Effectiveness of alcoholism treatment in partial versus inpatient settings: Twenty-four month outcomes

Addictive Behaviors, Vol. 10, pp. 235-248, 1985 Printed in the USA. All rights reserved.

0306-4603/85 $3.00 + .OO Copyright e 1985 Pergamon Press Ltd

EFFECTIVENESS OF ALCOHOLISM TREATMENT IN PARTIAL VERSUS INPATIENT SETTINGS:

TWENTY-FOUR MONTH OUTCOMES

EDWARD B. FINK, RICHARD LONGABAUGH, BARBARA M. McCRADY, ROBERT L. STOUT, MARTHA BEATTIE, ANN RUGGIERI-AUTHELET,

and DWIGHT MCNEIL Brown University Medical School and Butler Hospital

Abstract-The authors examined the effectiveness of the partial hospital setting, in contrast to the inpatient setting, for the rehabilitation of alcohol abusers and alcoholics. Outcomes after 24 months in five life health areas indicated marked improvement from baseline for the entire follow-up period on almost all measures. There also was a strong relationship between drinking outcomes and outcomes in the other health areas. Although there were few between group differences on the clinical outcome measures. differences which were found favored the partial hospital setting. Furthermore, cost of treatment over two years clearly favored the partial hospital.

The clinical and cost effectiveness of the partial hospital setting, in contrast to the inpatient setting, for the rehabilitation of alcohol abusers and alcoholics is examined in this report. Numerous well-controlled studies have demonstrated that partial hospitals are clinically as effective or more effective than inpatient treatment for a wide range of psychiatric patients (Penk, Charles, & Van-Hoose, 1978; Wilder, Levin, & Zwerling, 1966) and that treatment goals can be achieved at a lower cost (Washburn, Vannicelli, Longabaugh, et al., 1976; Fink, Longabaugh, & Stout, 1978). Some studies also have demonstrated that the cost savings of the initial hospitalization are maintained during follow-up because partial hospital patients are rehospitalized no more often than inpa- tients (Herz & Endicott, 1971; Michaux, Chelst, Foster, Pruim, & Dasinger 1973).

The partial hospital setting also has been used for the treatment of alcohol abusers (Pallet, 1976; Powell & Viamontes, 1974). Not until 1982, however, several years after this study was begun, was there a report of a controlled comparison of inpatient and day care treatment for alcohol abusers (McLachlan & Stein, 1982). One-year outcome data indicated clinical effectiveness and a cost advantage of the day care setting for this patient group. Subjects in the two experimental groups, however, did not receive the same specialized treatment program; other differences from the study reported here have been discussed (Longabaugh, McCrady, Fink, et al., 1983).

In this study, the first 6-month outcomes suggested that the partial hospital setting provided a cost advantage over inpatient treatment (Longabaugh, McCrady, Fink, et al., 1983). There also was considerable improvement in both groups, compared with baseline, in abstinence rates and other drinking measures, as well as in interpersonal functioning and psychological health. By the end of twelve months we found evidence of greater stability in the subject population (McCrady, Longabaugh, Fink, et al.,

Supported in part by grants from the U.S. National Institute on Alcohol Abuse and Alcoholism, RO 1 AA 04403 and from Blue Cross/Blue Shield of Rhode Island, (Richard Longabaugh, Ed.D., principal in- vestigator). Without the administrative support of Nathan Epstein, M.D., Medical Director, and Duane Bishop, M.D., Clinical Director of Butler Hospital, this study could not have been successfully completed. 7 (Reprint requests to Dr. Fink at the Department of Psychiatry and Human Behavior, Brown University Medical School and Butler Hospital, 345 Blackstone Boulevard, Providence, RI 02906.

235

236 EDWARD B. FINK et al.

1984). There was more employment and fewer rehospitalizations during the second 1 -,P ,-,I- T ~~ ~.I~ .d- ~~ I -I . ~~ ~.l narr 01 rne year. rmprovemenrs evraencea at six monrns were maintained; aiso, the cost advantage of the partial hospital setting continued.

The present report focuses on the outcomes after 24 months in the areas of drinking behavior, life task performance, interpersonal and psychological functioning, and physical health, as well as subsequent treatment costs. The relationship between abstinence versus intermediate drinking (defined below) and outcomes in other life health areas also is explored.

METHODS

Subjects All of the subjects were patients in a private, university-affiliated, psychiatric hospital.

Screening criteria included: (a) diagnosis of alcohol abuse or alcohol dependence or a problem list entry of alcohol abuse in the problem-oriented medical record (Ryback, Longabaugh, & Fowler, 1981); (b) a score of 5 or greater on the Michigan Alcoholism Screening Test (Selzer, 1971); (c) not acutely suicidal or significantly disorganized by psychosis; and (d) no irreversible organic brain syndrome.

