2004 - Characteristics Of Teams, Staff And Patients, Associations With Outcomes Of Patients In Assertive Outreach Plao4

rv (r00.)i r0!i r0._rr I

Characteristics of teams, staff and patients: associationswith outcomes of patients in assertive outreach S T E F A NP R I E B EW. A L I D F A K H O U R YI ,A N W H I T E ,] O A N N A W A T T S , P A U LB E B B I N G T O NI ,O A N N A B I L L I N G ST, O M S U R N S ,S O N I AJ O H N S O N , , l N R Y R I Ea n d C H R I S T I N EW R I G H Tf o r t h e P a n - L o n d o n l ' 4 A T Tl ' 1 U U E Nl A A s s e r i i v eO u t . e a c hS t u d yG r o u p Background

L tteB knownabout

whai characle. ti csofteams,nafl a.d patientsare assocatcdw th a favou.abe outcomeof seve.emcnta il ne* managed oy asserLiveolrffea.n Aims

To dentft predctoEolvo lntary

a n d . o m p u s o . y a d m s s o n sr o l t n e ase.tive outreach servi.csithe UK N nefeatlrcs olteam o r g a ns a t o na . d p o . y f v e v a r a b e s Method

a$es ng nafi satisfac I on and b!rn olt palientcha.acter|nics taken ande elen from the baseinedataofth. Pan London Asert ve O!lreach stldy weretned as predctorsofvolun!a.yandcompusofl a d m$ o n sw l h , na 9 m o n t h f o l o w! p

Results Weekendwork ng, staffb!.n ou! a.d lackof.onta.tolthe pal,enlw th olher se.v.es wefc asoc ated Independently w lh a h ghd probab ty of bothvolunia.t andcomp! soryadmrson n a d d i t o na d ms s o n sn t h c p a s t predcted further vo !ntafyand compuso.t admssons andteamsnot workingextendedho!E predcted compukoryadm$ ons nthe fo ow up

l .s oltcam Con
De
Flnding

provded by the Depa.lmeitol Heath.

Asseftiv€outrea.h ieaos h.vc bcen inrro duced in most pr.ts of Ensland ro assist in ihe nanasemc.t of severemenftl illne$. To improvc thcir effe(iveness,we need to know whar fadors are isscl.ted wirh favouable outcone. The Pan London Assertive Out.each Studr (PLAO) inve$i gated thc .outine practice of asseniveout' rcach relms in London. It e$ablishcd bow teans arc organised .rd opcratc (wridt et /1, 2003), assessedstdff burn out and saris{actun (Billin8s st dl, 2003). idcntified socio demographic and clhical characier istics of patients .nd .ecordcd rate! of hospitalisationand compuho.y admissions within a 9-nonrh follorv up pdiod (Priebe 'lhe present paper explorcs .r dl, 2003). rb€ baselinechancieristics of tcams, $aff and patienis that predict outconcs ov€r ihe 9 month follow-up.

METHOD chamcrristics of 24 desisn.ted ,scrive outreach teams, 187 s ff mcmbcF and 580 patientsFere rse$ed. A ccnsLsof al ieam p.tients on the cxe load was taker on 18 June 2001- Thc caseload for cach teifr *as dividcd into parients who had been for 3 lronrhs or lonser with thc tdn (c$ablsh€rl pati€ltt .nd those who had joined the case loal in the p.ev,ous 3 nonths lnew pati€nts). Newly a.cepted prtic's were overampled becaue thc inir'al $ise of as*nire outr.,ch provsDn nay be a thb'lisrion Friod wirh a relanvelypoor ouicome (Mccrcs !r dl, 19951.To increar the proportion of ncw parients in the sanple, anoth* censuswas raken on 18 Septmrb€r2001, whecby all paticnb who joine.l fie tean ln the preious 3 months were added to the sanplc. llE tot.lsample consisted ol 391 Nrablished !.tienis and 189 n.w paticnG(for more deiiils seerriebc et al,2OO3). Voluniary hospital adnnsion and com pulsory admi$ion Htes in rhe 9 nonth

