Anda di halaman 1dari 8

British Journal of Psychiatry (1994), 165, 204210

Home-Based Versus Out-Patient/In-Patient


Care for
People with Serious Mental Illness
Phase II of a Controlled Study
B. AUDINI,

I. M. MARKS,

R. E. LAWRENCE,

J. CONNOLLY

and V. WATTS

Background.The effect of a randomised controlledwithdrawal of home-based care was studied


for half of a sample of seriously mentally ill (SMI) patients from an inner London catchment

area, compared with the effects of continuing home-based care.

Method. Patients,aged 1864,


had enteredthe trial at month 0 when facing emergency
admission for SMI. After at least 20 months home-based care (Phase I), patients were
randomised at month 30 into Phase II (months 3045)to have either further home-based
care (DLPII, n = 33) or be transferred to out-/in-patient care (DLP-control, n = 33). They were

assessed at 30, 34, and 45 months. Phase I control patients (n = 70) were assessed again
at month 45. Measuresused were number and duration of in-patient admissions, independent
ratings of clinical and social function, and patients' and relatives' satisfaction.
Results. The slim clinical and social gains from home-basedv. out-tin-patient care during Phase
I were largely lost in Phase II. Duration of crisis admissions increased from Phase I to Phase
II in both DLPII and DLP-control patients. During Phase II, patients' and relatives' satisfaction
remained greater for home-based than out-fin-patient care patients. At 45 months, compared

with the Phase I controls, DLPII patients and relatives were more satisfied with care. Such
satisfaction was independent of clinical/social gains.
Conclusions.The lossof PhaseI gainswere perhapsdue to attenuationof home-basedcare
quality and to benefits of Phase I home-based care lingering into Phase II in DLP-controls.

The Phase II home-based care team suffered from low morale.

a controlled withdrawal study was conducted similar


to those done in some drug trials. With a slimmed
team, DLP care was continued into a second phase
for a randomised 50% of the original DLP cohort,
with the remaining 50lo
being transferred to control
out-/in-patient care.

Although several controlled studies have found more


benefit from community than hospital-based care in
seriousmental illness(SM!, i.e. schizophreniaand major
affective disorder), no controlled study has examined
how long such benefit can be sustained. The issue is
important

as SMI patients are difficult to treat, and

good community care is complex to organise and vul


nerable to many disruptions. In an uncontrolled study
in Madison, Wisconsin (Test eta!, 1991), most gains
were lost within 14months of the ending of community
care. In the absence of a community care comparison
group, however, it is unclear whether further community
care, too, might have lost its benefit during that time.
A chance to do such a study arose at the end of a
three-year randomised controlled study of community
care in London (Marks eta!, this issue, pp. 179-194).
This had compared home-based care (Daily Living
Programme,

DLP) with control in-/out-patient

care

for 189 SM! patients over a period of 20 months


from entry. DLP care was significantly superior to
control care, yielding slightly better clinical and social
outcome, and more patient and relative satisfaction.
From 3 to 20 months, DLP greatly reduced the
number of in-patient bed days, as long as the DLP
team was responsible for any in-patient phase its
patients had, saving the equivalent of ten beds.
DLP funding had been due to end three years after
the study began, by which time most patients had
been in the study for at least 18 months. Instead,

Method
From January 1990to December 1992,SM! patientswho
had begun DLP care in Phase! of the study (October
1987to December 1989)were randomised at 30months
after trial entry into Phase I! for a further 15 months
of study in the BethlemMaudsleyHospital while:
(a) continuing with DLP care (DLPII patients) or
(b) having control out-/in-patient care (DLP
control patients).
Phase II began a mean of 10 months after Phase
I ended (Fig. 1). Before entry into Phase II, DLP
patients and their relatives gave written agreement to
whichever care would be offered (DLPII or DLP
control). Patients who had been randomised in Phase I
to the original control in-/out-patient care group con
tinued with it in Phase I! and were followed up at a
mean of 45 months after trial entry (Phase I controls).