Table 1 details the sociodemographic and clinical characteristics of this severely alcohol impaired sample. During a 1Cmonth period, beginning February 1979, 174 subjects participated in the study. The original study design, however, included only 6-month follow-up data collection and therefore subjects had to sign a second consent form to continue in the study for the additional 18 months. Thirty-four percent of the subjects (n = 59) did not continue, leaving a sample of 115 subjects. The drop out rates for the two experimental groups were not significantly different, Extended Inpatient Treat- ment (EIT) = 38.3%, Partial Hospital Treatment (PHT) = 30.7Vo.

To determine whether there were differences between continuers and noncontinuers extensive univariate and multivariate comparisons were made using baseline data. The multivariate analysis of variance revealed no significant differences between baseline characteristics of continuers and noncontinuers, although one univariate test of 6-month outcome suggested that continuers were functioning better at that time (Global Assessment Scale- Gas) (Hargreaves & Fulwiler, 1977) comparison, t = 3.43, df = 42, p < .OOl). While the overall level of outcomes beyond 6 months may not be representative of the entire treated sample, the non-continuers were evenly distributed between the two groups and therefore experimental group comparisons should not be affected.

Procedures After patients signed an informed consent, comprehensive baseline data were col-

lected. They were randomly assigned to one of two experimental groups. All subjects received up to one week of detoxification and comprehensive medical and psychiatric assessment and inpatient treatment. Patients in both the PHT and EIT groups then participated in the same 5-day-per-week alcohol treatment program, located in the partial hospital.

Based upon a social learning model of alcohol abuse, treatment components included group sessions teaching behavioral analysis and behavior change skills and the application of these skills to the patients’ individual life situations. In addition, goal-setting meetings, patient and family member education sessions, modeling of successful abstinence by volunteers who previously had serious drinking problems, and social skills training were incorporated into the treatment program. Eighty percent of patients had family members or close friends involved in weekly multiple family group sessions and

Alcoholism treatment in partial versus inpatient settings 231

Table 1. Description of study sample.

Percent Standard

N Mean Deviation

I.

II

III.

Sociodemographic Characteristics Age (~-1 Sex (@lo male) Education (re 12 years) Ever married Living w/spouse Living alone Ethnic background (010 white)

Vocational Status Full-time employment Part-time Housewife No job Retired, disabled, other

Drinking Measures Diagnosis of alcoholism Diagnosis of alcoholism

or on problem list Prior alcoholism-related

hospitalizations ever Prior Alcoholics Anonymous

attendance (past 6 months) Prior use of outpatient services

for alcohol-related problems (past 6 months)

Alcohol-related arrests ever MAST Alcohol Impairment Index Quantity-Frequency Index Days abstinent, past 180 days Days with 6 drinks or more,

past 180 days Years drinking Global Assessment Scale

71.3 61.9 78.3 61.7 25.2 98.3

67.8 78 6.1 I 9.6 11 7.0 8 9.6 11

92.2 106

100.0 115

59.1 68

26.1 30

34.8 33.0

41.3 82 70 90 71 29

113

40 38

30.53 17.32 8.02

44.71

10.7 1.4 1.9

50.12

83.23 61.54 16.48 12.41 48.35 1.34

12.0

individual family sessions. Patients in the EIT group returned to their 24hour unit for nights and weekends. Further details of the structured treatment program are described elsewhere (McCrady et al., 1982).

Measures Baseline data, with a Bmonth window, were collected by direct patient interview;

monthly follow-up interviews for 2 years were conducted by telephone. Treatment cost data were abstracted from hospital bills. Drinking behavior was assessed at baseline by the measures of drinking listed in Table 2. Monthly, the QFI (Cahalan, Cisin, & Cross- ley, 1969) was readministered and subjects:wereqsked about drinking days and hospitalizations. Life tasks were assessed by the measures noted in Table 3 at baseline and monthly thereafter. Interpersonal functioning was assessed at baseline and every 6 months thereafter by the GAS and the role scales of the Psychosocial Functioning In- ventory (PFI) (Feragne, Longabaugh, & Stevenson, 1983) (Table 4). Psychological well-being was assessed at baseline and dmonth intervals thereafter by the PFI scales

Tab

le 2

. D

rink

ing

beha

vior

ov

er 2

4 m

onth

s.

Mea

sure

Tre

atm

ent

Tim

e C

ompa

riso

ns

Gro

up

Com

pari

sons

B

BB

B66

6

12

12

18

Bas

e-

Mon

ths

Mon

ths

Mon

ths

Mon

ths

MO

. M

O.

