The Evidence Base

It has not proved possible to carry out formal meta-analysis successfully (i.e., with pooling of data from a range of similar studies and estimations of effect sizes) for alcohol problems treatment as a whole. This is because the science of treatment in this field has not matured sufficiently to have arrived at standard forms of treatment modalities and control conditions, common use of outcome measures, consensually agreed reporting of client characteristics and other features necessary for formal meta-analysis to proceed. Instead, it is necessary to rely on the 'box-score' method in which individual studies are rated for the degree to which they either support or do not support a specific modality and an aggregate effectiveness score is arrived at by summing these ratings. This method of data synthesis has been criticised (for example, by Finney, 2000) but it is probably the best source of information available at present.

The leading and most recent box-score synthesis is an update on the so-called Mesa Grande ('large table' in Spanish) by Miller, Wilbourne & Hettema (2003) from the University of New Mexico. This included 381 controlled trials, mainly randomised controlled trials (RCTs), of different types or intensities of treatment or the same type of treatment with and without the addition of a special therapeutic component. Two independent raters judged the methodological quality of all studies included in the Mesa Grande on 12 dimensions, resulting in a methodological quality score (MQS) for each study. Outcome logic scores (OLS), reflecting the degree to which inferences about the effectiveness of the treatment could properly be derived from the study design, were arrived at by a similar rating process and resulted in a classification of each study as providing strong positive evidence (+2), positive evidence (+1), negative evidence (-1) or strong negative evidence (-2) for a particular treatment modality. The MQS and OLS were then multiplied to arrive at a weighting of the study's contribution to the evidence on treatment outcome by its methodological quality. These products were then summed across all studies bearing on the effectiveness of a specific treatment modality, resulting in the cumulative evidence score (CES) for each modality. Some interventions might be effective among problem drinkers identified in community samples but not in clinical samples, so separate ratings were calculated among those studies confined to clinical samples (see Table 15.1).

Table 15.1 shows a selection of treatment modalities from the Mesa Grande, with their respective CESs and other data. The selection is based on the 10 highest rated modalities and the five lowest but also includes other modalities for illustrative purposes. Two modalities from a separate section of the Mesa Grande devoted to treatments with only one or two investigations have been inserted in Table 15.1. The Mesa Grande will be the main source of evidence on treatment outcome in this chapter. The complete table and fuller details, including a listing of all studies comprising the Mesa Grande, will be found in Miller, Wilbourne & Hettema (2003).

Table 15.1 A selection of treatment modalities from the Mesa Grande

%+ Mean Mean % Excellent

Table 15.1 A selection of treatment modalities from the Mesa Grande

%+ Mean Mean % Excellent

Treatment modality

Rank

CES

N

MQS

severity

Brief intervention

i

39G

34

z4

i3.29

2.4z

S3

Motivational enhancement

2

iB9

iB

z2

i2.B3

2.z2

SG

GABA agonist (Acamprosate)

3

ii6

S

iGG

i i.6G

3.BG

2G

Community reinforcement

4.S

i iG

z

B6

i4.GG

3.43

zi

Self-change manual

4.S

i iG

iz

S9

i2.6S

2.S9

S3

(bibliotherapy)

Opiate antagonist (e.g.,

6

iGG

6

B3

ii.33

3.iz

G

Naltrexone)

Behavioural self-control

z

BS

3i

S2

i2.zz

2.9i

S2

training

Behaviour contracting

B

64

S

BG

iG.4G

3.6G

G

Social skills training

9

Sz

2G

SS

iG.9G

3.BG

2S

Marital therapy -

iG

44

9

S6

i2.33

3.44

44

behavioural

Cognitive therapy

i3

2i

iG

4G

iG.GG

3.zG

iG

Aversion therapy, covert

i4.S

iB

B

3B

iG.BB

3.SG

G

sensitisation

Client-centred counselling

iB

S

B

SG

ii.i3

3.3B

i3

Therapeutic community

-4

i

G

4.GG

3.GG

G

Antidipsotropic - Disulfiram

22

-6

2z

44

ii.Gz

3.69

26

Minnesota Model

-ii

i

G

ii.GG

4.GG

1GG

Antidepresssant - SSRI

23

-i6

is

S3

B.6G

2.6z

G

Problem solving

24

-26

4

2S

i2.2S

3.zS

SG

Lithium

2S

-32

z

43

ii.43

3.zi

29

Marital therapy -

26

-33

B

3B

i2.2S

3.63

2S

non-behavioral

Group process

2z

-34

3

G

B.GG

2.6z

G

psychotherapy

Functional analysis

2B

-36

3

G

i2.GG

2.6z

33

Relapse prevention

29

-3B

22

36

ii.z3

3.23

3i

Self-monitoring

3G

-39

6

33

i2.GG

3.iz

SG

Twelve-step facilitation

3z

-B2

6

iz

iS.GG

3.6z

B3

Alcoholics Anonymous

3B

-94

z

14

iG.zi

3.i4

29

Anxiolytic medication

39

-9B

is

2z

B.i3

3.4G

G

Milieu therapy

4G

-iG2

i4

2i

iG.B6

3.64

29

Antidepressant medication

42

-iG4

6

G

B.6z

3.iz

G

(non-SSRI)

