British Journal of Addiction (1992) 87, 1457-1467
RESEARCH REPORT
What happens to drinking after therapeutic intervention? F. DUCKERT,! A. AMUNDSEN^ & J. JOHNSEN^ ^National Institute for Alcohol and Drug Research, Oslo, Norway & ^Incognito Clinic, Oslo, Norway
Abstract Eighty-four men and 51 women who had been recruited to treatment through newspaper advertisements were matched pairwise and randomly assigned to either brief counselling or short-term group therapy. All participants were followed-up after 3, 9, 15 and 21 months by personal interviews and medical examinations. The average reduction in alcohol consumption was greatest to start with in the group attending group therapy, but this group experienced greater set-back during the following period. At 21 months there were no significant differences between the two groups as regards amount of alcohol consumed, GGT-levels, deaths or number of alcohol-related hospitalizations. The sample as a whole maintained a significant reduction in alcohol consumption throughout the follow-up period. At 21 months half of the sample had reduced their consumption with more than 50%, one-fourth had reduced 20-50%, one-tenth changed less than 20% and only 18% of the sample had increased their consumption with more than 20% compared with their baseline consumption. The reductions took place mainly by way of less frequent drinking, fewer episodes of heavy drinking and reduced consumption during weeks of moderate drinking. When heavy drinking episodes did occur, there were only minor changes in the amounts consumed.
Introduction With few exceptions, evaluation studies of the treatment of alcohol problems include little information about specific changes in drinking patterns of problem drinkers. Often the actual drinking behaviour is not described at all. When the drinking has been registered it has been usual to define categories as abstinent/non-abstinent (Ito et al., 1988) and to state consumption during typical days or weeks of drinking (Chick et al., 1988) or whether the person drank more or less than specified amounts, for instance, more or less than 3 oz pure ethanol per day (Maisto et al., 1982), or the number
Reprint requests should be addressed to Fanny Duckert, National Institute for Alcohol and Drug Research, Dannevigsveien 10, N-0463 Oslo 4, Norway.
of weeks including drinking episodes of more or less than 100/200 g alcohol per day of drinking (Orford & Edwards, 1977). These are rather static measures, however, and do not include intra-individual variations in drinking behaviour. Drinking patterns among problem drinkers are not uniform and may be more complicated than usually described. For instance, many vacillate between different types of drinking behaviour. In one study of hospitalized and outpatient problem drinkers, between 74% and 96% reported that their drinking repertoire included moderate drinking (Duckert, 1988). The amounts of alcohol actually consumed during 'moderate' and 'heavy' drinking episodes also varied considerably. Reported consumption during a typical week of moderation varied from 0.02-4.2 litres pure ethanol (men) and
1457
1458
F. Duckert et al.
0.02-2.1 litres (women). In a typical week of heavy drinking the amounts of alcohol drunk, varied from 0.3-6.7 litres (men) and 0.2-3.8 litres (women). These differences should be of interest in relation to criteria for change. For instance, after treatment one person may still drink more than is considered good for him, but having reduced his consumption down to half of the previous amount, he is quite satisfied with the result himself, while another person consumes small amounts and seemingly is more successful, but in reality is unchanged, since he had been drinking little to begin with. The necessity of more precise descriptions of pretreatment drinking has been pointed out by several researchers (i.e. Cooper et al, 1980; McCrady et al, 1984). Assessment of relative changes should be an important part of the criteria for improvement. Variations in drinking behaviour might also be clinically interesting. For instance, are heavy drinking and moderate drinking influenced by the same therapeutic techniques, or are they of a different nature and therefore responsive to different techniques? The present study attempts to compare two different methods of brief therapy: individual counselling and group therapy, with special focus on changes in drinking behaviour. In both alternatives the participants were allowed to choose their own goal for their future relationship to alcohol, in order to study how many would then choose abstinence and how many would actually stay abstinent during the project period. It was also important to study actual changes in drinking patterns over time, compared with the situation prior to entering treatment. Of special interest were the relationship between heavy and moderate drinking behaviour. It was also of interest to study differences between the sexes in regard to changes in drinking pattern, goal achievement, etc.
Methods The study had an experimental pre-test-post-test design with two matched groups, randomly assigned to different treatment alternatives.