follow-up pciod weie obtaiftd for 487 pariens. Dehih of rhc apprcach, the i'*rum€nts used aid thc $udy orsanisa tionhave beendcscribedinprevious p.Frs (Billinss ,r ,1, 2003j Priebc d al, 2003; For ihis analvls, a rohloI2S variables were selectedas rotenrial predic_rorvariables ot outcone. The vrriabics wcre choscn ro cov{ a wide ranec of feaiues of rhe rans ind padcnB wnhour inr.o du.ins multl-colli',carity, which would nake thc multiple resrcssioi resuhs hdrd to interprei, and also to frov,de sood predictive disdimin.uon (Harell ct dl, 1996). The selecrlonof the variablestook into a.tount the lnportance of ihc content covered by rhcs v.riables, as asiibcd by thc authoB o{ the three p.cvioDs ILAO paper from the baselincrcsults. In prelerenc. t. comlosiie neasures such ds fidclity scores, we selected iin. chaiacte.isrcsof teams that reflectedspr rate at$ibutesol tcln orsanis.tion and policy. I. this wa-r we intended to idcntify independo,r fcatures of teams rhat misht be rlevrnr fo. outcome.The tean, prcdi.ior va.iablcswere: ieam sia (the toral nmb€r of clinical i,ll time cquivalcnt staff)i desig nard psychiadisr inpur (full tine equivdlenr psychi,rrisi pef 100 patiens); lnt.snrjon oI health ard smial care non-intes.ation)j multi linresarion,disciplinarnr (numbcr of clinical di$ipliies repr.scntcd nr the t€am); tbe proportion ot face ro face consca th"r scre lcaied in rhe comrunityi ratio of tull time to paniime staffi wcckc.d wo*ing lwhether thc tean did or did not op-ar€ .i weekend$i out of-houn work (whetherthe tean l.om Mondays to lridats olerated out of hous or rormal office hours onl'); lnd casclond lthc avdase individu.l car load pcr $atr As polential p.edictus tflecting the views and work experi€nceof staff, we select€d thc three e,b sc.les of thc Maslach Burnout Inventofl (Maslach & Jackson, 1981): emotional €xhaustion (dephtion of cnotbnal resources)ideper sonahsation(nesative.nifudes and fcdinss about patients)i rnd pereonal acconplish nent (neg"iivc cvaluation of one's selr. especiallr rcsardin8 dealing with patie.6). A hish level of bum out is rcflcctd by a low sco.€ on pe6on.l accomplishm€ni,a hish score on e otional exhau$ion and . high scorc on d€person.lisrtion. we lko selcced No sub scalesol the Minncsota s,rnfactbn S.ale (Weiss et dl, 1967):

irtrinsi. lerte.! !o which rhey feel that their work firs their skilh) and exrinsic (satis faction w'th workins condit,ons and rewardt- A hish level of etisfaction is refl
nethod, .llowing ftx clusering within teams (Roge6, 1993). All analysesaliowed ror the samplins fraciion (i.e. 0.37 lor es bblih€d patients and I tor new patientsi Priebe .r al, 2003) by weishting by its inv6e (Horvitz & Thompson,1952). Thc tended to i.creasc the $andard e.rors bt To pcdi.t the vo dichotomous out come variables, both univariare and multiple logistic rsncno. was usd. Univariatc ana llss r.l"tcd ea.b outcome vi. logisric reges sion to each pr€dictor. Qudtiranve variables wcre entered as such, and ordered atgori cal variables were fltercd as conbnuous. Thc muldvanaie nodel was rlectd hon the teari, {aff and patent vadables.sdning

Table I

wirh .ll varirbles thar wcre univdiately sisnific.nt ard usins $epwisc sel€dbn to inchdc aU v.riables that were sisificant independent predictos of €irher or thc Mo outconeq conftoUins for the eif.c$ .f the orher vanabbs in dE dodel. For vdiable a liberal sisn,ficance seleclion, wc Nd l€vel of P<0.15- How€ver, the st2tistisl sisnificdncc of ssciations was t.Ien as as odds P<0.05. Results wcrc dprsed ratios for th. fr€stre ,. at6encc of a charac ter(ri., for r 10% increasc 1nthe lercodge of contads in the conmunity. for a lo-fear 1ncresei! age,for a onc standarddeviation incrcasein $oies of staff burn-ut md satn" facrio! and for a one unn ildease in oth.r

