3.22.: f: . .3 2.. .: u. . . . ray: .u .8 13%.? a, {z 3.! $rrbxl .13.. .2 .. . o .4: .9... fiwfifigw .. v\.. .1. tall...’ .5! un. J: L. i c. ; ikhiyarx l... L... : ...4. . .11-: 1.335.. . .1! 1'. ,I‘D‘il‘v u)“ Eartha. :91 .. \ . .1: . . IE 5 .9 r a . ‘ g twwwfiukfi .55.}... . h.l ~’:.: .q’h...‘l:..9\ .- '38.. , . A T m L amnes L WMiiiifln'ififfnmi.rmmmm 3 01026 2248 LIBRARY Michigan State .Unlverslty This is to certify that the dissertation entitled Social Ecologies and Addiction Relapse: An assessment of a newly deve10ped setting risk measure presented by Maureen Ann Walton has been accepted towards fulfillment of the requirements for Ph.D. Psychology degree in met/6.30% Majorprofessor Thomas M. Reichl Date X"; 7' 73 MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771 PLACE II RETURN BOX to remove We checkout from your record. TO AVOID FINES return on or bdoro duo duo. DATE DUE DATE DUE DATE DUE JANl 32mm 0301Q§ MSU I. An Affirmative Adlai/Equal Opportunity Inflation WWI SOCIAL ECOLOGIES AND ADDICTION RELAPSE: AN ASSESSMENT OF A NEWLY DEVELOPED SETTING RISK MEASURE BY Maureen Ann Walton A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirement for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1993 ABSTRACT SOCIAL ECOLOGIES AND ADDICTION RELAPSE: AN ASSESSMENT OF A NEWLY DEVELOPED SETTING RISK MEASURE BY Maureen Ann Walton Despite addiction theorists' acknowledgement of the impact of environmental factors on relapse, these factors have not been adequately assessed in the empirical literature. Lack of addiction based environmental assessment tools may explain why environmental influences of relapse have been overlooked. The purpose of this study was to assess the validity of a newly developed setting based relapse risk measure. The measure assessed participants perceptions of risk for relapse, exposure to substances, and involvement in reinforcing activities in their home, work, and community settings. Eighty-five participants were interviewed three times over the course of six months. In addition to the setting risk measure, self-efficacy, coping, social networks, addiction problem severity, and relapse data was also collected. Results for the validity of the measure were mixed. The setting risk variable showed the most evidence for its construct, concurrent, and predictive validity. Specific setting (home, work, community) factors were differentially related to other relapse indicators offering support for the measurement approach taken. These findings demonstrate the need for further investigation into the impact of social settings on addiction recovery. DEDICATION To Matt iii ACKNOWLEDGEMENTS Several persons at Michigan State University helped to ensure the quality of this product. First, the time and effort of the dissertation committee members is appreciated. Members included: William Davidson, Thomas Reischl, Chatapuram Ramanathan, Bertram Stoffelmayr, Ellen Strommen, and Thomas Connor. Special thanks goes to Tom Reischl who served as my doctoral advisor, the chair of my doctoral committee, and director of the Addiction Relapse Prevention Project from which this study was based. He made himself readily available for consultation, provided quick turnaround of paper versions, and was generally flexible and accommodating. These are commendable attributes for an advisor to have. Second, the emotional and technical support of several of my friends/eco-graduate students at MSU was indispensable. My office-mate, Hoa Nguyen, provided free statistical consulting as well as hilarious verbal distractions. Bonnie Boots served as my best friend throughout this effort providing emotional and instrumental support. Her endless compliments and belief in my abilities were of tremendous help. Finally, Suzi Pavich, the graduate psychology secretary, was enormously helpful by ensuring all the university hoops were jumped through at the appropriate iv times. The love of my friends and family deserves special thanks. My friend Alex sent me inspirational e-mail messages and late night phone consultations about the meaning of doctoral education. My roommate Kris listened attentively to my daily complaints and joined me in needed celebrations of school accomplishments. My parents gave me the stubbornness, sense of humor, and belief in my abilities. Most importantly, Matt has supported me unconditionally and without complaint despite the sacrifices he has made due to the imposition on our time together. Words of appreciation are not sufficient to adequately acknowledge how he has helped me complete this degree. Thank you, again, to all my colleges, friends, and family. TABLE OF CONTENTS List of Tables List of Figures Chapter 1: Introduction Organization of the Literature Review Theories of Relapse Motivational Models Motivation-Conditioning Models Behavioral Choice Model Cognitive-Behavioral Models Family Influence Model Lifestyle Balance Model Advocacy Models Outcomes Implied from Theories Theory Guiding Present Study Psychological Research and Relapse Self-efficacy Studies Studies of Self-efficacy and Coping Coping Studies Studies of Coping and the Environment Summary of Psychological Studies vi \lmU'l 10 10 11 13 13 14 14 16 16 Social/Environmental Factors and Relapse Social Support Studies Social Setting Involvement Studies Social Ecologies and Coping Summary of Social/Environmental Studies Comprehensive Addiction Indices Purpose of Present Study Measurement Development Hypotheses Tested in Present Study Chapter 2: Method Piloting Procedures Participants Recruitment Design Procedures Intervention Description Support Program Standard Treatment Comparison Measures Descriptive Measures Proximal Outcome Measures Multi-Setting Relapse Risk Indicators Distal Outcome Measures Data Analysis Power Analysis vii 17 17 19 20 21 21 24 25 27 31 31 31 33 34 34 36 36 36 37 37 37 39 42 