Selection criteria

204

Patients were accepted into the study if they:

HOME-BASEDV. OUT-/IN-PATIENT CARE


DLP n=92******@

189SMI

I- 1@33

66 DLP patients

randomlsed

patients
randomised

Phase

17
Pre-audit

L new controls

(DLP-control

patients)

Controla =97,

Month6@

I continuing
DIP(DLPIIpatients)

Phase

II

3@O

II

34

4'5

Of the 97 original controls (Phase I control


patients) who had entered Phase I, 70 patients were
rated at 45 months. Reasons for non-assessment
were: 4 abroad, 7 failed all appointments, 3 refused
consent, 3 had died from self-harm.
Each rating set took at least 4 hours to complete
as patients often failed appointments repeatedly,
requiring multiple visits.

Post-audit

Audit
Fig.I Studydesign.

(a)
(b)
(c)
(d)
(e)
(f)

205

Had received at least 18 months DLP care


Agreed to participate
Were aged 1864
Had no primary addiction
Had no primary organic brain damage
Were living in or connected to the deprived
inner city catchment area of south Southwark.

The sample included violent or suicidal cases.


In Phase I, 92 patients had been randomised to
DLP care. Of these, 26 could not be included in
Phase II: 5 had died (3 through suicide, 1 from
cervical cancer, 1 from cerebral haemorrhage), 5
were abroad, 4 were outside the catchment area in
the UK, 1 was in a Special Hospital, 1 was in long
term in-patient care, 4 refused consent, and 6 could
not be found at the start of Phase II.
The study sample in Phase II comprised 66 of the

Rating scales and evaluation


Details are in Marks et a! (this issue, pp. 179194).
Measures included the GAS, (score 0100;Endicott
et a!, 1976); the Present State Examination

(PSE;

Wing eta!, 1974); the Brief Psychiatric Rating Scale


(BPRS, 24 items; Overall & Gorham, 1962); Social
Adjustment Scale (SAS; Weissman eta!, 1971, 1974),
Client Satisfaction (Larsen eta!, 1979)and Relative's
Satisfaction.
Treatment

As before, DLP care was intended to minimise rather


than exclude in-patient admission. At the start of
Phase II, the DLP gave 24-hour cover (by phone at
night), case management at the site of breakdown
which should be problem-oriented (but was often not
so), brief in-patient admission if inevitable, assertive
follow-up, help to maintain or acquire daily living
skills, support and education of people important
original 92 Phase I DLP patients. The 66 patients
for the patient's well-being, and advocacy as needed.
were divided into two groups - those meeting DSM
As time went on the problem-oriented mode of
IIIRcriteria (American Psychiatric Association,
1987)
forschizophrenia
(n= 20)andthose
withother care became attenuated, there was a shortage of staff
and a drop in morale (see Discussion). Phase!! began
disorders. Within each of these two diagnostic groups,
with a registrar, senior registrar, senior nurse and
subjects were rank-ordered by their Global Assess
ment Schedule (GAS) rating at 20 months (or 11 four charge nurses, but later the team had only one
doctor, a senior nurse, and two charge nurses. In
months where a 20 months rating was unavailable) and
were then randomised in pairs to 15 months as in addition to having responsibility for the DLP
eitherthe DLPII or DLP-control condition (eachn = 33). patients, the team also cared for 16 new acute
SM! patients outside the study who were not evaluated.
A consultant (JC) gave one session a week to the
Assessment
DLP throughout Phase I!.
DLP-control and Phase I control patients had
There were two independent PSE-trained assessors
(a psychologist and a psychiatrist) who took no part standard BethlemMaudsleycare as out-patients
in care. At 30 months, ratings were available for 32 and/or in-patients as necessary.
DLPII patients (one refused) and 32 DLP-control
patients (1 committed suicide 2 weeks after entering Data analysis
the DLP-control group). At 34 months, ratings were
The DLP!I and DLP-control patient groups were
made of 31 DLPII and 30 DLP-control patients
compared at 30 months, the start of Phase!! (t-test).
as 2 subjects left the UK (1 DLPII, 1 DLP
control), and 2 refused contact (1 DLPII, 1 At 34 months and 45 months differences were tested
by analysis of covariance (ANCOVA) using the 30
DLP-control). At 45 months, 30 DLPII and 28 DLP
control patients were rated (3 refused-i DLPII and months score as covariate, and confidence intervals
2 DLP-controls); thus 88% of those who entered (95%) were calculated. Because Phase I control
patients had no ratings at 30 months and 34 months,
Phase II were rerated at its end.

