The Effect of Family Interventions on Relapseand Rehospitalization in Schizophrenia—
A Meta-analysis
by Qabi Pitschel'Wcdz, Stefan Leucht, Josef Bduml,Werner Kissling, and RolfR. Engel
AbstractTwenty-five intervention studies were meta-analyticallyexamined regarding the effect of including relatives inschizophrenia treatment. The studies investigated fam-ily intervention programs to educate relatives and helpthem cope better with the patient's illness. The patient'srelapse'rate, measured by either a significant worseningof symptoms or rehospitalization in the first years afterhospitalization, served as the main study criterion. Themain result of the meta-analysis was that the relapserate can be reduced by 20 percent if relatives of schizo-phrenia patients are included in the treatment If familyinterventions continued for longer than 3 months, theeffect was particularly marked. Furthermore, differenttypes of comprehensive family interventions have simi-lar results. The bifocal approach, which offers psy-chosocial support to relatives and schizophreniapatients in addition to medical treatment, was clearlysuperior to the medication-only standard treatment.The effects of family interventions and comprehensivepatient interventions were comparable, but the combi-nation did not yield significantly better results than dida treatment approach, which focused on either thepatient or the family. This meta-analysis indicates thatpsychoeducational interventions are essential to schizo-phrenia treatment
Keywords: Schizophrenia, psychoeducation, fam-ily intervention, relapse, meta-analysis.
Schizophrenia Bulletin, 27(l):73-92,2001.
The past decade has witnessed a growing interest in psy-choeducation and family participation in the treatment ofschizophrenia. Because of improved medication treatmentin the past 40 years, more patients can be treated today inan outpatient setting and the majority of the patients staywith their families (Schooler et al. 1995). Caring for aschizophrenia patient is often a burden for families. Abouttwo-thirds of the family caregivers feel considerably bur-dened (Creer and Wing 1974; Hatfield 1978; Fadden et al.
1987; Kuipers 1993; Winefield and Harvey 1993).Relatives involved are experiencing severe emotional andeconomic strain and often suffer from various healthproblems. Families with a member afflicted with such aserious illness need help to cope with this burden andrelated personal stress.
Several surveys indicate that relatives need moreinformation about the disease and how to deal with itmore effectively. The research of Brown and coworkers(1958, 1962, 1972) and the further work of Leff andVaughn on the concept of "expressed emotion" (Vaughn1986) strongly support the importance of psychoeduca-tional work with relatives of schizophrenia patients. Intheory and practice, the approach to caring for relativeshas gradually changed during the past two decades.Relatives of schizophrenia patients are no longer stigma-tized as having caused the illness; rather, they are consid-ered partners in treatment who need the proper tools.Mental health professionals have hoped that well-informed relatives could act as cotherapists (Lefley andJohnson 1990; Boker 1992; Bauml 1993) and might thushelp to improve patients' compliance (Corrigan et al.1990; Kissling 1994).
As a result, various family intervention programswere developed, such as family therapy in a single-familysetting (Esterson et al. 1965; Goldstein et al. 1978;Falloon et al. 1984; Tarrier et al. 1988; Hogarty et al.1991) or in a multifamily setting (McFarlane et al. 1995a,1995£>), psychoeducational relatives' groups (Leff et al.1990; Posner et al. 1992; Bauml et al. 1996), educationallectures for relatives (Smith and Birchwood 1987; Tarrieret al. 1988; Canive et al. 1993), counseling groups for rel-atives (Vaughan et al. 1992; Szmukler et al. 1996;Buchkremer et al. 1997), and group therapy for relatives(Schindler 1958; Kottgen et al. 1984; Lewandowski andBuchkremer 1988). Most of these interventions for rela-tives can be subsumed under the category of "psychoedu-cation" or at least contain psychoeducation as an essential
Send reprint requests to Dr. G. Pitschel-Walz, Klinik und Poliklinikfur Psychiatrie und Psychotherapie der Technischen UniversitatMiinchen, Ismaninger Str. 22, D-81675 Miinchen, Germany; e-mail:[email protected].
73
Schizophrenia Bulletin, Vol. 27, No. 1, 2001 G. Pitschel-Walz et al.
component. "Psychoeducation" is the most common col-lective designation for an intervention that combines theimparting of information with therapeutic elements, andthe term is internationally acknowledged.
Despite the positive results substantiated by numerousintervention studies, the inclusion of relatives in the treat-ment of schizophrenia patients is still not an integral partof routine procedures in many countries. This may resultfrom the initial work involved in establishing such inter-vention programs as well as general skepticism about theeffectiveness of such psychosocial treatment forms.Research results need to be summarized to determine theexact treatment effects achieved through family interven-tions. In previous narrative reviews (Barrowclough andTamer 1984; Strachan 1986; Waring et al. 1986; Tamer1989; Falloon et al. 1990; Smith and Birchwood 1990;Bellack and Mueser 1993; Dixon and Lehman 1995; Pennand Mueser 1996), the quantitative analysis was restrictedto the representation of the various statistical significancesof the individual studies or to counting the studies withsignificant results. Thus, information contained in the pri-mary literature was not fully exploited. Meta-analyticreviews are needed to assess quantitatively the efficacy offamily interventions and to summarize the current state ofscientific knowledge in this field. The first carefully con-ducted meta-analysis, by Mari and Streiner (1994),demonstrated a moderate effectiveness of family interven-tions in schizophrenia. The result was based on six pri-mary studies investigating community-oriented familyinterventions of more than five sessions. Updates of thismeta-analysis (Mari and Streiner 1996; Pharoah et al.1999) containing up to 13 studies confirmed the result thatfamily interventions may decrease the frequency of relapseand rehospitalization and that they may encourage compli-ance (Pharoah et al. 1999). But the authors point out thatfurther data are needed to consolidate these findings.
In our meta-analysis, some more studies (studies pub-lished up to 1997) could be added and the following ques-tions were studied in detail:
• Is treatment that includes relatives superior to theusual care?
• Do the treatment effects differ for followup periodsof different lengths?
• Do characteristics of the studies such as duration ofintervention, type of intervention, study criteria, andstudy sample influence the treatment effects?
• Does the combination of family intervention andpatient intervention produce better results than familyintervention alone?
• Is family intervention superior to psychosocialpatient intervention?
• Which type of family intervention achieves the bestresults?
MethodIdentification and Selection of Studies. A computerizedliterature search was performed using the data baseMedline (1966-December 1997) and the keywords"schizophrenia," "family," "psychoeducation," and"relapse." In addition, reference lists of previous reviewson the subject as well as references of other relevant arti-cles were sources of information. Only English andGerman studies were selected. Number of patients whorelapsed and number of patients who required rehospital-ization were used as outcome parameters. Thus, studiesthat investigated interventions on the relatives of schizo-phrenia patients but focused on other study criteria, suchas knowledge increase of the relatives or reduction ofburden and stress, were not considered. After carefullyscreening about 600 articles, 39 potentially suitable stud-ies were selected.• All these studies were coded with respect to relevant
variables corresponding to the suggestions byChalmers and coworkers (1981). The quality of eachstudy was rated "good," "sufficient," or "insuffi-cient." For this global rating, the quality of the studydesign, the presentation of the results, and the statisti-cal analyses were taken into account. Nonrandomizedstudies were generally considered "insufficient."Studies that received an "insufficient" for qualitywere excluded.
• To verify the reliability of the coding system, werecruited a second coder who was a specialist inmeta-analyses but not family interventions. Differingassessments and uncertainties were discussed andfinal agreement was established. A survey of the cod-ing variables and the final coding of the individualstudies may be obtained from the authors on request.
Data Analyses. The relapse rates or rehospitalizationrates achieved in the studies under the respective treat-ment conditions served as a basis for the present meta-analytical calculations. Several methods for the calcula-tion of effect sizes are available, but they generally do notyield greatly different results (Rosenthal 1991). In thismeta-analysis phi (())) was calculated as an effect size esti-mate according to the formula <|> = VX2/n, which corre-sponds to Pearson's correlation coefficient r applied todichotomous data. This effect size was applied because itis easy to interpret using the binomial effect size displaymethod (BESD: defined as a change from 0.5 - r/2 to 0.5+ r/2), developed by Rosenthal and Rubin (1982). All cal-culations were performed according to the proceduresproposed by Rosenthal (1991), which are acknowledgedand often used in social and medical research (BennettHerbert and Cohen 1993; Butzlaff and Hooley 1998;
74
Family Interventions Schizophrenia Bulletin, Vol. 27, No. 1, 2001
Leucht et al. 1999). The effect sizes were transformed intoFisher's zr values. The mean effect size 7 was then calcu-lated from zr the weighted mean of the zr values. All pooledcalculations included a test of heterogeneity using x2 tests.For the indirect comparison of two mean effect sizes withregard to a hypothesis, contrast weights were calculated.
Results are presented as (mean) effect sizes alongwith their 95 percent confidence interval (CI), with posi-tive values indicating effects in favor of the family inter-vention. Two-tailed significance tests were used, and forthe evaluation of the results a significance level of 5 per-cent was taken as the basis.
Studies With Repeated Measurement Several studiesprovided results for two or more followup periods (table1). Several effect sizes were calculated for each of thesestudies. The first-year effect size refers in each case to thelast assessment within the first year after the treatment;the second-year effect size refers to the assessment periodwithin the second year; and so on. For each year, a meaneffect size was calculated. When calculating the overallmean effect sizes, the mean values of the repeated mea-surements were entered. In the comparison of differentfollowup periods, only the last value of a study was usedto ensure independent measurement.
Treatment of Dropout Patients. Because dropoutpatients could not always be associated with the treatmentgroups involved in the individual comparisons, the figureswere used uniformly following the according-to-protocol(ATP) strategy. Two studies (Spencer et al. 1988; Kellyand Scott 1990) indicated the relapse rates of study com-pleters, which were used as a basis for the calculations. Inthe study of Tamer et al. (1989) the effect sizes for thesecond year were likewise calculated on the basis of thestudy completers only.
Results
Fourteen intervention studies had to be excluded becausetheir quality was rated "insufficient." Four studies (Baumlet al. 1991; McCreadie et al. 1991; Schneider et al. 1991;Boonen and Bockhorn 1992; McCreadie 1992) providedno control group, nine studies lacked a randomizedassignment to one of the treatment conditions (Schindler1958; Esterson et al. 1965; Snyder and Liberman 1981;Kottgen et al. 1984; Ehlert 1988; Lewandowski andBuchkremer 1988; MacCarthy et al. 1989; Scherrmannand Seizer 1989; Scherrmann et al. 1992; Rund et al.1994), and in one study data for group comparisons weremissing (Levene et al. 1989, 1990).
The selected studies enabled us to establish five basiccategories for comparisons.
1. Comparison I: Family intervention vs. usual care. Studiestesting a family intervention against standard treatment areincluded here. As comparison I encompasses a sufficientnumber of studies, contrast analyses could be performed.
2. Comparison II: Family intervention + patient interven-tion vs. usual care. Studies that combined a family inter-vention with a psychosocial patient intervention andtested this intervention package against the standard treat-ment are included in this category.
3. Comparison HI: Family intervention vs. patient interven-tion. Studies that compared family intervention to patient-oriented psychosocial intervention are part of this category.
