Measures / Baseline assessments and outcomes

4.2.2 DATA COLLECTION AND RANDOMIZATION

7UDLQHG UHVHDUFK QXUVHV ZHUH UHVSRQVLEOH IRU JDWKHULQJ WKH EDVHOLQH information for all the assessed residents and for SHUIRUPLQJDOOWKHIROORZ-up assessments. Those with limited communicative capabilities were assessed WRJHWKHUZLWKWKHLUQXUVHVRUSUR[LHVZKHQSRVVLEOH7KHUHVHDUFKQXUVHVZHUH blinded WR JURXS DOORFDWLRQ RI WKH UHVLGHQWV WKH\ DVVHVVHG. Baseline assessments were done before intervention started and follow-up assessments RIWKHRXWFRPHVZHUHPDGHDWVL[DQGmonths from the start of intervention.

,Q DGGLWLRQ DW PRQWKV UHVLGHQWV¶QoL was assessed and their hospital service use was retrieved from electronic patient records.

5HVLGHQW$VVHVVPHQW,QVWUXPHQW0LQLPXP'DWD6HW5$,0'60RUULV et al. 1990) information was available for 94 LTCF wards in Helsinki. It was used as a basis for pair-PDWFKLQJ VLPLODU /7&) ZDUGVThe RAI is a widely adopted international DVVHVVPHQWWRROXVHGLQPDMRULW\RILTCFs in Helsinki.

0'6DVVHVVPHQWVDUHPDQGDWRU\DWUHJXODULQWHUYDOVIRUDOOUHVLGHQWVLQWKH SDUWLFLSDWLQJ/7&)V7KHIROORZLQJ0'6variables were used pair-PDWFKLQJ JHQGHU DJH DQ\ GHJHQHUDWLYH EUDLQ GLVHDVH FDQFHU &36 –6 (poor FRJQLWLRQ$'/K –PDMRUGLIILFXOWLHVLQDFWLYLWLHVRIGDLO\OLYLQJ&+(66

! LQVWDELOLW\ RI KHDOWK LQGLFDWRUV KRVSLWDOL]HG ZLWKLQ WKUHHPRQWKV DQG HPHUJHQF\GHSDUWPHQWYLVLWZLWKRXWKRVSLWDOL]DWLRQ

&RQVLGHULQJ the estimated sample size requirements and available UHVRXUFHVZDUGVZLWKVLPLODU0'6FDVH-PL[esZHUHPDWFKHGDQGDWRWDORI NH wards and seven ALF wards were included in the trial. Computer-JHQHUDWHGUDQGRPQXPEHUVfrom a separate randomization center were used to randomize wards. Those involved in randomization were not aware of the characteristics of the participants.

4.2.3 MEASURES / BASELINE ASSESSMENTS AND OUTCOMES

Baseline characteristics

%DVHOLQHLQIRUPDWLRQIRUERWKVWXG\JURXSVZDVDVVessed between November DQG-DQXDU\,QDGGLWLRQWR GHPRJUDSKLFGDWDDJHVH[HGXFDWLRQ DFWLYHGLDJQRVHVZHUHUHWULHYHGIURPHOHFWURQLFSDWLHQWUHFRUGVDQG&KDUOVRQ FRPRUELGLW\ LQGH[HV ZHUH FRPSXWHG (Charlson et al. 1987). Current PHGLFDWLRQ ZDV H[DPLQHGthe number of UHJXODU PHGLFDWLRQV ZDV FRXQWHG and ERWK UHJXODU DQG ³SUR UH QDWD´ GUXJV FDWHJRUL]HGon the basis of the Anatomical Therapeutic Chemical (ATC) classification endorsed by the World +HDOWK RUJDQL]DWLRQ :+2 &ROODERUDWLQJ &HQWUH IRU 'UXJ 6WDWLVWLFV 0HWKRGRORJ\. Use of dUXJVW\SLFDOO\XVHd in palliative care (Jansen et al.

wasH[DPLQHGRSLRLGV(1$)SDUDFHWDPRO1%(VHOHFWLYHDQG nonselective nonsteroidal anti-LQIODPPDWRU\ GUXJV 16$,' 0$

antipsychotics (1$) DQ[LRO\WLFV(1%) DQWLHPHWLFV DQG GUXJV XVHG IRU nausea or death rattle (A04A or A03A).

7KH SUREDEOH SURJQRVLV IRU WKH UHVLGHQWV ZDV DVVHVVHG DFFRUGLQJ WR WKH GLDJQRVHVDQGGLVDELOLW\6HYHUHGHPHQWLDZDVGHWHUPLQHGDVDVFRUHRIthree LQWKH&OLQLFDO'HPHQWLD5DWLQJ&'5VFDOH+XJKHVHWDOor a score of underSRLQWVLQWKH0LQL-0HQWDO6WDWH([DPLQDWLRQ006((Folstein et al.

.