Subjects and assessment procedure The 84 male and 51 female participants in the study were recruited through newspaper advertisements headed: "Do you have alcohol problems and would like to do something about them?"
Respondents received initial information about the project over the phone. They were told that this was a research project, to test different treatment techniques. Participants would undergo a physiological and psychological examination. Afterwards they would be given a treatment regime and would be followed up for about two years. They would receive all treatment as outpatients. Criteria for acceptance into the project were that the applicants were resident in the Oslo area, had a stable job and housing, and were willing to participate throughout the whole project period. Furthermore, they should not have serious disorders (blindness, deafness, psychosis, etc.), which would hamper participation in group activities. Previous treatment for alcohol problems did not exclude participation. If the respondents, after receiving information were still interested, and if then they fulfilled the criteria, they were given an appointment for a personal assessment some days later. Participants were interviewed by a trained clinical psychologist. Special attention was given to the drinking pattern during the last 12 months. A modified Time-Line technique (Maisto et al, 1982) was used. This was done by means of a calendar on which the number of weeks of abstinence, of moderate drinking and heavy drinking were plotted. Afterwards the subjects were asked to describe their day-to-day drinking during a representative week of moderate and heavy drinking. What was 'moderate' and 'heavy' drinking was defined by the subjects themselves. The participants were also medically examined and asked to fill in two questionaires: Severity of Alcohol Dependency Questionnaire (SADQ) (Stockwell et al, 1983) and Short Michigan Alcoholism Screening Test (S-MAST) (Selzer et al, 1975). Blood specimens were also collected and analyzed for alcohol, drugs and the biological marker: GGT (gamma-glutamyl transpeptidase) (Reyes & Miller, 1980). After initial examination, the subjects were stratified for sex, age, alcohol consumption during last year, level of dependency and status of the GGT tests. Each variable was dichotomized as near the median as possible. Within the resulting groups, lots were drawn for pairwise randomization into one of the two intervention alternatives. In case of uneven numbers, a person from a nearby group was used. The procedure resulted in homogenous groups for all background variables. Afterwards they were also compared as regards various background variables not used in the randomizing process, such as
Drinking after therapeutic intervention education, housing situation and marital status. No significant differences were found between the two groups in these respects. All participants were given a new appointment, when they received feedback from the initial examination. They were also informed about the treatment alternative they were to join. An effort was made to present the two alternatives as equally significant and advantageous for the individual. The persons in the counselling group received an extended session of advice. The persons in the group therapy alternative received only brief advice before entering the groups. The groups were single-sexed and consisted of 9-11 participants who met weekly for lj-hour sessions for 12 weeks. The choice of treatment goal did not affect group membership. The main theme of the group therapy was systematic control training and training in alternative behaviours to drinking. The persons in the counselling alternative received a total of about 7 hours individual counselling. The persons in the group therapy alternative received less individual counselling, since the follow-up sessions were also conducted mainly in groups (apart from the medical examination and the follow-up interview). They received a total of about 25 hours therapeutic intervention. The main differences between the two alternatives were the number of therapeutic meetings and the presence/absence of other group members. Another difference was that the participants in counselling received the same intensity and type of attention throughout the whole project period, while the participants in group therapy experienced a decline in the frequency of contact after the end of therapy period. Both alternatives were based upon social learning theory, assuming problem drinking to be learned and modifiable behaviour (Miller & Mastria, 1977; Miller & Munoz, 1982). Therefore all participants received some basic information about control training and were given advice about how to attain their individual goal—also irrespectively of their goal was one of abstinence or_ reduction. The reason for this was the assumption that there is not a dramatic difference between no drinking and reduced consumption. Common problems have to be solved in both instances, for example: overcoming internal urges to drink, overcoming external pressures to drink, hindering a 'slip' to grow into a drinking binge and preventing serious relapses. Control techniques might thus be useful tools, also for the abstainers. The participants were followed up after 3, 9, 15
1459
and 21 months—each time by individual, personal interviews, medical examinations and a new session of advice (in groups for the group participants, individually for the rest). After each follow-up the participants were sent a written feedback report by mail, including comments on the present situation compared with the situation at the previous examinations. All procedures were carried out by the same team who consisted of one clinical psychologist, one physician, one medical nurse and one psychology student. This means that the project personnel were not blind to the conditions, something that of course increased the risk of biases. The project administrators were strongly aware of this risk and also were very concerned with giving all participants the best possible service, irrespectively of which treatment alternative they were subject to.
