, and ha o shel e ed housing (se ice housing, classi ied as non-ins i u ional ca e) has inc eased in Finland [10–12]. Ea lie s udies on ends in LTC use ha e ocused on old people in gene al, bu ends in LTC use among hose who use he se ices mos , old people a he end o li e, a e no known. As he numbe o such people is inc easing in many coun ies, knowing he ends is impo an o unde s anding he dis ibu ion o se ice use and o planning how o e- spond o inc easing needs in he nea u u e. * Co espondence: [email p o ec ed] 1 Facul y o Social Sciences (heal h sciences) and Ge on ology Resea ch Cen e (GEREC), Uni e si y o Tampe e, 33014 Tampe e, Finland Full lis o au ho in o ma ion is a ailable a he end o he a icle © The Au ho (s). 2017 Open Access This a icle is 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 p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Commons license, and indica e i changes we e made. The C ea i e Commons Public Domain Dedica ion wai e (h p://c ea i ecommons.o g/publicdomain/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed. Fo ma e al. BMC Heal h Se ices Resea ch (2017) 17:668 DOI 10.1186/s12913-017-2615-3 The e a e h ee ypes o ound- he-clock LTC in Finland: inpa ien ca e in heal h cen e wa ds (p ima y ca e hospi als, which also p o ide sho - e m ca e), esi- den ial homes, and shel e ed housing wi h 24-h assis - ance ( e e ed o he ea e as shel e ed housing). O hese, heal h cen e s and esiden ial homes a e ins i u- ional se ings, and shel e ed housing p o ides housing and closely ela ed se ices. In he pas , hese LTC acil- i ies we e p esen ed as o ming a hie a chy: people wi h he smalles ca e needs we e p edominan ly ca ed o in shel e ed housing, and hose wi h he g ea es needs in heal h cen e s [13]. Howe e , his hie a chy has been e- modeled, and nowadays esiden ial homes and shel e ed housing gene ally espond o simila needs. Those wi h he highes needs o in need o medical ca e a e ca ed o in heal h cen e s. Howe e , he clien p o iles in hese se ices o e lap o some ex en . Municipali ies a e esponsible o p o iding LTC o hei ci izens in Finland [14]. The la ges p opo ion o cos s is paid o by municipal axes (84% o heal h cen- e s and 72% o esiden ial homes), and he use s o se - ices pay he es ou o hei own pocke s [15]. Na ional in o ma ion on he unding o shel e ed housing is no a ailable, bu in hese se ings esiden s pay o a a ie y o day- o-day commodi ies, e.g. medicines, ou o hei own pocke s. Since hey a e en i led o apply o eim- bu semen s o medical and housing expenses om he Social Insu ance Ins i u ion (SII), esponsibili y o he unding o LTC is being shi ed om local- o na ional- le el wel a e sys ems, and o se ice use s hemsel es [16]. The uni cos s o heal h cen e s a e highes and hose o shel e ed housing lowes (Table 1), bu hese cos s do no include all he same cos i ems, and he e- o e he di e ences in cos a e no clea a he socie al le el. F om he iewpoin o municipali ies, shel e ed housing has been ound o be 22% cheape han ins i u- ional ca e, bu when SII eimbu semen is aken in o accoun , he o al cos s o shel e ed housing a e 9% cheape han hose o ins i u ional ca e [17]. S ill, he e is a lack o in o ma ion abou cos s paid by clien s, and consequen ly abou o al socie al cos s. We desc ibe and analyze he use and socie al cos s o ound- he-clock LTC among old people du ing he las wo yea s o li e, and how hose uses and cos s changed be ween 2002 and 2013. These we e analyzed o long- e m ca e in o al, and sepa a ely o heal h cen e in- pa ien wa ds, esiden ial homes, and shel e ed housing. Du ing he s udy pe iod he s uc u e o LTC con inued o change so ha ca e was shi ed om ins i u ions o he communi y, and dea h was being pos poned o in- c easingly old age. These changes a e cu en in many coun ies, and he esul s o his s udy add knowledge abou LTC use in a g oup wi h he highes ca e needs. This s udy was conduc ed as pa o he p ojec en i led “New Dynamics o Longe i y and he Changing Needs o Se ices”(COCTEL). Me hods S udy popula ion The s udy popula ion was d awn om he Causes o Dea h Regis e (S a is ics Finland). I consis s o all hose who died a he age o 70 yea s o o e in 2002–2013 in Finland. The cu o age o 70 was chosen because bo h he isk o dea h (S a is