44 45 Chapter 3: Results Reliability Hypothesis Hypothesis Hypothesis Hypothesis Construct Validity Predictive and Concurrent Validity Sensitivity to Interventions Conceptual Model Chapter 4: Discussion Reliability Construct Validity Concurrent Validity Predictive Validity Sensitivity to Interventions Conceptual Model Methodological Concerns Summary and Conclusions Appendix: Measures References viii 47 47 48 56 69 75 81 82 83 86 87 89 91 92 93 96 116 LIST OF TABLES Table 1. Settings and variable domains contained in SRI...27 Table 2. Predicted relationships of SRI summary variables to other relapse indicators...............................29 Table 3. Demographic profile of study participants........32 Table 4. Correlations: reliability of the SRI variables..43 Table 5. Correlations between SRI summary variables and other relapse indicators..................................49 Table 6.Correlations among SRI settings by domains and other relapse indicators..................................51 Table 7. Correlations among home, work, and community settings by variable domains..............................54 Table 8. Discriminant analyses: concurrent validity of SRI summary domains and reuse.............................59 Table 9. Discriminant analyses: predictive validity of pretest SRI and posttest one reuse........................60 Table 10. Discriminant analyses: concurrent validity of risk settings domains and reuse...........................62 Table 11. Discriminant analyses: pretest risk settings predicting posttest one reuse.............................63 Table 12. Discriminant analyses: concurrent validity of home setting and reuseOOOOOOOOOOOOOOOOOOO0.0.00.000000000065 Table 13. Discriminant analyses: pretest home setting predicting posttest one reuse.............................66 Table 14. Discriminant analyses: concurrent validity of community setting and reuse...............................68 ix Table 15. Discriminant analyses: pretest work setting predicting posttest one reuse............................70 Table 16. Means and standard deviations for summary SEI domains for support and comparison groups across three assessments...0.0.0.0....OOOOOOOOOOOOOOOOOOOOO ..... 72 Table 17. Means and standard deviations for risk settings for support and comparison groups across three assessmentSOOO0......OOOOOOOOOOOOOOOOIOO00.... ..... 73 Table 18. Means and standard deviations for exposure settings for support and comparison groups across three assessmentSOCOO0..OOOOOOOOIOOOOOOOOOOOOOI...00.0.0074 Table 19. Means and standard deviations for reinforcing activities settings for support and comparison groups across three assessments.................................76 Table 20. Discriminant analyses: concurrent validity of conceptual model and reuse at posttest one............78 Table 21. Discriminant analyses: concurrent validitiy of conceptual model and reuse at posttest two............79 Table 22. Discriminant analyses: pretest conceptual model predicting posttest one reuse......................80 LIST OF FIGURES Figure 1. Conceptual model of relapse process.............12 xi CHAPTER 1 Introduction Attempts to explain relapse following substance abuse treatment typically focus on individual factors such as adjustment, motivation, stress, and coping (Marlatt & Gordon, 1985; Shiffman & Wills, 1985). Yet, many addiction theorists acknowledge the importance of lifestyle changes in addiction recovery (Vaillant, 1988; Gorski, 1986; Marlatt & Gordon, 1985; Donovan & Marlatt, 1988) suggesting that environmental factors such as social networks and social settings are also important factors in relapse. Emphasis on these social or environmental aspects of relapse has been limited by the lack of environmental measurement instruments (Tucker, Vuchinich, & Gladsjo, 1991). Thus, researchers are calling for assessment tools measuring social environmental settings (Smith, Frawley, & Howard, 1991; Finney & Moos, 1984; Tucker, et al., 1991; Longabaugh, 1989). These assessments could explain multiple aspects of client functioning (Finney & Moos, 1984) as well as redirect intervention approaches (Smith et al., 1991). The purpose of this study was to: (a) develop a measure of recovering persons' involvements with social settings, and (b) examine the validity of this measurement approach. 2 Before describing the social setting measure developed and tested in this study, a justification is necessary for the conclusion that this measure is needed. The conclusion that a social setting measure is needed is based on theoretical and empirical grounds. Therefore, the theoretical and empirical relapse literature will be reviewed for the purpose of indicating how social ecologies might be useful in explaining addiction relapse. This is followed by a review of existing social ecology measures in order to provide a rationale for the measurement approach taken in this study. The subsequent paragraph describes how this literature was generated. Computer searches using Medline, Psychlit, and Sociofile (1976-1993) supplied the references included in the review. The reason for choosing this time period was that the majority of articles related to relapse were not published until the early 1980's; however, a few key studies appeared in the late 1970's thus they were also included. Key words entered were: relapse, relapse prevention, alcohol abuse treatment, drug abuse treatment, environment. From the articles generated, the following guidelines determined whether an article was included. First, theoretical articles about relapse, addiction treatment outcome, and conceptualizations of human environment were included. Theoretical articles about addiction etiology were excluded. Second, due to the enormous nature of the treatment outcome literature, only articles examining 3 variable of interest to the proposed