206

AUDINI ET AL

t-tests were used to test differences between Phase!


control v. DLP!! and DLP-control patients at 45
months.

admissions (29 v.27) and duration of admissions (22


days v. 21). During Phase I pre-audit (see Discussion,
below, and Marks eta!, this issue, pp. 185, 187), both
groups had shown similar durations of admissions
Results
(22 days v. 29). During the post-audit period near
The 33 DLPII and 33 DLP-control patients were the end of Phase I at 20 months and before Phase
assessed at 30, 34, and 45 months. Attrition rates I! began at 30 months, mean duration of admissions
were low -3% across both groups at 30 months, 6% had lengthened for both DLPII and DLP-control
for DLPII patients and 9% for DLP-control patients patients (respectively 56 v. 22 and 31 v. 21 days), but
at 34 months, 9% for DLP!! patients and 15% for more so for DLPI!.
DLP-control patients at 45 months. Of the 90 Phase
During Phase II, DLP!! and DLP-control groups
I controls alive and in the UK at 45 months, 70 were each had nine patients admitted, for similar mean
assessed (22lo
attrition). Dates and durations of in
durations (52 days v. 51)- constituting a 2.4-fold rise
patient admissions were available for 33 DLPII, 33 for DLPII and DLP-control patients compared with
DLP-control and 80 Phase! control patients.
the same patients in Phase I as a whole (P<0.005
and <0.008 respectively). Admission length for
Phase I control patients rose slightly from Phase I
Inter-rater reliability
to 11(76 days to 79). DLPII patients had 23lo
shorter
Two raters rated non-study SM! patients to obtain
admissions than did Phase I control patients in Phase
inter-rater reliability
one interviewed the patient
I!, whereas in Phase! the original DLP had 76lo
while the other was behind a one-way mirror. The shorter admissions than did the original controls.
measure of agreement
across all raters was intraclass
(Figures for Phase! control patients are less complete
correlation (r), which was satisfactory: GAS 0.76, than those for DLP!I and DLP-control patients.)
BPRS 0.75, PSE
total 0.83, DAH 0.83, BSO 0.79,
The total number of admissions was similar in DLPII
SNR 0.80, NSN 0.81, PSE Index of Definition 0.71. v. DLP-control patients (12 v. 14), and was 40 for
Phase! control patients.
In brief, from Phase Ito II admission duration
Month 30 (baseline)
increased significantly for DLPII and for DLP
DSM-I!!-R diagnoses were the same for DLP!! and control patients, but not for Phase I control patients.
DLP-control patients: 301o
schizophrenia (10 each), DLP!I and DLP-control patients had a similar
70% non-schizophrenia (23 each). PSE diagnosis for
number and duration of admissions; both groups had
Phase I controls had been: 471o
schizophrenia, 20lo shorter admissions than did Phase! control patients.
mania, 16% depression, l71o
neurosis.
Over the 15months of Phase 11,73% of DLPII, 73%
DLPII, DLP-control and Phase! control patients
of DLP-control and 76lo
of Phase I control patients
had virtually the same mean age (37), 010male (46) were not admitted.
and ethnic background (same as in south Southwark
65%
British/Irish, 26% Afro-Caribbean) and did
not differ significantly on first admission rates (61%, Clinical function
79%, 50lo) or proportion
without
home support
DLPII and DLP-control patients were compared
(alone/single with young children: 55lo,
79lo,
49%). at 30 months, 34 months, and 45 months. For each
Although before 30 months the future DLPII and time on each measure, calculation was made of
DLP-control patients and their relatives had similar means, 95% CI and ANCOVA, with the 30
satisfaction scores, at 30 months these scores began months score as covariate (Table 1). DLPII
to favour DLP II patients for the first time (P< 0.08 and DLP-control patients were each compared
for patients, P< 0.07 for relatives, t-tests). The 30 with Phase I controls at 45 months on t-test
months satisfaction ratings were made after patients
(Table2).
and relatives knew whether their continuing care
DLP!I
and DLP-control
patients
lost gains by 45
would be in the DLPII or DLP-control groups.
months on most clinical measures (GAS, PSE,
BPRS,