MO

. M

O.

line

l-6

7-12

13

-18

19-2

4 6

12

18

24

OA

v;

* y;

T

8 ‘2

”4 r

”;

vs8

‘2s;

Js

8.

“2”4

“2”

4

Qua

ntity

-fre

quen

cy

inde

x (M

ean)

Alc

ohol

ism

in

dex

(Mea

n)

Prop

ortio

n of

da

ys a

bstin

ent,

past

6 m

onth

s

Prop

ortio

n of

da

ys a

bstin

ent,

past

30

days

Cur

rent

ab

stin

ence

Con

tinuo

usly

ab

stin

ent

Hos

pita

lizat

ions

Ave

rage

nu

mbe

r of

day

s,

whe

n re

hosp

italiz

ed

8.02

0.

73

0.72

0.

72

0.85

*

17.3

2 na

. n.

a.

n.a.

24%

88

.4%

82

.1%

80

.7%

(2

8)

(93)

(8

%

(73)

2.18

<

___

____

n,a,

79.6

%

(71)

>

<__

I__,

<

__*_

_>

<---

n.a.

--->

+

<

___

____

____

__

n.a.

___

____

_ _

____

>

* <

__*_

_> <

__*_

_>

n.a.

86

.1%

82

.2%

80

.4%

84

.2%

(9

0)

(83)

(7

6)

(76)

0%

0%

68.6

%

(72)

-

60.4

%

(61)

27.9

%

(29)

16.5

%

(10)

60.3

%

(63)

71

.1%

(6

4)

22.8

%

(21)

11.4

%

(10)

P P

P <

___

-___

n,

a ._

____

_,

* *

* *

*

< _

____

____

-___

____

____

____

__n.

a.__

____

____

___.

.___

____

____

_ >

n.a.

(2

1)

20.0

%

(17)

11

.0%

(1

0)

11.6

0

< -

-- n

.a.

--- >

n.a.

6.

71

12.0

0 10

.30

< _

____

____

____

____

____

____

__n.

a._

____

__

____

__

____

_ ____

__

___ >

Key

for

Tab

les

2-5

l

= Si

gnif

ican

t at

.0

5 le

vel

or

belo

w

OA

=

O

vera

ll av

erag

e of

al

l da

ta

poin

ts

for

24 m

onth

s

P

=

Part

ial

hosp

ital

sign

ific

antly

be

tter

at

.0

5 le

vel

or

belo

w

E

=

Ext

ende

d in

patie

nt

sign

ific

antly

be

tter

at

.0

5 le

vel

or

belo

w

I =

T

reat

men

t by

tim

e in

tera

ctio

n

< -

-*--

>

=

Thi

s sy

mbo

l re

pres

ents

si

gnif

ican

ce

at t

he

.05

leve

l or

bel

ow,

whe

n ba

selin

e 1s

com

pare

d to

the

ave

rage

of

the

tim

e pe

riod

s in

dica

ted

(eith

er

6 an

d 12

mon

ths,

or

18

and

24 m

onth

s).

n.a.

=

N

ot

asce

rtai

ned

Tab

le 3

. L

ife

Tas

k Pe

rfor

man

ce

Ove

r 24

Mon

ths

Tre

atm

ent

Tim

e C

ompa

riso

ns

P G

roup

C

ompa

riso

ns

BB

BB

66

6 12

12

18

F

Bas

e-

Mon

ths

Mon

ths

Mon

ths

Mon

ths

MO

. M

O.

MO

. M

O.

0’

_;

Mea

sure

lin

e l-

6 7-

12

13-1

8 19

-24

6 12

18

24

O

A

“6”.

y;

‘1

”8 ‘

2”4

“1”;

y8.

‘2

”4 ;

“8

;4

“2”6

3 :

Full-

time

67.8

%

57.1

%

57.3

%

60.4

%

62.5

%

* +

G

em

ploy

men

t (1

070)

(7

8)

(60)

(5

9)

(55)

(5

5)

$j

Full-

time

77.4

%

66.7

%

67.0

%

69.2

%

70.5

%

5’

role

(T

o)

(89)

(7

0)

(69)

(6

3)

(62)

P $

Wor

k m

isse

d du

e 53

.0%

15

.2%

13

.6%

15

.4%

0.

0%

* *

* *

* *

* p

to d

rink

ing

(61)

(1

6)

(14)

(1

4)

(0)

<

9

Job

loss

due

15

.7%

8.

6%

4.9%

0.

0%

2.3%

*

:

to d

rink

ing

(18)

(9

) (5

) (0

) (2

) 5’

D

E

A

ny a

rres

ts

8.8%

3.

5%

1.9%

0.

0%

1.1%

i’

2

(10)

(4

) (2

) (0

) (1

) v)

z c.

