- Email: [email protected]
Effectiveness of brief alcohol interventions for general practice patients with problematic drinking behavior and comorbid anxiety or depressive disorders
Available online at www.sciencedirect.com
Drug and Alcohol Dependence 94 (2008) 214–220
Effectiveness of brief alcohol interventions for general practice patients with problematic drinking behavior and comorbid anxiety or depressive disorders Janina M. Grothues a,∗ , Gallus Bischof a , Susa Reinhardt a , Christian Meyer b , Ulrich John b , Hans-J¨urgen Rumpf a a
University of L¨ubeck, Department of Psychiatry and Psychotherapy, Research Group S:TEP (Substance Abuse: Treatment, Epidemiology and Prevention), Ratzeburger Allee 160, 23538 L¨ubeck, Germany b Ernst-Moritz Arndt University of Greifswald, Department of Epidemiology and Social Medicine, Walther-Rathenau-Str. 48, 17487 Greifswald, Germany Received 28 June 2007; received in revised form 23 November 2007; accepted 26 November 2007 Available online 22 January 2008
Abstract Background: Brief interventions (BIs) are effective methods to reduce problematic drinking. It is not known, if the effectiveness of BI differs between patients with or without comorbid depression or anxiety disorders. Methods: In a randomized controlled BI study with two intervention groups and one control condition, data were collected from 408 general practice (GP) patients with alcohol use disorders, at-risk drinking or binge drinking. 88 participants were diagnosed with comorbid anxiety and/or depressive disorders. The effectiveness of BI was assessed at a 12-month follow-up in relation to the presence and absence of comorbidity. Reduction of drinking in six ordered categories (g/alcohol) between baseline and follow-up served as the outcome variable. Results: BI were significantly related to reduction of drinking in the non-comorbid (−2.64 g/alcohol vs. −8.61 g/alcohol; p = .03) but not in the comorbid subsample (−22.06 g/alcohol vs. −22.09 g/alcohol; p = .76). Compared to non-comorbid participants, a significantly higher reduction of drinking was found for comorbid individuals (−6.55 g/alcohol vs. −22.08 g/alcohol; p = .01). An ordinal regression analysis revealed comorbidity to be a positive predictor for reduction of drinking (estimator = .594; CI = .175–1.013; p < .01). When entering the variables amount of drinking at baseline, intervention and classification of problematic drinking, these became significant predictors, whereas comorbidity showed only a tendency. Conclusion: BI did not significantly effect a reduction of drinking in comorbid patients. As BI are known to be less effective for dependent drinkers, a larger proportion of dependents among the comorbid might have limited the effectiveness of BI. Future studies with larger sample sizes of comorbid problem drinkers are necessary to confirm the results. © 2008 Elsevier Ireland Ltd. All rights reserved. Keywords: Brief intervention; Alcohol use disorders; Problematic drinking; Comorbidity; Anxiety disorders; Depressive disorders
1. Introduction While the majority of alcohol-related harm is not attributable to drinkers with severe alcohol dependence but to a much larger group of excessive or hazardous drinkers whose consumption exceeds recommended drinking limits (Anderson, 1991), brief interventions (BIs) aim to reduce consumption behavior and related harm. Motivational Interviewing (MI; Miller
∗
Corresponding author. Tel.: +49 451 500 2965; fax: +49 451 500 3480. E-mail address: [email protected] (J.M. Grothues).
0376-8716/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.drugalcdep.2007.11.015
and Rollnick, 1991) poses an effective technique to enhance motivation to change problematic drinking behavior. As the majority of problem drinkers do not seek help (Grant, 1997), primary care settings provide a suitable channel to reach a large number of the target group. A recent Cochrane review provided strong evidence that BI in primary care settings are effective to reduce problematic alcohol consumption (Kaner et al., 2007). However, Kaner et al. (2007) also describe “a clear need to characterise the types of drinkers for whom BI have a positive impact and any subgroups that have not been represented in the trials to date” (p. 3). While high rates of anxiety and depression have been found in various samples of individuals with alcohol
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
use disorders (Kessler et al., 1997; Regier et al., 1990; Wittchen et al., 1996) or at-risk drinking (Bott et al., 2005) as well as ¨ un and Sartorius, 1995), within the primary care setting (Ust¨ little is known about the effectiveness of BI for individuals with problematic drinking behavior and comorbid anxiety and/or depressive disorders. Further investigation is needed to improve pro-active intervention strategies for this subgroup. Regarding the outcome of alcohol specific treatment for comorbid individuals with alcohol use disorders and comorbid depression or anxiety, previous studies have provided heterogeneous results (Greenfield et al., 1998; Hasin and Grant, 2002). On the motivational level, a higher readiness to change problematic drinking behavior has been found for comorbid individuals in an outpatient dual diagnoses sample (Velasquez et al., 1999) and in pro-actively recruited general practice (GP) patients (Grothues et al., 2005). Comorbidity has also been associated with increased utilization of formal help (Lynskey, 1998; Regier et al., 1993). However, previous studies also show increased temptation to drink (Grothues et al., 2005; Velasquez et al., 1999) and lower self-efficacy to abstain from drinking (Grothues et al., 2005). The key objective of this paper is to assess the effectiveness of BI to reduce problematic drinking behavior on grounds of longitudinal data which has been derived from a sample of proactively recruited GP patients with a range of drinking problems and comorbid depressive and/or anxiety disorders. The following aspects are addressed: Does the effectiveness of BI differ according to the presence or absence of comorbid anxiety and/or depressive disorders? Which implication does this yield for the delivery of BI to comorbid individuals who do not seek help for their drinking? 2. Method 2.1. Procedure Within the randomized controlled BI study “Stepped Interventions for Problem Drinkers (SIP)”, data were collected by trained project staff in 81 general practices in the north German city of L¨ubeck and its 46 surrounding communities and also in four practices in the north German city of Kiel during the period 2001 and 2003. A power analysis based on data by Fleming (1997) indicated an 80% statistical power for a sample size of 114 participants per group when applying a 5%, one-way significance level to compare each of the two intervention groups with the control group. Assuming a 25% attrition rate resulted in 150 participants per group. To minimise time for data collection within the practice, the procedure was threefold: screening within the practice and administering telephone diagnostic assessments outside the practice. Additional data which are not subject for this analysis were collected via postal questionnaire. GP patients aged 18–64, attending the practice for a GP consultation, were contacted in the waiting room by a trained study nurse and asked to fill out a screening questionnaire. Patients with a positive screening result (see Section 2.2) were asked for written informed consent to participate further in the study. Study nurses undertook intense training to ensure high quality and standardized procedure. They were individually supervised within the practice setting by study staff and received fortnightly supervision sessions. On average 2 days after screening, participants who had agreed to participate further were sent a questionnaire on alcohol related problems and readiness to change variables. Two to four days after sending the questionnaire, participants were contacted by postgraduate psychologists to partake in a telephone diagnostic baseline assessment of alcohol related disorders and problematic drinking. Patients who had been in alcohol specific treatment within the previous 4 weeks were excluded from the study. Participants, meeting specified criteria for alco-
215
Fig. 1. Stepped care intervention group.
hol dependence, abuse, at-risk drinking or binge drinking, were included in the final study sample and a standardized diagnostic assessment of anxiety and/or depressive disorders was administered. Additional data, including utilization of formal help for drinking problems, were collected during the assessment. The baseline telephone contacts lasted 10–90 min, the average being 30 min. Final study participants were randomly allocated, using coloured cards in sealed, non-transparent envelopes (drawn envelopes were not put back), to either the control group or to one of two intervention groups. These were applied to serve a separate research aim of the study SIP: to assess the cost-effectiveness of a stepped care approach (SC; if drinking reduction below the study criteria for atrisk and binge drinking within the previous 4 weeks had been reached, scheduled counseling sessions were discontinued until the follow-up assessment; see Fig. 1) in comparison to a full care (FC) intervention (all scheduled interventions were applied, regardless of drinking reduction) (for details see Bischof et al., in pressa, in press-b). Participants in the control group were sent a booklet on health behavior; no alcohol related intervention was provided. Directly following the baseline assessment, both intervention groups received a computerized feedback (Bischof et al., in press-a, in press-b). In addition, intervention group 1 (FC) received a standardized number of four BI sessions directly after the baseline assessment, as well as after 1, 3 and 6 months. The scheduled time for each BI session for the FC condition was 30 min. Intervention group 2 (SC) received a maximum of three BI sessions at the time stages of 1, 3 and 6 months after baseline, depending upon success of the previous intervention. The scheduled time for each session in the SC condition was 30–40 min, thus enabling the maximum of all possible interventions to equal the standardized effort applied to the FC group. All counseling sessions were administered by telephone. They were fully based on MI spirit and principles (Miller and Rollnick, 1991) and structured by using the following elements of behavioral change counseling (BCC; Rollnick et al., 1999): decisional balance (routinely) and readiness- and self-efficacy rulers (if indicated). Counseling was carried out by postgraduate psychologists who were trained and experienced in MI, and was conducted according to standardized instruments and protocols including structure, procedure/methods and goals. Fidelity of interventions was monitored using standardized monitoring forms. Weekly supervision and standardized protocols ensured the consistent quality of the interventions and the adherence to an MI-consistent intervention procedure. Follow-up assessments of both intervention groups and the control group were carried out via telephone contact 12 months after the baseline assessment. Follow-ups were undertaken by a different psychologist other than the one conducting the diagnostic assessment and interventions. The allocation was randomized and blinded after the exclusion of the psychologist, who had initially conducted the baseline-assessment and counseling sessions.