Outcome (after baseline)


Length and number of in-patient admissions

During Phase I as a whole, both DLPII and DLP


control patients had had similar numbers of

SAS),

except

for DLP!!

patients

on PSE-SNR

and PSE-NSN and for DLP-control patients on


SAS-marital. The few between-group significant
differences favoured DLP!I over DLP-control
patients and were at 45 months, except for SAS
marital, which favoured DLP-control patients at 34
months. Outcome never favoured Phase! controls

HOME-BASED

V. OUT-/IN-PATIENT

CARE

207

Table 1
Outcome of clinical and satisfaction measures

(s.e.)nGAS30

I control

CIANCOVA

MonthDLPII

PPhase

mean(se.)rol n95%
mean(se.)nDLP-cont

66.4 (2.6) 31 66.6 (3.2) 30


34
61.5 (3.9)32 28NS
62.0 (4.0)32 3071.8(3.1)
4568.2(3.1)
(2.2)
(1.9)
39.1 (1.8) 31 39.5 (2.2) 30
34
(1.0)70PSE-total30 4537.2 42.3 (2.7)32 3036.5
41.4 (2.3)32 28NS
(2.2)70BPRS30

mean

NS58.4
NS41.2

8.4(1.7) 30
7.2(1.3) 31
34
(2.4)70PSE-DAH30457.0(1.6)
28NS
10.6(2.3)32
7.6(1.5)32
308.2(1.7)
(0.5)
(0.5)
1.6 (0.7) 30
1.6 (0.6) 31
34
(1.0)70PSE-BSO30451.3 2.6 (0.9)32 300.9 2.1 (0.9)32 28NS
(0.4)
(0.7)
1.1 (0.4) 30
1.5 (0.5) 31
34
(0.7)70PSE-SNR30451.7 2.5 (0.6)32 301.4 1.7 (0.6)32 28NS
(0.5)
(0.3)
30
1.3(0.4)
31
1.6(0.4)
34
2.5)0.01'4.6(0.8)70PSE-NSN30
281.5
451.1 0.5(0.2)32
301.6 2.0(0.6)32
(1.1)
(0.7)
4.1 (1.0) 30
31
34
2.9(0.6)
(1.05)70Patient's3027.4
452.8 2.1 (0.5)32 304.3 4.7 (1.3)32 282.4
(1.8)32satisfaction34
(0.9)3223.8

NS19.0
NS3.7
NS3.3
(0.4to
(0.2to 4.9)0.07'7.3

to 1.5)
(0.6)70Relative's3058.4
0.007'23.4
3.3
(5.6to 1.0)0.001'
22.9 (1.7)30 263.3(5.1
26.6 (0.9)31 2821.4(1.7)
4527.6(0.7)
(2.3)550.0(3.5)11satisfaction34
(3.4)
(3.5)
42.6 (3.2)14 1319.1 (32.8to 5.2)0.03'45.7
4552.3 58.0 (1.4)7 944.7

(2.8)70

DAH, delusional and hallucinatory; BSO, behaviour, speech and other; SNR, specific neurotic; NSN. non-specific neurotic.
in
favour of DLPII patients.

over DLPII patients, and almost never over DLPcontrol patients.

little better than the mean score for Phase I control


patients of 58 at 45 months.