R

esid

entia

l st

atus

0.

90

0.71

0.

69

0.80

0.

81

* *

2 an

d st

abili

ty

<__

+__

>

<__

I__>

v)

Tab

le 4

. In

terp

erso

nal

Func

tioni

ng

Ove

r 24

Mon

ths

Tre

atm

ent

Tim

e C

ompa

riso

ns

Gro

up

Com

pari

sons

B

BB

B66

6

12

12

18

Bas

e-

Mon

ths

Mon

ths

Mon

ths

Mon

ths

MO

. M

O.

MO

. M

O.

Mea

sure

lin

e l-

6 7-

12

13-1

8 19

-24

6 12

18

24

O

A

Vi’

r”

; ‘1

; v;

‘1

”;

‘l”s

’ ‘2

”4 y

8’

‘2”6

yi

Soci

al b

ehav

ior

2.70

3.

31

3.21

3.

30

3.42

*

z

scal

e (m

ean)

<

__*_

_>

<__

*__>

Pare

nt

role

(m

ean)

3.

22

3.51

3.

46

3.40

3.

56

Hou

sem

ate

role

2.

27

2.64

1.

65

2.78

2.

51

(me@

Spou

se r

ole

(mea

n)

2.48

2.

87

2.82

2.

84

2.98

E

E

I *

<__

I__>

<

__+

__>

l

<__*__> <._*__>

Subj

ectiv

e ro

le

3.07

4.

17

4.61

4.

11

4.04

.

*

perf

orm

ance

(m

ean)

<

__*_

_>

<__

*__>

Glo

bal

asse

ssm

ent

48.3

5 72

.79

71.4

0 76

.49

77.1

9 P

E

I <

__I_

_>

l

l

scal

e (m

ean)

<

__*_

_>

<__

*__>


listed in Table 5. Physical health was assessed by asking subjects each month whether or not they have experienced any health problems.

RESULTS

Study participation is high, with a mean of 21.2 interviews per patient over 24 months. Six patients died in the first 12 months and five in months 13-24. There are no dif-

ferences between PHT and EIT subjects regarding study participation or mortality.

Two year outcomes

Drinking behavior. Overall, positive changes observed earlier are maintained or further improved (Table 2). This is apparent on all measures for successive 6-month intervals, except for a slight linear decrease in percent abstinent days (t = 2.02, a” = 225, two-tailed p < 0.05). Fewer subjects are rehospitalized during the second year at risk. The PHT group has some advantage in abstinence rates over 2 years but less than is apparent at 12 months; no other between group differences are found.

Life tasks. Table 3 reveals that although during the first 6 months the employment rate (57.1%) is significantly lower than at baseline (67.8%), it increases over the suc- ceeding months and by the second half of the second year the employment rate (61.5%) is not significantly different from the baseline rate (McNemar x2 Test = 1.24, df = 1, p = 0.265). Significant gains also are apparent on other variables during the second year, including fewer days missed from work due to drinking and greater job stability. Legal problems are very infrequent in months 13-24 and residential stability (Sobell & Sobell, 1976) improves in the second year, F(l, 91) = 10.05, p < 0.01, although it re- mains lower than at baseline. There are no differences between the PHT and EIT groups.

Interpersonal functioning. Table 4 reflects the maintenance over 2 years of significantly improved functioning, in comparison with baseline. In addition, social behavior during the last 6 months of follow-up is significantly better than during months 7-12, and the month 7-12 deterioration in housemate role functioning is significantly reversed during year 2. Functioning in the parent and spouse roles remain stable for 2 years and improved from baseline. There are few differences, which do not persist over time, between the two experimental groups on these variables: the EIT group is favored at month 12 for the spouse role and at month 18 for the housemate role. Improvements during the first 6 months on the GAS are maintained during year 2. At month 6 the PHT group scores higher, whereas at month 18 the EIT group is favored; at months 12 and 24 there are no between group differences.

Psychological well-being. Significant decreases from month 6 to 12 in positive affect and subjective well-being are reversed during the second follow-up year when there also is a significant increase in life satisfaction and a significant decrease in negative affect (Table 5). Between-group differences at 6 months on life satisfaction, negative affect, and subjective well-being favor the PHT subjects. Furthermore, on the latter two measures the PHT group is favored over the entire 2 years.

Cost outcomes. Treatment cost data for two thirds of the 95 rehospitalization episodes are available. Rehospitalization occurred for 36% (n = 62) of the sample of 174 patients, including those who did not continue in the study beyond 6 months. For the 2-year period after the index hospitalization there is no significant difference between groups in mean total treatment costs: $1,032 per subject in the PHT group and

Tab

le 5

. Ps

ycho

logi

cal

Wel

l-B

eing

Ove

r 24

Mon

ths

Mea

sure

Tre

atm

ent

Tim

e C

ompa

riso

ns

Gro

up

Com

pari

sons

B

B

BB

666

12

12

18

Bas

e-

Mon

ths

Mon

ths

Mon

ths

Mon

ths

MO

. M

O.