4. Comparison IV: Family intervention + patient interven-tion vs. patient intervention. Studies that investigated whatadditional effects family intervention has compared to merepsychosocial patient intervention fall within this category.
5. Comparison V: Family intervention A vs. family inter-vention B. Studies that compared two different familyintervention programs appear in this category.
Seven studies were based on more than two interven-tion strategies (i.e., different types of family interventionsor different medication strategies). Because studies withmore than two treatment conditions cannot be reasonablysummarized under one specific hypothesis, these studieswere included several times in the meta-analysis. Eitherpartial samples were assigned to different comparisons(Tarrier et al. 1988, 1989; Kelly and Scott 1990; Hogartyet al. 1991; Hogarty et al. 1991b) or the studies wereincluded twice in a single comparison with two differentpartial samples (Goldstein et al. 1978; Tarrier et al. 1988,1989; McFarlane et al. 1995a; Schooler et al. 1997). Inthe final stage, 25 studies with a total of 40 individualcomparisons were included in the meta-analysis. Someimportant variables of the studies are shown in table 1.
Description of Studies. All of the studies included wereconducted between 1977 and 1997. In the majority of thestudies family interventions were performed in outpatientsettings (62%). The sample sizes of the studies variedfrom n = 23 to n = 418. The average age of the patients inthe individual studies lay between 16 and 40 years.Approximately one-third of the patients included werefemale. The percentage of females ranged from 0 to 65percent. Patients with schizophrenia or schizoaffectivepsychoses were included. Four studies also relied onpatients with different diagnoses according to one of theestablished diagnostic systems. The study patientsincluded those with first episodes and those with multipleepisodes. The average number of hospital stays of the
75
Tabl
e 1
. S
urv
ey o
f th
e in
terv
enti
on
stu
die
s in
clu
ded
Stu
dy
Dia
gn
osi
sF
ollo
wu
p p
erio
dD
ura
tion
of
fam
ilyin
terv
enti
on
Co
mp
aris
on
sn
Buc
hkre
mer
et a
l.19
97S
chiz
ophr
enia
12 a
nd 2
4 m
os8
mo
s
i
Cra
nach
198
1
Fal
loon
et a
l.19
82,1
985
Gol
dste
in e
t al.
1978
Hog
arty
et a
l.19
86,
1991
Sch
izop
hren
ia
Sch
izop
hren
ia
Sch
izop
hren
ia
Sch
izop
hren
ia
12 m
os
9 an
d 24
mos
6 m
os
12 a
nd 2
4 m
os
12 m
os
9 m
os
6 w
ks
24 m
os
Hog
arty
et
al.
1997
i>S
chiz
ophr
enia
12, 2
4, a
nd 3
6 m
os36
mos
Key
per
son
coun
selin
g 68
+ p
sych
oedu
catio
nal
med
icat
ion
man
agem
ent
+ c
ogni
tive
psyc
ho-
ther
apy
for
patie
nts
vs.
usua
l car
e(c
ompa
rison
II)
Pro
foun
d co
unse
ling
of r
elat
ives
+
86co
mpr
ehen
sive
out
patie
nt t
reat
men
tfo
r pa
tient
s vs
. usu
al c
are
(com
paris
on I
I)
Beh
avio
ral f
amily
int
erve
ntio
n vs
. 36
indi
vidu
al s
uppo
rtiv
e ps
ycho
ther
apy
for
patie
nts
(com
paris
on I
II)
Cris
is-o
rient
ed f
amily
the
rapy
vs.
usua
l car
e1.
Hig
h do
sage
(co
mpa
rison
I)
522.
Low
dos
age
(com
paris
on I
) 4
4
Psy
choe
duca
tiona
l fa
mily
tra
inin
g vs
. 50
usua
l car
e (c
ompa
rison
I)
Psy
choe
duca
tiona
l fam
ily t
rain
ing
+
49so
cial
ski
lls tr
aini
ng fo
r pa
tient
svs
. us
ual c
are
(com
paris
on I
I)P
sych
oedu
catio
nal
fam
ily t
rain
ing
vs.
41so
cial
ski
lls tr
aini
ng fo
r pa
tient
s(c
ompa
rison
III)
Psy
choe
duca
tiona
l fa
mily
tra
inin
g +
40
soci
al s
kills
trai
ning
for
patie
nts
vs.
soci
al s
kills
tra
inin
g fo
r pa
tient
s(c
ompa
rison
IV
)
1. P
sych
oedu
catio
nal f
amily
the
rapy
vs.
47
pers
onal
ther
apy
for
patie
nts
(com
paris
on I
II)2.
Psy
choe
duca
tiona
l fam
ily t
hera
py v
s.
48su
ppor
tive
ther
apy
for
patie
nts
(com
paris
on I
II)1.
Psy
choe
duca
tiona
l fa
mily
the
rapy
+
49pe
rson
al t
hera
py f
or p
atie
nts
vs.
pers
onal
ther
apy
for
patie
nts
(com
paris
on I
V)
2. P
sych
oedu
catio
nal f
amily
the
rapy
+
50pe
rson
al t
hera
py f
or p
atie
nts
vs.
i CD O P
Kel
ly a
ndS
cott
1990
McF
arla
ne e
t al.
1995
b
Pos
ner
etal
. 19
92
Sch
izop
hren
iaan
d ot
hers
Sch
izop
hren
ia
Sch
izop
hren
ia
6m
os
2 m
os
Kis
slin
g et
al.
1996
;P
itsch
el-W
alz
eta
l. 19
98
Left
et a
l.,19
82,
1985
Leff
et a
l.19
90
Lins
zen
et a
l.19
96
McF
arla
ne e
t al.
1995
a
Sch
izop
hren
ia
Sch
izop
hren
ia
Sch
izop
hren
ia
Sch
izop
hren
iaan
d ot
hers
Sch
izop
hren
ia
12 a
nd 2
4 m
os
9 an
d 24
mos
9 an
d 24
mos
12 m
os
12
,24
,36
,an
d 48
mos
4 to
5 m
os
9 m
os
9 m
os
12 m
os(1
8 se
ssio
ns)
4yr
s
12 a
nd 2
4 m
os
12 m
os
24 m
os
8 w
ks
supp
ortiv
e th
erap
y fo
r pa
tient
s(c
ompa
rison
IV
)
Psy
choe
duca
tiona
l fam
ily v
isits
at
hom
e 17
9vs
. us
ual c
are
(com
paris
on I)
Psy
choe
duca
tiona
l fa
mily
vis
its a
t ho
me
168
+ p
sych
oedu
catio
n fo
r pa
tient
s vs
. us
ual
care
(co
mpa
rison
II)
Psy
choe
duca
tiona
l fam
ily v
isits
at
hom
e 16
5vs
.psy
choe
duca
tion
for
patie
nts
in th
ecl
inic
(co
mpa
rison
III)
Psy
choe
duca
tiona
l gr
oups
for
163
rela
tives
and
pat
ient
s (b
ifoca
l) vs
.us
ual c
are
(com
paris
on I
I)
Brie
f ed
ucat
iona
l pr
ogra
m +
rel
ativ
es'
24gr
oups
and
fam
ily m
eetin
gs v
s.us
ual c
are
(com
paris
on I)
Psy
choe
duca
tiona
l pr
ogra
m +
fam
ily
23th
erap
y vs
. psy
choe
duca
tiona
l pr
ogra
m+
rel
ativ
es' g
roup
(co
mpa
rison
V)
Psy
choe
duca
tiona
l pr
ogra
m +
psy
cho-
76
soci
al in
terv
entio
n fo
r pa
tient
s +
beh
av-
iora
l fam
ily i
nter
vent
ion
vs. p
sych
o-ed
ucat
iona
l pro
gram
+ p
sych
osoc
ial
inte
rven
tion
for
patie
nts
(com
paris
on I
V)
1. P
sych
oedu
catio
nal
mul
tifam
ily g
roup
vs.
34
psyc
hoed
ucat
iona
l si
ngle
-fam
ily t
reat
men
t(c
ompa
rison
V)
2. P
sych
oedu
catio
nal
mul
tifam
ily g
roup
vs.
23
fam
ily d
ynam
ic m
ultif
amily
gro
up (
com
pari-
son
V)
3. P
sych
oedu
catio
nal
sing
le-f
amily
tre
at-
25m
ent
vs. f
amily
dyn
amic
mul
tifam
ily g
roup
(com
paris
on V
)
Psy
choe
duca
tiona
l m
ultif
amily
gro
up v
s.
172
psyc
hoed
ucat
iona
l si
ngle
-fam
ily t
reat
men
t(c
ompa
rison
V)
Psy
choe
duca
tiona
l su
ppor
t gr
oup
for
rela
- 39
tives
vs.
usu
al c
are
(com
paris
on I)
3 § I S'
o 8
Tabl
e 1
. S
urv
ey o
f th
e in
terv
enti
on
stu
die
s in
clu
ded
—C
on
tin
ued
Stu
dy
Dia
gn
osi
sF
ollo
wu
p p
erio
dD
ura
tion
of
fam
ilyin
terv
enti
on
Co
mp
aris
on
sn! Co £ O
Ran
dolp
h et
al.
1994
Ro-
Tro
ck e
tal
. 19
77
Sch
oole
r et
al.
1997
Sch
izop
hren
ia
Sch
izop
hren
iaan
d ot
hers
Sch
izop
hren
ia
12 m
os
3 m
os
24 m
os
12 m
os
10 s
essi
ons
24 m
os
Spe
ncer
et
al.
1988
Spi
egel
and
Wis
sler
1987
Tar
rier
etal
. 19
88,
1989
Sch
izop
hren
iaan
d ot
hers
Sch
izop
hren
ia
Sch
izop
hren
ia
6 an
d 18
mos
3 an
d 12
mos
9 an
d 24
mos
1-3
mos
6-1
2 w
ks
9 m
os
2 w
ks
Tel
les
etal
. 19
95
Vau
ghan
et
al.
1992
Xio
ng e
tal
. 19
94
Zas
tow
ny e
tal
. 19
92
Zha
ng e
tal
. 19
94
Sch
izop
hren
ia
Sch
izop
hren
ia
Sch
izop
hren
ia
Sch
izop
hren
ia
Sch
izop
hren
ia
12 m
os
9 m
os
12 m
os
18 m
os
12 m
os
10 w
ks
6, 1
2, a
nd 1
8 m
os
18 m
os
16 m
os
18 m
os
Beh
avio
ral f
amily
man
agem
ent
vs. u
sual
41
care
(co
mpa
rison
I)
Fam
ily t
hera
py v
s. in
divi
dual
psy
cho-
28
ther
apy
for
patie
nts
(com
paris
on I
II)
Inte
nsiv
e ap
plie
d fa
mily
man
agem
ent
vs.
supp
ortiv
e fa
mily
man
agem
ent
1. M
oder
ate
dose
(co
mpa
rison
V)
107
2. L
ow d
ose
(com
paris
on V
) 10
63.