Groups overall

worsened slightly from a mean of 37 at 30 months to


about 42 at 45 months, almost the same as the mean

BPRS: In DLPII and in DLP-control patients scores

GAS: Mean scoreforDLPII and forDLP-control scoreof41 forPhaseIcontrol


patients
at45 months.

patients worsened from about 70 (slight impairment)


at baseline to about 62 (mild to moderate problems),

PSE: On total score, the marginal DLP I!


patients' superiority over DLP-control patients was

Table 2
Outcome of all measures for DLPII and DLP-control v. Phase I controls

mean (se.)nDLP-controlmean (se.)nPhase


global
SAS parental
PSE-total
PSE-SNR
PSE-NSN

(0.3)
1.6 (0.2)
7.6(1.5)
0.5(0.2)
2.1 (0.5)

45
45
45
45

CIt-test

I con

MonthDLPII
P'SAS

mean (se.)trol

a95%

7
30
30
30

(0.2)
2.0 (0.5)
10.6 (2.3)
2.0(0.6)
4.7 (1.3)

6
28
28
28

(0.2)
2.7 (0.3)
19.0(2.4)
4.6(0.8)
7.3 (0.8)

23
70
70
70

28

22.9 (0.7)

26

23.5 (0.6)

64

(1.3to 0.1)
1.1
(2.1to 0.1) 0.04
11
(18
to 4) 0.004
4.2
(6.6
to 1.8) 0.001
5.2
(8.3to 2.1) 0.001

Patient's

satisfaction
Relative's
0.005,in
satisfaction45

45

26.6 (0.9)

3.2 (1.0 to 5.4)

0.005

12.3(4.1to20.5)0.02
453.0 58.0(1.4)30 92.9 42.6(3.3)28 133.7 45.7(2.8)66 170.7

favour of DLPII patients.

208

AUDINI ET AL

greater at 45 months than at 34 months or 30 months


(P= NS).

Compared with Phase ! control patients, both


DLP!!
@

and DLP-control

patients

were much better

at 45 months (P= 0.001, 0.02 respectively), mainly


due to their fewer neurotic symptoms (see below).
On syndrome subscores, at 45 months DLPI!
patients were superior to DLP-control and Phase!
control patients on SNR (0.01 and 0.001, respectively)
and NSN (0.07 and 0.001, respectively). DLP-control
were superior to Phase I control patients on SNR
(0.05) at 45 months.
SAS:

Both DLPI!

and DLP-control

patients

worsened slightly from 30 months to 45 months, and


at 45 months did not differ significantly. The only
difference was marital, favouring DLP-control
patients (0.02) at 34 months.
At 45 months, Phase I controls were slightly worse
than DLPI! or DLP-control patients on most subscales
except for economic, on which Phase I controls were
better than DLPII patients (P<0.008). A paid job
was held at 45 months by only 10(30lo)DLPI!,927lo)
DLP-control, and 22(31%) Phase I control patients.
Satisfaction

(a) Patient's satisfaction

@30

(0.005)

CD25
20

0.001

0.001

o.@b'i' 0.007

Phase I

Phase II

:15
@10
-5
R.ndo,,, OLP
i

0
0
@60

20

-@l

30

34

(b)Relative's
satisfaction

45
(0.005)

@.

CD50

.-.

*..

0.002

0.03

Phase I

Phase II

dip controls
(n=23) (n=15)

dipli
(n=9)

dlp.controls
14

10

Patients and relatives were satisfied with both forms


of care. However, satisfaction was even greater with
DLPII than DLP-control patients care. Mean scores
for satisfaction appear in Fig. 2.