study were reviewed: coping, social involvements, social networks, and self- efficacy. Studies describing, or matching clients to 1 treatment were not included unless they examined long-term outcome issues. Further, biological studies testing new drug therapies were beyond the scope of this project. Scrutiny of the reference sections of articles generated additional studies. Organizetien ef the LiEereteze Review Addiction relapse theories are presented first with particular attention to psychological and environmental influences of relapse and the implications for relapse prevention. In this regard, application of these relapse theories requires that researchers consider measuring environmental variables in addition to psychological variables. Further, the theorized relationships between psychological and environmental constructs provides the basis for hypothesizing how a social ecological measure should be related to other relapse indicators for it to be valid. W Meeixetienal Models In addiction treatment and self-help settings, client motivation (e.g. denial, hasn't hit bottom) is a frequently used idea to describe why substance users do not succeed at remaining abstinent (Gorski, 1986). While motivation measures predict relapse, they are simplistic in that they 4 fail to point out specifically why the relapse occurred in order to provide targets for relapse prevention interventions (Wilkenson & LeBreten, 1986; Marlatt, 1977). MotiveEion-Conditioning Medele In contrast to individualistic ideas of motivation, motivation-conditioning models employ social learning theory to explain addiction relapse. First, operant principles are used to explain a person’s motivation for substance use or reuse. One hypothesized motivator is the idea that substances are used to experience positive effects (Stewart, de Wit, & Eikelboom, 1984); in operant conditioning terms, substances serve as positive reinforcers. Another hypothesized motivator is the idea that substances are used to avoid withdrawal symptoms (Ludwig 8 Wikler, 1974); this is explained by negative reinforcement as drug use is pleasant because it removes unpleasant withdrawal feelings. Second, classical conditioning is used to explain the mechanism that triggers a person's desire to experience positive effects or avoid withdrawal. Hypothesized triggers, of either the euphoria or the withdrawal symptoms, are environmental cues. In classical conditioning terms, drug use (unconditioned stimulus) is repeatedly paired with different environmental cues (neutral stimuli) which is followed by the drug’s euphoric or withdrawal effects (unconditioned response). Over time, the environmental cues alone (now the conditioned stimuli) elicit the euphoria or withdrawal (now the conditioned response). So, when 5 recovering substance abusers pass a place where they previously used substances, they experience craving for these substances. Despite the importance motivation—conditioning models give to the environment, proponents of this theory emphasize relapse prevention interventions that target changing the individual’s response to the environment, and not the environment itself, through cue extinction interventions (Tucker, et al. 1991; Childress, Ehrman, McLellan, 8 O'Brien, 1988; Childress, McLellan, 8 O'Brien, 1986). This focus on targeting the individual and not the environment is similar to the motivation models described earlier. 831W]. The behavioral choice model also emphasizes the importance of the environment and ideas of reinforcement to explain relapse. Theories of choice behavior state that: (1) substance use is dependent upon the availability of alternative reinforcing activities; and (2) by examining the reinforcement conditions in a variety of life areas under which substance use becomes the behavior of choice, one would discover that environmental constraints exist that prevent the person from finding reinforcement by other means (Vuchinich 8 Tucker, 1988: Tucker et al., 1991). The underlying assumption for this theory is that a recovering person will remain abstinent if they have substance free reinforcing activities in their life. Like the motivational-conditioning theory, the 6 behavioral choice theory focuses on environmental contingencies. However, instead of relapse prevention interventions attempting to change the individual's response to the environment, supporters of this theory suggest changing the environment to supply new activities that provide the desired euphoric or reinforcing state. Exactly how to assess information about such reinforcing events and how to implement such a change is not clear. Tucker et a1. (1991) suggested measuring the frequency of valued life events and interruption of these events by previous substance use. Treatment focus would then be on re- establishing access to these life events. One limitation to this approach is that the reinforcement value of an activity or event varies widely according to the behavior setting in which it occurs (Wicker, 1972). Therefore, the reinforcement value of an event needs to be examined within the behavior settings in which it occurs. Qegn1§iye—Behaviorel Medels The most popular theory regarding relapse is based on Lazarus's (1966) stress-coping model in which substance use is viewed as a coping response to a stressful or risky situation (Marlatt 8 Gordon, 1985; Annis 8 Davis, 1988: Shiffman 8 Wills, 1985; Shiffman, 1989; McCrady, 1989). Psychological processes (cognitions) are considered the primary influences of relapse such as: self-efficacy, positive or negative expectations, and attribution of cause (Marlatt 8 Gordon, 1985) or life stress, substance use cues, 7 and problem severity (Shiffman, 1989). Once in a high risk situation, relapse is determined by an interaction between cognition and the availability of coping responses (Marlatt 8 Gordon, 