Tar
gete
d ea
rly in
terv
entio
n (c
ompa
ri-
100
son
V)
Fam
ily i
nter
vent
ion
with
psy
cho-
16
8ed
ucat
iona
l cha
ract
er v
s. u
sual
car
e(c
ompa
rison
I)
Fam
ily v
isits
(ps
ycho
educ
atio
n, e
lem
ents
36
of f
amily
the
rapy
, cr
isis
inte
rven
tion)
vs.
usua
l car
e (c
ompa
rison
I)
1. F
amily
inte
rven
tion
(ena
ctiv
e or
sym
bolic
) 40
vs.
usua
l car
e (c
ompa
rison
I)2.
Brie
f ps
ycho
educ
atio
nal
prog
ram
vs.
29
usua
l car
e (c
ompa
rison
I)F
amily
inte
rven
tion
vs. b
rief
psyc
ho-
39ed
ucat
iona
l pro
gram
(co
mpa
rison
V)
Beh
avio
ral f
amily
int
erve
ntio
n vs
. 42
case
man
agem
ent
+ s
uppo
rtiv
e ps
ycho
-th
erap
y fo
r th
e pa
tient
(co
mpa
rison
III)
Beh
avio
r- a
nd th
erap
eutic
ally
orie
nted
34
coun
selin
g gr
oup
for
rela
tives
vs.
usua
l car
e (c
ompa
rison
I)
Fam
ily-b
ased
psy
choe
duca
tiona
l 63
inte
rven
tion
vs.
usua
l car
e (c
ompa
rison
I)
Beh
avio
ral f
amily
man
agem
ent
trai
ning
30
vs. s
uppo
rtiv
e fa
mily
man
agem
ent
appr
oach
(co
mpa
rison
V)
Psy
choe
duca
tiona
l fam
ily in
terv
entio
n 78
vs.
usua
l car
e (c
ompa
rison
I)
O I i
Family Interventions Schizophrenia Bulletin, Vol. 27, No. 1, 2001
patients per study was between 1 and 7. The relativesincluded in the treatment were usually the patient's par-ents (range: 50%-100%). In four studies only high EE("Expressed Emotion" status assessed with theCamberwell Family Interview) families were involved.With the exception of the study of Schooler et al. (1997),who tested different medication strategies, and the studyof Goldstein et al. (1978), who varied the dose of theantipsychotic treatment, family and patient interventionswere carried out in addition to the standard medicationtreatment (antipsychotic relapse prevention). The durationof the family interventions ranged from 2 weeks to 4years. Most of the family interventions studied were clas-sified as psychoeducational. The psychosocial patientinterventions in the comparative studies included psy-choeducational interventions, social skills training, sup-portive or cognitive psychotherapy, and personal therapy.Because the selected studies and their intervention pro-grams have already been described in numerous reviews(Mosher and Keith 1980; Barrowclough and Tarrier 1984;
McGill and Lee 1986; Schooler 1986; Waring et al. 1986;Falloon et al. 1990; Smith and Birchwood 1990; Tarrierand Barrowclough 1990; Lam 1991; Bellack and Mueser1993; Dixon and Lehman 1995; Goldstein 1995, 1996a;Penn and Mueser 1996; Schaub and Brenner 1996), amore detailed description of each study has been omittedin this article. The interested reader may refer to the origi-nal literature and the reviews.
Family Intervention vs. Usual Care (Comparison I).The effect sizes from 12 different studies (of 14 individ-ual comparisons) were compared with each other. Theheterogeneity test was not significant. A mean effect size7 = 0.20 (p < 0.0001; CI = 0.14-0.27) was calculated inthe comparison of family intervention and usual care. Atreatment that includes family intervention is clearlysuperior to the usual care of schizophrenia patients.
Table 2 shows the studies included, the relapse ratesfor the first and second years, and the result of the effectsize calculations.
Table 2. Relapse rates and effect sizes (comparison I: family intervention vs. usual care)
Relapse RatesEffect size, Effect size,
Study n^/n2 Family intervention (%) Usual care (%) first year second year
Goldstein et al. 1978:high dosage
Goldstein et al. 1978:low dosage
Hogartyetal. 1991
Kelly and Scott 1990Leffetal. 1985
Posneretal. 1992Randolph etal. 1994Spencer etal. 1988
Spiegel and Wissler 1987Tarrier etal. 1989: long
interventionTarrier etal. 1989: short
interventionVaughan etal. 1992Xiong etal. 1994
Zhang et al. 1994Mean effect sizes
23/29
23/21
21/29
75/10412/12
19/2021/2079/89
77/81
14/2225/15
14/1557/6017/1734/2932/2839/39
0
22
1929359402638Global outcome:0.0214*Global outcome:0.0723*571233436041121315
x2
X2
14
48
38624550784050
= 5.29, ct f=1,p =
= 3.23, df=1,p =
50536053
65363654
0.26
0.27
0.20
0.110.46
0.150.120.18
-0.070.45
0.10
0.240.28
7 =0.19
0.33
0.38
0.14
0.26
0.03
0.270.407 = 0.25
Note.—Mean effect size (overall) r = 0.20; n = 874; p < 0.0001; 95% confidence interval = 0.14-0.27.* The Global Assessment Scale was used; the cutoff score was 60.** Negative effect sizes indicate a result in favor of the usual care treatment condition.
79
Schizophrenia Bulletin, Vol. 27, No. 1, 2001 G. Pitschel-Walz et al.
Different followup periods. The effect sizes wereexamined 6, 9, 12, 18, and 24 months after treatment. Thedifferences between the mean effect sizes independentlymeasured during these five different followup periodswere not statistically significant (7 6 months = 0.16, r 9
months = °-24'J 12 months = °- 0 7 ' ' 18 months = °-24> ~r 24months = 0.26, r total = 0.20; y} = 3.45; p > 0.1). The successrate remained approximately the same at all points in timewithin the first 2 years after treatment as long as relativeswere included in the treatment strategy.
Duration of family interventions. It was investi-gated whether short-term interventions (interventions last-ing less than 3 months) were just as effective as interven-tions of longer duration. The group of short-terminterventions included seven studies (table 1) with a treat-ment duration of 2 to 10 weeks. The group of long-terminterventions encompassed six studies with a treatmentduration of 9 to 24 months. The effect sizes within bothtypes of interventions were homogeneous. As shown intable 3, a mean effect size of 0.14 (CI = 0.06-0.22) wasfound for short-term interventions. For long-term inter-ventions the mean effect size was 0.30 (CI = 0.19-0.41).Both short-term and long-term family interventions inaddition to the usual care were superior to standard med-ical care alone (p < 0.005 and p < 0.0001). However,long-term interventions appeared to be more successful (z= 2.36; p < 0.05) than short-term interventions.
Type of family interventions. We tested whetherthe type of family intervention had any influence on theeffect size. For this analysis, each study was assessed asto whether the family intervention was more psychoed-ucationally or more therapeutically oriented. A meaneffect size of 0.18 (CI = 0.11-0.26) was calculated forpsychoeducational interventions. For therapeutic inter-ventions the mean effect size was 0.23 (CI =0.10-0.36). No matter whether the interventions werepsychoeducationally or therapeutically oriented, signifi-cantly better results were attained through additionalfamily interventions than with the usual standard med-ical care (p < 0.0001 and p < 0.001). Even though themean effect size for therapeutically oriented interven-tions was greater in magnitude, the contrast calculationproved that the type of family intervention did not sig-nificantly influence the effect size (z = 0.61; p > 0.5).
Study criteria. We examined whether different defi-nitions of the study criterion also led to different results.We averaged the effect sizes of studies that used worsen-ing of the patient's symptoms as a criterion for the relapserate. The effect sizes of studies that used rehospitalizationas a criterion were combined separately. In studies withtwo criteria (worsening of symptoms, rehospitalization),the figures for rehospitalization were selected preferen-tially. As listed in table 3, the mean effect sizes were simi-
(A
o
I
5<D
ICD
d)
"5oCO
oO)'•6CD (0Q . CD
8«N CD« £« O
is?sE •O Q-» E•O (0C «CO -oO) C
c «!E «liO o
CD
« E| 8
CO
c/>_ oX CS eg
'•5
Ia.|o(/)
o
1aQ.COOCD
<n
o
2CD
CO
o
I•oIu>a
.ong
oCO
CO O
N °! rin oo i -
m
S ss §A d V
£ gT- ^ °oo
in ^ qd CM dA d V
8CMi n
CO
oO5
0.1
A
en
d
0.0
V
enod
indA
^ CO • O .g] " O CM Oco co in A d V
° 8CM
5 Oq
o CM d<=> "* A d V
23
cp ind d
A
i- 00
CM CM i n
_ CO^ 00
w ri N
oo^ oo 9o ™ oA O V
oo
">. O 9O C O OA d V
co inco qCM d
indA
CD
CDC3> CMO >>
in
T - O
d V
oN05
CD T3N CD
o c co
,8 « £
mea
n r •-C X t) Q.
COCD O O O
o. N a
80
Family Interventions Schizophrenia Bulletin, Vol. 27, No. 1, 2001
lar in magnitude: 0.20 (p < 0.001; CI = 0.09-0.30) and0.19 (p < 0.0001; CI = 0.11-0.28). The calculated z valuewas not significant (z = 0.09; p > 0.5).
Sample selection. In most of the studies the sampleconsisted solely of patients with the diagnosis of schizo-phrenia or schizoaffective psychosis (according toDSM-III, DSM-III-R, Present State Examination, ICD-9,Research Diagnostic Criteria, and other diagnostic sys-tems). In two cases interventions were also offered to rel-atives of patients with other diagnoses. To check whetherthis difference in sample composition had any impact onthe effect size, the mean effect size of all studies based ona sample of only schizophrenia patients was contrastedwith the mean effect size of those studies with mixedsamples. As shown in table 3, a mean effect size of 0.24(p < 0.0001; CI = 0.17-0.29) was calculated for the stud-ies with only schizophrenia patients. The mean effect sizeof studies with mixed samples was 0.12 (p < 0.05; CI =0.02-0.23). Significant effects resulted for both sampletypes. The mean effect size of the studies with schizophre-nia patients only was higher, but contrast calculationsfailed to be statistically significant (z = 1.75; p ^ 0.08).
Family Intervention + Patient Intervention vs. UsualCare (Comparison II). In comparison II the effective-ness of combined interventions for schizophrenia patientsand their relatives was tested and compared to administer-ing only standard medical care. Five studies were part ofcomparison II. Calculations were based on a total of 523patients.
Table 4 lists the studies that fall into this category, therelapse rates for the first and second years, and the calcu-lated effect sizes.
Combining effect sizes. Combining the five calcu-lated effect sizes yielded a mean effect size of 7 = 0.18 (p
< 0.0001; CI = 0.16-0.29). As expected, the bifocalapproach offering support to relatives as well as schizo-phrenia patients turned out to be more successful than thesimple drug treatment of schizophrenia alone.
Independent measurements for the two different fol-lowup periods indicated that the mean effect size for thesecond year (7 = 0.23; p < 0.0001; CI = 0.12-0.33) wasgreater than in the first year (7 = 0.13; p = 0.03; CI =0.04-0.25).