Characr€ristics ot 24 a<
Total number orclinicallull-time equivalenr$afi : medi:n (ran8e) D€si8natedp:tchiaric time: medim (range) Whdh€rornortheteam

hd integratedhealthand soc alcafe: %

Nrmber ol cli^ical disclplines:median(ran8€) Proportion of .lient.ontad invivo: mediu (range) Rar o ot full-limel part time nafi: m€dhn tMse) Team op€rats weekends * wek days only: % Team operat6 ciended hoursv normd ofi@ hou.s: % r'4eanindividualc6er6d: Stafichr&t€ri*i.s

median(ringe)

Exfinsic jobsnidac.ion: medLan(range)

analysis

Paiients werc the unir of analrsis, so paiien$ i. ihe samc t€lm sharedrhe slmc tern cha.actenstics.They were ako auo cated thc samcstalf characteristics.follow ins thc tcdD lppro.ch of a$ertivc outreach wherebypari€trisare carcd lor by the whole tcam and not bv onc individual staff

7l .1(2-s) 1 5 ( r 66 6 ) l.r (0.48.0) 50 38 eF-r4)

(2verasedd term evel)

Intrlnsi. job etislidion: median(ranse)

Statistical

7.3(l.l 5 r) 2r) 0.3{0.0

Emdional €xhru$ion (bu rn our nventory): medan tahse) oeper$nalisation (bum out inveitory): mediantanse) Personala..omp shment (burn-om rnventory)rmediantange)

ase (yeart: medi tm8e)

4t 111-47) 70lt7 23) r3(l r0) 4(2r) 3 s ( 2 9l t ) 3 6 ( 1 67 l ) 36 52 5l

Previols hospnaladm ision s: %

Data Bere analysedusing STATA 7 0 for !0indows (Staiacorp, 1999). len paticnt variables had up to 9% missing values, .ni 24% oi patiens had missnrs vahes on 3t lca$ ore variable. To lvoid los of prdsion, we imputcd the nissing b,seline valucs using muhipl€ impDt tion (Clark & Altnan, 2001), so rhat a1l analyses were based on all subjects with thc outcone obserued.Bccausepatients in rhc s.me team may not b€ indePend.nr,$and ard $adstical r€chniqueswould produce nrored standard errors. We therelore conputeJ all stinda.d c.rors by the robust

Contrct with oth€rseryies: %

3 33 16 t9 72 55 2A 11 20 3,1

Admifted to hosptal nthelo ow up Perio!:% Compusor y admitted to hosptalin the lollow-up Period:%

25

0

t-3 4 9 H6piialisation inlhe ld 2 /earsi % Compukory admissionih rhe ldt 2 yeaB: % Alcoholordrus misuseordepcndenc/ in the l6t 2 y.ars: % V i o l e n @n t h e l a s 2 / e a r s : % Arre*ed inrhe la* 2yeatr: %

RESULTS

asoiated

T.ble ? shoss rsults ofthe univariateand muhivanate prediction of hospnal admissions of any rype in thc follow up period. Out of the nine te$ed team cbracteristics, four were as$ciated si8lificartly wiih ont cone in the univariar analysis. H.vins more clinic.l st ff and more dsienated psychiatiist 1nput, wo.kins !t weckcnds and workins our of office houB ach predicteda higler probabiliry of admission. In a mlltlvadate adjusted nodel, only we€kend wo.kins rcDained , sisnificant predictur of hi8her admissionrates. Hisher rores of stalf on pemnal acconpl'shmeni gedicied lower adnission rats i. both uiiva.iate aid nultivarlaF analysis.ln rhe multiva.iatc amlysn, hishcr