ics Finland) and heal hca e expendi- u es pe esiden s a o inc ease a a ound 70 yea s o age in Finland [18]. Use o LTC was examined o he las 730 days o li e. Thus he da a include deceden s o 12 yea s and se ice use o 14 yea s since 2000. Da a sou ces The da a on LTC use we e de i ed om he Ca e Regis- e o Heal hca e and Ca e Regis e o Social Wel a e (Na ional Ins i u e o Heal h and Wel a e). The in o - ma ion om hese egis e s was linked using unique Pe - sonal Iden i ica ion Codes. A mo e de ailed desc ip ion o he da a collec ion has been gi en elsewhe e [19]. Table 1 Uni cos s o di e en ypes o LTC Cos i ems €2011 b €2013 c To al cos s in €2013 Heal h cen e S a , adminis a ion, meals, clo hes, ca e supplies, housing, medicines, a TVs, hygiene p oduc s, phones 257 271 271 Residen ial home S a , adminis a ion, meals, clo hes, ca e supplies, housing, medicines, a TVs, hygiene p oduc s, phones 185 195 195 Shel e ed housing S a , meals, ca e supplies a 131 138 156 + housing cos s 13.60 d + medicine cos s 4.00 e a [21] The uni cos s o he ca e o old people epo ed by Kapiainen e al. [20] a e mainly based on his epo b Cos s pe day in LTC p o ided by municipali y [20] c Cos s we e con e ed o hei 2013 equi alen alues acco ding o he p ice index o public expendi u e o heal h and social se ices (S a is ics Finland) d [22] e COCTEL da a: medicine cos s pe day in he las wo yea s o li e among hose who died in 2013 and we e communi y-dwelling o a leas one day Fo ma e al. BMC Heal h Se ices Resea ch (2017) 17:668 Page 2 o 9 Days in ca e we e calcula ed o each indi idual on he basis o da es o admission o and discha ge om ca e. Pe mission o access and use he egis e da a was ob- ained om bo h egis e au ho i ies. The esea ch plan was app o ed by he Pi kanmaa Hospi al Dis ic E hics Commi ee. Measu es Use and cos s o ound- he-clock LTC we e analyzed in o al and sepa a ely o h ee ypes o LTC: (1) heal h cen e inpa ien wa ds, i he pe son had a con inuous leng h o s ay o 90 days o o e ; (2) esiden ial homes; (3) shel e ed housing wi h 24-h assis ance. The LTC in o al is he sum o hese h ee ypes o LTC. Th ee ou come a iables we e c ea ed o LTC in o al and o each ype o LTC: (1) any use, whe e 1 = used a leas once in he las 730 days o li e, and 0 = did no use in he las 730 days o li e; (2) days in ca e ( o heal h cen e s days a y om 90 o 730 and o o he se ice ypes om 1 o 730); (3) cos s o ca e in he las 730 days o li e. We mul iplied he numbe o days in di e en ypes o LTC by hei daily uni cos s, de i ed om a na ional e- po [20] (Table 1). The uni cos s o esiden ial homes and heal h cen e s a e g oss cos s caused by he use o se ices. We used he uni cos s o he yea 2013 o all yea s. Cos s we e con e ed om yea 2011 alues o hei 2013 equi alen alues acco ding o he p ice index o public expendi u e o heal h and social se ices (S a- is ics Finland). The uni cos o shel e ed housing included s a , meals, and ma e ials, bu excluded housing, medicines, and some pu chased se ices [21]. Ins ead, hese we e included in he uni cos s o heal h cen e s and esiden- ial homes (cos i ems a e desc ibed in Table 1). We es i- ma ed he cos o housing (an a e age o €13.60 pe day in shel e ed housing) using s a is ics om he SII [22]. The mean cos o p esc ibed ou pa ien medicines was de i ed om he SII’s indi idual-le el egis e da a. This was €4 pe day among hose who died in 2013 and who we e communi y-dwelling 1 o a leas one day in he las wo yea s o li e. We added hese cos i ems o he uni cos o shel e ed housing o make i mo e compa - able wi h o he ypes o LTC and o ep esen he soci- e al cos s, like he cos s o o he ypes o LTC do. None heless, many o he cos s paid by use s o shel e ed housing we e excluded, such as o clo hing, TVs, hy- giene p oduc s, and phones. In esiden ial homes and heal h cen e s hese cos s a e co e ed (see Table 1). The uni cos o shel e ed housing is an unde es ima ion, bu i is he bes es ima ion a ailable, and his limi a ion mus be kep in mind when in e p e ing he esul s. The uni cos s o shel e ed housing a e di icul o es ima e, since he e is a ia ion in he con en o se ice packages o e ed o clien s. Analyses Analyses we e pe o med o he whole s udy popula ion and by age g oup (70–79, 80–89 and 90+ yea s). Chi- squa e es