Relaxation training

44

-iS2

iB

iz

iG.S6

3.G6

iz

Confrontational counselling

4S

-iB3

i2

G

iG.2S

3.GG

33

Psychotherapy

46

-2GZ

i9

i6

iG.B9

3.26

2i

General alcoholism

4z

-2B4

23

9

11.26

3.22

22

counselling

Education (tapes, lectures or

4B

-443

39

i3

9.zz

2.44

CES = Cumulative Evidence Score.

N = Total number of studies evaluating modality.

% + = Percent of studies with positive finding for modality.

Mean MQS = Average methodological quality (0-17) score of studies.

Mean Severity = Average severity rating (1-4) oftreated populations.

% Excellent = Percentage of studies with MQS > 13.

Source: Adapted from Miller, Wilbourne & Hettema (2003).

films)

CES = Cumulative Evidence Score.

N = Total number of studies evaluating modality.

% + = Percent of studies with positive finding for modality.

Mean MQS = Average methodological quality (0-17) score of studies.

Mean Severity = Average severity rating (1-4) oftreated populations.

% Excellent = Percentage of studies with MQS > 13.

Source: Adapted from Miller, Wilbourne & Hettema (2003).

One of the main advantages of conducting a meta-analysis of treatment for a particular disorder is that pooling of data from numerous studies allows a large sample of clients to be considered and reduces the risk of Type 2 error - of accepting the null hypothesis of no effect for a treatment when that hypothesis is false and the treatment is in fact effective. However, the alcohol problems field is remarkable for existence of the largest multi-centre trial of psychosocial treatment ever conducted, not only of treatment for alcohol problems but for any kind of behavioural disorder.

This was Project MATCH (Matching Alcoholism Treatment to Client heterogeneity) (Project MATCH Research Group, 1997a,b; 1998a), which included 1726 individuals seeking treatment for alcohol problems, divided into two parallel but independent clinical trials - an outpatient arm (n = 952) and an aftercare arm (n = 774). Project MATCH was a rigorously designed and executed study and, in view particularly of its size, is guaranteed a special place in the evidence base on treatment for alcohol problems. Its findings will be used here to supplement those of the Mesa Grande.

As its name suggests, the project was designed to test the 'matching hypothesis' that appropriate allocation of clients to the form of treatment best suited to them would produce an overall improvement in outcome of treatment. Three types of psychosocial (i.e. non-drug) treatments were compared:

• cognitive-behavioural therapy (CBT);

• motivational enhancement therapy (MET); and

• twelve-step facilitation therapy (TSF) based on the philosophy and methods of AA and its famous '12-step' recovery programme.

The main conclusion of Project MATCH was that, whereas some clinically valuable client-treatment matches were found, the general matching hypothesis referred to above was not confirmed: matching on a priori grounds did not increase overall treatment effectiveness. The large sample size used in Project MATCH was necessary to provide sufficient statistical power to test interactions in the data between client characteristics and treatment modalities but it also provided very high power for the detection of main effects of treatment and these will be discussed below.

More recently, there has been a multi-centre RCT of alcohol problems treatment in the UK, with a sample size (n = 742) approaching that of Project MATCH or, at least, one of its arms. This is the United Kingdom Alcohol Treatment Trial (UKATT) (UKATT Research Team, 2001), which compared the effects of Social Behaviour and Network Therapy (SBNT) (Copello et al., 2002), a new form of treatment integrating a range of successful or promising social/ interpersonal approaches reported in the treatment literature, with MET (Miller et al., 1992), one of the modalities investigated in Project MATCH. Findings from UKATT will also be discussed in this chapter.

Beat The Battle With The Bottle

Beat The Battle With The Bottle

Alcoholism is something that can't be formed in easy terms. Alcoholism as a whole refers to the circumstance whereby there's an obsession in man to keep ingesting beverages with alcohol content which is injurious to health. The circumstance of alcoholism doesn't let the person addicted have any command over ingestion despite being cognizant of the damaging consequences ensuing from it.

Get My Free Ebook


Post a comment