although there was no significatrt rsciation .t the univariatel€vel.li n b b€ noted rhat high leMnal accomplishmentcorrelat€d sis fi.antly wrth los deper$nalisation. Five patien! characterisrics were correlated wnb adnnsions ar the univariare level, thre€ of which remai.ed sisrificant in ihe multivariate model. Patienis wirh more .dnissiotu in then history and, inde pendcntly,more adnissions within rhe lasi 2 years were nore likcly to bc admiited again, whereascont.d with other snices was assciared with lower adnission raiesThe univariareand multivariate predic tior ofcompukory admission in ihe followup F€riod is summa.'sd in Tablc 3. In rhc univariate alalysis. five ream character istics were aseciared with o clinicrl staff, more psychiarist inpui,

suks

Df

deFMmlehun

also

wE

T.bl€ 2

Prcdicbr olprt enr h6ptali:ations (all)ar 9 month foll*

wnh bwer a
integlatior of health and socialca!e,week cnd workins and workins out of officc hous each p.€diced a hisher probabiliry of compulsory admissbn to hospnalwithin th€ follow up period- In the multivariat€ model, only worldng on weekends and o!! of office ho!6 remaiied significart predictors. ln this model. however, the dn€cion of efiect of out of hours workira was reversed compared with the univariat€ aralysis. When the influence of all other varilblcs had ben adjusted for, out-ofhours working was asoiat€d wnh lowcr not hisher conpulsory admi sion .ates, wheieas weekend workins continued to predici a hider probabiliry of compukory admissions.staff scoreson d€person,lisarionand personalaccomphh' ment predictd compulsory admhsions in the sme way as ihey did lor adnission of

up

H6pitalietion

at 9-modh rollo*up

(y6irc)

Full/.djunedG.d.or teamNerrse)

(95%Cr) Total number of clinic full-tim€ ialt in team (m€ unit inoese) &siSnat€n pttchiahnt tifr€ (one unit in.reas.) Wh.ther

6r rct the rean hd inregrded h€lrh and sid

c.re

oddsrnio

(r.0o-r.12) o.0!l (r.0rr.66) 0,029

t.05 t.30 t.8l

(0.991.29)

t.0a t.05

(0.33-r.12) (0.90r 2l)

t.02 tia |.17 0.92

(0.9f r.r0) ( t . 2 62 7 9 )

0.051

(9s% cD

(03r-r.0 r) (0.87r.29) {0.72t.65)

0.089 0.ss9 0.676

(0.3s,r.r3) (0.e5r.B)

0,939

(0.86r.r) (r.323.26) (0.520.95) (0.89-r.06)

0.4'7 O.O(n 0.023 0,177

(0.32-r.35) (087 r.,r€) 10.67t.26) (0.63-0.r) (0.7t0.98)

0.694 0.359 0.619 0.001 0,03|

{0.75-r.r4) (0.83t.%)

0.4s4 0.269

(0.5rr.16)

o.,rs6

(inre3rationv. non-ituqrnion) Nunb€r otclinic.l di*iplines (one unit increa*) Proponion ot clier 6nt.d

in vivo (10%poift increde in

t.00

percentre€ofcoftads in rhe ommunty) Rariooflull-tine/pafr rime $afi(d€ Term opertd

trhit ii.rde)

wekends r week days onlyl

Term operat6 eiended t4un individual6e-lo.d

hours i nofmal offi.e hours (one unit incred€)

Infinslc job satisr&tion (ones.d. in.ree) Etnnsi. job s*iddion

(m3.d.

Emdionalexhaustion burn-odinveftory(ones.d. in.ree) Depeenalhation - burn ort inveftot, {one s.d. inrce) Persnal e@mplEhmed

t.05

incre6e)

buh-od

inv€ntory(one3.d. increAe)

|0l t.06 0.75 0.92

Living alon€v. livinSwith dh€rs Prdio6

h6phal adm ssio6 (one unlt increM) H6pitalietion in.h€ ld 2 /ears (y6 y. no) Conpuhory rdmksion inthe ld

2 yeaB (ye v. no)

Alcoholor drlg misuseord€pendency in rhe ldr 2 ydrs (yes f, no) viole.c Arre$e! cont'd

inthe l6t 2 yea6 (/es v, no) in the ld

2 yeaB (y6 v. no)

with dher s€rvies (y6 v. no)

r. v.rue5of P in bordryP{rre < or5. 3. B5r rul moder:*ek€nd houB.