NutritionDOVWDWXVZDVDVVHVVHGXVLQJ 0LQL1XWULWLRQDO$VVHVVPHQW01$

6FRUHVRIXQGHUSRLQWVZHUHFRQVLGHUHGWRLQGLFDWHPDOQXWULWLRQ–

points at-ULVNRIPDOQXWULWLRQDQGDERYHSRLQWVwell-nourished *XLJR]

HWDO. 1HHGLQJDVVLVWDQFHLQDFWLYLWLHVRIGDLO\OLYLQJZDVGHILQHGDV&'5 VFDOHµSHUVRQDOFDUH¶VFRUH•

³'R-not-UHVXVFLWDWH´'15³'R-not-KRVSLWDOL]H´'1+) orders as well as ADs were retrieved from medical records. Data on earlier hospitalizations were retrieved from the residents’ medical records.

Primary outcomes

The primary outcome measures per-protocol were: FKDQJH LQ +54R/

DFFRUGLQJWRWKH'LQVWUXPHQWand the difference in the number of hospital GD\VGXULQJPRQWKVIURPEDVHOLQHRUXQWLO the resident’sGHDWK7KH' instrument was used to evaluate GLIIHUHQW GLPensions of HRQoL to construct aZHLJKWHGLQGH[EHWZHHQDQGZLWKODUJHUYDOXHVUHSUHVHQWLQJ better HRQoL. The assessed GLPHQVLRQV RI +54R/ DUH PRELOLW\ VLJKW KHDULQJ EUHDWKLQJ VOHHSLQJ HDWLQJ VSHDNLQJ H[FUHWLRQ XVXDO DFWLYLWLHV PHQWDOIXQFWLRQGLVFRPIRUWDQGV\PSWRPVDQ[LHW\GHSUHVVLRQYLWDOLW\DQG VH[XDO DFWLYLW\ 7KH ³VH[XDO DFWLYLW\´ GLPHQVLRQ VKRZHG a lDUJH QXPEHU RI PLVVLQJYDOXHV7KHUHIRUHWKH\ZHUHLPSXWHGE\WKHORZHVWYDOXH'FDQEH HYDOXDWHGHLWKHUE\WKHSDWLHQWRUFRPSOHWHGE\SUR[\6LQWRQHQ(OOLRWW HW DO . The 'assessments were carried out by study nurses. For residents with coJQLWLYH RU FRPPXQLFDWLYH GLIILFXOWLHV D VWDII RU IDPLO\

PHPEHU IDPLOLDU ZLWK WKH UHVLGHQW ZDV XVHG DV D SUR[\ LQIRUPDQW The instrument has in a previous study shown sensitivity WR FKDQJH LQ /7&) VHWWLQJV3LWNlOlHWDO. 'follow-up assessments were performed at VL[ DQG PRQWKV IURP WKH start of intervention. '-instrument is LQFOXGHG DV $SSHQGL[ The amount of hospital days was retrieved from HOHFWURQLFSDWLHQWUHFRUGVPRQWKVDIWHULQWHUYHQWLRQ

Secondary outcomes

In addition WR KRVSLWDO GD\V RWKHU KHDOWK-service use was retrieved from electronic patiHQW UHFRUGV 7KHVH VHFRQGDU\ RXWFRPHV LQFOXGHG HPHUJHQF\

GHSDUWPHQW('YLVLWVKRVSital inpatient days in various types of facilityDQG

hospital-at-home service use. To determine health-FDUHFRVWVPHDQXQLWFRVWV IURPQDWLRQDOUHJLVWULHVIURPZHUH used .DSLDLQHQHWDO with an appropriate correction for inflation rate. All costs were calculated in euros (€) and WUDQVIRUPHGWRHTXLYDOHQWVXVLQJDSSURSULDWHLQIODWLRQUDWHV

6HFRQGDU\ RXWFRPHV RI V\PSWRPV SDLQ DQG SUR[\ VDWLVIDFWLRQ ZHre PHDVXUHGDWVL[ DQGPRQWKVIURPWKH start of LQWHUYHQWLRQ7KH(GPRQWRQ 6\PSWRP$VVHVVPHQW6ystem (6$6DQGthe Pain Assessment in Advanced Dementia (PAINAD) were used in residents’ symptom assessment. (6$6DQG 3$,1$' IRUPV FDQ EH IRXQG DV $SSHQGL[ DQG (6$6 ZDV RULJLQDOO\

developed to be used for repeated symptom assessment of palliative patients with cancer (Bruera et al. 1991). )RUUHVLGHQWVZLWKFRJQLWLYHRUFRPPXQLFDWLYH GLIILFXOWLHVDVWDIIRUIDPLO\PHPEHU familiar with the resident was used as a SUR[\ LQIRUPDQW %HLQJ ZLGHO\ DGRSWHG WR FOLQLFDO SUDFWLFH LQWHUQDWLRQDOO\