L
0

20

R.ndom OLP

@ithd@@i
]

30

34

45

Monthsafter
entry

Fig.2 Satisfaction
scales:
mean scores
(a)Patients'
satisfaction.
Patient's scale: The difference favouring DLPII
Phase I:, DLP(n=68); --, controls (n=67). Phase II: -,
patients on the total score became significant by 34 DLPII (n= 32);$@4DLP-controls
(n= 32).(b)Relatives'
satisfaction. Phase I: ,DLP (n = 23); ,
controls (n = 15).
months (P< 0.001) and remained so at 45 months
@4, DLP-controls (n= 14).
(P< 0.007);
itwas alsosignificant
on eachoftheeight Phase II: -,DLPII (n=9);
P=0.005, for DLPII v. PhaseI controls. P=0.00I etc. for DLP
scale items. DLPI! patients' satisfaction was greater v. Phase I controls and DLPII v. DLP-controls in Phase II.

than that for Phase I control patients (P<0.005).


Phase I control patients' satisfaction was like that
of DLP-control patients.
Re!ative's scale: The difference favouring DLPI!

patients on the total score was significant at 45


months (P< 0.03); it was significant on each of the
13 scale items. If relatives' satisfaction is an indirect
guide to family burden, then the burden was less in
DLPII than DLP-control patients. DLP!! relatives'
satisfaction was also better than for Phase I control
relatives at 45 months (P< 0.005).
There is a caveat about results with this scale. At
45 months, 33% (22) of all rated patients were living
with a relative. Among these, a relative's satisfaction
rating was obtained in 50lo
(5/9) of DLPI! patients
and 80% (11/13) of DLP-control patients.
Subgroups

(ANCOVAs

Schizophrenics:

DLPII

on all measures)
and dip-control

patients

differed little at 34 months and 45 months. At 34

months,
patients.

SAS global

Non-schizophrenics:

(0.04) favoured

DLPII

DLPII were superior to DLP

control patients at 34 months on SAS marital (0.02)


and patient's satisfaction (0.001), and at 45 months
on PSE-total (0.03), PSE-SNR (0.007), PSENSN
(0.03), and patient's satisfaction (0.001).
New and old patients:

patients'

For the subgroup

new

as identified in Phase I, DLPII were

superior to DLP-control

patients at 34 months on

patients' satisfaction (0.001), SAS social (0.08), and


at 45 months on GAS (0.07) PSESNR(0.03),
patients' satisfaction (0.06) and relatives' satisfaction
(0.08). For old
patients' the only difference was for
patient satisfaction (0.05).
In brief, both DLPII and DLP-control patients
worsened somewhat by 45 months. DLP-control
were almost never significantly better than DLPII
patients.

The few significant

differences

favoured

HOME-BASED

DLPII

patients,

symptoms,

and then mainly

largely

V. OUT-/IN-PATIENT

on neurotic

in the non-schizophrenic

subgroup. Patient and relative satisfaction favoured


DLPII patients significantly.
Discussion

The comparisons in this study are well controlled for


the DLP in Phase II (DLPII patients) v. the ex-DLP
patients who became controls (DLP-control patients),
to which assignment had been properly randomised.
The comparison of DLPII and of DLP-control
patients v. the original controls (Phase I control
patients), however, is not on a randomised basis, as
an appreciable number of Phase I DLP patients
could not be included in Phase II.
Our controlled results are rather chastening. The
anticipated superiority of DLPII over DLP-control
patients did not materialise on most measures. Many
of the gains from community care which had been
present at month 20 were lost during months 3045
(Phase II) in both groups. There was no DLP!I v.
DLP-control patients superiority on duration of
admissions or on most clinical or any social
measures. Patients and relatives nevertheless con
tinued to be more satisfied with DLP than control
care, and some continued to try to get DLP care
despite being discouraged from having it (see below).
The mean duration of in-patient admissions rose
from 21 days during DLP care to 52 days during
Phase II DLP care (DLPII patients), becoming
similar to that of the new (ex-DLP) controls (DLP
control patients, 51 days). In both groups the mean
duration of in-patient admissions was almost a third
less than that for the original controls (Phase I
control patients) during Phases!(75 days) and 11(79
days). This lowering relative to Phase I control
patients might partly reflect two of the original DLP
patients having to be excluded from the DLP!I and
DLP-control patients groups, one as he was in a
Special Hospital and another as he was in long-term
in-patient care. The DLPI! patients' team was not
responsible for any in-patient phase of care, which
probably prolonged admission in DLPII patients
(Marks et a!, this issue, pp. 179194).
A different factor may have reduced the duration
of DLP-control patients' admissions. During Phase
II, some DLP-control patients and their relatives
insisted on writing to and phoning the DLP team,