0.42 t.l9 t.,13 4.02 2.95 t.0,1 t.82

(r.00 2.rs) (0.34-r.0 r) (0.84r.r0) (0.73-r.r7) (0.79t.2r) (0.86r.r2) (0.62-!.9r) (0.80r.07) (0.7s-r.72) (0.55r.23) {0.8,1t.68) (t.t8t.7,t) (2.241.71'J (r.96-44t) (0.66r.55) (t.2t-2.74) (0.9r2.37) {0.320.7,1)

0.640 0.002 0,050 0.045

0.954 0.589 0.003 0.244 0.535 0.140 0.3t7

0.855 0.txx 0.1t1 0,ool

2.07

t.05 t.t3 0.93

t.77 0.83 t.05 t.14, rie 0.88 0.98 0ie

10.69t.6t) (.09 r.65) 1.49-4.52) (0,892.r5) (0.55r.,rr) (0.9r-?.r9) (0.6rr.58) (0.2t 0.63)

0.8t6 O.t06 0,001 0.r5s 0.s97 0.r2 0.9,+l
Tr6leI

PredicbBof paiient.ompuhor/admnsion:at9 monthLlld

up

compul$ry admksionat nine-month follow-!p(yetno) Fulry adjlned (s.d. or team-averag€) Ondsratio

195%CD Totalnumberofclini.alfull-tim€ *alt in team(on€unit increae) (one D€siSnated unit incfease) Fych ahkt rime Whetheror not rhe teamha nt€Brdedhalth andtridcare Numberol clinicaldi*iplines(ore unit lnseae) Proponionolcliemcortad in v,vo(10%polntin.r€asein pereitaSeoronta.c ii the communny) Ratioolrull-tlme/padimes. r(ore uit moe6e)

t .t 2 |.t9 t.30

t.ol 2.59

Teim operates w€€k€ndsv. weekdrFonly Teanopenresetended houBv.nornalofiicehou6 M€anndividual c&-l@d (oe unit incr€ase)

0.91

(r.06 r.20) (r02 r.88) (r.2r-6.96) (0.94 .78) ( 0 . 9 0r . l r ) (0.9rr 12) ( 4t 4 72) (r.0r-2.73) (08 r.07)

0,0J5 0.0t7 0lt0

0.002 0.323

Insinsiciob satsracrion (o^es.d.increrse) E{rinsicjob srisfa
t.t2

(0.34r.s0) (0.72-r.23)

Emotional exhau*ion bum-or invetory (oes.d. inaes€) Dep€rcnalsarion- burn od inv€ibr/ (de s.d.inciese) Personal a@mplchmeft buri-od invenbry(de s.d.inftse)

t.05

(0.7rr.4r)

470

(0.94-r.43) (0.s10.r2)

0.t66 0,0t0 0.884

|.t7 t.44 1.43

(0.82r.2s) (0.49-r.34) (0.73r.52) (0.852.04) (r.r9-r.73) (2.r19.5)

6.98

(r.r r6.00)

2.34 t90

(060 r.92) (r.,o 3.91) ( r . r 01 . 2 8 ) (0.100.76)

|.02

r.o5 LivnEalon€r livinaw!h others Previous hospiral idmksions(oneun't incree) Hospnalisation in the las 2 years(yesr no) Compuleryadmission in the lat 2 y€aE(yesr no) Al@holordrugmku* or dependenct inthe la3tzyeaE (yesv. no) Violenceh rhe l5t 2 /ears(ys v. no) Arre$ed in ihe 1612 yeam(yesx no) ()€sr no) Contactwith o.herservices

alltypes. With resp€ctto patient character istics.the totilnunber ofadnissions in the patienCs hntory as wcll as admissions, compuhory admissons, v,olence and a.ie$s in thc la$ 2 yca^ each predicted hisher conpulsoq admission rate!, Bhereas coniact with oiher serrices wrs r$ocirt.d ivnh . lowcr r.obabiliry oi bens section€d. ln rh€ multivariate nodel only two variables renained sisnificant rrcdictors {i.c. conpulso.y a.lnissions in thc la$ 2 yca6 lnd con ta.t wnh oth€r senicet and phrsical violence in the last 2 rca6 aplroachcd $ati$ical sisnificance. Paieise lnterictions {ere Ested bcdccn rhor lariabhs that arc sisnificant pr€dictors nr thc final nodel. Ahoseder 58 irtera.tbns were tested, 34 fd ea.h