Statistical procedures At end of the follow-up period, information about the drinking status and the social situation of the whole group was collected and analyzed. The analyses of relative changes in drinking patterns over time were restricted to the persons who had participated in all follow-ups. These persons were compared for age, education, work situation, duration of problem drinking, alcohol consumption, GGT-levels and scores on SADQ and S-MAST with the ones who had dropped out of one or more follow-ups. Mean scores for both SADQ and S-MAST ( - 4 . 8 on SADQ and - 1 . 1 on SMAST) were lower among the group who had attended all follow-ups than among the participants who had lost one or more follow-ups. Despite these differences we chose to continue this method of analyzing relative changes, since it was important that the same persons participated all the time. The original design was one of matched-pair randomization. However, because of the effects of drop-outs, the subjects were treated statistically as two unrelated samples. Otherwise, if one member of the pair dropped out the whole pair would have had to be excluded from statistical analyses. This would have reduced the number of pairs to only 27. Furthermore, the correlations within the pairs were low. Each follow-up was analysed separately and compared with the situation at the previous followup and the situation during the last year before entering the project. Changes from one period to the next were tested for statistical significance by Wilcoxon matched-pairs signed-ranks tests. Differ-
1460
P. Duckert et al.
ences between the two kinds of intervention were tested by chi-square and Mann-Whitney U tests for each sex separately. The kind of test used is stated in each table.
Results At intake, the mean age of the men was 43.1 years and of the women 42.8 years. Mean duration of drinking problems was 9.7 years for the men and 6.9 years for the women. Mean score on SADQ was 26.0 for the men and 18.6 for the women. About onethird of the men and one-fifth of the women had been treated before for their alcohol problems, most as outpatients. As a group, the participants in this study were middle class and fairly well educated, and compared with groups of clients served in the ordinary alcoholism treatment units in Oslo, had relatively moderate alcohol problems. The sample is described in more detail elsewhere (Duckert, 1988). All the women and 96% of the men completed the group treatment. The number of persons attending the different follow-ups varied somewhat: at 3 months: 90%, at 9 months: 83%, at 15 months: 77% and at 21 months: 89%. It was not uncommon for a person to drop-out of one follow-up, but to reappear in the next one. Only 3% were permanently lost.
Ninety-two persons (68%) participated in all follow-ups. It turned out that official registration varied a great deal between the different offices in question. As a consequence, we were able to obtain complete registration information for only 92% of the total group.
Status at the end of the project period
At intake, one-fourth of the total sample expressed a wish for abstinence, while more than half wanted to reduce their alcohol consumption without abstaining completely. At end of project, there were fewer persons in almost all subgroups who wished to become or remain abstinent in future. The number of persons who actually remained abstinent varied over time, and was highest in the beginning. There were large fiuctuations as regards which persons were abstinent in the different periods. Only two persons were totally abstinent throughout the whole period. At the 21 months follow-up 69% of the participants had achieved a reduction in relation to consumption during the year before entering the project. This reduction was statistically significant in all subgroups. However, there were no significant
Table 1. Expressed goal of treatment at intake, percentage of the groups who were abstinent atfollow-ups, percentage of the group wanting abstinence, and drinking status at 21 months' follow-up Women
Men Total (135)
Difference between counsel/group
Group (42)
Counsel. (24)
Group (27)
Men
Women
26 50 24
25 54 21
11 67 27
NS
p < 0.05
Percentage of group who had been abstinent during each follow-up period: 0-3 Months 16 21 17 19 3-9 Months 6 8 6 5 9-15 Months 9 12 10 9 15-21 Months 10 10 9 9
4 0 0 4
NS NS NS NS
p < 0.05 NS
Percentage of group who wanted abstinence as a goal at end of project: 19 26 21
13
11
NS
NS
Changes in consumption at end offollow-up period (% of group): More than 50% reduction 51 41 55 20-49% reduction 20 18 21 0-19% change 10 13 6 20-49% increase 4 8 3 More than 50% increase 14 21 15
61 17 13 0 9
54 25 8 0 13
NS
NS
n
Counsel. (42)
Expressed goal for alcohol use at intake (% of group): Abstinence 25 31 Reduction 56 55 Unchanged/Unspecified 19 14
Test used: chi-square.