MO

. M

O.

vs.

vs.

vs.

vs.

vs.

vs.

vs.

vs.

vs.

vs.

line

l-6

7-12

13

-18

19-2

4 6

12

18

24

OA

6

12

18

24

12

18

24

18

24

24

Posi

tive

affe

ct

2.04

2.

24

2.09

2.

20

2.25

*

*

(mea

n)

<__

*__>

<

__‘_

_>

Neg

ativ

e af

fect

2.

26

1.76

1.

83

1.70

1.

69

P P

+

(mea

n)

<__

*__>

<

__*_

_>

Lif

e sa

tisfa

ctio

n 1.

82

2.38

2.

29

2.33

2.

36

P *

(mea

n)

<__

*__>

<

__*_

_>

Subj

ectiv

e w

ell-

bein

g 0.

1 1.

1 .0

8 1.

0 1.

1 P

P <

__I_

_>

* (m

ean)

<

__+

__>

<

__*_

_>


$988 per subject in the EIT group. For the PHT group, costs tend to be lower in the second 12-month interval ($496) than in the first year ($625). The significant differences in treatment costs for the index hospitalization which favor the PHT group therefore are maintained over the subsequent 24 months.

Relationships between drinking and other health areas. Subjects were categorized for the 24-month follow-up period as always abstinent, in-

termediate, or uncontrolled (McCrady, Longabaugh, Fink, et al., 1984). The intermediate group included patients who had a QFI between zero and three for each month and whose daily consumption never exceeded two quarts of beer or five (4 oz.) glasses of wine or four (1% oz.) drinks of liquor. Subjects completing less than 80% of the inverviews were excluded from this analysis.

Variables were selected to represent outcomes in five major health areas. Good physical and psychological health outcomes were defined as no problems reported in that area by the patient during 80% or more of the monthly interviews. Vocational health was defined as good if the subject reported a full-time occupational role at least 80% of the time and experienced no job losses over the 2 years. Social health was categorized as good if there were no arrests for two years; the need for therapeutic supports was scored as good if there were no rehospitalizations over the 24 months. Several con- tinuous variables also were examined in the areas of psychological health and social health.

Table 6 reports on these variables for each drinking outcome category. Physical health scores decrease significantly from the best to the poorest drinking outcome category. In the psychological health area, the uncontrolled drinking category has significantly fewer subjects with no problems, and lower subjective well-being. Social health is poorest for the uncontrolled category. There are no differences across drinking categories for spouse role (which has a smaller sample) or subjective role performance (a less objective measure than the others in this health area). From the best to the poorest drinking outcome category there is a significant decrease in the use of Alcoholics Anonymous (AA) as a therapeutic support and a significant increase in the percentage of subjects requiring rehospitalization. Only in the vocational health area are there no significant differences across drinking outcome categories, although trends are in the expected direction.

In view of the significant differences found in the overall analyses, further analyses were conducted to discern between which specific outcome categories the differences lay. The Scheffe Multiple Comparison Procedure and chi-square analyses, as appro- priate, indicate that the only significant differences (p < .05 on all measures) are between the intermediate and uncontrolled outcome categories. Subjects in the latter category have poorer physical health, more psychological problems and poorer subjective well-being. They have more arrests, poorer social behavior, more rehospitalizations and a nonsignificant trend toward less AA use. No differences between the always abstinent and intermediate categories were found, probably because of the smaller sample size in the abstinent category and the resultant loss of power to detect differences.

Next, chi-square tests comparing the two treatment conditions reveal that AA use is greater for PHT than EIT subjects (x’ = 4.26, df = 1, p c 0.05) within each outcome category. There are no differences between the EIT and PHT groups on the other health area outcome measures.