1985). For example, whether coping responses are used is thought to be determined by the person's self- efficacy or perceived capabilities for executing the coping response. Expectations regarding the consequences of using or not using the coping response as well as expectations surrounding substance use also influence reuse. Once the person uses alcohol/drugs, their attributions for the cause of this initial use determine future use. Internal attributions are thought to predict reuse whereas external attributions are thought to predict a return to abstinence. Further, cognitive dissonance (e.g., I am an abstainer but I just used) and affective reactions to using (e.g., guilt) are called the "abstinence violation effect" and are hypothesized to predict continued use (Marlatt 8 Gordon, 1985). While the stress-coping model of the relapse process allows for some environmental impacts of stress, cue, or high risk event on relapse (Marlatt 8 Gordon, 1985), the environment is given secondary importance as managed through one’s cognitions and coping repertoires (Tucker et at., 1991). Thus, interventions springing from these approaches are based on increasing an individual’s coping repertoire (Chaney, O'Leary, 8 Marlatt, 1978). Femily Igfluence Mode; Family influences are given primary attention in the 8 study of addiction etiology (Fitzgerald, Davies, Zucker, 8 Klinger, in press) but are seldom mentioned in the relapse area. McCrady (1989) offered a promising expansion of stress-coping theories by including family influences, particularly those of spouses. McCrady noted the importance that significant others have in a substance user's life and hypothesized that the user really faces two high risk situations: (1) the high risk situation itself, and (2) the spouse's reaction to the substance user’s relapse or abstinence following the high risk situation. The spouse's coping reaction is thought to depend on similar variables as described under the cognitive-behavioral model. These include: attributions about why the user has quit using; outcome expectancies for the user’ relapse; and their own self-efficacy for dealing with the user's relapse. The spouse’s coping reaction can either facilitate or attenuate the user’s recovery. The user's coping response is affected by expected reinforcement from the spouse for abstinence or expected loss of reinforcement from the spouse for relapsing. Thus, like the cognitive-behavioral model, cognitions are viewed as important and like the behavior choice model, reinforcement is viewed also as important but its source is a person and not an activity. No one to date has explicitly used this specific model as a basis for a relapse prevention intervention. This model, however, is implied in social network therapies that involve providing therapy for substance abusers and their social networks 9 concomitantly (Galanter, 1987). st e ala ce ode Theories by clinicians and researchers as well as self- help ideologies assert that maintenance of sobriety requires a lifestyle change before treatment effects can be sustained (Gorski, 1986; Living Sober, 1975; Marlatt 8 Gordon, 1985). Substance abusers are renown to live an "addictive lifestyle" of which substance use is only one part. Furthermore, according to Alcoholics Anonymous ideology, former "people, places, and things" can exert negative influences especially on early recovery. Before abstinence can be established users must change their social networks, social activities, and social settings from substance using to primarily non-using (Living Sober, 1975). Marlatt (1985) espoused a lifestyle balance concept which involved the substance abuser balancing out "shoulds" or obligations, and "wants" or leisure activities in their new sober life. This concept is similar to the behavior choice theory which places importance on the availability of reinforcing activities. For the substance abusers, this involves replacing negative addictions with positive addictions such as exercise or meditation so that the person still has enjoyable "wants" in their life. The emphasis of these lifestyle models is on both psychological (stress, bad habits) and environmental influences (social influences). Surprisingly, few empirically controlled studies have applied this theory. Marlatt (1985) described several case 10 studies in which the substance user's lifestyle was changed through altering the environment to a lifestyle of activities incompatible with substance use. For example, a person who drank large quantities of alcohol at lunch was directed to take up exercise during lunch instead of frequenting a social setting (restaurant) in which alcohol was present. AQ¥Q§§Q¥_MQQ§1§ Advocacy theory has viewed relapse as a function of a shortage of various environmental resources, e.g., social support, income/employment, drug-free recreational activities, etc. (Fagan 8 Mauss, 1986). The mechanism by which the resources prevent relapse is not made explicit; implicit is the idea that resources buffer life stress. Relapse prevention strategies, according to these models, would include providing a case advocate to assist the person in obtaining resources and eventually empower the person to be their own advocate. As in the behavior choice and lifestyle change models, this model focuses on the environment and not psychological influences of relapse. O omes ' o i s Consistent in all of the above theories is the inclusion of multiples levels of outcomes following substance abuse treatment. While all of the theories described above focus on relapse as a distal treatment outcome, the theories vary according to whether they include proximal outcomes that are psychological or environmental in 11 nature (Martin 8 Wilkenson, 1989; Nathan 8 Skinstad, 1987; Leukefeld 8 Tims, 1989; Eriksen, Bjornstad, 8 Gotestam, 1986). The conditioning and