NS NS
Drinking after therapeutic intervention differences between participants in the two therapeutic alternatives. One half of the total sample had decreased their alcohol consumption with more than 50%. Twenty per cent had reduced consumption between 20% and 50%. Ten per cent had changed less than 20% and 18% had increased consumption with more than 20%, in relation to alcohol consumption prior to entering the project. As to degree of satisfaction with the outcome, in general, a higher proportion of women than men reported 'better than expected'. But if we combine 'better than expected' with 'as expected', the rate of satisfaction was higher for men in group therapy and for women in individual counselling (62% for both subgroups), but lower for men in individual counselling (45%). Forty-eight per cent of the women who had participated in group therapy were satisfied with the outcome. The most frequently reported subjectively experienced changes in drinking pattern in all subgroups was drinking less often and more frequently and longer periods of abstinence. The second most often reported change, was lower consumption during drinking episodes. There were no significant differences between groups as regards the kind of changes reported in drinking pattern. However, in general the men who attended group therapy reported the most positive changes and the men who received individual counselling the fewest positive changes of all. The women who received individual counselling reported more positive changes than the women who attended group therapy.
The development in drinking pattern during the project period Table 2 demonstrates that at the first follow-up, all groups (except the women who had received individual counselling) had reduced their alcohol consumption significantly and afterwards maintained a significant reduction throughout the whole project period. At the second follow-up (9 months), the women who had received individual counselling also had reduced their consumption significantly and continued to reduce, so that at the end of the project period these women had achieved a mean reduction of 51 % in relation to their baseline consumption. For the men who had received individual counselling the pattern fiuctuated more from follow-up to followup. At the 21 month follow-up this group had achieved a mean reduction of 36%.
1461
At the first follow-up, the group therapy participants showed the greatest decrease in consumption (men: 57%, women: 60%), compared with the others (men: 52%, women: 25%). The difference between the two treatment alternatives was significant for the women, but not for the men. At the next followup, however, consumption had increased again among the participants in group therapy to a higher level than at the first follow-up. Afterwards, this group showed a tendency to stabilize or reduce consumption again (but the changes in consumption registered at later follow-ups were not significant). GGT-levels were reduced in all groups in the beginning, but (except among the women who received individual counselling) increased again after the next follow-up period. The fiuctuations over time had a tendency to follow the same trend as the self-reported drinking. At the end of the observation period, there was a general reduction in GGT-levels for all subgroups, compared with the situation at intake. The correlation between alcohol consumption and GGT-levels was as high, or a little higher, than usually found in studies of this kind (Chick et al., 1981; Bernadt et al., 1982; SanchezCraig & Annis, 1982; Gjerde et al., 1987).
Specific changes in the drinking pattern Tables 3-4 indicate that for the group as a whole there was a significant reduction in time spent on heavy drinking and an increase in time spent abstinent throughout the whole project period, even if there were some fiuctuations from follow-up to follow-up. The share of the participants who reported episodes of heavy drinking was reduced from 63% the year before entrance into the project to 28% at first follow-up. Afterwards the percentage increased somewhat: 32% at 9 months, 39% at 15 months and 42% at 21 months. The tendency to decrease time spent on heavy drinking was significant only for the counselling receivers (except on the last follow-up (men) and on the second followup (women)). The time spent on moderate drinking did not change. The number of persons reporting moderate drinking remained high throughout the observation period for all groups, especially for the women. There was a general tendency for all groups to decrease consumption during weeks of moderate drinking. The decreases were largest at the first follow-up (except for the women who received individual counselling). As mentioned, the number of persons reporting heavy drinking decreased.