Finally, a summary score for overall life functioning was computed by counting the number of the five life areas in which the subject was functioning successfully, This

Tab

le 6

. R

elat

ions

hip

Bet

wee

n D

rink

ing

Out

com

es

and

Num

ber

of S

ucce

ssfu

l L

ife

Are

as

Dri

nkin

g O

utco

mes

Lif

e Fu

nctio

ning

Con

tinuo

us

Abs

tinen

ce

or

Abs

tinen

ce

or

Abs

tinen

ce

Mod

erat

e 2

80%

M

oder

ate

< 8

0%

(n

=

22)

(n

=

38)

(n

=

35)

n Sc

ore

n Sc

ore

n Sc

ore

Sign

ific

ance

T

est

Phy

scia

l H

ealt

h N

o ph

ysic

al h

ealth

pr

oble

ms

z 80

% o

f tim

e

Psy

chol

ogic

al

Hea

lth

No

psyc

holo

gica

l he

alth

pr

oble

ms

2 80

% o

f tim

e Su

bjec

tive

wel

l-be

ing

Voc

atio

nal

Hea

lth

Fullt

ime

role

~8

0%

of t

ime

No

job

loss

es

Soci

al H

ealt

h N

o ar

rest

s R

SSI

Spou

se r

ole

Soci

al b

ehav

ior

scal

e Su

bjec

tive

role

per

form

ance

Ther

apeu

tic

Supp

orts

N

o ho

spita

lizat

ions

A

A u

se

1E

81.8

%

16

72.7

%

22

1.21

15

68.2

%

20

90.9

%

22

lOO

.Qo

22

.89

11

2.97

22

3.

41

14

4.02

18

81.8

%

9 40

.9%

30

78.9

%

32

84.2

%

37

1.38

19

50.0

%

33

86.8

%

37

97.4

%

38

0.83

22

3.

03

36

3.56

31

4.

25

25

65.8

%

10

26.3

%

16

51.6

%

16

51.6

%

30

0.66

19

61.3

%

25

80.6

%

28

80.0

%

34

0.72

14

2.

91

29

3.24

25

3.

79

10

28.6

%

3 9.

7%

x’

=

7.95

, df

=

2,

p

<0.

05

x1=

877

df=

2p<

OO

l F

(2.8

6)

=’

5.20

, p

‘<O

.OO

i

n.s.

n.

s.

x1 =

9.

76,

df

= 2,

~ <

O.O

l F

(2,9

1)

=

3.29

, <

0.05

p

F(2

,84)

=

4.0

3:‘;

<

0.05

F

(2,6

7)

=

2.44

, p

<0.

095

XI

=

18 1

0 df

=

2,

<

0 00

01

. ,

p x’

=

7.

01,

df

=

2,~

<O

.d5


score is strongly correlated with drinking outcome category (Kendall’s tau b = 0.414, p < 0.0001). Patients with the best drinking outcome have more areas of successful functioning; for example, 36.4% have all five, whereas only 5.7% of patients in the uncontrolled drinking outcome category have that many. A log-linear categorical analysis confirms a strong association between drinking outcome and life functioning, but no treatment group differences or interactions are found.

In summary, the results of these analyses clearly demonstrate that at 2 years post- treatment drinking outcomes are interrelated with outcomes in four of five health areas, with only vocational health being unrelated. Also, the overall relationship between drinking outcome and the number of areas of successful life functioning first evi- dent after 12 months is maintained and appears to be considerably stronger. Subjects with better drinking outcomes clearly have more areas of successful life functioning.

DISCUSSION

Study limitations The study design necessitates some limitation in the interpretation of results. First,

subjects in the EIT group stayed on a psychiatric unit with patients with nonalcohol- related problems. Therefore, unlike most 24-hour alcoholism rehabilitation programs, evening and weekend treatment components were not specific to the clinical needs of the EIT subjects.

The generalizability of these results to other groups of alcohol abusers and alcoholics is limited because the sample excluded patients who were acutely suicidal, had severe medical problems, or were very agitated as a result of psychosis or a severe organic brain syndrome. Also, most subjects had some social support potentially available (e.g., family members, employment or other occupational role), and were not public inebriates. Finally, there is some evidence from the GAS scores at month 6 that subjects who participated in the entire 2 years of follow-up were not fully representative of the initially treated study sample.

Two-year outcomes Analyses of the 24-month outcomes for drinking behavior, life task performance, in-

terpersonal and psychological functioning, physical health, and treatment costs revealed several important findings. For subjects continuing beyond 6 months there was little deterioration in drinking outcome over the two years. Almost one quarter of the subjects (23%) remained continuously abstinent for the entire 24 months and those who did drink reported a significant decrease in drinking days and quantity consumed compared with the 6 months prior to treatment. Fewer rehospitalizations took place during the second 12 months than in the first year. This evidence of better functioning during the second year is consistent with the report of Caddy and Addington (1978). On almost all measures in the other health areas there was significant improvement from baseline to one year, which was maintained or further improved during the second year. Overall, these findings suggest considerable stability over two years for all areas examined and marked improvement from the six months prior to treatment. The deterioration found by others (Finney & Moos, 1981; Jellinek, 1960) was not confirmed, perhaps because of the better premorbid functioning of our study sample.

Differences in clinical outcomes between the two experimental groups were small. PHT subjects averaged significantly more abstinent days than did EIT subjects throughout the 24 months. However, this difference was marginal; it was significant only at the


12-month interval. Other drinking measures and time intervals did not show this difference.