Bifocal vs. unifocal approach. We wanted to knowwhether the combination family intervention + patientintervention yielded better results than family interventionalone (without special psychosocial intervention on behalfof the patient). For this purpose, an indirect comparisonwas carried out by means of contrasts of Fisher's zs. Thedifference between the weighted mean effect size of com-parison I (7 = 0.20) and the weighted mean effect size ofcomparison II (7 = 0.18) proved to be statistically non-significant (z = 0.47; p > 0.5). The effects of combinedinterventions were comparable to those of a family inter-vention not accompanied by a special patient intervention.
Family Intervention vs. Patient Intervention(Comparison III). ' In comparison III the question wasasked whether family interventions might lead to a lowerrelapse rate than interventions solely focusing on theschizophrenia patients and excluding relatives. Six studieswere part of this group and included a total of 407patients. Table 5 lists all studies included; the relapse ratesfor the first, second, and third years; and the calculatedeffect sizes.
Combining effect sizes. The averaging process ofthe effect sizes resulted in a value of 7 = 0.01 (p > 0.5; CI= -0.09-0.11). Thus, the effect of a family interventionseems to be comparable to that of an intervention
Table 4. Relapse rates and effect sizes (comparison II: family intervention + patient intervention vs.usual care)
Study
Buchkremer et al. 1997
Cranach 1981Hogartyetal. 1991
Kelly and Scott 1990Pitschel-Walz et al. 1998
Mean effect sizes
n-f/n2
33/3533/3444/4220/29
64/10481/8279/74
Relapse Rates
Family intervention+ patient intervention
(%)
1524180
25282141
Usual care(%)2250213862453858
Effect size,first year
0.10
0.040.45
0.170.18
7=0.17
Effect size,second year
0.27
0.37
0.187 = 0.23
Note.—Mean effect size (overall) r = 0.18; n = 523; p < 0.0001; 95% confidence interval = 0.09-0.26.
81
Schizophrenia Bulletin, Vol. 27, No. 1, 2001 G. Pitschel-Walz et al.
Table 5. Relapse rates and effect sizes (comparison III: family intervention vs. patient intervention)
Study
Falloon et al.1982, 1985
Hogarty et al.1991
Hogarty et al.1997, 1 "
Hogarty et al.1997,2"
Kelly and Scott1990
Ro-Trock et al.1977
Tellesetal. 1995
Mean effect sizes
nA/n2
18/18
21/20
24/2322/2319/22
24/2422/1819/16
75/90
14/14
n =
42
Relapse Rates
Family
intervention
(%)6
17
1929
334153
334153
35
0
Survival analysis: ̂
0.015*
Patient
intervention
(%)4483
2050
91314
213344
35
43
: 2 =5.91, df=1,p =
7
Effect size,
first year
0.45
0.00
-0.30*
-0.14*
0.00
0.52
-0.38*
= -0.01* 7
Effect size,
second year
0.67
0.22
-0.32*
-0.08*
= 0.12
Effect size,
third year
-0.42*
-0.09*
7 =-0.28*
Note.—Mean effect size (overall) r =0.01; n = 407;p>0.5; 95% confidence interval = -0.09-0.11.
* Negative effect sizes indicate a result in favor of the patient intervention.
** Numbers after authors and date correspond to description of study in table 1.
addressed directly to the patients. When it comes todrawing conclusions one should proceed with caution,because only a small number of studies were conductedin this field and each of the studies was quite different.Averaging the effect sizes of these studies producedstrong heterogeneity (x2 = 35.1; df= 6; p < 0.001). Therewere three studies with a result clearly in favor of thefamily intervention, one study with no differencebetween family and patient interventions, and two stud-ies with a result clearly in favor of the patient interven-tion.
Different followup periods. The mean effect sizesfor the three defined measurement periods were calcu-lated based on independent sampling.Whereas theeffects of the interventions with families and withpatients did not differ at all during the first year (7 =0.00; p = 1.0; CI = -0.13-0.12), a highly significant dif-ference was found during the second year after dis-charge (7 = 0.46; p < 0.0001; CI = 0.26-0.62). For thethird year a significant result in favor of the patientintervention was calculated (7 = -0.28; p < 0.0001; CI= -0.48-0.05).
These results must also be interpreted with cautionbecause the effect sizes are largely heterogeneous in thefirst (x2 = 14.6; df= 2;p< 0.001) and second year (x2 =6.09; df= 1; p < 0.05) and the number of studies avail-able varies across the followup periods. Although fam-ily interventions on behalf of schizophrenia patientsseem to be as successful as patient interventions duringthe first year, the second year after discharge shows adifferent picture. Here it becomes apparent how effec-tive family interventions can be and that such measuresactually reduce the relapse rate among schizophreniapatients significantly. On the other hand the latestHogarty et al. (1991b) study, which provided the onlydata for the third year, demonstrated the superiority ofpersonal therapy, a comprehensive patient intervention,over the family intervention.
Family Intervention + Patient Intervention vs. PatientIntervention (Comparison IV). Hogarty et al. (1991,1991b) and Linszen et al. (1996) investigated additionaleffects of family interventions. All study patients receivedcomprehensive psychosocial treatment, and some of them
82
Family Interventions Schizophrenia Bulletin, Vol. 27, No. 1, 2001
were subject to a family intervention. Table 6 lists allstudies included; the relapse rates for the first, second, andthird years; and the calculated effect sizes.
Combining effect sizes. The mean effect size calcu-lation produced an 7 of 0.04 (p > 0.5; CI = -0.10-0.17).After 1, 2, and 3 years no difference in results wasdetected for the two treatment conditions. Thus, no addi-tional effect of a family intervention could be found, oncondition that the patients received a comprehensive treat-ment.
Famiiy Intervention A vs. Family Intervention B(Comparison V). Several research groups set out toinvestigate which form of family intervention is mosteffective. They conducted comparative studies on thetopic—mostly without involving a control group. Sixstudies could be assigned to comparison V, whichincluded data of 659 patients. Table 7 lists all studiesincluded, the different types of family intervention thatwere tested, and the calculated effect sizes.
Combining effect sizes. Although very different typesof interventions were tested, we tried to combine themunder one single explicit hypothesis. What the studies had incommon was that the authors wanted to investigate whetherthere is a significant difference between family interventionsthat differ in their intensity. Thus, more and less intensiveinterventions were compared. An overall mean effect size of7 = 0.10 (p < 0.01; CI = 0.03-0.18) resulted. This indicatesthat a more intensive family treatment approach is statisti-cally significantly superior to a more limited approach.
Taking a closer look at the individual studies, it can be con-cluded for comparison V that good results were obtainedwith comprehensive interventions of different types.Differences in the effect sizes of family interventions referto the studies testing brief interventions, which made ingeneral a poorer showing, and studies testing the multifam-ily format, which seems to be more successful in the longrun than the single-family approach.
Discussion
This meta-analysis clearly indicates that including the rel-atives in treatment programs is an effective way of reduc-ing relapse and rehospitalization rates in schizophreniapatients. The mean effect size of 0.20 attained in thismeta-analysis appears low in numerical terms but shouldbe considered quite substantial when compared to thequestion posed. We are aware of the difficulty of inter-preting a meta-analytic effect size. For a better under-standing we can use BESD. The 7 of 0.20 would corre-spond to a decrease in relapse rate of 20 percent. Anaverage relapse rate of 60 percent thus can be reduced toabout 40 percent. This definitely is an important findingof clinical relevance. However, some points require amore detailed discussion.
Duration of Family Interventions. The differentiatedanalysis of treatment effects in comparison I revealed thatthe effect size varied depending on the duration of the
Table 6. Relapse rates and effect sizes (comparison IV: family intervention + patient intervention vs.patient intervention)
Study
Hogarty et al.1991Hogarty et al.1997 ,1 "
Hogarty et al.1997,2"
Linszen et al.1996Mean effect sizes
n,/n2
20/20
26/2325/2324/2226/2425/1824/16
37/39
Relapse
Family intervention+ patient intervention
(%)025122838122838
16
Rates
Patientintervention
(%)
205091314213344
15
Effect size,first year
0.33
-0.05*
0.13
-0.01*7 = 0.08
Effect size,second year
0.26
-0.18*
0.06
7 = 0.03
Effect size,third year
-0.27*
0.06
7 =-0.12
Note.—Mean effect size (overall) r = 0.04; n = 215; p>0.5; 95% confidence interval =-0.10-0.17.* Negative effect sizes indicate a result in favor of the patient intervention alone treatment condition." Numbers after authors and date correspond to description of study in table 1.
83
Schizophrenia Bulletin, Vol. 27, No. 1, 2001 G. Pitschel-Walz et al.
Table 7. Studies and effect sizes (comparison V: family intervention A vs. family intervention B)
Study Type of family interventions
Effectsize, first
year
Effectsize,
secondyear
Effectsize, third
year
Effectsize,
fourthyear
Left etal. 1990
McFarlane et al.1995a, 1 "McFarlane et al.1995a, 2 "McFarlane et al.1995a, 3 "McFarlane et al.1995bSchooler et al.1997
Tarrieret al. 1989
Zastowny et al.1992Mean effect sizes
Psychoeducational program + family therapy 0.34 0.03vs. psychoeducational program + relatives'groupPsychoeducational multifamily group vs. 0.13psychoeducational single-family treatmentPsychoeducational multifamily group vs. 0.34family dynamic multifamily groupPsychoeducational single-family treatment vs. 0.21family dynamic multifamily groupPsychoeducational multifamily group vs. 0.15psychoeducational single-family treatmentIntensive applied family management vs.supportive family managementModerate doseLow doseTargeted early interventionFamily intervention vs. brief psycho- 0.35educational programBehavioral family management training vs. -O.07*supportive family management approach
7 = 0.18 7=0.10
0.20
0.18
-0.01*
0.15
0.13-0.01*
0.060.23
0.24
0.05
-0.17*
0.35
0.12
-0.21
7 = 0.07 7=0.12
Note.—Mean effect size (overall) r = 0.10; n = 659; p < 0.01; 95% confidence interval = 0.03-0.18.* Negative effect size indicates a result in favor of the family intervention approach that was considered more limited by the authors." Numbers after authors and date correspond to description of study in table 1.
interventions. The success of family interventions wasparticularly evident when these interventions lasted longerthan 3 months. Not only in this contrast analysis but alsoin comparison V of this meta-analysis, where variousforms of family interventions were directly compared,brief interventions proved inferior. It was clearly demon-strated that a few lessons on schizophrenia for relativeswas simply not sufficient to substantially influence therelapse rate. These findings are in line with the conclu-sions drawn by Fiedler et al. (1986), Tarrier andBarrowclough (1990), and Mari and Streiner (1994), andthey confirm Hogarty's criticism of the Videka-Shermanmeta-analysis (Hogarty 1989). The author of the meta-analysis (Videka-Sherman 1988) had concluded that briefpsychosocial interventions including family interventionsare more effective than long-term interventions. Hogarty(1989) argued that low-quality studies might have undulyinfluenced this result and pointed out that data from betterdesigned studies lead to the opposite conclusion.