cognitive-behavioral models focus primarily on psychological level outcomes. Proximal outcomes according to the conditioning models are defined as changes in the person's reaction to environmental cues. According to the cognitive-behavioral models, proximal outcomes consist of measuring expectancies, attributions, and coping skills. The behavioral choice, lifestyle, and advocacy theories focus more on social factors. Proximal outcomes, according to the behavioral choice model, are defined as decreased access to substances and increased access to drug-free pleasurable life activities. Proximal outcome measures based on the lifestyle model would consist of measuring frequency of enjoyable leisure activities, social network sobriety membership, and social setting support for abstinence. For the advocacy relapse model, proximal outcomes include availability of social support, employment, housing, and drug-free recreational activities. I] 3 .3. E ! E! l The model guiding this research was a synthesis of the above theories. The central organizing theme was that human behavior, in this case substance abuse, is a product of fit between individual person factors and social environmental factors (Lewin, 1935; Bronfenbrenner, 1979; Pargament, 12 1986). Further, multiple levels of proximal and distal outcomes determine the relapse process. Thus, relapse following substance abuse treatment was hypothesized as a function of several of the psychological and environmental resources highlighted in the various models discussed (see Figure 1). Psychological resources influencing relapse examined included self-efficacy and coping strategies. These were drawn from the cognitive-behavioral theories. Environmental resources that were hypothesized to influence reuse were various social involvements (social problems, social settings, social networks). These were drawn from the advocacy, conditioning, behavioral choice, and lifestyle theories. Rationale for inclusion of these constructs was also based on their influence on relapse as described by the empirical relapse literature discussed below. icky Setting lndicatore 9 Relapse Environmental Resource. Individual Reeourcee Figure 1. Conceptual model of relapse proceee. 13 According to the model proposed, environmental resources (social networks and setting risk indicators) will be correlated. Setting risk indicators are exogenous variables predicting relapse directly and indirectly through coping and self-efficacy. Coping and self-efficacy are endogenous variables, directly predicting relapse. This model provides the conceptual basis for hypothesizing how a social setting measure might be related to other relapse indicators. This model will not be tested as a causal model of relapse since this was beyond the focus of this study which was to determine the validity of the setting risk measure. Before testing the validity of the setting measure in predicting other relapse indicators, the validity of the other indicators in predicting relapse needs to be established. The research describing the relationship between relapse indicators and relapse is grouped into three sections: psychological research, social/environmental research, and comprehensive addiction indices. Ps cho 1c 5 s Self-Efficacy SEediee According to cognitive-behavior theories, one factor that may be critical in determining whether a person will cope with a risky situation by turning to substance use is self-efficacy (Bandura, 1977; Marlatt 8 Gordon, 1985). Using retrospective and prospective designs, researchers have determined: that greater self-efficacy was related to lower relapse rates (Yates 8 Thain, 1985; Barrios 8 Niehaus, 14 1985; Walton, 1988); that greater self-efficacy differentiated a person who has a minor lapse from one who has a major relapse (Condiotte 8 Lichtenstein, 1981); and that greater change in self-efficacy over the course of treatment was more predictive of abstinence following a relapse crisis than those with initial high or low self- efficacy that does not change over the course of treatment (Burling, Reilly, Moltzen, 8 Ziff, 1989; Brandon, Tiffany, Obremski, 8 Baker, 1990). Thus, self-efficacy is related to reuse as predicted by the cognitive-behavioral models. Studiee_2;_Self:£ffisasx.end.§eeins Theoretically, self-efficacy is thought to determine whether coping strategies are used. Timmer, Veroff, 8 Colten (1985) found that under conditions of high stress, persons who had high self-efficacy were less likely to use substances to cope than people with low self-efficacy. However, this study did not examine the impact of self- efficacy on use of other more positive coping strategies. Qopieg §§udies Most studies of coping and relapse do not include use of drugs or alcohol as a coping mechanism. Therefore when the coping research is described below, "coping" refers to non-using strategies. A vast number of studies explored the theorized effect of coping and relapse and found that regardless of how coping is measured, it is consistently predictive of abstinence. For example, relapsers score lower on coping measures than do abstainers (Rosenberg, 15 1983). Other authors found the number of coping responses was predictive of abstinence following a relapse crisis (Bliss, Garvey, Heinold, 8 Hitchcock, 1989; Litman, Eiser, Rawson, 8 Oppenheim, 1979); and that perceived effectiveness of coping behaviors were predictive of reuse (Litman et al., 1979; Litman, Stapleton, Oppenheim, Peleg, 8 Jackson, 1984). No single coping response was found to be consistently superior to another in preventing relapse; although, Billings and Moos (1983) found that the crucial coping strategy predictive of abstinence was avoidance strategies (along the lifestyle change model). Several authors have noted that abstinence following a potential relapse crisis is influenced by the use of