s we e used o es he di e ences in he p o- po ions o se ice use s and gende s be ween he yea s o dea h. One-way ANOVA was used o assess he change in he mean age a dea h du ing he s udy pe iod. Independen samples median es s we e used o es he di e ences in he numbe o days in ca e, and in he cos s be ween hose who died in di e en yea s. The numbe o days in and cos s o LTC we e analyzed among he use s o se ices. In addi ion, he sum o days in LTC is p esen ed. Bina y logis ic eg ession analyses we e pe o med o assess he p obabili y o using LTC by yea o dea h. Age and gende we e adjus ed o . The cos s o LTC in ou da a include a lo o ze oes, and he dis ibu ions a e U-shaped, meaning ha he e we e many indi iduals wi h ew o no days in ca e ( o heal h cen e s 0 o ≥90) and many who we e in ca e o 730 o al- mos 730 days. Su i al me hods such as Cox p opo ion haza d models may be employed when he da a a e skewed o mul imodal, ha e hea y ails, o consis o excess ze oes [23]. These models ha e been used e en when he censo - ing does no ha e o be co ec ed [24]. In compa isons o di e en eg ession models o analyzing he cos da a, p opo ional haza d models ha e been shown o pe o m well [25] when he p opo ional haza ds assump ion is me [24].The p opo ional haza ds assump ion means ha he su i al cu es o wo di e en le els o a co a ia e a e p opo ional o e ime (i.e. cons an ela i e haza d) [26]. We employed Cox p opo ional haza d models o analyze he de elopmen o he cos s o LTC pe use be- ween 2002 and 2013. Cos s we e conside ed a “su i al ime” a iable, and hose who did no ha e “su i al ime”, i.e. whose cos s we e 0, we e d opped om he models. None o he obse a ions was ea ed as censo ed, as o all he end poin was dea h. Age, gende , and yea o dea h we e independen a iables. In addi ion, models including an in e ac ion e m (age * yea o dea h) we e un o ind ou whe he he e ec o age on cos s di e ed be ween he yea s. The nega i e coe icien es ima e (haza d a io (HR) <1.00) o he p opo ional haza d model indica es a dec eased haza d o eaching o al cos s, hence an in- c ease in o al cos [25]. The e o e, we p esen he in e ses o HRs o make he in e p e a ion easie , i.e. a highe alue means highe cos s. Resul s Desc ip i es The da a included 427,078 pe sons. The mean age a dea h inc eased om 82.3 o 83.8 yea s be ween 2002 Fo ma e al. BMC Heal h Se ices Resea ch (2017) 17:668 Page 3 o 9 and 2013 (p< .001), and he p opo ion o women de- c eased om 59.1% o 56.5% (p< .001). The numbe o people who died a he age o 70–79 yea s dec eased and ha o olde people inc eased du ing he s udy pe iod (Fig. 1). The p opo ion o LTC use s The p opo ion o hose who used LTC in o al in- c eased (p< .001) o e he s udy pe iod among hose who died a he age o 70+. I inc eased in younge age g oups, bu dec eased (p< .001) among he oldes (90+) (Fig. 2). The use o ins i u ional ca e (heal h cen e s and esiden ial homes) dec eased (p< .001), while he use o shel e ed housing inc eased (p< .001) in all age g oups. The likelihood o using LTC in o al emained app oxi- ma ely a he same le el be ween 2002 and 2013 when age was adjus ed o (s epwise analyses no shown) (Table 2). The dec ease in he use o ins i u ional ca e and he inc ease in use o shel e ed housing we e also shown in hese analyses. Use o all ypes o LTC was mo e common among olde han younge deceden s and among women han men (Table 2). The numbe o days in, and he cos s o , long- e m ca e The numbe o dea hs a he age o 70+ dec eased om 2002 o 2005, and inc eased ema kably om 2006 o 2013. The e o e he sum o days in LTC in o al i s de- c eased and hen inc eased (Fig. 1). The sum o days in LTC dec eased among he younges (70–79) bu in- c eased in he olde age g oups. Among hose who used LTC, he numbe o days in LTC in o al inc eased sligh ly (mean 450 in 2002 and 448 in 2013, median 540 and 549 espec i ely, p< .001) in he s udy pe iod, bu he cos s pe use dec eased (p< .001) (Fig. 2). The numbe o days, as well as he cos s o ca e in heal h cen e s and esiden ial homes, de- c eased (p< .001) om 2002 o 2013. Con e sely, he numbe o days in and cos s o shel e ed housing in- c eased (p< .001). We an Cox p opo ional haza d models o ind ou how he cos s o LTC de eloped du ing he s udy pe iod among se ice use