Type of Family Interventions. In comparison I it wasindirectly examined whether psychoeducational interven-tions are just as effective as intensive therapeutic inter-ventions. When grouping effect sizes according to thetype of intervention no statistically significant difference
was found. Closer inspection revealed that the effectswere confounded with those of the variable duration. Thesomewhat lower mean effect size for psychoeducationalinterventions particularly derived from studies that per-formed a brief psychoeducational intervention, whereaslong-term psychoeducational interventions such as thoseof Hogarty et al. (1991) or Leff et al. (1985) yielded largeeffect sizes. On the whole the long-term participation ofrelatives in the treatment of schizophrenia patients provedvery successful, with the type of intervention being rathersecondary. An explanation of this finding could be thatgeneral therapeutic factors as defined for example byYalom (1984) for group psychotherapy may also becomeeffective in psychoeducational interventions over time.
The results of the contrast analyses (comparison I)were confirmed by comparison V, where various forms offamily interventions were directly compared. Taking thestudies together it turned out that there are significant dif-ferences in the effectiveness of family interventions.Again, long-term interventions appeared to be more effec-tive. Under the condition that they are long-term interven-tions, the psychotherapeutic intensity or orientation seemsto be secondary. Comparison V adds another interestingfinding concerning the format of family interventions. Inthe two McFarlane et al. studies (1995a, 1995fc) the multi-
84
Family Interventions Schizophrenia Bulletin, Vol. 27, No. 1, 2001
family approach yielded better results than the treatmentin a single-family setting. However, conclusive results onthis issue can be derived only from future studies.
Methodological Variables of Family InterventionStudies. The impact of methodological variables of thestudies on the effect sizes was also investigated in thismeta-analysis (comparison I). Criteria were selected suchas followup period, study criterion, and choice of sample(which had been discussed as influencing variables in therelevant literature) to analyze their specific influence onthe effect sizes in intervention studies in more detail.Considering followup period, several authors hypothe-sized that the effectiveness of family interventions wouldbe visible more clearly during later measurements, forexample, during the second year after the interventiontook place. This assumption was based on the results ofthe Hogarty et al. (1991) and Falloon et al. (1985) studies.The meta-analysis revealed that the treatment effects didnot differ significantly for followup periods of differentlengths. The question of whether the effect size remainsthe same for followups beyond the 24-month followupcannot yet be answered adequately. Hogarty et al. (1991)found some evidence that family intervention effects donot persist long after the treatment ends. On the otherhand, two recent studies (Bauml et al. 1997; Hornung etal. 1999) provided positive results concerning the long-term effects achieved through bifocal interventions. Therehospitalization rates in both studies increased with thenumber of years, but the difference between the interven-tion group and the control group remained constant. The5- and 8-year followup data of Tarrier et al. (1994) alsosupport the positive long-term effects of family interven-tions. Thus, psychoeducational family interventions mayhave a positive effect not only during the treatment andshortly afterward but also months and years thereafter. Toget a definite answer to this question, more long-termstudies are needed. In addition, further studies shouldinvestigate whether regular booster sessions can help tostabilize the effects of psychoeducational family interven-tions over a longer period of time.
The impact of another criterion, the study criterion,was investigated because one group of intervention stud-ies classified the worsening of symptoms as a relapse,whereas another group used rehospitalization as an essen-tial outcome criterion. The literature supports both vari-ants. However, rehospitalization has become a dominantcriterion in today's research, as it is easy and clear to mea-sure. The effect size did not vary in dependence on thestudy criterion. This means that family interventions didnot only improve the competence of relatives in dealingwith the patients, implying that fewer rehospitalizationsbecame necessary due to acute crises, but did actuallydecrease the occurrence of acute crises considerably.
Separate averaging of the effect sizes for differentsamples (schizophrenia/mixed diagnoses) indicated thatthe mean effect size tended to be higher for homogeneousschizophrenia samples than for diagnostically heteroge-neous samples. It definitely makes sense in all cases ofpsychiatric illnesses to include the patient's relatives inthe treatment process, to provide them with informationabout the illness and explain specific psychosocial aids tothem. Because patients suffering from schizophrenia areparticularly dependent on the support of their families, itis important to enhance the involvement of relatives andto test its effectiveness. It can be concluded from theresults of this meta-analysis that family interventions areparticularly effective when they are tailored directly to theneeds of this special target group. If such interventions areoffered only to relatives of schizophrenia patients, thera-pists can go into detail on their specific problems moreintensively. Participants might benefit better from thistype of intervention than from interventions that cover theentire spectrum of questions and problems concerningvarious psychiatric diseases in the same number of ses-sions.
It is also important to remember that the averagerelapse rate observed during the first followup year mightbe lower for patients with diagnoses other than schizo-phrenia. Therefore, the actual margin for a reduction inthis rate was less to start with.
The Impact of Family Interventions in Comparison toPsychosocial Patient Interventions. Even though somestudies suggest an additive effect of family interventionand patient intervention (Kelly and Scott 1990; Hogarty etal. 1991; Bauml et al. 1997; Buchkremer et al. 1997), theindirect comparison (comparison II) did not prove that.Bifocal intervention approaches did not produce a signifi-cantly greater effect size than family interventions alone.This does not mean that a psychosocial interventionfocused on the schizophrenia patient alone has no positiveeffect on the relapse rate. When effect sizes were com-bined in comparison III, the results of patient interven-tions were comparable to the results of family interven-tions. The effect sizes of the individual studies, however,were very heterogeneous. Three studies showed the supe-riority of family interventions, one study found no differ-ence, and two studies brought better results under thepatient intervention condition. The reason for this hetero-geneity may be that there are differences in the effective-ness of the individual patient intervention approaches.Whereas in a previous meta-analysis on social skills train-ing only moderate effects on the relapse rate were found(Benton and Schroeder 1990), the meta-analysis ofWunderlich et al. (1996) on the effectiveness of psychoso-cial interventions for schizophrenia patients revealedequally large effect sizes for family therapy and cognitive
85
Schizophrenia Bulletin, Vol. 27, No. 1, 2001 G. Pitschel-Walz et al.
therapy for patients. The positive results of the latestHogarty et al. study (1997a, 1991b), which was includedin our meta-analysis, could indicate a particularly goodeffectiveness of personal therapy, although it should bementioned that the positive effects of personal therapy onrelapse rates were found only among patients who livedwith family.
Furthermore, in comparison IV the additional effect offamily interventions for schizophrenia patients whoreceived a comprehensive psychosocial treatment wasinvestigated. The relapse rate could not be lowered throughan additional family intervention. In the Linszen et al. 1996study all patients received outstanding medical and psy-chosocial treatment during the entire study period. It wouldhave therefore been difficult to further improve the out-come through additional family intervention. The psychoe-ducational sessions offered at the beginning of the interven-tion under both treatment conditions probably leveled thedifferences out. For some families, characterized as low-EEin several studies, it might be sufficient to get basic psy-choeducation to support the patient adequately. Hogarty etal. showed in their 1991 study the superiority of the com-bined treatment over an intervention approach focused onthe patient alone, whereas in their 1997Z? study personaltherapy with the patient alone brought better long-termresults than the combination with a family intervention. Asthe studies with very positive results for patient interven-tions are recent studies, it might be imaginable that the psy-chosocial treatment of schizophrenia patients has beenimproved considerably in recent years. The meta-analysisof Mojtabai et al. (1998) on psychosocial treatments forpatients with schizophrenia would support this assumption,showing that more recent studies tended to produce largereffect sizes. Furthermore, psychoeducation has graduallybecome a basic treatment for all schizophrenia patients andtheir relatives and part of the usual care, at least in theUnited States. It might therefore get more and more diffi-cult to produce a significant effect through an additionalfamily treatment.
Limitations. Despite the clearly positive results and newperspectives for the future, some limitations should bekept in mind. First, our meta-analytical calculations werenot based on intention to treat (ITT). Because dropoutpatients could not always be associated with the treatmentgroups involved in the individual comparisons, ATPanalyses were presented. Most of the authors noted thatdropout rates in intervention and control groups were notsignificantly different. Therefore, a systematic errorcaused by patients dropping out of the study groups wasnot to be expected. In the Kelly and Scott (1990) study,which had higher dropout rates, especially in the controlgroup (family intervention: 29%, patient intervention:
21%, combination: 30%, usual care: 42%), a small effectsize was calculated. In the Hogarty et al. (1997&) studyalso a particularly high dropout rate was found for thesupportive therapy condition (family therapy: 21%, per-sonal therapy: 4%, supportive therapy: 33%) and a smalleffect size was calculated as well. Thus, an overestimationof the family intervention as a result of the evaluationstrategy cannot be assumed. It appeared to be appropriateto use the ATP strategy also considering the content of themeta-analysis, because it should answer the basic questionabout the additional effect achieved by family interven-tions. Therefore, the relatives who were randomlyassigned to the intervention groups should have gotten theintervention in fact. Other important questions concerningthe acceptance among patients, relatives, and profession-als or the implementation in routine care should be inves-tigated independently. Nevertheless the possibility that anevaluation according to the ITT strategy would yield find-ings of less significance cannot be ruled out.
Secondly, unpublished studies remained outside thescope of this meta-analysis. As negative results have ahigher tendency to remain unpublished, a bias cannot beexcluded, even if there are few indications of unpublishedstudies concerning the topics treated here.
A third general limitation derives from the fact thatthe reported findings relate to only schizophrenia patients(1) who have relatives, (2) who are in regular contact withtheir relatives, and (3) who have relatives willing tobecome involved on behalf of the schizophrenia patient.At least one-third of all schizophrenia patients (Hogarty1993; Pitschel-Walz 1997) do not meet these basic prereq-uisites and, therefore, simply cannot be supported throughpsychoeducational family interventions. Other treatmentstrategies must be developed for these patients in theoryand practice to actively support them in dealing effec-tively with their illness. New developments are encourag-ing, showing that schizophrenia patients can absolutelybenefit from psychosocial treatments when they are tai-lored to the patients' special needs (Hogarty et al. 1997a;Mojtabai et al. 1998; Schaub 1998). In addition, hopes areraised by the introduction of the new atypical antipsy-chotic drugs, which have a more favorable side effect pro-file and may improve the patient's compliance and con-tribute to further reduction of the relapse rate (Weiden etal. 1996; Leucht et al. 1999).
Conclusions
This meta-analysis could clearly confirm the hypothesisthat psychoeducational family interventions reduce therelapse and rehospitalization rates of schizophreniapatients. In the coming decade the results of this researchneed to be applied in practical settings to enable as many
86
Family Interventions Schizophrenia Bulletin, Vol. 27, No. 1, 2001
patients as possible to benefit from these findings.Psychoeducation for patients and their families shouldbecome a basic part of a comprehensive psychosocialtreatment package that is offered to all schizophreniapatients. Future research should focus on and evaluate theprocess of integrating family interventions into the clinicalroutine and examine the long-term effects in more detail.