A stronger difference favoring PHT patients was in their negative affect and subjective well-being across the 24 months. These differences were statistically significant at the 6-month interval, but not at subsequent intervals, suggesting that the greatest dif- ferential impact of the two modalities was short-term. Interpersonal functioning was the one area in which EIT patients might have done somewhat better. Differences in spouse role at 12 months and housemate role and GAS at 18 months, however, may be spurious since they were not observed at other intervals nor for the average across intervals. The housemate and GAS scores also reveal complex treatment by time period interactions (Table 4). We have no plausible interpretations.

While differences in clinical effectiveness were slight, differences in treatment costs were not. The PHT group established savings in treatment costs during the criterion hospitalization which were preserved throughout the two years. This advantage of the partial hospital has been found by the present authors (Fink et al., 1978; Longabaugh, 1978; Washburn et al., 1976) and others (Herz et al., 1971; Michauz et al., 1973) in the study of other patient populations, as well as by McLachlan and Stein (1982) for alcoholics.

Relationship between drinking and other health areas The strong overall relationships between drinking outcomes and outcomes in other

health areas after 24 months were consistent with each other and in the expected direction. Better drinking outcome was associated with a greater number of areas of successful life functioning, and with less rehospitalization and more use of AA. These findings are similar to those of Pettinati, Sugerman, DiDonato, et al. (1982), and Bris- sett et al. (1980).

The differences between groups categorized by drinking outcome on eight of twelve measures in four of the five life health areas were not apparent at 12 months when they differed only in residential stability (where the intermediate drinking group was most stable). Therefore, our conclusion based upon 12 months of observation that a considerable number of subjects function well despite fairly heavy alcohol consumption must now be tempered when outcome at 2 years is considered. This is especially true in areas of physical health, psychological health, and rehospitalization rates. Pettinati et al. (1982) also report that over their 4-year follow-up period a decreasing percentage of patients were able to drink and also maintain good functioning in other health areas.

Analyses to discern differences between treatment conditions within drinking categories revealed a greater use of AA by PHT patients. This finding is consistent with the nature of partial hospital treatment programs: PHT subjects had a greater oppor- tunity to practice the use of AA evenings and weekends, with staff support, during the index hospitalization. Also, unlike EIT patients who might have gone on pass to an AA meeting near the hospital, the PHT subjects would be more likely to attend an AA meeting near their homes and find an abstinent social network which would continue after treatment. This small number of between group differences is not surprising given the duration of the follow-up period.

The finding that outcome scores for those always abstinent was not significantly different from those in the intermediate drinking category must be interpreted with cau- tion in view of the clearly significant findings relating better drinking outcome with more areas of successful life functioning and less rehospitalization. These findings lend only weak support to Miller’s (1983) conclusion that controlled drinking is no more or less likely than abstinence to end in relapse or to bring about improvement in other


health areas; they are similar to others (e.g., Brisset et al., 1980; Pettinati et al., 1982) who found that few individuals are able to drink without any problems. These inconsistencies may relate to the fact that Miller (1983) reported on treatment programs specifically designed to teach controlled drinking. Miller (1983) and Polich et al. (1980) also suggested that controlled drinking may be successful as a treatment goal only with a subgroup of their sample (viz. younger, less dependent problem drinkers),

These inconsistencies in the literature may be resolved with further research on how best to match patients with optimal treatment settings and strategies. Our research shows that a wide range of patients can benefit from partial hospital treatment at least as much as from inpatient treatment, thus widening the range of options for those seek- ing treatment. We are conducting further research that we expect to advance our understanding of how patient characteristics and treatment approaches should be com- bined.

REFERENCES

Armor, D.J., Polich, J.M., & Stambul, H.B. (1976), Alcoholism and Treatment. Santa Monica, CA: Rand Corporation.

Brisset, D., Laundergan, J. Clark, & Biele, M. (1980). Drinkers and nondrinkers at three and a half years after treatment: Attitudes and growth. Journal of Studies on Alcohol, 41, 945-952.

Caddy, G.R., Addington, H.J., & Perkins, D. (1978). Individualized behavior therapy for alcoholics: A third year independent double-blind follow-up Behavior Research and Therapy, 16, 345-362.

Cahalan, D., Cisin, I., & Crossley, H. (1969). American drinking prucfices: A national study of drinking behuvior and patterns. New Brunswick, NJ: Rutgers Center for Alcohol Studies.

Feragne, M.A., Longabaugh, R., & Stevenson, J.F. (1983). The psychosocial functioning inventory. Evuluu- fion and the Heulth Professions, 6, 25-48.