1462
F. Duckert et al.
c/l
00
(ooooocoooooeo
z z zzzzzzz 4J
I
il Q
^
C/1
00
C/3t/5t/5(/3C/J(y3(/5C/2
(ZIOOMOOOOCZlOO
z zzzzzzz
z zzzzzzzz
o •S
CO
« I2Z2Z2Z '
I (
I 00
o "^ qKq9(NgTi.j^
RESEARCH REPORT
What happens to drinking after therapeutic intervention? F. DUCKERT,! A. AMUNDSEN^ & J. JOHNSEN^ ^National Institute for Alcohol and Drug Research, Oslo, Norway & ^Incognito Clinic, Oslo, Norway
Abstract Eighty-four men and 51 women who had been recruited to treatment through newspaper advertisements were matched pairwise and randomly assigned to either brief counselling or short-term group therapy. All participants were followed-up after 3, 9, 15 and 21 months by personal interviews and medical examinations. The average reduction in alcohol consumption was greatest to start with in the group attending group therapy, but this group experienced greater set-back during the following period. At 21 months there were no significant differences between the two groups as regards amount of alcohol consumed, GGT-levels, deaths or number of alcohol-related hospitalizations. The sample as a whole maintained a significant reduction in alcohol consumption throughout the follow-up period. At 21 months half of the sample had reduced their consumption with more than 50%, one-fourth had reduced 20-50%, one-tenth changed less than 20% and only 18% of the sample had increased their consumption with more than 20% compared with their baseline consumption. The reductions took place mainly by way of less frequent drinking, fewer episodes of heavy drinking and reduced consumption during weeks of moderate drinking. When heavy drinking episodes did occur, there were only minor changes in the amounts consumed.
Introduction With few exceptions, evaluation studies of the treatment of alcohol problems include little information about specific changes in drinking patterns of problem drinkers. Often the actual drinking behaviour is not described at all. When the drinking has been registered it has been usual to define categories as abstinent/non-abstinent (Ito et al., 1988) and to state consumption during typical days or weeks of drinking (Chick et al., 1988) or whether the person drank more or less than specified amounts, for instance, more or less than 3 oz pure ethanol per day (Maisto et al., 1982), or the number
Reprint requests should be addressed to Fanny Duckert, National Institute for Alcohol and Drug Research, Dannevigsveien 10, N-0463 Oslo 4, Norway.
of weeks including drinking episodes of more or less than 100/200 g alcohol per day of drinking (Orford & Edwards, 1977). These are rather static measures, however, and do not include intra-individual variations in drinking behaviour. Drinking patterns among problem drinkers are not uniform and may be more complicated than usually described. For instance, many vacillate between different types of drinking behaviour. In one study of hospitalized and outpatient problem drinkers, between 74% and 96% reported that their drinking repertoire included moderate drinking (Duckert, 1988). The amounts of alcohol actually consumed during 'moderate' and 'heavy' drinking episodes also varied considerably. Reported consumption during a typical week of moderation varied from 0.02-4.2 litres pure ethanol (men) and
1457
1458
F. Duckert et al.
0.02-2.1 litres (women). In a typical week of heavy drinking the amounts of alcohol drunk, varied from 0.3-6.7 litres (men) and 0.2-3.8 litres (women). These differences should be of interest in relation to criteria for change. For instance, after treatment one person may still drink more than is considered good for him, but having reduced his consumption down to half of the previous amount, he is quite satisfied with the result himself, while another person consumes small amounts and seemingly is more successful, but in reality is unchanged, since he had been drinking little to begin with. The necessity of more precise descriptions of pretreatment drinking has been pointed out by several researchers (i.e. Cooper et al, 1980; McCrady et al, 1984). Assessment of relative changes should be an important part of the criteria for improvement. Variations in drinking behaviour might also be clinically interesting. For instance, are heavy drinking and moderate drinking influenced by the same therapeutic techniques, or are they of a different nature and therefore responsive to different techniques? The present study attempts to compare two different methods of brief therapy: individual counselling and group therapy, with special focus on changes in drinking behaviour. In both alternatives the participants were allowed to choose their own goal for their future relationship to alcohol, in order to study how many would then choose abstinence and how many would actually stay abstinent during the project period. It was also important to study actual changes in drinking patterns over time, compared with the situation prior to entering treatment. Of special interest were the relationship between heavy and moderate drinking behaviour. It was also of interest to study differences between the sexes in regard to changes in drinking pattern, goal achievement, etc.
Methods The study had an experimental pre-test-post-test design with two matched groups, randomly assigned to different treatment alternatives.