s. The cos s o LTC in o al dec eased sys ema ically om 2002 o 2013 (Table 3). The cos s o heal h cen e s dec eased, and he cos s o shel e ed housing inc eased. The cos s o esiden ial homes did no change du ing he s udy pe iod. Olde people and women had a highe p obabili y o highe cos s o each ype o LTC han younge people and men (Table 3). We also an Cox p opo ional haza d models includ- ing he in e ac ion e m (age * yea o dea h) o ind ou whe he he e ec o age on LTC cos s changed du ing he s udy pe iod (analyses no shown). The e ec o his in e ac ion e m was s a is ically signi ican ly associa ed wi h he cos s o LTC in o al o he yea s 2010–2013 ( o all yea s in e se o HR 0.996, 95% con idence in e - als (CIs) 0.993, 0.999), and wi h he cos s o esiden ial homes in 2012 and 2013 ( o bo h yea s in e se o HR 0.993, 95% CIs 0.989, 0.998). This indica es ha in hese yea s he e ec o age on LTC cos s was 0.4% and on esiden ial home cos s 0.7% weake han in 2002. Discussion The aim o his s udy was o desc ibe and analyze how he use and cos s o ound- he-clock LTC among old people in he las wo yea s o li e changed be ween 2002 and 2013. We ound ha he p opo ion o LTC Fig. 1 The sum o days in LTC in he las wo yea s o li e and he numbe o deceden s by age g oup om 2002 o 2013: N= 427,078 Fo ma e al. BMC Heal h Se ices Resea ch (2017) 17:668 Page 4 o 9 Fig. 2 The p opo ion o LTC use s ou o all deceden s, and he mean cos s o LTC (€2013) among use s by age g oup om 2002 o 2013 ( he cos s o heal h cen e s a e highe han hose o LTC in o al, since he numbe o days in a heal h cen e a ies om 90 o 730 and o o he ypes o LTC om 1 o 730) Fo ma e al. BMC Heal h Se ices Resea ch (2017) 17:668 Page 5 o 9 use s inc eased among hose who died a he age o 70 yea s o o e , bu he inc ease was no clea when age was adjus ed o . This implies ha he inc ease was due o he change in he age s uc u e o old people, i.e. he pos ponemen o dea h o olde ages. Use and cos s o LTC a e highe among olde old han younge old people, and he e ec o age on use and cos s o LTC did no change much du ing he s udy pe iod. As he las yea s o li e a e being li ed a a g ea e age han be- o e, unc ional and cogni i e disabili y is p obably Table 2 The associa ion o age, gende , and yea o dea h wi h any use o LTC, logis ic eg ession analyses: N= 427,078 LTC Heal h cen e Residen ial home Shel e ed housing OR (95% CIs) OR (95% CIs) OR (95% CIs) OR (95% CIs) Age 1.10 *** (1.10, 1.10) 1.04 *** (1.04, 1.04) 1.07 *** (1.07, 1.07) 1.06 *** (1.06, 1.06) Gende ( e . man) 1.54 *** (1.52, 1.56) 1.48 *** (1.45, 1.51) 1.23 *** (1.21, 1.25) 1.29 *** (1.27, 1.31) Yea o dea h 2002 ( e .) 2003 0.98 (0.95, 1.01) 0.97 (0.93, 1.00) 0.93 *** (0.90, 0.97) 1.22 *** (1.16, 1.29) 2004 0.94 *** (0.91, 0.97) 0.92 *** (0.89, 0.96) 0.88 *** (0.85, 0.91) 1.33 *** (1.27, 1.41) 2005 0.95 *** (0.92, 0.98) 0.92 *** (0.88, 0.96) 0.83 *** (0.80, 0.86) 1.58 *** (1.50, 1.66) 2006 0.98 (0.95, 1.01) 0.93 *** (0.90, 0.97) 0.81 *** (0.79, 0.84) 1.73 *** (1.64, 1.82) 2007 1.02 (0.99, 1.05) 0.92 *** (0.89, 0.96) 0.84 *** (0.81, 0.87) 1.98 *** (1.89, 2.08) 2008 1.03 (1.00, 1.06) 0.91 *** (0.87, 0.95) 0.77 *** (0.75, 0.80) 2.27 *** (2.17, 2.38) 2009 1.04 ** (1.01, 1.08) 0.86 *** (0.83, 0.90) 0.78 *** (0.76, 0.81) 2.51 *** (2.40, 2.64) 2010 1.03 * (1.00, 1.07) 0.82 *** (0.79, 0.85) 0.75 *** (0.72, 0.78) 2.88 *** (2.75, 3.01) 2011 1.04 (1.00, 1.07) 0.74 *** (0.71, 0.77) 0.71 *** (0.68, 0.73) 3.33 *** (3.18, 3.49) 2012 1.03 * (1.00, 1.06) 0.66 *** (0.63, 0.69) 0.66 *** (0.64, 0.68) 3.77 *** (3.60, 3.95) 2013 1.04 (1.00, 1.07) 0.57 *** (0.54, 0.59) 0.61 *** (0.59, 0.63) 4.38 *** (4.19, 4.58) Nagelke ke R 2 0.151 0.039 0.069 0.091 *p< .05, **p< .01, ***p< .001 LTC long- e m ca e, OR odds a io, CI con idence in e al Table 3 The associa ion o age, gende , and yea o dea h wi h cos s o LTC among use s, Cox p opo ion haza d analyses LTC in o al Heal h cen e Residen ial home Shel e ed housing N 196,461 69,958 96,722 71,404 iHR (95% CIs) iHR (95% CIs) iHR (95% CIs) iHR (95% CIs) Age 1.01 *** (1.01, 1.01) 1.01 *** (1.01, 1.01) 1.02 *** (1.02, 1.02) 1.02 *** (1.02, 1.02) Man ( e . woman) 1.26 *** (1.25, 1.27) 1.17 *** (1.15, 1.19) 1.29 *** (1.27, 1.31) 1.25 *** (1.23, 1.27) Yea o dea h ( e . 