attending its base, and seeking its advice for their


problems despite the team asking those patients to
attend the hospital service. The DLP had become
part of the network of those patients, who tried to
keep it rather than use the hospital service. Some
DLP-control patients thus had a diluted form of

DLP

209

CARE

care from months

3045, reducing

their

similarity to the original controls and their difference


from DLPII patients.
On clinical measures outcome was largely similar
across the DLPI! and DLP-control patients' groups.
At 45 months the only measure on which the DLPII

patients were significantly superior was PSE


neurotic symptoms, and the DLPII patients were also
superior on that measure and on the PSE-total when
compared with the original controls.
On social measures, at 45 months the DLPI! and
DLP-control groups did not differ significantly,
though DLP!I were better than the original control
on SAS overall and parental, but worse on SAS
economic.
DLPII patients only showed consistent superiority
to DLP-control patients at 45 months on patient's
and relative's satisfaction, on which DLPII were also
superior to Phase I control patients. It is unclear
which of the various facets special to home-based
care particularly appealed to patients and their
relatives. Satisfaction did not relate closely to
clinical/social outcome. Perhaps it related to the
greater continuity of care for a DLPII patient, or
to appreciation of home visits.
Blurring of distinction between DLPH and
DLP-control care
Several factors reduced differences between the two
groups:
(a) Lingering benefits in DLP-control patients
from previous DLP care in Phase I.
(i) At 30 months the new carers had an unusually
thorough handover and instruction on relevant
issues by DLP staff.
(ii) Some DLP-control
patients insisted on
continuing contact with and support from the
DLP team despite discouragement

from this.

(b) Loss of quality of DLP care during Phase II.


(i) There was loss of continuity between care in
the community and care during in-patient
admissions required by some DLPI! patients.
This followed an audit in the wake of adverse
media publicity (Marks et a!, this issue),
after which the hospital transferred responsi
bility for any crisis in-patient phase from
the DLP to the ward team. This transfer
had already trebled the durations of DLP
admissions towards the end of Phase I and
presumably accounted for much of the
lengthened duration of DLPI! patients'
admissions

during Phase I!.

AUDINI ET AL

210

(ii) Related to (a, i) was the low morale of the DLP


team during Phase II, a point concordant with
the team appearing to take more sick and study
leave than during Phase I. The low morale
reflected (i) the earlier media ordeal, (ii)
uncertainty about funding and the future of
DLP!! patients, and (iii) the loss of key staff,
due to which weekend staffmg was stopped. Low
morale was associated with an attenuation of the
problem-oriented form of work which the DLP
had largely employed in Phase I.
(c) There was more conflict about nursedoctor
ownership'of patients. A DLPII psychiatrist
thought that a DLPII nurse who disliked giving
medication was reluctant to have him see patients
who were on no medication but needed it, in his
opinion. Case management needs to be flexible and
allow for a variety of treatment options.
(d) The workload of the DLP!I team rose when
they took on acute cases outside the study. There

were more absences for sick and study leave, a


smaller number of staff members and a reduction
in their skill mix (no occupational therapist or social
worker at any time). Those team members who had
been in Phase I felt that the reduction of resources
in Phase II resulted in its service becoming less
comprehensive.
Perhaps some of the above problems could have
been diminished by good problem-oriented training
of team members who joined in Phase II. It is also
possible that however good the training, motivation
is difficult to sustain year in and year out with work
of this kind.