egg cognitive-behavioral coping response, regardless of type, as opposed to a specific coping response (Shiffman, 1982; Curry 8 Marlatt, 1985). No one coping strategy may be universally superior since a particular coping strategy may be more or less effective based on the social situation in which it occurs. Two studies concluded that their data did not support Marlatt's relapse model in that coping responses executed effie; an initial slip (lapse) did not predict abstinence or continued use (relapse) (Baer, Kamarack, Lichtenstein, 8 Ransom, 1989; Brandon, Tiffany, 8 Baker, 1988). Thus, upon lapsing, other variables may be more potent moderators. In this regard, McCrady's (1989) theory that the user faces a second high risk situation once lapsing, the reaction of the spouse (or other social network members), may apply; the 16 expectation of the significant other’s reaction could determine if user's lapse becomes a relapse. Thus, while evidence is fairly conclusive that coping responses are important determinants of initial lapses, it is not clear if coping determines continued use or relapse. §tudies_2f_Q2nins_and_the_finxir2nment Some theorists have acknowledged the importance of environmental variables on facilitating or inhibiting coping responses (Lazarus 8 Folkman, 1984). Thus, instead of viewing coping from a competence only standpoint, coping was proposed to be mediated by the environment which provides resources, constraints, and demands (Lazarus 8 Folkman, 1984; Sarason, 1977). Cronkite and Moos (1980) verified this conceptualization among alcoholics by finding the combined effects of coping and intake social resources were better at explaining relapse than coping alone. SummaIY_2£_E§29belegieal_§tudiee Therefore, there is some evidence to support the cognitive-behavioral theories of relapse. Both self- efficacy and coping are predictive of relapse following a high risk situation. Yet, these psychological factors do not exist in a vacuum; they exist in a social setting. It is likely then that the environment interacts with these psychological factors as exemplified by the finding that material resources affect coping responses (Cronkite 8 Moos, 1980). 17 o ‘ v s This section documents the theorized influence of social/environmental resources on relapse. Most of the published literature on environmental influences focuses either on social support/network factors or on social setting factors. Wise Research has shown that alcoholics and drug addicts tend to lack social skills and therefore are isolated from mainstream society (O’Leary 8 O'Leary, 1976). Their entire environment revolves around drug use (Fraser 8 Hawkins, 1984) so that after treatment, lack of environmental support attenuates recovery (Havassy, Hall, 8 Tschann, 1986; Page 8 Badgett, 1984; Joe 8 Simpson, 1983). Family involvement in treatment (Moberg, Krause, 8 Klien, 1982), involvement in aftercare support groups (Wallace 1989; Svanum 8 McAdoo, 1989), and support after treatment facilitates sobriety (Moos 8 Finney, 1983; Captain, 1989; Mermelstein, Lichtenstein, 8 McIntyre, 1983). Most treatment programs recognize the substance user's isolation and prescribe attendance to self-help groups (eg. Alcoholics Anonymous, AA) as settings for supportive sober leisure life (Catalano 8 Hawkins, 1985; Marlatt 8 Gordon, 1980). Yet, existing data documenting the benefits of AA are difficult to understand due to sampling bias. Correlational data shows modest support for AA facilitating sobriety maintenance (Williams, Stout, 8 Erickson, 1986; 18 Sheeren, 1987; Vaillant, 1988); however several prospective studies have not found such effects (McLatchie 8 Lomp, 1988; Thurstin, Alfano, 8 Nerviano, 1987). It could be that AA’s positive effects vary according to person variables such as motivation or differential setting variables such as reinforcement. Research consistently shows that the substance abusers’ social network influences relapse; although, the network factor studied varies. While some research suggests that network size is an important influence of relapse (Favazza 8 Jackson-Thompson, 1984), other research suggests different network factors are important such as perceived support and members’ substance use (Rosenberg, 1983; Brown et al., 1989; MacDonald, 1987). For example, in a study of twelve alcoholics, Favazza 8 Jackson-Thompson (1984) found that those who relapsed significantly reduced their network size as compared to abstainers. This result should be interpreted with caution due to the very small sample size. In studies of larger sample sizes, abstainers retrospectively reported greater perceived support for sobriety than do relapsers although the actual number of contacts was not different (Rosenberg, 1983). Prospectively, relapsers reported associating with pretreatment drug using friends (Brown et al., 1989) or with those non-supportive of sobriety (MacDonald, 1987). Thus, this research provides a basis for concluding that social support and social networks influence relapse. 