Subjects and assessment procedure The 84 male and 51 female participants in the study were recruited through newspaper advertisements headed: "Do you have alcohol problems and would like to do something about them?"
Respondents received initial information about the project over the phone. They were told that this was a research project, to test different treatment techniques. Participants would undergo a physiological and psychological examination. Afterwards they would be given a treatment regime and would be followed up for about two years. They would receive all treatment as outpatients. Criteria for acceptance into the project were that the applicants were resident in the Oslo area, had a stable job and housing, and were willing to participate throughout the whole project period. Furthermore, they should not have serious disorders (blindness, deafness, psychosis, etc.), which would hamper participation in group activities. Previous treatment for alcohol problems did not exclude participation. If the respondents, after receiving information were still interested, and if then they fulfilled the criteria, they were given an appointment for a personal assessment some days later. Participants were interviewed by a trained clinical psychologist. Special attention was given to the drinking pattern during the last 12 months. A modified Time-Line technique (Maisto et al, 1982) was used. This was done by means of a calendar on which the number of weeks of abstinence, of moderate drinking and heavy drinking were plotted. Afterwards the subjects were asked to describe their day-to-day drinking during a representative week of moderate and heavy drinking. What was 'moderate' and 'heavy' drinking was defined by the subjects themselves. The participants were also medically examined and asked to fill in two questionaires: Severity of Alcohol Dependency Questionnaire (SADQ) (Stockwell et al, 1983) and Short Michigan Alcoholism Screening Test (S-MAST) (Selzer et al, 1975). Blood specimens were also collected and analyzed for alcohol, drugs and the biological marker: GGT (gamma-glutamyl transpeptidase) (Reyes & Miller, 1980). After initial examination, the subjects were stratified for sex, age, alcohol consumption during last year, level of dependency and status of the GGT tests. Each variable was dichotomized as near the median as possible. Within the resulting groups, lots were drawn for pairwise randomization into one of the two intervention alternatives. In case of uneven numbers, a person from a nearby group was used. The procedure resulted in homogenous groups for all background variables. Afterwards they were also compared as regards various background variables not used in the randomizing process, such as
Drinking after therapeutic intervention education, housing situation and marital status. No significant differences were found between the two groups in these respects. All participants were given a new appointment, when they received feedback from the initial examination. They were also informed about the treatment alternative they were to join. An effort was made to present the two alternatives as equally significant and advantageous for the individual. The persons in the counselling group received an extended session of advice. The persons in the group therapy alternative received only brief advice before entering the groups. The groups were single-sexed and consisted of 9-11 participants who met weekly for lj-hour sessions for 12 weeks. The choice of treatment goal did not affect group membership. The main theme of the group therapy was systematic control training and training in alternative behaviours to drinking. The persons in the counselling alternative received a total of about 7 hours individual counselling. The persons in the group therapy alternative received less individual counselling, since the follow-up sessions were also conducted mainly in groups (apart from the medical examination and the follow-up interview). They received a total of about 25 hours therapeutic intervention. The main differences between the two alternatives were the number of therapeutic meetings and the presence/absence of other group members. Another difference was that the participants in counselling received the same intensity and type of attention throughout the whole project period, while the participants in group therapy experienced a decline in the frequency of contact after the end of therapy period. Both alternatives were based upon social learning theory, assuming problem drinking to be learned and modifiable behaviour (Miller & Mastria, 1977; Miller & Munoz, 1982). Therefore all participants received some basic information about control training and were given advice about how to attain their individual goal—also irrespectively of their goal was one of abstinence or_ reduction. The reason for this was the assumption that there is not a dramatic difference between no drinking and reduced consumption. Common problems have to be solved in both instances, for example: overcoming internal urges to drink, overcoming external pressures to drink, hindering a 'slip' to grow into a drinking binge and preventing serious relapses. Control techniques might thus be useful tools, also for the abstainers. The participants were followed up after 3, 9, 15
1459
and 21 months—each time by individual, personal interviews, medical examinations and a new session of advice (in groups for the group participants, individually for the rest). After each follow-up the participants were sent a written feedback report by mail, including comments on the present situation compared with the situation at the previous examinations. All procedures were carried out by the same team who consisted of one clinical psychologist, one physician, one medical nurse and one psychology student. This means that the project personnel were not blind to the conditions, something that of course increased the risk of biases. The project administrators were strongly aware of this risk and also were very concerned with giving all participants the best possible service, irrespectively of which treatment alternative they were subject to.