(6$6 HYDOXDWHV D VHW RI V\PSWRPV HDFK XVLQJ D11-SRLQW 156 IURP ³1R V\PSWRP´WR³:RUVWSRVVLEOHV\PSWRP´ The last domain of overall well-EHLQJLVUDWHGIURP³%HVW´WR³:RUVW´ 6HYHUDOYHUVLRns of the scale DUHXVHGDVWKHVFDOHKDVXQGHUJRQHYDULRXVGHYHORSPHQWVRYHUWKH\HDUV (Hui DQG%UXHUD. $VDPHDVXUHRIRYHUDOOV\PSWRPEXUGHQZe XVHG(6$6 total score instead of reportiQJLQGLYLGXDOV\PSWRPVHYHULWLHV(Hui and Bruera . The version used in the current study iQFOXGHVGRPDLQVTXHVWLRQV SDLQDWUHVWSDLQDWHIIRUWIDWLJXHVKRUWQHVVRIEUHDWKODFNRIDSSHWLWHQDXVHD GU\PRXWKFRQVWLSDWLRQGHSUHVVLRQDQ[LHW\DQd overall well-EHLQJ

,Q SDLQ DVVHVVPHQW XVLQJ the PAINAD scale DQ H[WHUQDO REVHUYHU QRWHV behavioral cues considered to imply possible pain. These cues are derived from ILYH GRPDLQV EUHDWKLQJ YRFDOL]DWLRQ IDFLDO H[SUHVVLRQ ERG\ODQJXDJH DQG FRQVRODELOLW\(DFKof the five domains LVUDWHGEHWZHHQDQG\LHOGLQJD WRWDO PD[LPXP VFRUH RI $FFRUGLQJ WRthe RULJLQDO GHYHORSHUV RIthe LQVWUXPHQW VFRUHV RI –1 DUH FRQVLGHUHG WR UHSUHVHQW VHYHUH SDLQ –6 PRGHUDWHSDLQDQG-3 mild pain :DUGHQHWDO. Later LQYHVWLJDWRUV haYH VXJJHVWHG WKDW VFRUHV DERYH WZR should prompt a reassessment of a patients pain treatment =ZDNKDOHQHWDO. The Finnish translation of the PAINAD scale has been widely used %M|UNPDQHWDO.

$VL[-TXHVWLRQTXHVWLRQQDLUHGHYHORSHGSUHYLRXVO\ZDVXVHGLQDVVHVVLQJ SV\FKRORJLFDOZHOO-EHLQJ (PWB) 5RXWDVDORHWDO. 7KHVHVL[TXHVWLRQV cover areas that are considered important in PWB :+242/*URXS.

The questionnaire KDVVKRZQJRRGWHVWဨretest reliability 6DYLNNRHWDO DQGSURJQRVWLFYDOLGLW\ (PitkällHWDOE. ,QWKLVWULDODVWXG\QXUVHILOOHG the questionnaires with those residents able to adequately communicate. The VL[ DUHDV FRYHUHG LQ WKH TXHVWLRQQDLUH DUH OLIH VDWLVIDFWLRQ IHHOLQJ QHHGHG KDYLQJ SODQV IRU WKH IXWXUH KDYLQJ ]HVW IRU OLIH IHHOLQJ GHSUHVVHGDQGVXIIHULQJIURPORQHOLQHVV4XHVWLRQV–DUHJUDGHGQR\HV SRLQWVDQGTXHVWLRQV–DVVHOGRPRUQHYHUVRPHWLPHVRIWHQRUDOZD\V 7KH WRWDO VXP RI WKH LWHPV LV GLYLGHG E\ WKH QXPEHU RI LWHPVto which the participant has responded. 7KH UHVXOWLQJ DYHUDJH IRUPV D 3:%

VFRUHZKHUHDVFRUHRILQGLFDWHVWKHEHVWSV\FKRORJLFDOZHOO-EHLQJDQGWKH

poorest. 0LQLPXP RI WZR DQVZHUHG TXHVWLRQV ZHUH UHTXLUHG IRU WKH calculation of this score. This questioQQDLUHFDQEHIRXQGLQ$SSHQGL[

7KH 6DWLVIDFWLRQ ZLWK &DUH– (QG-of-/LIH LQ 'HPHQWLD 6:&-(2/' DVVHVVPHQWWRROZDVXVHGWRHYDOXDWHSUR[LHV¶VDWLVIDFWLRQZLWKFDUH7KHVFDOH is part of a set of tools to evaluate end-of-life care for people with advanced dementia 9ROLFHUHWDO. It contains TXHVWLRQV each rated on a scale IURP WR ZLWK WKH PD[LPXP WRWDO VFRUH RI LPSO\LQJ FRPSOHWH satisfaction. The Finnish translatLRQ RI 6:&-(2/' LV IURP WKH (XURSHDQ 3$&( SURMHFW EXW QR YDOLGDWLRQ GDWD IRU WKHtranslated version is available 9DQGHQ%ORFNHWDO. This questionnaire LVLQFOXGHGDV$SSHQGL[