In this meta-analysis we focused on the relapse rateand the rehospitalization rate as outcome criteria.However, there are further important effects of familyinterventions that were not investigated here but should beconsidered when the implementation of family interven-tions is discussed. In addition to a reduced relapse rate,several studies demonstrated the following:
• a reduction of family burden (Pakenham and Dadds1987; Abramowitz and Coursey 1989; Birchwood etal. 1992),
• a change from high EE to low EE behavior (Snyderand Liberman 1981; Leff et al. 1982; Kottgen et al.1984; Hogarty et al. 1986; Tarrier et al. 1988;MacCarthy et al. 1989; McCreadie et al. 1991;Vaughan et al. 1992; Zastowny et al. 1992; Randolphet al. 1994; Rund et al. 1994),
• an improvement of knowledge about schizophrenia(Snyder and Liberman 1981; McGill et al. 1983;Berkowitz et al. 1984; Barrowclough et al. 1987;Pakenham and Dadds 1987; Smith and Birchwood1987; Cozolino et al. 1988; Cazzullo et al. 1989;Birchwood et al. 1992; Fowler 1992; Posner et al.1992; Zastowny et al. 1992; Canive et al. 1993;Bauml et al. 1996; Pitschel-Walz 1997),
• better compliance (Falloon et al. 1982; Kelly andScott 1990; Goldstein 1994; Mari and Streiner 1994;Xiang et al. 1994; Zhang et al. 1994; McFarlane et al.19956; Bauml et al. 1996),
• better patient social adjustment (Cranach 1981;Falloon et al. 1987; Spiegel and Wissler 1987;Levene et al. 1989; Barrowclough and Tarrier 1990;Hogarty et al. 1991; Vaughan et al. 1992; Xiang et al.1994; Hornung et al. 1995; Bauml et al. 1996),
• improved quality of life (Zastowny et al. 1992), and• reduced costs for society (Falloon et al. 1984; Cardin
et al. 1985; Liberman et al. 1987; Tarrier et al. 1991;Kissling 1993; Lindgens 1993; Mari and Streiner1994; Rund et al. 1994; Xiong et al. 1994; McFarlaneet al. 1995; Goldstein 1996fc)
Additional meta-analyses should be carried out in thefuture using these other outcome measures.
The overall goal of all scientific efforts finallyremains the effective improvement of long-term outcomesfor patients with schizophrenia based on antipsychoticdrug therapy; psychotherapy; family intervention meth-ods; and new, individualized concepts of psychiatric reha-bilitation.
References
Abramowitz, I.A., and Coursey, R.D. Impact of an educa-tional support group on family participants who take careof their schizophrenic relatives. Journal of Consulting andClinical Psychology, 57(2):232-236, 1989.
Barrowclough, C , and Tarrier, N. Psychosocial interven-tions with families and their effects on the course ofschizophrenia: A review. Psychological Medicine,14:629-642, 1984.
Barrowclough, C , and Tarrier, N. Social functioning inschizophrenic patients. Social Psychiatry and PsychiatricEpidemiology, 25:125-129, 1990.
Barrowclough, C ; Tarrier, N.; Watts, S.; Vaughn, C ;Bamrah, J.S.; and Freeman, H.L. Assessing the functionalvalue of relatives' knowledge about schizophrenia: A prelim-inary report. British Journal of Psychiatry, 151:1-8,1987.
Bauml, J. Der Umgang mit schizophrenen Patienten ausder Sicht der Angehorigen. In: Kockott, G., and Mb'ller,H.-J., eds. Sichtweisen der Psychiatric Miinchen,Germany: W. Zuckschwerdt Verlag, 1993. pp. 42-52.
Bauml, J.; Kissling, W.; Meurer, C ; Wais, A.; and Lauter,H. Informationszentrierte Angehorigengruppen zurComplianceverbesserung bei schizophrenen Patienten.Psychiatrische Praxis, 18:48-54, 1991.
Bauml, J.; Pitschel-Walz, G.; and Kissling, W.Psychoedukative Gruppen bei schizophrenen Psychosenfur Patienten und Angehorige. In: Stark, A., ed.Verhaltenstherapeutische und psychoedukative Ansatze imUmgang mit schizophren Erkrankten. Tubingen,Germany: dgvt-Verlag, 1996. pp. 217-255.
Bauml, J.; Pitschel-Walz, G.; and Kissl ing, W.Psychoedukative Gruppenarbeit bei schizophrenenPsychosen: Spezifische Auswirkungen eines bifokalenAnsatzes auf Krankheitsbewaltigung und Rezidivratenim 4-Jahreszeitraum. Ergebnisse der Miinchner PIP-Studie. In: Dittmar, V; Klein, H.E.; and Schon, D.,eds. Die Behandlung schizophrener Menschen—Integrative Therapiemodelle und ihre Wirksamkeit.Regensburg, Germany: S. Roderer Verlag, 1997. pp.169-195.
Bellack, A.S., and Mueser, K.T. Psychosocial treatmentfor schizophrenia. Schizophrenia Bulletin, 19(2):317—336,1993.
Bennett Herbert, T., and Cohen, S. Depression and immu-nity: A meta-analytic review. Psychological Bulletin,113(3):472-486, 1993.
Benton, M.K., and Schroeder, H.E. Social skill trainingwith schizophrenics: A meta-analytic evaluation. Journalof Consulting and Clinical Psychology, 58(6):741-747,1990.
87
Schizophrenia Bulletin, Vol. 27, No. 1, 2001 G. Pitschel-Walz et al.
Berkowitz, R.; Eberlein-Fries, R.; Kuipers, L.; and Leff,J.P. Educating relatives about schizophrenia.Schizophrenia Bulletin, 10(3):418-429, 1984.
Birchwood, M.J.; Smith, J.V.; and Cochrane, R. Specificand nonspecific effects of educational intervention forfamilies living with schizophrenia: A comparison of threemethods. British Journal of Psychiatry, 160:806-814,1992.
Boker, W. A call for partnership between schizophrenicpatients, relatives and professionals. British Journal ofPsychiatry, 161(Suppl 18): 10-12, 1992.
Boonen, M., and Bockhorn, M. Schizophreniebehandlungin der Familie. Psychiatrische Praxis, 19:76-80,1992.
Brown, G.W.; Birley, J.L.T.; and Wing, J.K. Influence offamily life on the course of schizophrenic disorders: Areplication. British Journal of Psychiatry, 121:214-258,1972.
Brown, G.W.; Carstairs, G.M.; and Topping, G. Post-hos-pital adjustment of chronic mental patients. The Lancet,2:685-689, 1958.
Brown, G.W.; Monck, E.M.; Carstairs, G.M.; and Wing,J.K. Influence of family life on the course of schizo-phrenic illness. British Journal of Prevention and SocialMedicine, 16:55-68, 1962.
Buchkremer, G.; Klingberg, S.; Holle, R.; Schulze-Monking, H.; and Hornung, P. Psychoeducational psy-chotherapy for schizophrenic patients and their key rela-tives or care-givers: Results of a 2-year follow-up. AdaPsychiatrica Scandinavica, 96:483-491, 1997.
Butzlaff, R.L., and Hooley, J.M. Expressed emotion andpsychiatric relapse—A meta-analysis. Archives of GeneralPsychiatry, 55:547-552, 1998.
Canive, J.M.; Sanz-Fuentenebro, J.; Tuason, V.B.;Vazquez, C ; Schrader, R.M.; Alberdi, J.; andFuentenebro, F. Psychoeducation in Spain. Hospital andCommunity Psychiatry, 44(7):679-681, 1993.
Cardin, V.A.; McGill, C.W.; and Falloon, I.R.H. An eco-nomic analysis: Costs, benefits and effectiveness. In:Falloon, I.R.H., ed. Family Management ofSchizophrenia. Baltimore, MD: Johns Hopkins UniversityPress, 1985. pp. 115-123.
Cazzullo, C.L.; Bertrando, P.; Clerici, C ; Bressi, C ; DaPonte, C ; and Albertini, E. The efficacy of an informationgroup intervention on relatives of schizophrenics.International Journal of Social Psychiatry,35(4):313-323, 1989.
Chalmers, T.C.; Smith, H.; Blackburn, B.; Silverman, B.;Schroeder, B.; Reitman, D.; and Ambroz, A. A method forassessing the quality of a randomized control trial.Controlled Clinical Trials, 2:31-49, 1981.
Corrigan, P.W.; Liberman, R.P.; and Engel, J.D. Fromnoncompliance to collaboration in the treatment of schiz-ophrenia. Hospital and Community Psychiatry,41(ll):1203-1211, 1990.
Cozolino, L.J.; Goldstein, M.J.; Nuechterlein, K.H.; West,L.W.; and Snyder, K. The impact of education aboutschizophrenia on relatives varying in expressed emotion.Schizophrenia Bulletin, 14(4):675-687, 1988.
Cranach, M.V. Psychiatrische Versorgung durch niederge-lassene Arzte und ambulante Dienste. In: Bauer, M, andRose, H.K., eds. Ambulante Dienste fur psychisch Kranke.Koln, Germany: Rheinland-Verlag, 1981. pp. 31-41.
Creer, C , and Wing, J.K. Schizophrenia at Home. Surrey,U.K.: National Schizophrenia Fellowship, 1974.
Dixon, L.B., and Lehman, A.F. Family interventions forschizophrenia. Schizophrenia Bulletin, 21(4):631-643,1995.
Ehlert, U. Psychosocial intervention for the relatives ofschizophrenic patients. In: Emmelkamp, P.; Everaed, W.;Kraaimaat, F.; and van Son, M., eds. Advances in Theoryand Practice in Behaviour Therapy. Amsterdam, TheNetherlands: Swets and Zeitlinger, 1988. S. 251-259.
Esterson, A.; Cooper, D.G.; and Laing, R.D. Results offamily-oriented therapy with hospitalized schizophrenics.British Medical Journal, 2:1462-1465, 1965.
Fadden, G.; Bebbington, P.; and Kuipers, L. The burdenof care: The impact of functional psychiatric illness on thepatient's family. British Journal of Psychiatry,150:285-292, 1987.
Falloon, I.R.H.; Boyd, J.L.; and McGill, C.W. Family Careof Schizophrenia. New York, NY: Guilford Press, 1984.
Falloon, I.R.H.; Boyd, J.L.; McGill, C.W.; Razani, J.;Moss, H.B.; and Gilderman, A.M. Family managementin the prevention of exacerbations of schizophrenia.New England Journal of Medicine, 306:1437-1440,1982.
Falloon, I.R.H.; Boyd, J.L.; McGill, C.W.; Williamson,M.; Razani, J.; Moss, H.B.; Gilderman, A.M.; andSimpson, G.M. Family management in the prevention ofmorbidity of schizophrenia. Archives of GeneralPsychiatry, 42:887-896, 1985.
Falloon, I.R.H.; Hahlweg, K.; and Tarrier, N. Familyinterventions in the community management of schizo-phrenia: Methods and results. In: Straube, E.R., andHahlweg, K., eds. Schizophrenia: Concepts, Vulnerability,and Intervention. Berlin, Germany: Springer-Verlag,1990. pp. 217-240.
Falloon, I.R.H.; McGill , C.W.; Boyd, J.L.; andPederson, J. Family management in the prevention ofmorbidity of schizophrenia: Social outcome of a two-year longitudinal study. Psychological Medicine,17:59-66, 1987.
88
Family Interventions Schizophrenia Bulletin, Vol. 27, No. 1, 2001
Fiedler, P.; Niedermeier, T; and Mundt, C. TherapeutischeGruppenarbeit mit Angehorigen und Familien schizo-phrener Patienten—Eine kritische Ubersicht.Gruppendynamik, 17(3):255-272, 1986.