Fink, E., Longabaugh, R., & Stout, R. (1978). The paradoxical underutilization of partial hospitalization. American Journal of Psychiatry, 135, 713-716.

Finney, J.W., & Moos, R.H. (1981). Characteristics and prognoses of alcoholics who become moderate drinkers and abstainers after treatment. Journal of Studies on Alcohol, 42, 94-105.

Hargreaves, W., & Fulwiler, R. (1977). A Global Assessment Scale for Substance Abusers. University of California at San Francisco: Langley Porter Neuropsychiatric Institute, San Francisco, CA.

Hertz, M.I., Endicott, J., Spitzer, R.L., & Mesnikoff, A. (1971). Day versus inpatient hospitalization: A controlled study. American Journal of Psychiatry, 127, 1371-1382.

Jellinek, E.M. (1960). The disease concept of alcoholism. Highland Park, NJ: Hillhouse Press. Longabaugh, R. (1978). A one year follow-up comparison of day hospital, day hospital transfer, and inpa-

tient samples with respect to cost effectiveness of psychiatric care. In Evaluation in practice: A source- book of program evaluation studies from mental health cure system in the United States, pp. 202-203. Washington, DC: NIMH.

Longabaugh, R., McCrady, B., Fink, E., Stout, R., McAuley, T., Doyle, C., & McNeill, D. (1983). Cost effectiveness of alcoholism treatment in partial versus inpatient settings: Six-month outcomes. Journal of Studies on Alcohol, 44, 1049-1071.

McCrady, B.S., Dean, L., Dubreuil, E., 8~ Swanson, S. (1982). The problem drinkers project: A progrum- mutic application of social learning bused treatment. In G.A. Marlatt & J. Gordon (Eds.), Relapse prevention. New York: Guilford Press.

McCrady, B.S., Longabaugh, R., Fink, E.B., Stout, R., McAuley, T., & McNeil, D. (1984). Cost effectiveness of alcoholism treatment in purtiul vs. inpatient settings: Twelve months outcomes. Submitted for publication.

McLachlan, J.F.C., & Stein, R.L. (1982). Evaluation of a day clinic for alcoholics. Journal of Studies on Alcohol, 43, 261-272.

Michaux, M.H., Chelst, M.R., Foster, S.A., Pruim, R. J., & Dasinger, E.M. (1973). Post-release adjustment of day and full-time psychiatric patients. Archives of General Psychiatry, 29, 647-651.

Miller, W.R. (1983). Controlled drinking: A history and a critical review. Journal of Studies on Alcohol, 44, 68-83.

Pallett, A. (1976). Alcoholic day treatment unit: Herbert Day Hospital, Bournemouth (Notes and News). British Journal of Addictions, 71, 99-100.

Pattison, E.M. (1976). A conceptual approach to alcoholism treatment goals. Addictive Behaviors, 1177-1192.

Penk, W.E., Charles, H.L., & Van-Hoose, T.A. (1978). Comparative effectiveness of day hospital and inpatient psychiatric treatment. Journal of Consulting and Clinical Psychology, 46, 94-101.

Pettinati, H.M., Sugarman, A.A., DiDonato, N., & Maurer, H.S. (1982). The natural history of alcoholism over four years after treatment. Journal of Studies on Alcohol, 43, 201-215.


Polich, J.M., Armor, D.J., & Braiker, H.B (1980). Patterns of alcoholism overfour years. Santa Monica, CA: Rand Corporation.

Powell, B.J. Jr Viamontes, J. (1974). Factors affecting attendance at an alcoholic day hospital. British Jour- nal of Addictions, 69, 339-342.

Ryback, R.S., Longabaugh, R., & Fowler, D.R. (1981). Theproblem oriented record inpsychiatryand mental health care. (Rev. ed.). New York: Grune and Stratton.

Seizer, M.L. (1971). The Michigan Alcoholism Screening Test: The quest for a new diagnostic instrument. American Journal of Psychiatry, 127, 1653-1658.

Sobell, M.B., & Sobell, L.C. (1976). Second year treatment outcome of alcoholics treated by individualized behavior therapy: Results. Behaviour Research & Therapy, 14, 195-215.

Washburn, S.L., Vannicelli, M., Longabaugh, R., & Scheff, B. (1976). A controlled comparison of psychiatric day treatment and inpatient hospitalization. Journal of Consulting and Clinical Psychology, 44, 665-675.

Wilder, J.F., Levin, G., & Zwerling, I. (1966). A two-year follow-up evaluation of acute psychotic patients treated in a day hospital. American Journal of Psychiatry, 122, 1095-l 101.

Effectiveness of alcoholism treatment in partial versus inpatient settings: Twenty-four month outcomes

Documents