0.48

outcome- Four of rhelr .rc sisnificanr at r<0.05j 3.,1suchrcNlts rrero be expecied b ) ' c h a n . c . n d n o n eo f t h e i n t e r d c t l o nws d s s i g . r f i c ! n ta t P < 0 . 0 1 .

DrscusstoN What predi
0.86 t.24 t.t2

1.03 .05 t.20

0.783 0.222

0.821

(9s%cD

(0.96-r.03) (0.69r.03) (0.78-r.98) (0.90-r.39) ( 0 8 8r r r ) (0.90-r.r4) ( .17.+.14) (0.49 0.9r) (0.91-r.B) (0.85r.30) {0.79r.16) (079 r.30) (0.7r0.96) {0.66 0.9,1)

0.549 0.t91 0.359 4.291

0.808 0.002 O.O,ll

0.651 0.745 0.0t2 0.007

(0.85r.44) (0.5sr.72)

O449 0.9t5

(0.64r.84)

0 772

t02 .08

(r06-r.57) 0.oro 10.26t.64)

0.3t/ 0.020 0.002

0.85

0.364

(2.5020.00)
this sttrdy does nor address the overall effeciirenessof .ssenlve o!t.e.ch teams. Rathcr. t utiliscs thc cxnring variltion bc dccn tcrnE, {aff and parients to asscss whcrhc. (and. if so. in what wat) such charade.isticspredict our.ome- Onlr No simple outcone diftria were used:whether patients were admilad and whthcr thcy wc.c adnined compulsorily wnhjn a 9 monih follow up period. Other outcome crireri!. such as taricnts psychopatholo$. quality of hfc lnd trcain€nt sarisfacrion, Dray be equally of ev€n more imlonalt ta.sets of a$erive otrrreachthan lreven. ins .dmi$ions. In some cases,voluntary hospnalirtion oighr even b. resardcd is a positivc ortcomc if it indicatcsa dcg.cc of c'l8ascnrnt wnh seric€s, alihoush coD puhory admissbn is atr .dve6e outconc

th.i servicestrr nr avoid. The advantascof rhese ourcomc critcria is lhar rhey are deadr op€rarionalncd and car bc est.b' lished on the basi of rccords alorc, rhls avoidinS selecdon bias throush non .csponse to research intcniews. 'rtc m6t imponantesult is that ceftain charachrnt]cs of eans, staff and laiients were all foMd to b. predidive of outcome. This held irue i! multivariate lmlyss whcn rhe influence of all oth€r variables was con tolled for. Belore concluding th.i thes ef fects are causd. w€ m!s! conremplaterhe ide. that ihey may be du€ to conJomdns by uinersured Yadables.