2002) 2003 0.97 ** (0.95, 0.99) 0.97 (0.94, 1.01) 0.99 (0.96, 1.02) 0.99 (0.94, 1.04) 2004 0.97 ** (0.95, 0.99) 0.98 (0.94, 1.01) 1.00 (0.97, 1.03) 1.04 (0.99, 1.09) 2005 0.95 *** (0.93, 0.98) 0.96 * (0.92, 0.99) 1.00 (0.97, 1.03) 1.04 (0.99, 1.09) 2006 0.96 *** (0.94, 0.98) 0.95 ** (0.92, 0.99) 1.02 (0.99, 1.05) 1.05 (1.00, 1.10) 2007 0.93 *** (0.91, 0.95) 0.94 *** (0.91, 0.98) 0.99 (0.96, 1.02) 1.04 (0.99, 1.09) 2008 0.93 *** (0.91, 0.95) 0.94 *** (0.91, 0.97) 1.01 (0.98, 1.05) 1.06 ** (1.02, 1.11) 2009 0.91 *** (0.89, 0.93) 0.93 *** (0.90, 0.96) 1.05 ** (1.02, 1.08) 1.06 * (1.01, 1.11) 2010 0.87 *** (0.85, 0.89) 0.87 *** (0.84, 0.90) 1.02 (0.99, 1.05) 1.07 *** (1.03, 1.12) 2011 0.82 *** (0.81, 0.84) 0.83 *** (0.80, 0.86) 1.02 (0.99, 1.05) 1.12 *** (1.07, 1.17) 2012 0.79 *** (0.77, 0.81) 0.77 *** (0.75, 0.80) 1.04 ** (1.01, 1.08) 1.20 *** (1.15, 1.26) 2013 0.73 *** (0.71, 0.74) 0.71 *** (0.68, 0.74) 0.99 (0.96, 1.02) 1.28 *** (1.23, 1.33) *p< .05, **p< .01, ***p< .001 LTC long- e m ca e, iHR in e se o haza d a io, CI con idence in e al Fo ma e al. BMC Heal h Se ices Resea ch (2017) 17:668 Page 6 o 9 g ea e han p e iously [27], and he pe iod du ing which ca e is needed may be longe [28]. The composi ion o LTC changed du ing he s udy pe iod. The use o ins i u ional LTC (heal h cen e s and esiden ial homes) dec eased, and he use o shel e ed housing inc eased. The numbe o days in LTC in o al among use s did no change much, bu he cos s o LTC in o al pe use dec eased. This was mainly due o he eplacemen o ins i u ional LTC wi h shel e ed housing, he cos s o which a e lowe han hose o ins i u ional LTC. Howe e , he uni cos s o shel e ed housing a e di icul o es ima e, as he cos s and con en s o se ices a y be ween shel e ed housing uni s. In pa icula , he cos s ha clien s pay ou o hei own pocke s a e no well known, and no all o hem a e included in he uni cos s used he e. We used he na ional uni cos s, bu since hese do no include he cos s o medicines and housing, we es ima ed hei impac on cos s. Conse- quen ly, he di e ence be ween he daily cos s o esi- den ial homes and shel e ed housing diminished om €57 o €39. Howe e , he uni cos o shel e ed housing is p obably an unde es ima ion. The cos s o LTC na u ally depend hea ily on he uni cos s used. We used he same uni cos s (2013) o all yea s, al hough cos s may ha e changed du ing he s udy pe iod. Fu he mo e, i he e we e signi ican changes in he con en o ca e, o in he need o ca e among use s, his may also ha e led cos s o change. The uni cos s used he e a e he na ional a e age cos s, while he ac ual cos s a y somewha be ween municipali ies and se ice p o ide s [29]. The ole o shel e ed housing clea ly inc eased du ing he s udy pe iod in he ca e o old people in hei las wo yea s o li e, as in he ca e o old people in gene al [10–12, 30]. This shi is mainly due o changes in he supply o se ices, a he han o he p e e ences o old people. I an indi idual in shel e ed housing needs as much ca e and as many se ices as someone in a esi- den ial home, he cos s o hei ca e a e likely o in- c ease, becoming nea o equal o he cos s o ca e in a esiden ial home [16]. Ea lie s udies ha e epo ed e- ma kable di e ences in ca e needs be ween esiden s in di e en ypes o LTC [30], bu in ou da a people using di e en ypes o LTC did no di e much in e ms o age, gende , o demen ia diagnosis (da a no shown). Acco ding o ou p e ious analyses, old people a e no li ing un il hei dea hs in shel e ed housing as o en as in esiden ial homes. Also, he numbe o ansi ions o di e en ca e acili ies (commonly hospi als) in he las yea o li e has been highe om shel e ed housing han om esiden ial homes [31, 32]. In addi ion o possible p oblems in he con inui y o ca e, mul iple admissions o hospi al ca e also inc ease end-o -li e ca e cos s. We also analyzed he a ia ion in he o al cos s o social and heal h se ices be ween hose li ing in esiden ial homes and hose in shel e ed housing, and no di e ence was ound [33]. The use o egis e s, which a e conside ed eliable [34, 35], is a s eng h o his s udy. In analyzing he cos s o LTC we employed Cox p opo ional haza d models, which a e no a e y es ablished way o make such an analysis. Howe e , acco ding o p e ious econome ic compa isons, i is a sui able me