19 Some tangential evidence exists for the impacts of social settings on relapse. Vaillant (1988) concluded that those who remain abstinent were those who change their entire pattern of living, particularly their social structure. Unfortunately, research documenting this conclusion is limited perhaps due to the lack of measures available to measure these lifestyle or environmental setting changes (Tucker et al., 1991). Moos 8 Bromet (1977) found that marital and employment stability at intake was related to less behavioral impairment from drinking and better psychological and social functioning at follow-up; they did not examine actual reuse. This highlights the importance of marital and work resources in addiction recovery. Billings and Moos (1983) measured reuse and found that recovering alcoholics had social resources similar to controls while relapsed alcoholics had less positive work, family, and informal support networks. Other studies that have examined vocational rehabilitation (Lowe 8 Thomas, 1976; Towle, 1974; Page 8 Badgett, 1984; Wanberg 8 Horn, 1983), or social/marital involvement (Wanberg 8 Horn, 1983) have found supportive work and social involvements were consistently predictive of sobriety. These findings support the advocacy model of relapse and suggests that resources are inversely related to reuse. When specifically examining social setting using the Family Environment Scale (FES) (Moos, 1974), Bromet and Moos 20 (1977) found at six months that a positive family milieux (high cohesion, low conflict, high support, recreation) at follow-up was related to better outcomes: less behavioral impairment; fewer self-rated problems; and better social and psychological functioning. The most important predictor of functioning was family involvement in active recreation (Moos, Bromet, Tsu, 8 Moos, 1979) with drinking persons perceiving family leisure time more negatively. Finney, Moos, and Mewborn (1980) however found that only family cohesion was related to less reuse at two year follow-up. By examining setting using the Work Environment Scale (WES), several studies found that married alcoholics’ (who resided with their families) work environments were not related to behavioral impairment, self-rated problem, and social or psychological functioning (Moos 8 Ingel, 1974; Bromet 8 Moos, 1977) or reuse (Finney, Moos, 8 Mewborn, 1980). Among non-married alcoholics, a more positive perception of the work environment was associated with better functioning (Moos 8 Ingel, 1974; Bromet 8 Moos, 1977). The authors concluded that location in families may buffer the negative impacts of work environments (Bromet 8 Moos, 1977) MW Cronkite and Moos (1980) found that family environment was related indirectly to reuse through stress and coping. High coping was related to positive family environment and high stress was related to low positive family environment. 21 u a of ' vi e e These data provide some support for the lifestyle and behavioral choice theories demonstrating the importance of measuring social ecologies both at home and work in order to best explain addiction relapse. Still, social/environmental setting factors (home, work, community) have not been as extensively studied. While the PBS and WES are two of the best in measuring different social ecologies they have limited utility since the items are not Specific-to substance abuse. They also fail to consider the social ecologies of neighborhoods in which the family is embedded and the community in which leisure activities take place. C e s'v c 'o s Addiction researchers are calling for comprehensive outcome assessment tools (Eriksen, Bjornstad, 8 Gotestam, 1986; Leukefeld 8 Tims, 1989; Wells, Hawkins, 8 Catalano 1988; Nathan 8 Skinstad, 1987) especially those tools examining environmental variables (Maisto 8 Connors, 1988; Moos 8 Finney, 1983; Maisto 8 Conners; Tucker et al., 1991). Yet, very few reliable and valid environmentally based addiction measures exist. Several early attempts at developing comprehensive measures were not successful since these measures were not related to reuse outcome; these measures did not assess environmental factors. They included the Clinical Outcome Score (Schuckit, Morrison, Gold, 1984; Schuckit, Schwei, 8 Gold, 1986), Background Information Form (Bromet 8 Moos, 22 1977), and the Multidimensional Index (Congdon 8 Holland, 1988). One of the most widely used reliable and valid comprehensive measurement tools is the Addiction Severity Index (ASI); however, the original severity ratings produced were meant as an intake addiction problem assessment, not an outcome measure (McLellan, et al., 1985) so more objective composite scores were developed. Several studies have investigated whether the ASI composites improve over the course of treatment. Results show trends toward improvement in medical, employment, social, drug, alcohol, legal, and psychiatric status (McLellan, Luborsky, Woody, O’Brien, 8 Kron, 1981; McLellan, O’Brien, Woody, Luborsky, 8 Druley, 1982; McLellan, Luborsky, 8 O’Brien, 1986; Sanchez-Carbonell, Cami, 8 Brigos, 1988); although, inconsistencies have been found in that improvements have not always been significant (McLellan, Luborsky, 8 O’Brien, 1986; Lesieur 8 Blume, 1991; Woody, McLellan, Luborsky, 8 O’Brien, 1987). These findings may be related to: sample characteristics, e.g., alcoholic or drug addict; low problem composites at intake implying floor effects might limit amount of improvement available to be measured (a higher score indicates more severe problems); type of treatment experienced; follow-up period as improvements following treatment typically dissipate over time. Improvement in follow-up composite scores is not always related to improvement in alcohol and drug use composites 23 (McLellan, Woody, Luborsky, O’Brien, 8 Druley, 1983). A few studies have interrelated the ASI functioning composites to determine if reuse is multidimensional. Using factor- analytic techniques, Kosten, Rounsaville, and Kleber (1987) found that at follow-up, the drug, alcohol, and legal composites were independent of the other four composite areas (medical, family, psychiatric, and employment, social). Alterman, Kushner, and Holahan (1990) performed canonical correlations with difference scores and found that the alcohol and drug use composites were independent of all of the other composite areas. The use of difference scores which compound the unreliability of measures makes confidence in this finding tentative. In another study, pretest alcohol and drug composites were related to the legal composite but not to other areas; at the posttest, both the alcohol and drug composites were related to the