Team chara
p.edict out one when th€ innuence of other factors is controlled for. l he* frcton rhceiore mar be les imlortanr for the cffNrvc nes ot ieans thm has ban sugg*ted on the basis of rviews (Mueser ct 41, 1998, Gny ur zl. 2002). The findings midt encou rage senicc proidere to be nore flexible ovd rhesc dspccls of as*ft]le oureach, and not n(€$a.jly adhcrc to dctailed prescdptions lackins rerarch evidence. Staff chara
Patient
This was cxpcctedbecdlse the two c.iteria are not ,ndependeni hoslihl admission ,ncludcd compukory adnissiois. A higher rotrl rumhd of previous adnissions, volurrary o. comtulsory admissionsin the la$ 2 y€aB, physical violencc in rhc la$ 2 years and no contact wnh other s€rvices predictedpoorer ourcome on both cliterii. In muhivarjate .nalr.ses,however,different lnd spccific cverts i'r the past seem to br the be$ predictoB of sinil,r cvenrs in rhe follow up period (i.€. hospihl adDissions in the past predia funher adnissions, and a hisro.y ot .ompukory hospital rdmis sions n rhe bcs prcd'cror of compulsory admissionsin lhc tururc). One misht conclude rhat wnere teatn€nt has failcd ,n the past it is nor€ lik€l/ to fail in the tutue, a.d thosepaiienrs{or whos care rhe as*r riv€ outreach reams havc bcen spcc'fically vt u! (i.e- those wnh a history of voluntary and .ompulsory admisio6), still have the Poorestout ome- Asserive outreach reams face the same problems with these paticns as generic codnuniry mental heahh ieams, despite then superior resources and tar seied approacb. This implies ihat t€ams wirh a hish pcrccnraseof this core group of ladents manased br as*rtiye outcach on their.ase load inevitablr tend to achieve a less favourabl€ average outcome, and what eams can .ealislicallylccomplish *i11 deperd on the history of th€n paticntCont ct with other seryices emers€d as ! vcry powerfLl, independe.t p.ediaor of favourable outcome. To som€ desrcc, patienis' .ontact with oth€r snices misht simply refl&r , higher ievcl of engagement, a greater willin€Fress to ac.epi suppoft and befter skilk to sek and receive it. Thus, patients' attnldes .nd skilk miy explain th€ predrrivc a$ociat'on. Nevenhele$, th€ facr rhat co.r,ct w'th orhe. seryices alone reduc€s the risk fd voluntary rnd .ompuhory adnissjons by aroud 50% may be *en as evidence for ihe inportance for mllti-asency *orking with this soup-

lmpli.ations

and future

research

The findinss of rhc study point ar thc complexities of Fediciins outcon€ under routide conditions- Aspects of how the team is oreanised,staff bd out, patienis' histo.y and lhcn@nlact with other ervices have be.n idcnrificd as indcpendentsisnifi cant predictoB and should becons'dcrcdnl reka.h as wcll .s clinical practice. La the UK, lbe decisionon wheiher asenive out rcach should be imllemented has been

talen, and asseftiveoutrcach teamswill be pan of establishedservicesfor somc timc ro cone. The challense now ls to cvalutc how the teams wo.k r.d to inprove rhen cffcdivenss. This $udr provides sone indi.ation abour whai factors may hdR to be hrgeted in the procc$cs of clinical sovemance.nd senice dcvclopment.

CLINICAL II{?LICATIONS

outr€ch teams wo.k, ttafi burn'out and patient chamcteristics are ass€iated indep€nde^$, whh outcone. r Asp€cts of how a$qdw

I Weekend rc.kint

ncnr ofthe nuiy Peterlrre, <evi C.lriat (;rl ruu Thori.lir T.i, Crag ri. AnB. r ajr.rtc! lh. iucr wis d.,r.p.d ana . mder ih€ rurDier o' The L.i.l Vrtual n\t'rrte f.r Reser{h and rrev..pm.ni

r Only 25 predictor variableswere tested. . Thh i! a naturalistic obsemtional study, and the as$.iatio8 confoundint rather than cdsal relationships.

The Pan-London Asse.tiv€ Outrea
J.. ,o6n,q, s.. BGMnsbn, P, d or (200r)

5 T E | A NP R L B Lf r l c r ' : / . h r ' , ' A ) f A K H O U R Y P h q U n t f o r 5 0 . r t r r d C o n n r t / r a / . h 1 i r r E r f r ! a n . T h . AN Wl rll +lr flRcBoiati.s u1r cJn,&Jle IoANNA wAlTs ll^ uri L.inor sdio. ..Medr.e ro. so.i and a.mmlnit r!/.hrir/-iafrs and Th. LDndorsllio! .i Hed. re P\Lr LltBtsNCTON Belv.orr 9.'.n..s Un,leEir Co ese Fftckr.h IO NN BLLNGS Ml DeprrhentorPirhrtryan. FRCP5y.h, Depr{menl L.ndoi and Camdenand nmebn l'1e.ta tseilh aid 50.alCrr.tnni TOI'1BLIRN5. on,l.n SONAIOHNsON MRCP!/.' .fGeErrrP![l]rt.r5r Ceotle! I ospia |]ed.a S.h.. Deparmeii.J r':r.h rhy rna B."rvourz 9.cn.cs Unve^ tr Co ege L.,.d.i rna a:r.dei ria s nebn f'1cniaH.ai.and So.u CtrreTruit |ATT IUIEN PhD AN RYRli.fA sansburyCcrfe f.r l'1e{: ).pafrm.ni.f G.ncrtr F\y.htrri 5: cellje! rlotpta f4..l.a Hea rr cHRr5TNt !!Rc rt FRa:rrh