2.2. Instruments 2.2.1. Screening within the practice. The Alcohol Use Disorder Identification Test (AUDIT; Saunders et al., 1993), a 10-item self-report questionnaire, was used for the initial screening. Based on data for a German general population sample (Rumpf et al., 2002), a cut-off value of five points was chosen. The retest reliability and validity of the AUDIT proved good combinations of sensitivity (.97 for alcohol dependence, alcohol use disorders (AUDs), and AUDs and/or
216
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
at-risk consumption) and specificity (.88 for alcohol dependence, .92 for AUDs, and .91 for AUDs and/or at-risk drinking) for the cut-off value five based upon analysis of the present GP sample, hence the lowering of the standard cutoff value has been strongly recommended (Dybek et al., 2006). Alternatively, a minimum of two points on the L¨ubeck Alcohol dependence and abuse Screening Test (LAST; Rumpf et al., 1997) indicated positive screening. 2.2.2. Baseline diagnostic assessment via telephone. The Munich Composite International Diagnostic Interview (M-CIDI; Wittchen et al., 1995), which is the German translation of the CIDI (Robins et al., 1988), was used for the baseline diagnostic identification of alcohol abuse or dependence according to the 4th and international version of the Diagnostic and Statistical Manual of psychiatric disorders (DSM-IV; American Psychiatric Association, 1995). Criteria for atrisk drinking were defined as an average consumption of >20/30 g of alcohol per day for women/men within the previous 4 weeks (British Medical Association, 1995) or binge drinking as >60/80 g of alcohol for women/men on at least two occasions within the previous 4 weeks (Babor et al., 1992). The definition of drinking measure used for analyses was defined as the average amount across all drinking days obtained from the information on quantity/frequency of drinking
within the previous 12 months. The measures of drinking for the pre-intervention assessment and follow-up were assessed and calculated according to identical procedures. The following comorbid disorders were assessed with the M-CIDI: major depressive disorder, dysthymic disorder, panic disorder, agoraphobia, social phobia, generalised anxiety disorder and anxiety disorder not otherwise specified. Lifetime utilization of formal help for drinking problems was assessed using the following categories: professional advice/treatment, alcohol detoxification/rehabilitation treatment and self-help group visits. 2.2.3. Follow-up diagnostic assessment via telephone. The follow-up diagnostic assessment of alcohol abuse or dependence was conducted using the M-CIDI. Utilization of formal help for drinking problems within the previous 12 months was assessed using the categories specified at baseline.
2.3. Analysis procedures The differential amount of alcohol consumption (g/alcohol per day) between baseline and follow-up assessments served as the outcome variable.
Fig. 2. CONSORT statement.
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
217
Applying an intention to treat analysis, participants who could not be reached at follow-up were defined as unchanged. Analysis groups 1 (non-comorbid participants) and 2 (participants with comorbid anxiety and/or depressive disorders) were compared. Due to lack of normal distribution of the outcome variable, non-parametric tests were used. Ordinal regression analyses were applied as a multivariate approach to assess the predictive value of comorbidity for drinking reduction in relation to variables, in which comorbid and non-comorbid participants were found to differ. For the ordinal regression analyses, the outcome variable reduction of drinking was split into six equally ordered categories: (1) increased consumption, (2) no reduction, (3) reduction <10 g, (4) reduction <30 g, (5) reduction <60 g, and (6) reduction >60 g.
Table 1 Drinking status of study groups at baseline (CG: control group; SC: stepped care; FC: full care)
3. Results
more often alcohol dependent as opposed to alcohol abusers, at-risk drinkers and binge drinkers (χ2 = 42.1; d.f. = 3; p < .001). At baseline, comorbid individuals had significantly higher mean values of alcohol consumption than non-comorbid participants (75.22 g/alcohol vs. 40.53 g/alcohol; Mann–Whitney U = 111974.000; p < .05). Comorbid participants were significantly more often female (χ2 = 17.0; d.f. = 1; p < .001), but did not differ from non-comorbid individuals in years of schooling (up to 9 years vs. more than 9 years; χ2 = 1.6; d.f. = 1; p = .21) and age (t = 1.5; d.f. = 152.2; p = .14). Lifetime and 12-month utilization of formal help for drinking were significantly higher for comorbid than for non-comorbid individuals (χ2 = 12.95; d.f. = 1; p < .001/χ2 = 31.75; d.f. = 1; p < .001). At 12-month follow-up, 3 (0.7%) participants of the baseline sample described above (n = 408), had died, 27 (6.6%) were not attainable, 4 individuals (1.0%) withdrew their further participation, and 374 individuals (91.7%) were re-interviewed. The proportion of individuals followed up successfully according to group status was: SC = 95.7%, FC = 88.5%, CG = 90.6% (χ2 = 4.725; d.f. = 2; p = .094).
3.1. Participants The CONSORT statement is presented in Fig. 2. In total, 10,803 GP patients were screened (refusal rate: 5.9%). Among these, 2239 (20.7%) were positive according to the AUDIT and/or the LAST criteria. Of these, 1410 patients subsequently agreed to participate further in the study (63.0%). Later, 7% of these withdrew from further participation and 13.6% had to be excluded for other reasons (e.g. no telephone access). With 1119 (79.4%) of positively screened patients with informed consent, diagnostic interviews could be conducted (response rate of screening positives = 50.0%). Screening positives, who refused to participate or were ineligible for the telephone interview, were older (38.7, S.D. = 13.7 vs. 36.8, S.D. = 13.6; p = .001) and less educated (more than 9 years of schooling 54.4% vs. 31.3%; p < .001) compared to study participants, but revealed a similar rate of females (28.7% vs. 31.3%; p = .180) and smokers (39.6% vs. 42.5%; p = .228) and similar mean scores on the AUDIT (M = 9.0, S.D. = 5.8 vs. M = 9.1, S.D. = 5.9; p = .702) and the LAST (M = 1.6, S.D. = 1.7 vs. M = 1.6, S.D. = 1.6; p = .966). Among those individuals with whom the baseline diagnostic interview could be conducted, 664 patients (59.3%) did not meet a diagnosis of alcohol use disorders or criteria of at-risk or binge drinking. Another 47 individuals refused study participation or did not return the baseline questionnaire. Finally, 408 participants fulfilled the study inclusion criteria of alcohol dependence, abuse, at-risk drinking or binge drinking. Of these, 278 participants (68.1%) were male and 130 (31.9%) female. The mean age was 36.9 (S.D. = 13.44; range 18–64). 124 (30.4%) participants were diagnosed with alcohol dependence and 59 (14.5%) with alcohol abuse, 112 (27.5%) fulfilled study criteria for atrisk drinking and 113 participants (27.7%) for binge drinking only. There were slight differences between the study groups in terms of drinking status, although not statistically significant (Table 1). Of the final study sample (n = 408), 320 participants (78.4%) did not meet diagnostic criteria for comorbidity, while 88 participants (21.6%) were diagnosed with comorbid anxiety and/or depressive disorder (35 participants with depression only, 30 with anxiety disorder only, and 23 with both anxiety and depressive disorder). The prevalence of comorbid participants did not differ between control group, FC or SC intervention (n = 29/29/30; χ2 = .203; d.f. = 2; p = .90). Compared to non-comorbid patients, comorbid participants were
Drinking status
CG (n = 139)
SC (n = 138)
FC (n = 131)
χ2 ; p
Alcohol dependence Alcohol abuse At-risk drinking only Heavy drinking episodes
25.9 15.1 33.2 28.8
27.5 11.6 31.2 29.7
38.2 16.8 20.6 24.4
9.295; .158
3.2. Reduction of drinking Overall, significantly stronger drinking reductions were obtained for the comorbid subsample, compared to noncomorbid participants (Mann–Whitney U = 11680; p = .01). An ordinal regression analysis showed comorbidity to be a significant predictor for drinking reduction (estimator = .594; CI = .175–1.013; p < .01). In order to interpret the influence of the different factors in a regression model, the estimator is used. Positive estimators indicate the respective category to act as a higher category of the dependent variable; negative estimators act as a lower category of the dependent variable. In these analysis models, positive values for the estimator thus indicate reduction of drinking. BI were significantly related to reduction of alcohol consumption between baseline and follow-up for non-comorbid (Mann–Whitney U = 9891.0; p = .03) but not for comorbid participants (Mann–Whitney U = 821.0; p = .76). Data show stronger drinking reductions for the comorbid intervention groups (SC/FC) as well as the comorbid control group, in comparison to the respective non-comorbid intervention groups (SC/FC) and control group, however these differences did not reach significance (Table 2). Since comorbid and non-comorbid participants significantly differed with respect to gender, classification of problematic
218
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
Table 2 Reduction of drinking according to intervention groups and comorbid disorders (Kruskall–Wallis H-test)
1. 2. 3.