Fowler, L. Family psychoeducation: Chronic psychiatri-cally ill Carribbean patients. Journal of PsychosocialNursing and Mental Health Services, 30(3):27-32, 1992.
Goldstein, M.J. Psychoeducational and family therapy inrelapse prevention. Ada Psychiatrica Scandinavica,89(Suppl 382):54-57, 1994.
Goldstein, M.J. Psychoeducation and relapse prevention.International Clinical Psycho pharmacology, 9(Suppl5):59-69, 1995.
Goldstein, M.J. Psychoeducation and family treatmentrelated to the phase of a psychotic disorder. InternationalClinical Psychopharmacology, l l (Suppl 2):77-83,1996a.
Goldstein, M.J. Psychoeducational family programs in theUnited States. In: Moscrelly, M.; Rupp, A.; and Sartorius,N.; eds. Handbook of Mental Health Economics, I.Schizophrenia. Chichester, U.K.: J. Wiley and Sons,\996b.
Goldstein, M.J.; Rodnick, E.H.; Evans, J.R.; May, R.P.A.;and Steinberg, M.R. Drug and family therapy in the after-care of acute schizophrenics. Archives of GeneralPsychiatry, 35:1169-1177, 1978.
Hatfield, A.B. Psychological costs of schizophrenia to thefamily. Social Work, 23:355-359, 1978.
Hogarty, G.E. Metaanalysis of the effects of practice withthe chronically mentally ill: A critique and reappraisal ofthe literature. Social Work, 34:363-373, 1989.
Hogarty, G.E. Prevention of relapse in chronic schizo-phrenic patients. Journal of Clinical Psychiatry,54(Suppl): 18-23, 1993.
Hogarty, G.E.; Anderson, CM.; Reiss, D.J.; Kornblith, S.J.;Greenwald, D.P.; Javan, CD.; and Madonia, M.J. Familypsychoeducation, social skills training, and maintenancechemotherapy in the aftercare treatment of schizophrenia.Archives of General Psychiatry, 43:633-642,1986.
Hogarty, G.E.; Anderson, CM.; Reiss, D.J.; Kornblith,S.J.; Greenwald, D.P.; Ulrich, R.F.; and Carter, M.Family psychoeducation, social skills training, andmaintenance chemotherapy in the aftercare treatment ofschizophrenia. Archives of General Psychiatry,48:340-347, 1991.
Hogarty, G.E.; Greenwald, D.P.; Ulrich, R.F.; Kornblith,S.J.; DiBarry, A.L.; Cooley, S.; Carter, M.; and Flesher, S.Three-year trials of personal therapy among schizophrenicpatients living with or independent of family: II. Effectson adjustment of patients. American Journal ofPsychiatry, 154(11): 1514-1524, 1997a.
Hogarty, G.E.; Kornblith, S.J.; Greenwald, D.P.; DiBarry,A.L.; Cooley, S.; Ulrich, R.F.; Carter, M.; and Flesher, S.Three-year trials of personal therapy among schizophrenicpatients living with or independent of family: I.Description of study and effects on relapse rates.American Journal of Psychiatry, 154(11): 1504—1513,1997*.
Hornung, W.P.; Feldmann, R.; Klingberg, S.; Buchkremer,G.; and Reker, T. Long-term effects of a psychoeduca-tional psychotherapeutic intervention for schizophrenicoutpatients and their key-persons—results of a five-yearfollow-up. European Archives of Psychiatry and ClinicalNeuroscience, 249(3): 162-167, 1999.
Hornung, W.P.; Holle, R.; Schulze-Monking, H.;Klingberg, S.; and Buchkremer, G. Psychoedukativ-psy-chotherapeutische Behandlung von schizophrenenPatienten und ihren Bezugspersonen. Nervenarzt,66:828-834, 1995.
Kelly, G.R., and Scott, J.E. Medication compliance andhealth education among outpatients with chronic mentaldisorders. Medical Care, 28(12):1181-1197, 1990.
Kissling, W. Schizophrenic: Jedes zweite Rezidiv warevermeidbar. Fortschritte der Medizin, ll(Suppl 147):3-5,1993.
Kissling, W. Compliance, quality assurance and standardsfor relapse prevention in schizophrenia. Acta PsychiatricaScandinavica, 89(Suppl 382): 16-24, 1994.
Kissling, W.; Bauml, J.; and Pitschel-Walz, G.Psychoedukation und Compliance bei der Schizophre-niebehandlung. In: Helmchen, H., and Hippius, H., eds.Psychiatrie fur die Praxis 23. Miinchen, Germany: MMVMedizin Verlag, 1996. pp. 14-23.
Kottgen, C ; Sonnichsen, I.; Mollenhauer, K.; and Jurth,R. Group therapy with the families of schizophrenicpatients: Results of the Hamburg Camberwell-family-interview study III. Journal of Family Psychiatry,5:84-94, 1984.
Kuipers, L. Family burden in schizophrenia: Implicationsfor services. Social Psychiatry and PsychiatricEpidemiology, 28(5):207-210, 1993.
Lam, D.H. Psychosocial family intervention in schizo-phrenia: A review of empirical studies. PsychologicalMedicine, 21:423-441, 1991.
Leff, J.; Berkowitz, R.; Shavit, N.; Strachan, A.M.; Glass,I.; and Vaughn, C. A trial of family therapy versus a rela-tives' group for schizophrenia. British Journal ofPsychiatry, 157:571-577,1990.
Leff, J.; Kuipers, L.; Berkowitz, R.; Eberlein-Vries, R.;and Sturgeon, D. A controlled trial of social interventionin the families of schizophrenic patients. British Journalof Psychiatry, 141:121-134, 1982.
89
Schizophrenia Bulletin, Vol. 27, No. 1, 2001 G. Pitschel-Walz et al.
Leff, J.; Kuipers, L.; Berkowitz, R.; and Sturgeon, D. Acontrolled trial of social intervention in the families ofschizophrenic patients: Two year follow-up. BritishJournal of Psychiatry, 146:594-600, 1985.
Lefley, H.P., and Johnson, D.L., eds. Families as Allies inTreatment of the Mentally III: New Directions for MentalHealth Professionals. Washington, DC: AmericanPsychiatric Press, 1990.
Leucht, S.; Pitschel-Walz, G.; Abraham, D.; and Kissling,W. Efficacy and extrapyramidal side effects of the newantipsychotics olanzapine, quetiapine, risperidone, andsertindole compared to conventional antipsychotics andplacebo: A meta-analysis of randomized controlled trials.Schizophrenia Research, 35:51-68, 1999.
Levene, J.E.; Newman, F.; and Jeffries, J.J. Focal familytherapy outcome study: I. Patient and family functioning.Canadian Journal of Psychiatry, 34(7):641-647, 1989.
Levene, J.E.; Newman, F.; and Jeffries, J.J. Focal familytherapy: Theory and practice. Family Process,29(l):73-86, 1990.
Lewandowski, L., and Buchkremer, G. TherapeutischeGruppenarbeit mit Angehorigen schizophrener Patienten—Ergebnisse zweijahriger Verlaufsuntersuchungen. Zeitschriftfur Klinische Psychologie, 17:2,1988.
Liberman, R.P.; Cardin, V.A.; McGill, C.W.; Falloon,I.R.H.; and Evans, CD. Behavioral family managementof schizophrenia: Clinical outcome and costs. PsychiatricAnnals, 17(9):610-619, 1987.
Lindgens, M. Ambulante Familienintervention zahlt sichaus: Kosten durch Prophylaxe minimiert. Fortschritte derMedizin, ll(Suppl 147):7-9, 1993.
Linszen, D.; Dingemans, P.; Van der Does, J.W.; Nugter,A.; Scholte, P.; Lenior, R.; and Goldstein, M.J. Treatment,expressed emotion and relapse in recent onset schizo-phrenic disorders. Psychological Medicine, 26:333-342,1996.
MacCarthy, B.; Kuipers, L.; Hurry, J.; Harper, R.; andLeSage, A. Counselling the relatives of the long-termadult mentally ill. British Journal of Psychiatry,154:768-775, 1989.
Mari, J.J., and Streiner, D.L. An overview of family inter-ventions and relapse on schizophrenia: Meta-analysis ofresearch findings. Psychological Medicine, 24:565-578,1994.
Mari, J.J., and Streiner, D.L. Family intervention for peo-ple with schizophrenia (Cochrane Review). In: TheCochrane Library, Iss. 1. Oxford, U.K.: Update Software,1996.
McCreadie, R.G. The Nithsdale Schizophrenia Surveys:An overview. Social Psychiatry and PsychiatricEpidemiology, 27:40-45, 1992.
McCreadie, R.G.; Phillips, K.; Harvey, J.A.; Waldron, G.;Stewart, M.; and Baird, D. The Nithsdale SchizophreniaSurveys: VJH. Do relatives want family intervention—anddoes it help? British Journal of Psychiatry, 158:110-113,1991.
McFarlane, W.R.; Link, B.; Dushay, R.; Marchal, J.; andCrilly, J. Psychoeducational multiple family groups: Four-year relapse outcome in schizophrenia. Family Process,34:127-144,1995a.
McFarlane, W.R.; Lukens, E.; Link, B.; Dushay, R.;Deakins, S.A.; Newmark, M.; Dunne, E.J.; Horen, B.; andToran, J. Multiple-family groups and psychoeducation inthe treatment of schizophrenia. Archives of GeneralPsychiatry, 52:679-687, 1995&.
McGill, C ; Falloon, I.R.H.; Boyd, J.L.; and Wood-Siverio, C. Family educational intervention in the treat-ment of schizophrenia. Hospital and CommunityPsychiatry, 34(10):934-938, 1983.
McGill, C, and Lee, E. Family psychoeducational inter-vention in the treatment of schizophrenia. Bulletin of theMenninger Clinic, 50(3):269-286, 1986.
Mojtabai, R.; Nicholson, R.A.; and Carpenter, B.N. Roleof psychosocial treatments in management of schizophre-nia: A meta-analytic review of controlled outcome studies.Schizophrenia Bulletin, 24(4):569-587, 1998.
Mosher, L.R., and Keith, S.J. Psychosocial treatment:Individual, group, family, and community supportapproaches. Schizophrenia Bulletin, 6(l):10-41, 1980.
Pakenham, K.I., and Dadds, M.R. Family care and schizo-phrenia: The effects of a supportive educational programon relatives' personal and social adjustment. Australianand New Zealand Journal of Psychiatry, 21:580-590,1987.
Penn, D.L., and Mueser, K.T. Research update on the psy-chosocial treatment of schizophrenia. American Journalof Psychiatry, 153(5):607-617, 1996.
Pharoah, F.M.; Mari, J.J.; and Streiner, D.L. Family interven-tion for schizophrenia (Cochrane Review). In: The CochraneLibrary, Iss. 4. Oxford, U.K.: Update Software, 1999.
Pitschel-Walz, G. Die Einbeziehung der Angehorigen indie Behandlung schizophrener Patienten und ihr Einflufiaufden Krankheitsverlauf. Frankfurt am Main, Germany:Peter Lang, 1997.