hod o he analysis o complex cos dis ibu ions, al hough an essen ial equi emen is ha he p opo ional haza ds assump ion be me [23, 25]. We examined his assump- ion by conside ing he “su i al cu es”o cos s be- ween yea s o dea h, and ound ha hey we e p opo ional (pa allel) a all poin s o he cos dis ibu- ion. In addi ion, Cox eg ession has been ound o be alid o analyzing cos s when he da a a e no censo ed [36]; his was he case in ou s udy, whe e he ollow- up was wo yea s o all. The ocus o his s udy was on ound- he-clock LTC; hus home ca e and in o mal ca e, which a e impo an pa s o LTC, we e no included. The p opo ion o home ca e use s in he las wo yea s o li e inc eased om 19% in 2002 o 21% in 2008 in Finland [37]. Al- hough he e is a policy emphasis on li ing a home o as long as possible, he s abili y o he numbe o days in LTC among he use s in his s udy sugges s ha old people may no be able o li e longe in hei own homes wi h he cu en le el o home ca e. Indi idual cha ac e is ics, o he han age and gende , we e no con olled o in ou analyses, al hough disease and disabili y [38, 39], li ing a angemen s and socio- economic s a us [40, 41], and a ailabili y o in o mal ca e [42, 43] a e known o be impo an de e minan s o he use and cos s o LTC. In his s udy, howe e , he ocus was on ime ends in he use and cos s o LTC, no on indi idual de e minan s. Conclusions A ema kable change in he a angemen o ound- he- clock LTC o old people in he las wo yea s o li e ook place in Finland be ween 2002 and 2013. The num- be o days in LTC inc eased sligh ly, bu he cos s o LTC in o al pe use dec eased du ing he s udy pe iod. The e was a shi om ins i u ional ca e o shel e ed housing. Howe e , eliably compa ing he cos s o di e - en ypes o LTC is di icul , because hey do no include he same cos i ems. We es ima ed he socie al cos s o shel e ed housing, bu s ill could no include hem all. In pa icula , he cos s paid by he se ice use s a e di icul o es ima e. The inc eased use o shel e ed housing aises conce ns abou shi ing he inancial esponsibili y on o clien s. Use o LTC in o al inc eased, as he age s uc u e o he deceden s changed. As dea h con inues Fo ma e al. BMC Heal h Se ices Resea ch (2017) 17:668 Page 7 o 9 o be pos poned o olde ages, he need o , and he use and cos s o , LTC will p obably inc ease. Choices should be made abou how o supply good ca e o old people in ways ha will sha e he cos s o LTC equi ably. Endno es 1 People we e conside ed communi y-dwelling i hey li ed in a p i a e home o shel e ed housing, which is classi ied as non-ins i u ional ca e. Abb e ia ions CI: Con idence in e al; COCTEL: New Dynamics o Longe i y and he Changing Needs o Se ices p ojec ; HR: Haza d a io; iHR: in e se o haza d a io; LTC: Long- e m ca e; OR: Odds a io; SII: Social Insu ance Ins i u ion o Finland Acknowledgemen s No applicable. Funding This wo k was suppo ed by a g an o Leena Fo ma om he Academy o Finland (276189), a g an o P o esso Pekka Rissanen om he Founda ion o Municipal De elopmen and g an s o P o esso Ma ja Jylhä om he Academy o Finland (250602 and 287372). The unding bodies did no in luence any aspec o he design o he s udy, analysis o in e p e a ion o da a. A ailabili y o da a and ma e ials The indings o his s udy a e based on egis e da a a ailable om S a is ics Finland and he Na ional Ins i u e o Heal h and Wel a e. The da a used a e co e ed by he pe mission g an ed o he cu en s udy, and a e no publicly a ailable. Au ho s’con ibu ions All au ho s con ibu ed o he s udy design, in e p e a ion, w i ing and/o e ision o his publica ion. LF and JR pe o med he s a is ical analyses. LF d a ed he manusc ip . MJ and PR o iginally planned he COCTEL s udy. All au ho s ead and app o ed he inal manusc ip . 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. Consen o publica ion No applicable. Compe ing in e es s The au ho s decla e ha hey ha e no compe ing in e es s. Publishe ’sNo e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps and ins i u ional a ilia ions. Au ho de ails 1 Facul y o Social Sciences (heal h sciences) and Ge on ology Resea ch Cen e (GEREC), Uni e si y o Tampe e, 33014 Tampe e, Finland. 2 Ins i u e o Ad anced Social Resea ch, Uni e si y o Tampe e, Tampe e, Finland. 