Comorbid disorder
Control group (mean value: g/alcohol)
SC intervention group (mean value: g/alcohol)
FC intervention group (mean value: g/alcohol)
Comparisons between controland intervention groups χ2 ; significance (p)
No comorbidity Comorbidity No depression Depression only No anxiety Anxiety only
−2.64 −22.06a −2.96 0.45d −2.96 −31.68g
−8.88 −14.93b −9.21 −13.08e −9.21 −33.94h
−8.31 −29.00c −9.23 −35.50f −10.15 .05i
χ2 = 7.26; p = .03 χ2 = .165; p = .92 χ2 = 7.26; p = .03 χ2 = .59; p = .75 χ2 = 7.26; p = .03 χ2 = .67; p = .72
Comparisons between comorbid and non-comorbid participants within control- or intervention groups. a U = 1246.50; p = .07. b U = 1471.50; p = .57. c U = 1178.00; p = .06. d U = 532.50; p = .511. e U = 747.00; p = 90. f U = 386.00; p = .22. g U = 434.00; p = .12. h U = 227.50; p = .21. i U = 572.50; p = .45. Table 3 Ordinal logistic regression analysis to predict reduction of drinking (dependent variable: reduction of drinking (g/alcohol) in six ordered categories Independent variables
Estimator
Confidence interval
Model 1 Comorbidity Gender
.416 −.047
−.077 to .910 −.461 to .808
.098 .825
Drinking amount at baseline Utilization of help
.200
−.408 to .808
.519
Problematic drinking (reference category dependence) Binge drinking At-risk drinking Alcohol abuse
.088 .562 .252
−.478 to .654 .048 to 1.075 −.369 to .874
.759 .032 .426
.425 .014
−.024 to .874 .009 to .018
.063 <.001
Intervention (reference category control group) Fixed care group Stepped care group
.203 .533
−.230 to .635 .108 to .957
.358 .014
Problematic drinking (reference category dependence) Binge drinking At-risk drinking Alcohol abuse
.072 .736 .268
−.463 to .607 .250 to 1.222 −.322 to .857
.793 .003 .373
Model 2 Comorbidity Drinking amount at baseline
drinking (dependence, abuse, at-risk drinking, binge drinking), drinking amount at baseline and utilization of formal help for drinking problems, we assessed the predictive value of comorbidity for drinking reduction in relation to these independent variables in another ordinal regression analysis (Table 3, model 1). Here, gender and utilization of help were not predictive for reduction of drinking. Drinking amount at baseline and atrisk drinking were found to be significant positive predictors of reduction of drinking. Comorbidity showed a tendency towards significance. In a second ordinal regression model (Table 3, model 2), we entered those variables with a significant predictive value together with comorbidity and group status (FC/SC/CG) as independent variables. The variables amount of drinking at baseline, at-risk drinking and SC intervention were found to be
Significance (p)
significant predictors for drinking reduction. For comorbidity, a tendency towards significance was found. 4. Discussion This is the first paper examining an interaction between the effectiveness of BI for problematic drinking and comorbid depression and/or anxiety disorders. Although BI were significantly related to reduction of drinking in non-comorbid participants, this was not found for comorbid individuals. Regardless of the intervention, comorbidity was associated with a stronger drinking reduction than was found for the non-comorbid subsample. When comorbidity was entered as a single independent variable into an ordinal
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
regression model, it became a significant predictor for drinking reduction. However, when further independent variables were considered, the amount of drinking at baseline, at-risk drinking and SC intervention were found to be predictors for reduction of drinking, while comorbidity lost its predictive value. The following limitations ought to be taken into account when interpreting the results: first, this study is a secondary analysis to a trial assessing the effectiveness of a stepped care approach for at-risk consumers and individuals with alcohol use disorders in primary health care settings (Bischof et al., in press-a, in pressb). Statistical power had been calculated for the original trial and not for the present subgroup analysis of comorbidity, this work therefore presents an explanatory analysis. However, while lack of power would have led to overlook differences between groups; the present study shows differences between comorbid and noncomorbid groups. As comorbidity showed a tendency towards significance when predicting reduction of drinking, future studies with a larger sample size might be able to reassess this relationship. Secondly, higher percentages of dependent individuals within the comorbid subsample naturally enable stronger drinking reductions for the comorbid. By splitting the outcome variable into six ordered categories and including drinking at baseline in the analyses, the effect of regression to the mean has been reduced but cannot be fully excluded. Thirdly, a reduction of drinking even in control arms has been reported in a Cochrane review which has included a number of BI studies (Kaner et al., 2007). While the reductions of alcohol consumption which occurred in the comorbid control condition but not the noncomorbid control condition of this study are unlikely to be due to regression to the mean, given the low cut-off point on AUDIT and the relatively low threshold of risk in terms of daily drinking and the use of a sliding scale of risk measures to help explore this issue, the findings suggest that screening alone might represent an impetus for behavioral change in patients with problematic drinking and comorbid anxiety and/or depression. The extensive baseline assessment, which was similar to the length of the BI sessions, might have contributed to this effect. Fourth, previous review articles (Bien et al., 1993) show BI to be effective for alcohol dependents. A more recent meta-analysis (Moyer et al., 2002) postulates effectiveness mainly for at-risk drinkers. This is confirmed by outcome data of this study described in Bischof et al. (in press-a, in press-b), which show effectiveness of BI for at-risk but not for dependent drinkers. Hence, it can be assumed that the higher percentage of dependent drinkers in the comorbid subsample is likely to have limited the effectiveness of BI for the comorbid. However, excluding dependent individuals from the analyses would have undermined statistical power. The following aspects might be considered when regarding the methodology: the assessment procedure was designed to take up the least time necessary within the practice setting, therefore additional assessments were done via mail and telephone. However, the possible bias related to this procedure is likely to be marginal. This study is based on self-reports of alcohol consumption. One may critically assume, that self-reports by comorbid individuals may be influenced by over- or underreporting due to social acceptance bias or even more realistic reporting. However, a study by Stasiewicz et al. (1997) reported no signifi-
219
cant difference in the validity of self-report data between alcohol dependents with or without comorbid psychiatric disorder. Only half of all individuals screening positive consented to participate in the study. Although refusers did not differ from participants in terms of screening characteristics, our sample might not be entirely representative of the basic population. In conclusion, a series of BI did not significantly predict drinking reduction in comorbid individuals who did not seek treatment for problematic drinking. Further studies with larger sample sizes of problematic drinkers ought to be conducted in order to gain further insight into the effectiveness of BI for problematic drinkers with comorbid disorders. Previous data (Grothues et al., 2005) highlight the need to adjust BI to comorbid individuals. Due to a higher temptation to drink and lower self-efficacy to abstain from drinking compared to non-comorbid individuals, special emphasis should be given to coping strategies and increasing self-efficacy. In addition, referral to specialized services, offering formal help for problematic drinking and psychiatric disorders, seems to be indicated.
Conflict of interest All authors declare that they have not conflict of interest.