Consumer satisfaction apart, the benefits of


good community care for serious mental illness are
not easy to sustain indefinitely.
Acknowledgements
Grateful

thanks

are

extended

to

the

Wolfson

Foundation

for

funding the study's evaluation, the Department of Health for


funding its clinical part, Dr Matt Muijen for facilitating access to
data from Phase I, and Dr P. Mascunan for assistance first with
evaluation and later with treatment.

References
AMERICAN PSYCHIA1RIC AasOCwnON

(1987) Diagnostic

and Statistical

Manual of Mental Disorders (3rd edn, revised) (DSMIIIR).


Washington, DC: APA.
ENDIcorr, J., Sprrz.aa, R. L., Fwss, J. L., et al(l976) The Global
Assessment Scale. A procedure for measuring overall severity of
psychiatric disturbance. Archives of General Psychiatry, 33,
766771.
L4utsui, D. H., ArriusoN, C. C., HAROREAVES,
W. A., ci al (1979)
Assessment of client/patient satisfaction: development of a
general
scale. Evaluation
OVERALL, J. E. & Goa@,

and Program
Planning,
2, 197207.
D. R. (1962) The Brief Psychiatric

Rating Scale. Psychological

Reports,

10, 799812.

Tarr, M. A., Ki'ioimut&,W. H., ALLNESS,


D. J., et al(l99l) Long
term community care through an assertive continuous treatment
team. Advances in Neuropsychiatry, 1, 239-246.
WEISSMAN,M. M., PAYKEL, E. S., Siaovi,

R., et al (1971) The

social role performance of depressed women: comparisons witha


normal group. American Journal of Ckthopsychiaby, 41,390-405.
Kit@,
G. L., PAYKEL,E. S., et al (1974) Treatment
effects on the social adjustment of depressed patients. Archives
of General Psychiatry, 30, 771778.
WINO,

J.

K.,

COOPER,

J.

B.

&

SARTORIUS,

N.

(1974)

The

Measurement and Class@(ication of Psychiatric Symptoms.


Cambridge: Cambridge University Press.

Bernard Audim, Bsc, Research Unit, Royal College of Psychiatrists; Isaac M. Marks, FRCPsych,Institute
of Psychiatry and Bethlem-Maudsley Hospital; Robin Lawrence, MRCPsych,West Lambeth Community
Care Trust, London; Joseph Connolly, FRCPsych,Bethlem-Maudsley Hospital; V. Wafts, MRCPsych,
Maudsley Hospital, London
Correspondence:

Professor

(First received November

I. M. Marks, Institute of Psychiatry,

Dc Crespigny Park, Denmark

1993, final revision March 1994, accepted April 1994)

Hill, London SE5 8AF

Home-based versus out-patient/in-patient care for people with


serious mental illness. Phase II of a controlled study.
B Audini, I M Marks, R E Lawrence, J Connolly and V Watts
BJP 1994, 165:204-210.
Access the most recent version at DOI: 10.1192/bjp.165.2.204

References
Reprints/
permissions
You can respond
to this article at
Downloaded
from

This article cites 0 articles, 0 of which you can access for free at:
http://bjp.rcpsych.org/content/165/2/204#BIBL
To obtain reprints or permission to reproduce material from this paper, please write
to permissions@rcpsych.ac.uk
/letters/submit/bjprcpsych;165/2/204
http://bjp.rcpsych.org/ on December 8, 2016
Published by The Royal College of Psychiatrists

To subscribe to The British Journal of Psychiatry go to:


http://bjp.rcpsych.org/site/subscriptions/

Anda mungkin juga menyukai