3 UKK-Ins i u e o Heal h P omo ion, Tampe e, Finland. Recei ed: 30 Ma ch 2017 Accep ed: 12 Sep embe 2017 Re e ences 1. Wea e F, S ea ns SC, No on EC, Spec o W. P oximi y o dea h and pa icipa ion in he long- e m ca e ma ke . Heal h Econ. 2009;18(8):867–83. 2. La sson K, Kå ehol I, Tho slund M. Ca e u ilisa ion in he las yea s o li e in ela ion o age and ime o dea h: esul s om a Swedish u ban popula ion o he oldes old. Eu J Ageing. 2008;5:349–57. 3. We blow A, Felde S, Zwei el P. Popula ion ageing and heal h ca e expendi u e: a school o ' ed he ings'? Heal h Econ. 2007;16(10):1109–26. 4. Fo ma L, Rissanen P, Aal onen M, Rai anen J, Jylhä M. 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 Pub Heal h. 2009;19(3):313–8. 5. Mu phy M, Ma ikainen P. Use o hospi al and long- e m ins i u ional ca e se ices in ela ion o p oximi y o dea h among olde people in Finland. Soc Sci Med. 2013;88:39–47. 6. Minis y o Social A ai s and Heal h. The memo andum o wo king g oup o se ice s uc u e. Pal elu akenne yö yhmän muis io: The Minis y o Social A ai s and Heal h; 1992. 7. Minis y o Social A ai s and Heal h, Associa ion o Finnish Local and Regional Au ho i ies. Na ional F amewo k o High-Quali y Se ices o Olde People: Helsinki: Minis y o Social A ai s and Heal h publica ions; 2008. 8. Landaue M, Kohle PA, Ko ek Y, Ross F, Walse C. Long- e m ca e bene i s and se ices in Eu ope. Ge on ology. 2009;55(5):481–90. 9. 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 Spec No 3:118–124. 10. Kokko S, Val onen H. Kunna ja anhuspal eluiden pi käaikaishoidon akennemuu okse . Local au ho i ies and s uc u al changes in long- e m ca e se ices o he elde ly. Yh eiskun apoli iikka. 2008;73(1):12–23. 11. O icial S a is ics o Finland (OSF). Ins i u ional ca e and housing se ices in social ca e 2011. Social p o ec ion ed. Na ional Ins i u e o Heal h and Wel a e: Helsinki; 2012. 12. Blomg en J, Einiö E. Lai oshoidon ähenemisen yh eys ikään yneiden muihin pi käaikaishoi an pal eluihin ja sai aus akuu uksen ko aamien lääkkeiden kus annuksiin uosina 2000-2013. The deins i u ionalisa ion o old age ca e: i s associa ion wi h o he long- e m ca e se ices and ou pa ien medical expenses in 2000–2013. Yh eiskun apoli iikka. 2015;80(4):334–48. 13. No o A, Finne-So e i H. A o- ja lai oshoidon suhde - ka saus nyky ilaan ja ule aan. In: Ilmakunnas S, edi o . Hy in oin ipal eluja en is ä ehokkaammin. Uudis us en mahdollisuuksia ja keinoja. Helsinki: Go e nmen Ins i u e o Economic Resea ch; 2008. p. 115–140. 14. Vuo enkoski L, Mlado sky P, Mossialos E. Finland: heal h sys em e iew. Heal h Sys ems in T ansi ion. 2008;10(4):1–168. 15. Peku inen M, E hola M, Häkkinen U, Jonsson PM, Keskimäki I, Kokko S. Sosiaali- ja e eydenhuollon monikana aisen ahoi uksen edu , hai a ja kehi ämis a pee . Repo 4. Na ional Ins i u e o Heal h and Wel a e: Helsinki; 2011. 16. Pehkonen-Elmi T, Ke unen A, Pulliainen M. Lai oshoidos a omaan ko iin. Asumisen ahoi us- ja kehi ämiskeskuksen apo eja 5;2013. 17. Leskelä R, Komssi V, Sands öm S, Mikkola H, Ahola E, Pikkujämsä S, Olli S, Ha e inen A, Yli alo-Ka ajis o K, Huu e E. E i ahoi uskana ien ooli oululais en sosiaali- ja e eyspal eluissa. The oles o a ious sou ces o unding in social and heal hca e se ices in Oulu. Suomen Lääkä ileh i, Finnish Medical Jou nal. 2016;71(11):809–15. 18. Kapiainen S, Eskelinen J. Mies en ja nais en e eysmeno ikä yhmi äin 2011: Helsinki: Na ional Ins i u e o Heal h and Wel a e; 2014. 19. Fo ma L, Rissanen P, No o A, Rai anen J, Jylhä M. Heal h and social se ice use among old people in he las 2 yea s o li e. Eu J Ageing. 2007;4:145–54. 20. Kapiainen S, Väisänen A, Haula T. Te eyden- ja sosiaalihuollon yksikkökus annukse Suomessa uonna 2011. The uni cos s o heal h and social ca e in Finland in 2011. Repo s 3/2014. Helsinki: Na ional Ins i u e o Heal h and Wel a e; 2014. 21. Pasila A. Kuuden suu imman kaupungin anhus en sosiaali- ja e eyspal elujen ja kus annus en e ailu 2011: Helsinki: Kuusikko- yö yhmä; 2012. 22. O icial S a is ics o Finland (OSF). S a is ics o housing allowance 2013: Social Insu ance Ins i u ion o Finland; 2014. 23. Mihaylo a B, B iggs A, O'Hagan A, Thompson SG. Re iew o s a is ical me hods o analysing heal hca e esou ces and cos s. Heal h Econ. 2011; 20(8):897–916. 