Acknowledgements We thank all participating general practices for enabling data collection on their premises, and all participants for providing their data for our analyses. We thank Mrs. Bluhm, Mrs. Fahndrich, Mrs. Tychsen-Witt, Mrs. Wolf, Mrs. Willudt, Mrs. Madaler, Mr. Kruse, Mrs. F¨orst, Mr. W¨ustenberg, Mr. Biedendieck, Mrs. Dybeck, Mrs. Macke, Mr. Mommsen and Mr. Sommer for assisting us with data collection. We thank Mrs. Kreutzer for data entry and correction. Role of funding source: this study is part of the German research network EARLINT (EARLy substance use INTervention) and was funded by the German Federal Ministry of Research and Education. The present analysis is part of the project SIP (Stepped Interventions for Problem drinkers; grant no. 01 EB 0121). The funding source had no further role in study design; in the collection, analysis or interpretation of data; in the writing of the report; or in the decision to submit the paper for publication. Contributors: all authors contributed to the planning, evaluation, and reporting of this study. HJR was responsible for the conception of the study and the design, secured funding, established and coordinated the study and contributed to data analysis. GB contributed to the study design, data collection/BI and study coordination, was responsible for data editing and contributed to the analysis. JMG was responsible for data analysis, writing of the manuscript and will act as guarantor. JMG and SR contributed to the study design and data collection/BI and assisted in study coordination. CM gave valuable input to statistical analysis. CM and UJ contributed to the study design. All authors gave valuable contribution to the revision of the manuscript.
220
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
References American Psychiatric Association (Ed.), 1995. Diagnostic and Statistical Manual of Mental Disorders, fourth edition, international version. American Psychiatric Association, Washington, D.C. Anderson, P., 1991. Alcohol as a key area. Br. Med. J. 303 (6805), 766–769. Babor, T.F., Hodgson, B., Ritson, B., McRee, G., Ernberg, K., Connor, K., Grant, M., 1992. Experimental design and project administration. In: Babor, T.F., Grant, M. (Eds.), Program on Substance Abuse. Project on Identification and Management of Alcohol-related Problems. Report on Phase II: A Randomized Clinical Trial of Brief Interventions in Primary Health Care. World Health Organization, Geneva, pp. 15–56. Bien, T.H., Miller, W.R., Tonigan, J.S., 1993. Brief interventions for alcohol problems: a review. Addiction 88, 315–336. Bischof, G., Reinhardt, S., Meyer, C., John, U., Rumpf, H.-J., in press-a. The expert test and report on alcohol (extra): development and evaluation of a computerized software program for problem drinkers. In: Columbus, F. (Ed.), Progress in Alcoholism Research. Nova Publishers, New York. Bischof, G., Grothues, J., Reinhardt, S., Meyer, C., John, U., Rumpf, H.J., in press-b. Evaluation of a telephone-based stepped care intervention for alcohol-related disorders: a randomized controlled trial. Drug Alcohol Depend.,doi:10.1016/j.drugalcdep.2007.10.003. Bott, K., Meyer, C., Rumpf, H.-J., Hapke, U., John, U., 2005. Psychiatric disorders among at-risk consumers of alcohol in the general population. J. Stud. Alcohol 66, 246–253. British Medical Association, 1995. Guidelines on Sensible Drinking. British Medical Association, London. Dybek, I., Bischof, G., Grothues, J., Reinhardt, S., Meyer, C., John, U., Brooks, A., Hohagen, F., Rumpf, H.-J., 2006. The reliability and validity of the alcohol use disorder identification test (AUDIT) in a German general practice population sample. J. Stud. Alcohol 67 (3), 473–481. Fleming, M.F., 1997. Strategies to increase alcohol screening in health care settings. Alcohol Health Res. World 21 (4), 340–347. Grant, B.F., 1997. Barriers to alcoholism treatment: reasons for not seeking treatment in a general population sample. J. Stud. Alcohol 58 (4), 365– 371. Greenfield, S.F., Weiss, R.D., Muenz, L.R., Vagge, L.M., Kelly, J.F., Bello, L.R., Michael, J., 1998. The effect of depression on return to drinking: a prospective study. Arch. Gen. Psychiatry 55, 259–265. Grothues, J., Bischof, G., Reinhardt, S., Hapke, U., John, U., Meyer, C., Rumpf, H.-J., 2005. Intention to change drinking behavior in general practice patients with problematic drinking and comorbid depression and anxiety. Alcohol Alcohol. 40 (5), 394–400. Hasin, D.S., Grant, B.F., 2002. Major depression in 6050 former drinkers: association with past alcohol dependence. Arch. Gen. Psychiatry 59, 794– 800. Kaner, E.F.S., Beyer, F., Dickinson, H.O., Piennaar, E., Campbell, F., Schlesinger, C., Heather, N., Saunders, J., Burnand, B., 2007. Effectiveness of brief alcohol interventions in primary care populations. Cochrane Database Syst Rev. (2), CD004148.
Kessler, R.C., Crum, R.M., Warner, L.A., Nelson, C.B., Schulenberg, J., Anthony, J.C., 1997. Lifetime co-occurrence of DSM-III-R alcohol abuse with other psychiatric disorders in the national comorbidity survey. Arch. Gen. Psychiatry (54), 313–321. Lynskey, M.T., 1998. The comorbidity of alcohol dependence and affective disorders: treatment implications. Drug Alcohol Depend. 52, 201–209. Miller, W.R., Rollnick, S. (Eds.), 1991. Motivational Interviewing. Preparing People to Change Addictive Behavior. The Guilford Press, New York. Moyer, A., Finney, J.W., Swearingen, C.E., Vergun, P., 2002. Brief interventions for alcohol problems: a meta-analytic review of controlled investigations in treatment-seeking and non-treatment-seeking populations. Addiction 97, 279–292. Regier, D.A., Farmer, M.E., Rae, D.S., Locke, B.Z., Keith, S.J., Judd, L.L., Goodwin, F.K., 1990. Comorbidity of mental disorders with alcohol and other drug abuse: results from the epidemiologic catchment area (ECA) study. JAMA 264, 2511–2518. Regier, D.A., Narrow, W.E., Rae, D.S., Manderscheid, R.W., Locke, B.Z., Goodwin, F.K., 1993. The de facto us mental and addictive disorders service system. Arch. Gen. Psychiatry 50, 85–94. Robins, L.N., Wing, J., Wittchen, H.U., 1988. The composite international diagnostic interview: an epidemiological instrument suitable for use in conjunction with different diagnostic systems and in different cultures. Arch. Gen. Psychiatry 45, 1069–1077. Rollnick, S., Mason, P., Butler, C., 1999. Health Behaviour Change. Churchill Livingstone, Kent. Rumpf, H.-J., Hapke, U., Hill, A., John, U., 1997. Development of a screening questionnaire for the general hospital and general practices. Alcohol. Clin. Exp. Res. 21, 894–898. Rumpf, H.-J., Hapke, U., Meyer, C., John, U., 2002. Screening for alcohol use disorders and at-risk drinking in the general population: psychometric performance of three questionnaires. Alcohol Alcohol. 37, 261–268. Saunders, J.B., Aasland, O.G., Babor, T.F., DeLaFuente, J.R., Grant, M., 1993. Development of the alcohol use disorders identification test (AUDIT): who collaborative project on early detection of persons with harmful alcohol consumption-II. Addiction 88, 617–629. Stasiewicz, P.R., Bradizza, C.M., Connors, G.J., 1997. Subject-collateral reports of drinking in inpatient alcoholics with comorbid mental disorders. Alcohol. Clin. Exp. Res. 21, 530–536. ¨ un, T.B., Sartorius, N., 1995. Mental Health in General Health Care Across Ust¨ the World. An International Study. John Wiley & Sons, New York. Velasquez, M.M., Carbonari, J.P., DiClemente, C.C., 1999. Psychiatric severity and behavior change in alcoholism: the relation of the transtheoretical model variables to psychiatric distress in dually diagnosed patients. Addict. Behav. 24 (4), 481–496. Wittchen, H.-U., Beloch, E., Garczynski, E., Holly, A., Lachner, G., Perkonigg, A., Vodermaier, A., Vossen, A., Wunderlich, U., Zieglg¨ansberger, S., 1995. M¨unchener composite international diagnostic interview (m-cidi), version 2.2. Max-Planck-Institut of Psychiatry, Munich. Wittchen, H.-U., Perkonigg, A., Reed, V., 1996. Comorbidity of mental disorders and substance use disorders. Eur. Addict. Res. 2, 36–47.