tf^i

c.tty,I.i Bur.s.r, Kmea i.. ddr(uoou) H.ic

M.G|d, j, H,, Bond,G, R.,Di€t.n,1.,dor(!ets)

ryEi P, i.non.

M*$n, s.,rrEr, P,oi'€r, s.,d dr(rn) rar/ ( r , a r r r f n ( - r3 3 1l

(, d d (200r) ^

rrBr, P, Morgan,l.,v.n Horn. E-.- d (tes) A

h .i rda i,ai4.ncni

Ii t!,x!1,{Pe^q f,i| D{,rss E E.,L., L. K. & krk,

15 e4 0l

5.h rprrf.i.

rrE4 P, Ev$s. K-,Grdhr, N.. d d (et8)

Pri€be,s., F:rdoury,w,,w:as, l., .r .r (200r)

Fy.rdr.

nted5 ,/l

rrren P, i.nl€y, c.,v'

rbr.,

}16 0+ 09

E.. - d (4oo)

o B.0$6)

ode.

mo'les. evr !r: r; rsurdons iftrrdsru;

, studr Pr-: I a4rshr.xi.r ol r1rror!

Ro8€E,w.H.CeB)

Hornd, Dc. & rbF,oi,

loi.d 3at /16 759

!.Ndbr.pri.hnr.par.ft;

d{hrr{ H.r.r,

(ett)

r1

MreseaK.r,,sond,G.R.e Dnh, R.E.(.ett) F! ew !r E.e4

or.r

s..ory€d, s.,dd

32 lgl rC

cErL,aG. & artman, dG. (206) D.!. op i!:

N..ryror. P. R.E.

c

ra.avea l0 N.r.mb.r 2001 lra f.rson27l'1try200.1 i(epled 76,uie 7004)

l33 l9 17

Pri I srrif lo'iorrlPv.hd.y

crdhr.

found may reflect

C.i r : o f f e l p . i a e i . . , r . J . * . r S i . l a nP r . b c U n t L r 5 0 . r r r d C r n n u r : , P \ r . h r _ r r \ . i h r r +44(.4.141t4C2916 l o r l ' 1 c n i r H . r i r L o n d o nE I S 5 P U KT e + 1 4 i 0 r 2 1 l 7 5 1 r r r l 0 t ' e t r D i p r e b e @ q m ut r . u L

R EF ER E N C E S

ft rioroskdrl1rj.i.

outrea.h hwe a

r Onb/ voluntary and compukor/ admissionswithin a9'month foll@-up period were taken as out me crite.ia. Other (ireria such at p6tients' viM and quality of life *re not @nsidered.

APPENDIX

r!.w

admision rates.

r Patients who ar€ in contact with 3enices other than drertire mu.h lorer rhk of voluntary and compulssy adnission.

ACKNOWLEDGEI{ENTS

uflins,

predicti hither admissio./@pulsory

dl.0rt2)

a

\lfdrrs,

i

.us€.ed'mp6 !r/r/ rerh..rrlnrft, rl 9 rl st.66p

r nr lnLveE lorrdr orrh. An-..r

RegRiq

103

0rtD rdd sbrntri s.tMr rirrrcr /r

!d il , rr

ryrer, P & simmonds.s. (20q r. id
w.iss, D j., o{L, R.v,En3bnd,G-w, dd(r%7) //drNdrlor ir. , rie ' o t ira.r,r q,aruidr wrisht, c., Bqrns,r.,lan.s, P, d rr (2oor) A'{

!.

/'rdjrd.h

ra4.rari ds.dar 6rt sr rrridr.l fti.hror/ l3r

]tl