Statistical procedures At end of the follow-up period, information about the drinking status and the social situation of the whole group was collected and analyzed. The analyses of relative changes in drinking patterns over time were restricted to the persons who had participated in all follow-ups. These persons were compared for age, education, work situation, duration of problem drinking, alcohol consumption, GGT-levels and scores on SADQ and S-MAST with the ones who had dropped out of one or more follow-ups. Mean scores for both SADQ and S-MAST ( - 4 . 8 on SADQ and - 1 . 1 on SMAST) were lower among the group who had attended all follow-ups than among the participants who had lost one or more follow-ups. Despite these differences we chose to continue this method of analyzing relative changes, since it was important that the same persons participated all the time. The original design was one of matched-pair randomization. However, because of the effects of drop-outs, the subjects were treated statistically as two unrelated samples. Otherwise, if one member of the pair dropped out the whole pair would have had to be excluded from statistical analyses. This would have reduced the number of pairs to only 27. Furthermore, the correlations within the pairs were low. Each follow-up was analysed separately and compared with the situation at the previous followup and the situation during the last year before entering the project. Changes from one period to the next were tested for statistical significance by Wilcoxon matched-pairs signed-ranks tests. Differ-
1460
P. Duckert et al.
ences between the two kinds of intervention were tested by chi-square and Mann-Whitney U tests for each sex separately. The kind of test used is stated in each table.
Results At intake, the mean age of the men was 43.1 years and of the women 42.8 years. Mean duration of drinking problems was 9.7 years for the men and 6.9 years for the women. Mean score on SADQ was 26.0 for the men and 18.6 for the women. About onethird of the men and one-fifth of the women had been treated before for their alcohol problems, most as outpatients. As a group, the participants in this study were middle class and fairly well educated, and compared with groups of clients served in the ordinary alcoholism treatment units in Oslo, had relatively moderate alcohol problems. The sample is described in more detail elsewhere (Duckert, 1988). All the women and 96% of the men completed the group treatment. The number of persons attending the different follow-ups varied somewhat: at 3 months: 90%, at 9 months: 83%, at 15 months: 77% and at 21 months: 89%. It was not uncommon for a person to drop-out of one follow-up, but to reappear in the next one. Only 3% were permanently lost.
Ninety-two persons (68%) participated in all follow-ups. It turned out that official registration varied a great deal between the different offices in question. As a consequence, we were able to obtain complete registration information for only 92% of the total group.
Status at the end of the project period
At intake, one-fourth of the total sample expressed a wish for abstinence, while more than half wanted to reduce their alcohol consumption without abstaining completely. At end of project, there were fewer persons in almost all subgroups who wished to become or remain abstinent in future. The number of persons who actually remained abstinent varied over time, and was highest in the beginning. There were large fiuctuations as regards which persons were abstinent in the different periods. Only two persons were totally abstinent throughout the whole period. At the 21 months follow-up 69% of the participants had achieved a reduction in relation to consumption during the year before entering the project. This reduction was statistically significant in all subgroups. However, there were no significant
Table 1. Expressed goal of treatment at intake, percentage of the groups who were abstinent atfollow-ups, percentage of the group wanting abstinence, and drinking status at 21 months' follow-up Women
Men Total (135)
Difference between counsel/group
Group (42)
Counsel. (24)
Group (27)
Men
Women
26 50 24
25 54 21
11 67 27
NS
p < 0.05
Percentage of group who had been abstinent during each follow-up period: 0-3 Months 16 21 17 19 3-9 Months 6 8 6 5 9-15 Months 9 12 10 9 15-21 Months 10 10 9 9
4 0 0 4
NS NS NS NS
p < 0.05 NS
Percentage of group who wanted abstinence as a goal at end of project: 19 26 21
13
11
NS
NS
Changes in consumption at end offollow-up period (% of group): More than 50% reduction 51 41 55 20-49% reduction 20 18 21 0-19% change 10 13 6 20-49% increase 4 8 3 More than 50% increase 14 21 15
61 17 13 0 9
54 25 8 0 13
NS
NS
n
Counsel. (42)
Expressed goal for alcohol use at intake (% of group): Abstinence 25 31 Reduction 56 55 Unchanged/Unspecified 19 14
Test used: chi-square.