Drug and Alcohol Dependence 94 (2008) 214–220
Effectiveness of brief alcohol interventions for general practice patients with problematic drinking behavior and comorbid anxiety or depressive disorders Janina M. Grothues a,∗ , Gallus Bischof a , Susa Reinhardt a , Christian Meyer b , Ulrich John b , Hans-J¨urgen Rumpf a a
University of L¨ubeck, Department of Psychiatry and Psychotherapy, Research Group S:TEP (Substance Abuse: Treatment, Epidemiology and Prevention), Ratzeburger Allee 160, 23538 L¨ubeck, Germany b Ernst-Moritz Arndt University of Greifswald, Department of Epidemiology and Social Medicine, Walther-Rathenau-Str. 48, 17487 Greifswald, Germany Received 28 June 2007; received in revised form 23 November 2007; accepted 26 November 2007 Available online 22 January 2008
Abstract Background: Brief interventions (BIs) are effective methods to reduce problematic drinking. It is not known, if the effectiveness of BI differs between patients with or without comorbid depression or anxiety disorders. Methods: In a randomized controlled BI study with two intervention groups and one control condition, data were collected from 408 general practice (GP) patients with alcohol use disorders, at-risk drinking or binge drinking. 88 participants were diagnosed with comorbid anxiety and/or depressive disorders. The effectiveness of BI was assessed at a 12-month follow-up in relation to the presence and absence of comorbidity. Reduction of drinking in six ordered categories (g/alcohol) between baseline and follow-up served as the outcome variable. Results: BI were significantly related to reduction of drinking in the non-comorbid (−2.64 g/alcohol vs. −8.61 g/alcohol; p = .03) but not in the comorbid subsample (−22.06 g/alcohol vs. −22.09 g/alcohol; p = .76). Compared to non-comorbid participants, a significantly higher reduction of drinking was found for comorbid individuals (−6.55 g/alcohol vs. −22.08 g/alcohol; p = .01). An ordinal regression analysis revealed comorbidity to be a positive predictor for reduction of drinking (estimator = .594; CI = .175–1.013; p < .01). When entering the variables amount of drinking at baseline, intervention and classification of problematic drinking, these became significant predictors, whereas comorbidity showed only a tendency. Conclusion: BI did not significantly effect a reduction of drinking in comorbid patients. As BI are known to be less effective for dependent drinkers, a larger proportion of dependents among the comorbid might have limited the effectiveness of BI. Future studies with larger sample sizes of comorbid problem drinkers are necessary to confirm the results. © 2008 Elsevier Ireland Ltd. All rights reserved. Keywords: Brief intervention; Alcohol use disorders; Problematic drinking; Comorbidity; Anxiety disorders; Depressive disorders
1. Introduction While the majority of alcohol-related harm is not attributable to drinkers with severe alcohol dependence but to a much larger group of excessive or hazardous drinkers whose consumption exceeds recommended drinking limits (Anderson, 1991), brief interventions (BIs) aim to reduce consumption behavior and related harm. Motivational Interviewing (MI; Miller
∗
Corresponding author. Tel.: +49 451 500 2965; fax: +49 451 500 3480. E-mail address: [email protected] (J.M. Grothues).
0376-8716/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.drugalcdep.2007.11.015
and Rollnick, 1991) poses an effective technique to enhance motivation to change problematic drinking behavior. As the majority of problem drinkers do not seek help (Grant, 1997), primary care settings provide a suitable channel to reach a large number of the target group. A recent Cochrane review provided strong evidence that BI in primary care settings are effective to reduce problematic alcohol consumption (Kaner et al., 2007). However, Kaner et al. (2007) also describe “a clear need to characterise the types of drinkers for whom BI have a positive impact and any subgroups that have not been represented in the trials to date” (p. 3). While high rates of anxiety and depression have been found in various samples of individuals with alcohol
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
use disorders (Kessler et al., 1997; Regier et al., 1990; Wittchen et al., 1996) or at-risk drinking (Bott et al., 2005) as well as ¨ un and Sartorius, 1995), within the primary care setting (Ust¨ little is known about the effectiveness of BI for individuals with problematic drinking behavior and comorbid anxiety and/or depressive disorders. Further investigation is needed to improve pro-active intervention strategies for this subgroup. Regarding the outcome of alcohol specific treatment for comorbid individuals with alcohol use disorders and comorbid depression or anxiety, previous studies have provided heterogeneous results (Greenfield et al., 1998; Hasin and Grant, 2002). On the motivational level, a higher readiness to change problematic drinking behavior has been found for comorbid individuals in an outpatient dual diagnoses sample (Velasquez et al., 1999) and in pro-actively recruited general practice (GP) patients (Grothues et al., 2005). Comorbidity has also been associated with increased utilization of formal help (Lynskey, 1998; Regier et al., 1993). However, previous studies also show increased temptation to drink (Grothues et al., 2005; Velasquez et al., 1999) and lower self-efficacy to abstain from drinking (Grothues et al., 2005). The key objective of this paper is to assess the effectiveness of BI to reduce problematic drinking behavior on grounds of longitudinal data which has been derived from a sample of proactively recruited GP patients with a range of drinking problems and comorbid depressive and/or anxiety disorders. The following aspects are addressed: Does the effectiveness of BI differ according to the presence or absence of comorbid anxiety and/or depressive disorders? Which implication does this yield for the delivery of BI to comorbid individuals who do not seek help for their drinking? 2. Method 2.1. Procedure Within the randomized controlled BI study “Stepped Interventions for Problem Drinkers (SIP)”, data were collected by trained project staff in 81 general practices in the north German city of L¨ubeck and its 46 surrounding communities and also in four practices in the north German city of Kiel during the period 2001 and 2003. A power analysis based on data by Fleming (1997) indicated an 80% statistical power for a sample size of 114 participants per group when applying a 5%, one-way significance level to compare each of the two intervention groups with the control group. Assuming a 25% attrition rate resulted in 150 participants per group. To minimise time for data collection within the practice, the procedure was threefold: screening within the practice and administering telephone diagnostic assessments outside the practice. Additional data which are not subject for this analysis were collected via postal questionnaire. GP patients aged 18–64, attending the practice for a GP consultation, were contacted in the waiting room by a trained study nurse and asked to fill out a screening questionnaire. Patients with a positive screening result (see Section 2.2) were asked for written informed consent to participate further in the study. Study nurses undertook intense training to ensure high quality and standardized procedure. They were individually supervised within the practice setting by study staff and received fortnightly supervision sessions. On average 2 days after screening, participants who had agreed to participate further were sent a questionnaire on alcohol related problems and readiness to change variables. Two to four days after sending the questionnaire, participants were contacted by postgraduate psychologists to partake in a telephone diagnostic baseline assessment of alcohol related disorders and problematic drinking. Patients who had been in alcohol specific treatment within the previous 4 weeks were excluded from the study. Participants, meeting specified criteria for alco-
215
Fig. 1. Stepped care intervention group.
hol dependence, abuse, at-risk drinking or binge drinking, were included in the final study sample and a standardized diagnostic assessment of anxiety and/or depressive disorders was administered. Additional data, including utilization of formal help for drinking problems, were collected during the assessment. The baseline telephone contacts lasted 10–90 min, the average being 30 min. Final study participants were randomly allocated, using coloured cards in sealed, non-transparent envelopes (drawn envelopes were not put back), to either the control group or to one of two intervention groups. These were applied to serve a separate research aim of the study SIP: to assess the cost-effectiveness of a stepped care approach (SC; if drinking reduction below the study criteria for atrisk and binge drinking within the previous 4 weeks had been reached, scheduled counseling sessions were discontinued until the follow-up assessment; see Fig. 1) in comparison to a full care (FC) intervention (all scheduled interventions were applied, regardless of drinking reduction) (for details see Bischof et al., in pressa, in press-b). Participants in the control group were sent a booklet on health behavior; no alcohol related intervention was provided. Directly following the baseline assessment, both intervention groups received a computerized feedback (Bischof et al., in press-a, in press-b). In addition, intervention group 1 (FC) received a standardized number of four BI sessions directly after the baseline assessment, as well as after 1, 3 and 6 months. The scheduled time for each BI session for the FC condition was 30 min. Intervention group 2 (SC) received a maximum of three BI sessions at the time stages of 1, 3 and 6 months after baseline, depending upon success of the previous intervention. The scheduled time for each session in the SC condition was 30–40 min, thus enabling the maximum of all possible interventions to equal the standardized effort applied to the FC group. All counseling sessions were administered by telephone. They were fully based on MI spirit and principles (Miller and Rollnick, 1991) and structured by using the following elements of behavioral change counseling (BCC; Rollnick et al., 1999): decisional balance (routinely) and readiness- and self-efficacy rulers (if indicated). Counseling was carried out by postgraduate psychologists who were trained and experienced in MI, and was conducted according to standardized instruments and protocols including structure, procedure/methods and goals. Fidelity of interventions was monitored using standardized monitoring forms. Weekly supervision and standardized protocols ensured the consistent quality of the interventions and the adherence to an MI-consistent intervention procedure. Follow-up assessments of both intervention groups and the control group were carried out via telephone contact 12 months after the baseline assessment. Follow-ups were undertaken by a different psychologist other than the one conducting the diagnostic assessment and interventions. The allocation was randomized and blinded after the exclusion of the psychologist, who had initially conducted the baseline-assessment and counseling sessions.