Pitschel-Walz, G.; Kissling, W.; and Bauml, J.Psychoeducation. In: The Workshop Manager: EvidenceBased Medicine in Schizophrenia (CD-ROM). London,U.K.: OCC Multimedia, 1998.
Posner, CM.; Wilson, K.G.; Krai, M.J.; Lander, S.; andMcllwraith, R.D. Family psychoeducational supportgroups in schizophrenia. American Journal ofOrthopsychiatry, 62(2):206-218, 1992.
90
Family Interventions Schizophrenia Bulletin, Vol. 27, No. 1, 2001
Randolph, E.T.; Spencer, E.T.H.; Glynn, S.M.; Paz, G.G.;Leong, G.B.; Shaner, A.L.; Strachan, A.; Van Vort, W.;Escobar, J.I.; and Liberman, R.P. Behavioural familymanagement in schizophrenia—outcome of a clinic-basedintervention. British Journal of Psychiatry, 164:501-506,1994.
Rosenthal, R. Meta-analytic procedures for socialresearch. Rev. ed. Applied Social Research MethodsSeries. Vol. 6. Newbury Park, CA: Sage, 1991.
Rosenthal, R., and Rubin, D.B. A simple, general purposedisplay of magnitude of experimental effect. Journal ofEducational Psychology, 74:166-169, 1982.
Ro-Trock, G.K.; Wellisch, D.K.; and Schoolar, J.C. Afamily therapy outcome study in an inpatient setting.American Journal of Orthopsychiatry, 47(3):514-522,1977.
Rund, B.R.; Moe, L.; Sollien, T.; Fjell, A.; Borchgrevink,T.; Hallert, M.; and Naess, P.O. The Psychosis Project:Outcome and cost-effectiveness of a psychoeducationaltreatment programme for schizophrenic adolescents. ActaPsychiatrica Scandinavica, 89:211-218, 1994.
Schaub, A. Cognitive-behavioural coping-orientated ther-apy for schizophrenia: A new treatment model for clinicalservice and research. In: Perris, C., and McGorry, P., eds.Cognitive Psychotherapy of Psychotic and PersonalityDisorders: Handbook of Theory and Practice. Chichester,U.K.: John Wiley and Sons, 1998. pp. 91-109.
Schaub, A., and Brenner, H.D. Aktuelle verhaltensthera-peutische Ansatze zur Behandlung schizophren Erkrankter.In: Stark, A., ed. Verhaltenstherapeutische und psychoe-dukative Ansatze im Umgang mit schizophren Erkrankten.Tiibingen, Germany: dgvt-Verlag, 1996. pp. 37-65.
Scherrmann, T., and Seizer, H.-U. Das TiibingerAngehorigenprojekt: Eine Pilotstudie iiber die Konzeptionund Evaluation einer Informationsgruppe fur Angehorigeschizophrener Patienten. In: Buchkremer, G., and Rath, N.,eds. Therapeutische Arbeit mit Angehorigen schizophrenerPatienten. Bern, Switzerland: Hans Huber, 1989. pp. 87-91.
Scherrmann, T.E.; Seizer, H.-U.; Rutow, R.; and Vieten,C. Psychoedukative Angehorigengruppe zurBelastungsreduktion und Riickfallprophylaxe in Familienschizophrener Patienten. Psychiatrische Praxis, 19:66-71,1992.
Schindler, R. Ergebnisse und Erfolge der Gruppenpsy-chotherapie mit Schizophrenen nach den Methoden derWiener Klinik. Wiener Zeitschrift fur Nervenheilkunde u.deren Grenzgebiete, 15:250-261, 1958.
Schneider, F.; Leitner, D.; and Heimann, H.Psychoedukative Angehorigengruppe bei psychiatrischenPatienten verschiedener Diagnosen. Schweizer Archiv furNeurologie und Psychiatrie, 142(3):247-258, 1991.
Schooler, N.R. The efficacy of antipsychotic drugs andfamily therapy in the maintenance treatment of schizo-phrenia. Journal of Clinical Psychopharmacology,6(l)(Suppl):llS-19S, 1986.
Schooler, N.R.; Keith, S.J.; Severe, J.B.; and Matthews,S.M. Maintenance treatment of schizophrenia: A reviewof dose reduction and family treatment strategies.Psychiatric Quarterly, 66(4):279-292, 1995.
Schooler, N.R.; Keith, S.J.; Severe, J.B.; Matthews, S.M.;Bellack, A.S.; Glick, I.D.; Hargreaves, W.A.; Kane, J.M.;Ninan, P.T.; Frances, A.; Jacobs, M.; Lieberman, J.A.;Mance, R.; Simpson, G.M.; and Woerner, M.G. Relapse andrehospitalization during maintenance treatment of schizo-phrenia. Archives of General Psychiatry, 54:453—463, 1997.
Smith, J.V., and Birchwood, M.J. Specific and non-spe-cific effects of educational intervention with families liv-ing with a schizophrenic relative. British Journal ofPsychiatry, 150:645-652,1987.
Smith, J.V., and Birchwood, M.J. Relatives and patients aspartners in the management of schizophrenia. BritishJournal of Psychiatry, 156:654-660, 1990.
Snyder, K.S., and Liberman, R.P. Family assessment andintervention with schizophrenics at risk for relapse. In:Goldstein, M.J., ed. New Developments in InterventionsWith Families of Schizophrenics. San Francisco, CA:Jossey-Bass, 1981. pp. 49-60.
Spencer, J.H.; Glick, I.D.; Haas, G.L.; Clarkin, J.F.;Lewis, A.B.; Peyser, J.; DeMane, N.; and Good-Ellis, M.A randomized clinical trial of inpatient family interven-tion: III. Effects at 6-month and 18-month follow-ups.American Journal of Psychiatry, 145(9): 1115—1121, 1988.
Spiegel, D., and Wissler, T. Using family consultation aspsychiatric aftercare for schizophrenic patients. Hospitaland Community Psychiatry, 38(10):1096-1099, 1987.
Strachan, A.M. Family intervention for the rehabilitationof schizophrenia: Toward protection and coping.Schizophrenia Bulletin, 12(4):678-698, 1986.
Szmukler, G.I.; Herrman, H.; Colusa, S.; Benson, A.; andBloch, S. A controlled trial of a counselling interventionfor caregivers of relatives with schizophrenia. SocialPsychiatry and Psychiatric Epidemiology, 31(3-4):-149-155, 1996.
Tarrier, N. Effect of treating the family to reduce relapsein schizophrenia: A review. Journal of the Royal Societyof Medicine, 82(7):423^24, 1989.
Tarrier, N., and Barrowclough, C. Family interventions forschizophrenia Behavior Modification, 14(4):408-440,1990.
Tarrier, N.; Barrowclough, C ; Porceddu, K.; andFitzpatrick, E. The Salford Family Intervention Project:Relapse rates of schizophrenia at five and eight years.British Journal of Psychiatry, 165:829-832, 1994.
91
Schizophrenia Bulletin, Vol. 27, No. 1, 2001 G. Pitschel-Walz et al.
Tarrier, N.; Barrowclough, C ; Vaughn, C; Bamrah, J.S.;Porceddu, K.; Watts, S.; and Freeman, H. The communitymanagement of schizophrenia—a controlled trial of abehavioural intervention with families to reduce relapse.British Journal of Psychiatry, 153:532-542, 1988.
Tarrier, N.; Barrowclough, C ; Vaughn, C ; Bamrah, J.S.;Porceddu, K.; Watts, S.; and Freeman, H. Communitymanagement of schizophrenia—a two-year follow-up of abehavioural intervention with families. British Journal ofPsychiatry, 154:625-628, 1989.
Tarrier, N.; Lowson, K.; and Barrowclough, C. Someaspects of family interventions in schizophrenia: II.Financial considerations. British Journal of Psychiatry,159:481-484, 1991.
Telles, C ; Karno, M.; Mintz, J.; Paz, G.; Arias, M.;Tucker, D.; and Lopez, S. Immigrant families coping withschizophrenia—behavioural family intervention v. casemanagement with a low-income Spanish-speaking popu-lation. British Journal of Psychiatry, 167:473-479, 1995.
Vaughan, K.; Doyle, M.; McConaghy, N.; Blaszczynski,A.; Fox, A.; and Tarrier, N. The relationship between rela-tives' expressed emotion and schizophrenic relapse: AnAustralian replication. Social Psychiatry and PsychiatricEpidemiology, 27:10-15, 1992.
Vaughn, C.E. Patterns of emotional response in the fami-lies of schizophrenic patients. In: Goldstein, M.J., Hand,I., and Hahlweg, K., eds. Treatment of Schizophrenia:Family Assessment and Intervention. Berlin, Germany:Springer-Verlag, 1986. pp. 97-106.
Videka-Sherman, L. Metaanalysis of research on social workpractice in mental health. Social Work, 33:325-338,1988.
Waring, E.; Carver, C ; Moran, P.; and Lefcoe, D.H.Family therapy and schizophrenia: Recent develop-ments. Canadian Journal of Psychiatry, 31(2): 154—160,1986.
Weiden, P.; Aquila, R.; and Standard, J. Atypicalantipsychotic drugs and long-term outcome in schizo-phrenia. Journal of Clinical Psychiatry, 57(Supplll):53-60, 1996.
Winefield, H.R., and Harvey, E.J. Determinants of psy-chological distress in relatives of people with chronicschizophrenia. Schizophrenia Bulletin, 19(3):619—625,1993.
Wunderlich, U.; Wiedemann, G.; and Buchkremer, G. Arepsychosocial methods of intervention effective in schizo-phrenic patients? A meta-analysis. Verhaltenstherapie,6:4-13, 1996.
Xiang, M.: Ran, M.: and Li, S. A contolled evaluation ofpsychoeducational family intervention in a rural Chinesecommunity. British Journal of Psychiatry, 165:544—548,1994.
Xiong, W.; Phillips, R.; Hu, X.; Wang, R.; Dai, Q.;Kleinman, J.; and Kleinman, A. Family-based interven-tion for schizophrenic patients in China. British Journalof Psychiatry, 165:239-247, 1994.
Yalom, I.D. Theorie und Praxis der Gruppenpsychother-apie. Miinchen, Germany: Verlag J. Pfeiffer, 1984.
Zastowny, T.R.; Lehman, A.F.; Cole, R.E.; and Kane, C.Family management of schizophrenia: A comparison ofbehavioral and supportive family treatment. PsychiatricQuarterly, 62(2): 159-186, 1992.
Zhang, M.; Wang, M.; Li, J.; and Philips, M.R. BritishJournal of Psychiatry, 165(Suppl 24):96-102, 1994.
The Authors
Gabi Pitschel-Walz, Dr. rer. biol. hum., is ClinicalPsychologist, Department of Psychiatry, TechnicalUniversity, Munich. Stefan Leucht, M.D., is Registrar,Department of Psychiatry, Technical University Munich.Josef Bauml is Consultant Professor, Department ofPsychiatry, Technical University Munich. WernerKissling is Consultant Professor, Department ofPsychiatry, Technical University Munich. Rolf R. Engel,Dr. rer. nat., is Head of the Psychological Section,Department of Psychiatry, University Munich, Munich,Germany.
92