NS NS
Drinking after therapeutic intervention differences between participants in the two therapeutic alternatives. One half of the total sample had decreased their alcohol consumption with more than 50%. Twenty per cent had reduced consumption between 20% and 50%. Ten per cent had changed less than 20% and 18% had increased consumption with more than 20%, in relation to alcohol consumption prior to entering the project. As to degree of satisfaction with the outcome, in general, a higher proportion of women than men reported 'better than expected'. But if we combine 'better than expected' with 'as expected', the rate of satisfaction was higher for men in group therapy and for women in individual counselling (62% for both subgroups), but lower for men in individual counselling (45%). Forty-eight per cent of the women who had participated in group therapy were satisfied with the outcome. The most frequently reported subjectively experienced changes in drinking pattern in all subgroups was drinking less often and more frequently and longer periods of abstinence. The second most often reported change, was lower consumption during drinking episodes. There were no significant differences between groups as regards the kind of changes reported in drinking pattern. However, in general the men who attended group therapy reported the most positive changes and the men who received individual counselling the fewest positive changes of all. The women who received individual counselling reported more positive changes than the women who attended group therapy.
The development in drinking pattern during the project period Table 2 demonstrates that at the first follow-up, all groups (except the women who had received individual counselling) had reduced their alcohol consumption significantly and afterwards maintained a significant reduction throughout the whole project period. At the second follow-up (9 months), the women who had received individual counselling also had reduced their consumption significantly and continued to reduce, so that at the end of the project period these women had achieved a mean reduction of 51 % in relation to their baseline consumption. For the men who had received individual counselling the pattern fiuctuated more from follow-up to followup. At the 21 month follow-up this group had achieved a mean reduction of 36%.
1461
At the first follow-up, the group therapy participants showed the greatest decrease in consumption (men: 57%, women: 60%), compared with the others (men: 52%, women: 25%). The difference between the two treatment alternatives was significant for the women, but not for the men. At the next followup, however, consumption had increased again among the participants in group therapy to a higher level than at the first follow-up. Afterwards, this group showed a tendency to stabilize or reduce consumption again (but the changes in consumption registered at later follow-ups were not significant). GGT-levels were reduced in all groups in the beginning, but (except among the women who received individual counselling) increased again after the next follow-up period. The fiuctuations over time had a tendency to follow the same trend as the self-reported drinking. At the end of the observation period, there was a general reduction in GGT-levels for all subgroups, compared with the situation at intake. The correlation between alcohol consumption and GGT-levels was as high, or a little higher, than usually found in studies of this kind (Chick et al., 1981; Bernadt et al., 1982; SanchezCraig & Annis, 1982; Gjerde et al., 1987).
Specific changes in the drinking pattern Tables 3-4 indicate that for the group as a whole there was a significant reduction in time spent on heavy drinking and an increase in time spent abstinent throughout the whole project period, even if there were some fiuctuations from follow-up to follow-up. The share of the participants who reported episodes of heavy drinking was reduced from 63% the year before entrance into the project to 28% at first follow-up. Afterwards the percentage increased somewhat: 32% at 9 months, 39% at 15 months and 42% at 21 months. The tendency to decrease time spent on heavy drinking was significant only for the counselling receivers (except on the last follow-up (men) and on the second followup (women)). The time spent on moderate drinking did not change. The number of persons reporting moderate drinking remained high throughout the observation period for all groups, especially for the women. There was a general tendency for all groups to decrease consumption during weeks of moderate drinking. The decreases were largest at the first follow-up (except for the women who received individual counselling). As mentioned, the number of persons reporting heavy drinking decreased.
1462
F. Duckert et al.
c/l
00
(ooooocoooooeo
z z zzzzzzz 4J
I
il Q
^
C/1
00
C/3t/5t/5(/3C/J(y3(/5C/2
(ZIOOMOOOOCZlOO
z zzzzzzz
z zzzzzzzz
o •S
CO
« I2Z2Z2Z '
I (
I 00
o "^ qKq9(NgTi.j^