2.2. Instruments 2.2.1. Screening within the practice. The Alcohol Use Disorder Identification Test (AUDIT; Saunders et al., 1993), a 10-item self-report questionnaire, was used for the initial screening. Based on data for a German general population sample (Rumpf et al., 2002), a cut-off value of five points was chosen. The retest reliability and validity of the AUDIT proved good combinations of sensitivity (.97 for alcohol dependence, alcohol use disorders (AUDs), and AUDs and/or
216
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
at-risk consumption) and specificity (.88 for alcohol dependence, .92 for AUDs, and .91 for AUDs and/or at-risk drinking) for the cut-off value five based upon analysis of the present GP sample, hence the lowering of the standard cutoff value has been strongly recommended (Dybek et al., 2006). Alternatively, a minimum of two points on the L¨ubeck Alcohol dependence and abuse Screening Test (LAST; Rumpf et al., 1997) indicated positive screening. 2.2.2. Baseline diagnostic assessment via telephone. The Munich Composite International Diagnostic Interview (M-CIDI; Wittchen et al., 1995), which is the German translation of the CIDI (Robins et al., 1988), was used for the baseline diagnostic identification of alcohol abuse or dependence according to the 4th and international version of the Diagnostic and Statistical Manual of psychiatric disorders (DSM-IV; American Psychiatric Association, 1995). Criteria for atrisk drinking were defined as an average consumption of >20/30 g of alcohol per day for women/men within the previous 4 weeks (British Medical Association, 1995) or binge drinking as >60/80 g of alcohol for women/men on at least two occasions within the previous 4 weeks (Babor et al., 1992). The definition of drinking measure used for analyses was defined as the average amount across all drinking days obtained from the information on quantity/frequency of drinking
within the previous 12 months. The measures of drinking for the pre-intervention assessment and follow-up were assessed and calculated according to identical procedures. The following comorbid disorders were assessed with the M-CIDI: major depressive disorder, dysthymic disorder, panic disorder, agoraphobia, social phobia, generalised anxiety disorder and anxiety disorder not otherwise specified. Lifetime utilization of formal help for drinking problems was assessed using the following categories: professional advice/treatment, alcohol detoxification/rehabilitation treatment and self-help group visits. 2.2.3. Follow-up diagnostic assessment via telephone. The follow-up diagnostic assessment of alcohol abuse or dependence was conducted using the M-CIDI. Utilization of formal help for drinking problems within the previous 12 months was assessed using the categories specified at baseline.
2.3. Analysis procedures The differential amount of alcohol consumption (g/alcohol per day) between baseline and follow-up assessments served as the outcome variable.
Fig. 2. CONSORT statement.
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
217
Applying an intention to treat analysis, participants who could not be reached at follow-up were defined as unchanged. Analysis groups 1 (non-comorbid participants) and 2 (participants with comorbid anxiety and/or depressive disorders) were compared. Due to lack of normal distribution of the outcome variable, non-parametric tests were used. Ordinal regression analyses were applied as a multivariate approach to assess the predictive value of comorbidity for drinking reduction in relation to variables, in which comorbid and non-comorbid participants were found to differ. For the ordinal regression analyses, the outcome variable reduction of drinking was split into six equally ordered categories: (1) increased consumption, (2) no reduction, (3) reduction <10 g, (4) reduction <30 g, (5) reduction <60 g, and (6) reduction >60 g.
Table 1 Drinking status of study groups at baseline (CG: control group; SC: stepped care; FC: full care)
3. Results
more often alcohol dependent as opposed to alcohol abusers, at-risk drinkers and binge drinkers (χ2 = 42.1; d.f. = 3; p < .001). At baseline, comorbid individuals had significantly higher mean values of alcohol consumption than non-comorbid participants (75.22 g/alcohol vs. 40.53 g/alcohol; Mann–Whitney U = 111974.000; p < .05). Comorbid participants were significantly more often female (χ2 = 17.0; d.f. = 1; p < .001), but did not differ from non-comorbid individuals in years of schooling (up to 9 years vs. more than 9 years; χ2 = 1.6; d.f. = 1; p = .21) and age (t = 1.5; d.f. = 152.2; p = .14). Lifetime and 12-month utilization of formal help for drinking were significantly higher for comorbid than for non-comorbid individuals (χ2 = 12.95; d.f. = 1; p < .001/χ2 = 31.75; d.f. = 1; p < .001). At 12-month follow-up, 3 (0.7%) participants of the baseline sample described above (n = 408), had died, 27 (6.6%) were not attainable, 4 individuals (1.0%) withdrew their further participation, and 374 individuals (91.7%) were re-interviewed. The proportion of individuals followed up successfully according to group status was: SC = 95.7%, FC = 88.5%, CG = 90.6% (χ2 = 4.725; d.f. = 2; p = .094).
3.1. Participants The CONSORT statement is presented in Fig. 2. In total, 10,803 GP patients were screened (refusal rate: 5.9%). Among these, 2239 (20.7%) were positive according to the AUDIT and/or the LAST criteria. Of these, 1410 patients subsequently agreed to participate further in the study (63.0%). Later, 7% of these withdrew from further participation and 13.6% had to be excluded for other reasons (e.g. no telephone access). With 1119 (79.4%) of positively screened patients with informed consent, diagnostic interviews could be conducted (response rate of screening positives = 50.0%). Screening positives, who refused to participate or were ineligible for the telephone interview, were older (38.7, S.D. = 13.7 vs. 36.8, S.D. = 13.6; p = .001) and less educated (more than 9 years of schooling 54.4% vs. 31.3%; p < .001) compared to study participants, but revealed a similar rate of females (28.7% vs. 31.3%; p = .180) and smokers (39.6% vs. 42.5%; p = .228) and similar mean scores on the AUDIT (M = 9.0, S.D. = 5.8 vs. M = 9.1, S.D. = 5.9; p = .702) and the LAST (M = 1.6, S.D. = 1.7 vs. M = 1.6, S.D. = 1.6; p = .966). Among those individuals with whom the baseline diagnostic interview could be conducted, 664 patients (59.3%) did not meet a diagnosis of alcohol use disorders or criteria of at-risk or binge drinking. Another 47 individuals refused study participation or did not return the baseline questionnaire. Finally, 408 participants fulfilled the study inclusion criteria of alcohol dependence, abuse, at-risk drinking or binge drinking. Of these, 278 participants (68.1%) were male and 130 (31.9%) female. The mean age was 36.9 (S.D. = 13.44; range 18–64). 124 (30.4%) participants were diagnosed with alcohol dependence and 59 (14.5%) with alcohol abuse, 112 (27.5%) fulfilled study criteria for atrisk drinking and 113 participants (27.7%) for binge drinking only. There were slight differences between the study groups in terms of drinking status, although not statistically significant (Table 1). Of the final study sample (n = 408), 320 participants (78.4%) did not meet diagnostic criteria for comorbidity, while 88 participants (21.6%) were diagnosed with comorbid anxiety and/or depressive disorder (35 participants with depression only, 30 with anxiety disorder only, and 23 with both anxiety and depressive disorder). The prevalence of comorbid participants did not differ between control group, FC or SC intervention (n = 29/29/30; χ2 = .203; d.f. = 2; p = .90). Compared to non-comorbid patients, comorbid participants were
Drinking status
CG (n = 139)
SC (n = 138)
FC (n = 131)
χ2 ; p
Alcohol dependence Alcohol abuse At-risk drinking only Heavy drinking episodes
25.9 15.1 33.2 28.8
27.5 11.6 31.2 29.7
38.2 16.8 20.6 24.4
9.295; .158
3.2. Reduction of drinking Overall, significantly stronger drinking reductions were obtained for the comorbid subsample, compared to noncomorbid participants (Mann–Whitney U = 11680; p = .01). An ordinal regression analysis showed comorbidity to be a significant predictor for drinking reduction (estimator = .594; CI = .175–1.013; p < .01). In order to interpret the influence of the different factors in a regression model, the estimator is used. Positive estimators indicate the respective category to act as a higher category of the dependent variable; negative estimators act as a lower category of the dependent variable. In these analysis models, positive values for the estimator thus indicate reduction of drinking. BI were significantly related to reduction of alcohol consumption between baseline and follow-up for non-comorbid (Mann–Whitney U = 9891.0; p = .03) but not for comorbid participants (Mann–Whitney U = 821.0; p = .76). Data show stronger drinking reductions for the comorbid intervention groups (SC/FC) as well as the comorbid control group, in comparison to the respective non-comorbid intervention groups (SC/FC) and control group, however these differences did not reach significance (Table 2). Since comorbid and non-comorbid participants significantly differed with respect to gender, classification of problematic
218
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
Table 2 Reduction of drinking according to intervention groups and comorbid disorders (Kruskall–Wallis H-test)
1. 2. 3.