24. Basu A, Manning WG, Mullahy J. Compa ing al e na i e models: log s cox p opo ional haza d? Heal h Econ. 2004;13(8):749–65. 25. Aus in PC, Ghali WA, Tu JV. A compa ison o se e al eg ession models o analysing cos o CABG su ge y. S a Med. 2003;22(17):2799–815. 26. Sp uance SL, Reid JE, G ace M, Samo e M. Haza d a io in clinical ials. An imic ob Agen s Chemo he . 2004;48(8):2787–92. 27. Zhao J, Ba clay S, Fa quha M, Kinmon h AL, B ayne C, Fleming J. Camb idge Ci y o e -75s coho s udy collabo a ion. The oldes old in he las yea o li e: popula ion-based indings om Camb idge ci y o e -75s Fo ma e al. BMC Heal h Se ices Resea ch (2017) 17:668 Page 8 o 9 coho s udy pa icipan s aged 85 and olde a dea h. J Am Ge ia Soc. 2010;58(1):1–11. 28. Payne G, Lapo e A, Debe R, Coy e PC. Coun ing backwa d o heal h ca e's u u e: using ime- o-dea h modeling o iden i y changes in end-o -li e mo bidi y and he impac o aging on heal h ca e expendi u es. Milbank Q. 2007;85(2):213–57. 29. Väisänen A, Hujanen T, Sosiaalihuollon yksikkökus annukse Suomessa uonna 2007. The uni cos s o social wel a e in Finland 2007. Helsinki: Na ional Ins i u e o Heal h and Wel a e; 2010. 30. Na ional Resea ch and De elopmen Cen e o Wel a e and Heal h. Ikään yneiden sosiaali- ja e eyspal elu 2005. Ca e and se ices o olde people 2005. STAKES: Helsinki; 2007. 31. Fo ma L, Jylhä M, Aal onen M, Rai anen J, Rissanen P. Vanhuuden iimeise uode . Pi käaikaishoi o ja sii ymä hoi opaikkojen älillä. The las yea s o li e in old age. Long- e m ca e and ansi ions be ween ca e acili ies. Kunnallisalan kehi ämissää iön u kimuksia 69, 2012. 32. Aal onen M. Pa e ns o ca e in he las wo yea s o li e. Ca e ansi ions and places o dea h o old people: Tampe e: Ac a Uni e si a is Tampe ensis 2094; 2015. 33. Rissanen P, Fo ma L, Aal onen M, Rai anen J, Pulkki J, Jylhä M. Onko ehos e u pal eluasuminen hal empaa kuin anhainko iasuminen? In: Mäklin S, edi o . Te eys alous iede 2015. Te eys alous ie een päi ä 6.2. 2015. Helsinki: Te eyden ja hy in oinnin lai os; 2015. p. 34–39. 34. Sund R, Nu mi-Lü hje I, Lü hje P, Tanninen S, Na inen A, Keskimäki I. Compa ing p ope ies o audi da a and ou inely collec ed egis e da a in case o pe o mance assessmen o hip ac u e ea men in Finland. Me hods In Med. 2007;46(5):558–66. 35. Keskimäki I, A o S. Accu acy o da a on diagnoses, p ocedu es and acciden s in he Finnish hospi al discha ge egis e . In J Heal h Sciences. 1991;2(1):15– 21. 36. Dieh P, Yanez D, Ash A, Ho nb ook M, Lin DY. Me hods o analyzing heal h ca e u iliza ion and cos s. Annu Re Public Heal h. 1999;20:125–44. 37. Pulkki J, Aal onen M, Fo ma L, Jylhä M, Rai anen J, Rissanen P. Pi käaikaishoi o kunnissa anhuuden iimeisinä uosina. Pal elujen käy ön e o ja muu okse uosina 2002-2008. Kunnallis ie eellinen aikakauski ja. 2014;42(1):27–43. 38. Po AM, Po ai F, Visse G, Pu s M, an G oenou MI, Deeg DJ. 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. 39. de Meije C, Koopmanschap M, D'U a TB, an Doo slae E. De e minan s o long- e m ca e spending: age, ime o dea h o disabili y? J Heal h Econ. 2011;30(2):425–38. 40. Ma ikainen P, Mous gaa d H, Mu phy M, Einiö E, Koskinen S, Ma elin T, No o A. Gende , li ing a angemen s, and social ci cums ances as de e minan s o en y in o and exi om long- e m ins i u ional ca e a olde ages: a 6-yea ollow-up s udy o olde Finns. Ge on ologis . 2009; 49(1):34–45. 41. Nih ilä E, Ma ikainen P. Household income and o he socio-economic de e minan s o long- e m ins i u ional ca e among olde adul s in Finland. Popul S ud (Camb). 2007;61(3):299–314. 42. Kehusmaa S, Au i-Rämö I, Rissanen P. Omaishoidon aiku us ikään yneiden hoidon menoihin. Yh eiskun apoli iikka. 2013;78(2):138–51. 43. Van Hou en CH, No on EC. In o mal ca e and heal h ca e use o olde adul s. J Heal h Econ. 2004;23(6):1159–80. • We accep p e-submission inqui ies • Ou selec o ool helps you o ind he mos ele an jou nal • We p o ide ound he clock cus ome suppo • Con enien online submission • Tho ough pee e iew • Inclusion in PubMed and all majo indexing se ices • Maximum isibili y o you esea ch Submi you manusc ip a www.biomedcen al.com/submi Submi you nex manusc ip o BioMed Cen al and we will help you a e e y s ep: Fo ma e al. BMC Heal h Se ices Resea ch (2017) 17:668 Page 9 o 9