Comorbid disorder
Control group (mean value: g/alcohol)
SC intervention group (mean value: g/alcohol)
FC intervention group (mean value: g/alcohol)
Comparisons between controland intervention groups χ2 ; significance (p)
No comorbidity Comorbidity No depression Depression only No anxiety Anxiety only
−2.64 −22.06a −2.96 0.45d −2.96 −31.68g
−8.88 −14.93b −9.21 −13.08e −9.21 −33.94h
−8.31 −29.00c −9.23 −35.50f −10.15 .05i
χ2 = 7.26; p = .03 χ2 = .165; p = .92 χ2 = 7.26; p = .03 χ2 = .59; p = .75 χ2 = 7.26; p = .03 χ2 = .67; p = .72
Comparisons between comorbid and non-comorbid participants within control- or intervention groups. a U = 1246.50; p = .07. b U = 1471.50; p = .57. c U = 1178.00; p = .06. d U = 532.50; p = .511. e U = 747.00; p = 90. f U = 386.00; p = .22. g U = 434.00; p = .12. h U = 227.50; p = .21. i U = 572.50; p = .45. Table 3 Ordinal logistic regression analysis to predict reduction of drinking (dependent variable: reduction of drinking (g/alcohol) in six ordered categories Independent variables
Estimator
Confidence interval
Model 1 Comorbidity Gender
.416 −.047
−.077 to .910 −.461 to .808
.098 .825
Drinking amount at baseline Utilization of help
.200
−.408 to .808
.519
Problematic drinking (reference category dependence) Binge drinking At-risk drinking Alcohol abuse
.088 .562 .252
−.478 to .654 .048 to 1.075 −.369 to .874
.759 .032 .426
.425 .014
−.024 to .874 .009 to .018
.063 <.001
Intervention (reference category control group) Fixed care group Stepped care group
.203 .533
−.230 to .635 .108 to .957
.358 .014
Problematic drinking (reference category dependence) Binge drinking At-risk drinking Alcohol abuse
.072 .736 .268
−.463 to .607 .250 to 1.222 −.322 to .857
.793 .003 .373
Model 2 Comorbidity Drinking amount at baseline
drinking (dependence, abuse, at-risk drinking, binge drinking), drinking amount at baseline and utilization of formal help for drinking problems, we assessed the predictive value of comorbidity for drinking reduction in relation to these independent variables in another ordinal regression analysis (Table 3, model 1). Here, gender and utilization of help were not predictive for reduction of drinking. Drinking amount at baseline and atrisk drinking were found to be significant positive predictors of reduction of drinking. Comorbidity showed a tendency towards significance. In a second ordinal regression model (Table 3, model 2), we entered those variables with a significant predictive value together with comorbidity and group status (FC/SC/CG) as independent variables. The variables amount of drinking at baseline, at-risk drinking and SC intervention were found to be
Significance (p)
significant predictors for drinking reduction. For comorbidity, a tendency towards significance was found. 4. Discussion This is the first paper examining an interaction between the effectiveness of BI for problematic drinking and comorbid depression and/or anxiety disorders. Although BI were significantly related to reduction of drinking in non-comorbid participants, this was not found for comorbid individuals. Regardless of the intervention, comorbidity was associated with a stronger drinking reduction than was found for the non-comorbid subsample. When comorbidity was entered as a single independent variable into an ordinal
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
regression model, it became a significant predictor for drinking reduction. However, when further independent variables were considered, the amount of drinking at baseline, at-risk drinking and SC intervention were found to be predictors for reduction of drinking, while comorbidity lost its predictive value. The following limitations ought to be taken into account when interpreting the results: first, this study is a secondary analysis to a trial assessing the effectiveness of a stepped care approach for at-risk consumers and individuals with alcohol use disorders in primary health care settings (Bischof et al., in press-a, in pressb). Statistical power had been calculated for the original trial and not for the present subgroup analysis of comorbidity, this work therefore presents an explanatory analysis. However, while lack of power would have led to overlook differences between groups; the present study shows differences between comorbid and noncomorbid groups. As comorbidity showed a tendency towards significance when predicting reduction of drinking, future studies with a larger sample size might be able to reassess this relationship. Secondly, higher percentages of dependent individuals within the comorbid subsample naturally enable stronger drinking reductions for the comorbid. By splitting the outcome variable into six ordered categories and including drinking at baseline in the analyses, the effect of regression to the mean has been reduced but cannot be fully excluded. Thirdly, a reduction of drinking even in control arms has been reported in a Cochrane review which has included a number of BI studies (Kaner et al., 2007). While the reductions of alcohol consumption which occurred in the comorbid control condition but not the noncomorbid control condition of this study are unlikely to be due to regression to the mean, given the low cut-off point on AUDIT and the relatively low threshold of risk in terms of daily drinking and the use of a sliding scale of risk measures to help explore this issue, the findings suggest that screening alone might represent an impetus for behavioral change in patients with problematic drinking and comorbid anxiety and/or depression. The extensive baseline assessment, which was similar to the length of the BI sessions, might have contributed to this effect. Fourth, previous review articles (Bien et al., 1993) show BI to be effective for alcohol dependents. A more recent meta-analysis (Moyer et al., 2002) postulates effectiveness mainly for at-risk drinkers. This is confirmed by outcome data of this study described in Bischof et al. (in press-a, in press-b), which show effectiveness of BI for at-risk but not for dependent drinkers. Hence, it can be assumed that the higher percentage of dependent drinkers in the comorbid subsample is likely to have limited the effectiveness of BI for the comorbid. However, excluding dependent individuals from the analyses would have undermined statistical power. The following aspects might be considered when regarding the methodology: the assessment procedure was designed to take up the least time necessary within the practice setting, therefore additional assessments were done via mail and telephone. However, the possible bias related to this procedure is likely to be marginal. This study is based on self-reports of alcohol consumption. One may critically assume, that self-reports by comorbid individuals may be influenced by over- or underreporting due to social acceptance bias or even more realistic reporting. However, a study by Stasiewicz et al. (1997) reported no signifi-
219
cant difference in the validity of self-report data between alcohol dependents with or without comorbid psychiatric disorder. Only half of all individuals screening positive consented to participate in the study. Although refusers did not differ from participants in terms of screening characteristics, our sample might not be entirely representative of the basic population. In conclusion, a series of BI did not significantly predict drinking reduction in comorbid individuals who did not seek treatment for problematic drinking. Further studies with larger sample sizes of problematic drinkers ought to be conducted in order to gain further insight into the effectiveness of BI for problematic drinkers with comorbid disorders. Previous data (Grothues et al., 2005) highlight the need to adjust BI to comorbid individuals. Due to a higher temptation to drink and lower self-efficacy to abstain from drinking compared to non-comorbid individuals, special emphasis should be given to coping strategies and increasing self-efficacy. In addition, referral to specialized services, offering formal help for problematic drinking and psychiatric disorders, seems to be indicated.
Conflict of interest All authors declare that they have not conflict of interest.
Acknowledgements We thank all participating general practices for enabling data collection on their premises, and all participants for providing their data for our analyses. We thank Mrs. Bluhm, Mrs. Fahndrich, Mrs. Tychsen-Witt, Mrs. Wolf, Mrs. Willudt, Mrs. Madaler, Mr. Kruse, Mrs. F¨orst, Mr. W¨ustenberg, Mr. Biedendieck, Mrs. Dybeck, Mrs. Macke, Mr. Mommsen and Mr. Sommer for assisting us with data collection. We thank Mrs. Kreutzer for data entry and correction. Role of funding source: this study is part of the German research network EARLINT (EARLy substance use INTervention) and was funded by the German Federal Ministry of Research and Education. The present analysis is part of the project SIP (Stepped Interventions for Problem drinkers; grant no. 01 EB 0121). The funding source had no further role in study design; in the collection, analysis or interpretation of data; in the writing of the report; or in the decision to submit the paper for publication. Contributors: all authors contributed to the planning, evaluation, and reporting of this study. HJR was responsible for the conception of the study and the design, secured funding, established and coordinated the study and contributed to data analysis. GB contributed to the study design, data collection/BI and study coordination, was responsible for data editing and contributed to the analysis. JMG was responsible for data analysis, writing of the manuscript and will act as guarantor. JMG and SR contributed to the study design and data collection/BI and assisted in study coordination. CM gave valuable input to statistical analysis. CM and UJ contributed to the study design. All authors gave valuable contribution to the revision of the manuscript.
220
J.M. Grothues et al. / Drug and Alcohol Dependence 94 (2008) 214–220
References American Psychiatric Association (Ed.), 1995. Diagnostic and Statistical Manual of Mental Disorders, fourth edition, international version. American Psychiatric Association, Washington, D.C. Anderson, P., 1991. Alcohol as a key area. Br. Med. J. 303 (6805), 766–769. Babor, T.F., Hodgson, B., Ritson, B., McRee, G., Ernberg, K., Connor, K., Grant, M., 1992. Experimental design and project administration. In: Babor, T.F., Grant, M. (Eds.), Program on Substance Abuse. Project on Identification and Management of Alcohol-related Problems. Report on Phase II: A Randomized Clinical Trial of Brief Interventions in Primary Health Care. World Health Organization, Geneva, pp. 15–56. Bien, T.H., Miller, W.R., Tonigan, J.S., 1993. Brief interventions for alcohol problems: a review. Addiction 88, 315–336. Bischof, G., Reinhardt, S., Meyer, C., John, U., Rumpf, H.-J., in press-a. The expert test and report on alcohol (extra): development and evaluation of a computerized software program for problem drinkers. In: Columbus, F. (Ed.), Progress in Alcoholism Research. Nova Publishers, New York. Bischof, G., Grothues, J., Reinhardt, S., Meyer, C., John, U., Rumpf, H.J., in press-b. Evaluation of a telephone-based stepped care intervention for alcohol-related disorders: a randomized controlled trial. Drug Alcohol Depend.,doi:10.1016/j.drugalcdep.2007.10.003. Bott, K., Meyer, C., Rumpf, H.-J., Hapke, U., John, U., 2005. Psychiatric disorders among at-risk consumers of alcohol in the general population. J. Stud. Alcohol 66, 246–253. British Medical Association, 1995. Guidelines on Sensible Drinking. British Medical Association, London. Dybek, I., Bischof, G., Grothues, J., Reinhardt, S., Meyer, C., John, U., Brooks, A., Hohagen, F., Rumpf, H.-J., 2006. The reliability and validity of the alcohol use disorder identification test (AUDIT) in a German general practice population sample. J. Stud. Alcohol 67 (3), 473–481. Fleming, M.F., 1997. Strategies to increase alcohol screening in health care settings. Alcohol Health Res. World 21 (4), 340–347. Grant, B.F., 1997. Barriers to alcoholism treatment: reasons for not seeking treatment in a general population sample. J. Stud. Alcohol 58 (4), 365– 371. Greenfield, S.F., Weiss, R.D., Muenz, L.R., Vagge, L.M., Kelly, J.F., Bello, L.R., Michael, J., 1998. The effect of depression on return to drinking: a prospective study. Arch. Gen. Psychiatry 55, 259–265. Grothues, J., Bischof, G., Reinhardt, S., Hapke, U., John, U., Meyer, C., Rumpf, H.-J., 2005. Intention to change drinking behavior in general practice patients with problematic drinking and comorbid depression and anxiety. Alcohol Alcohol. 40 (5), 394–400. Hasin, D.S., Grant, B.F., 2002. Major depression in 6050 former drinkers: association with past alcohol dependence. Arch. Gen. Psychiatry 59, 794– 800. Kaner, E.F.S., Beyer, F., Dickinson, H.O., Piennaar, E., Campbell, F., Schlesinger, C., Heather, N., Saunders, J., Burnand, B., 2007. Effectiveness of brief alcohol interventions in primary care populations. Cochrane Database Syst Rev. (2), CD004148.
Kessler, R.C., Crum, R.M., Warner, L.A., Nelson, C.B., Schulenberg, J., Anthony, J.C., 1997. Lifetime co-occurrence of DSM-III-R alcohol abuse with other psychiatric disorders in the national comorbidity survey. Arch. Gen. Psychiatry (54), 313–321. Lynskey, M.T., 1998. The comorbidity of alcohol dependence and affective disorders: treatment implications. Drug Alcohol Depend. 52, 201–209. Miller, W.R., Rollnick, S. (Eds.), 1991. Motivational Interviewing. Preparing People to Change Addictive Behavior. The Guilford Press, New York. Moyer, A., Finney, J.W., Swearingen, C.E., Vergun, P., 2002. Brief interventions for alcohol problems: a meta-analytic review of controlled investigations in treatment-seeking and non-treatment-seeking populations. Addiction 97, 279–292. Regier, D.A., Farmer, M.E., Rae, D.S., Locke, B.Z., Keith, S.J., Judd, L.L., Goodwin, F.K., 1990. Comorbidity of mental disorders with alcohol and other drug abuse: results from the epidemiologic catchment area (ECA) study. JAMA 264, 2511–2518. Regier, D.A., Narrow, W.E., Rae, D.S., Manderscheid, R.W., Locke, B.Z., Goodwin, F.K., 1993. The de facto us mental and addictive disorders service system. Arch. Gen. Psychiatry 50, 85–94. Robins, L.N., Wing, J., Wittchen, H.U., 1988. The composite international diagnostic interview: an epidemiological instrument suitable for use in conjunction with different diagnostic systems and in different cultures. Arch. Gen. Psychiatry 45, 1069–1077. Rollnick, S., Mason, P., Butler, C., 1999. Health Behaviour Change. Churchill Livingstone, Kent. Rumpf, H.-J., Hapke, U., Hill, A., John, U., 1997. Development of a screening questionnaire for the general hospital and general practices. Alcohol. Clin. Exp. Res. 21, 894–898. Rumpf, H.-J., Hapke, U., Meyer, C., John, U., 2002. Screening for alcohol use disorders and at-risk drinking in the general population: psychometric performance of three questionnaires. Alcohol Alcohol. 37, 261–268. Saunders, J.B., Aasland, O.G., Babor, T.F., DeLaFuente, J.R., Grant, M., 1993. Development of the alcohol use disorders identification test (AUDIT): who collaborative project on early detection of persons with harmful alcohol consumption-II. Addiction 88, 617–629. Stasiewicz, P.R., Bradizza, C.M., Connors, G.J., 1997. Subject-collateral reports of drinking in inpatient alcoholics with comorbid mental disorders. Alcohol. Clin. Exp. Res. 21, 530–536. ¨ un, T.B., Sartorius, N., 1995. Mental Health in General Health Care Across Ust¨ the World. An International Study. John Wiley & Sons, New York. Velasquez, M.M., Carbonari, J.P., DiClemente, C.C., 1999. Psychiatric severity and behavior change in alcoholism: the relation of the transtheoretical model variables to psychiatric distress in dually diagnosed patients. Addict. Behav. 24 (4), 481–496. Wittchen, H.-U., Beloch, E., Garczynski, E., Holly, A., Lachner, G., Perkonigg, A., Vodermaier, A., Vossen, A., Wunderlich, U., Zieglg¨ansberger, S., 1995. M¨unchener composite international diagnostic interview (m-cidi), version 2.2. Max-Planck-Institut of Psychiatry, Munich. Wittchen, H.-U., Perkonigg, A., Reed, V., 1996. Comorbidity of mental disorders and substance use disorders. Eur. Addict. Res. 2, 36–47.