psychiatric composites but not to the other composites (McLellan, Luborsky, Woody, O’Brien, 8 Kron, 1981). Together, these results suggest that the addiction related problems for a group of substance abusers are generally not related and vary with each case. Only two studies compared the ASI composites to independent measures of reuse._ Kosten et al. (1987) found that using difference scores (which are notoriously unreliable) only the drug and legal composites were related to reuse. Kadden, Getter, Cooney and Litt (1989) compared only the psychiatric and employment composites with reuse of 24 alcohol and found that the psychiatric composite was predictive of reuse; the employment composite was not related to drinking. In summary, the ASI problem composites typically do show improvement over the course of treatment; when the drug and alcohol composites are compared to the other composites they are generally not related to one another and are not necessarily related to relapse. The variation observed may be because many of the items used in the composite indexes are not directly related to drug or alcohol use such as "having a car available for use" or "net income". Other variables may have more consistent impact, such as the social environment and drug involvement in various work, home, and community settings (Cronkite 8 Moos, 1980; Moos 8 Finney, 1983). Alternatively, some argue that the various composites are not correlated because they represent separate dimensions of functioning (Martin 8 Wilkinson, 1989). s h ud The present proposal anticipates filling the gap in the measurement of social settings by producing a reliable and valid multi-setting relapse risk indicators measure. In general, validity of the measure will be determined according to whether it is: (1) related to variables that are predictors of addiction outcomes, (2) related to reuse retrospectively and prospectively, and (3) sensitive to the effects of interventions. The measure will be used as part 25 of a larger longitudinal study of a skills building social support relapse prevention intervention. as e t v e The multi-setting relapse indicator (SRI) measure developed for this study assessed individuals’ perceived quality of social involvement in three social settings: work/school, home/residence, and community/leisure. These three settings were chosen to obtain maximum ecological coverage of a person’s microsystems (Bronfenbrenner, 1979). There are several reasons why the measure focuses on a person’s perceptions of their social environments and not observation of their behavior settings. First, several prominent theorists argue that the meaning a person gives to different aspects of the environment feneeien is more powerful in determining that person’s growth than the objective physical conditions fern (Lewin, 1935; Bronfenbrenner, 1979; Perkins, Burns, Perry, 8 Nielson, 1988) regardless of whether these perceptions are accurate (Wicker, 1987). Further, the measure of environmental attributes requires non-intrusive observers of the physical, temporal, and behavioral aspects of a setting (Wicker, 1972). This is extremely costly and time intensive as well as impractical for substance abusers whose use may take place in private settings, e.g. home (Perkins et al., 1988). For these reasons, a self-report measure was chosen to include the substance users’ perception of social setting relapse indicators. 26 Within each of the three settings (home, work, community) variables included in the SRI were: (1) exposure in setting to drugs or alcohol, (2) perceived risk for relapse in setting, (3) availability of reinforcing activities in the setting (see Table 1). These constructs were based on a synthesis of several ideas from social ecologists regarding the commonalities found in setting; they were then tailored to be salient for addiction recovery. Exposure to substances in a setting is based on the influence of interpersonal connections that take place in settings (Bronfenbrenner, 1979; Moos, 1973); it is also similar to the social network literature. Perceived risk for relapse relates to the stress-coping theories and also acknowledges the influence of role expectations (Bronfenbrenner, 1979). The reinforcement construct is also similar to the lifestyle balance theory, where "shoulds" are equal to "wants" and although the exact label varies, most ecologists note that personal needs, self-enhancement or satisfaction are important aspects of settings (Moos, 1973; Wicker, 1987; Insel 8 Moos, 1974; Barker, 1963). Setting substance use is merely a physical/behavioral aspects of the environments. The SRI variables can also be combined across setting to create summary domains. 27 table 1. Settings and variable domains contained in SRI. Variable Domains Setting Risk for Exposure to Drugs or Settings Relapse Reinforcing activities alcohol Home Home Risk Home Reinforcing Rome Exposure Activities Work Work Risk Work Reinforcing Uork Exposure Activities Community Community Risk Community Reinforcing Community Exposure Activities All settings Statuary SUIIIIII’Y Summary Exposure Risk Reinforcing Activities e s st ' es Once again, the purpose of this study was to determine the reliability and validity of a newly developed multi- setting relapse risk indicator measure in the context of a larger longitudinal study of a support group relapse prevention intervention. Validity was examined by analyzing the pattern of relationships with other variables. In addition, the sensitivity of the measure to distinguish change in skill level (between intervention groups) and to show change in relapse status over three and six month posttest assessments were examined. Detailed hypotheses that were tested as part of this research are as follows. 28 HypoEQesie 1. The SRI will demonstrate convergent and discriminant validity. The correlates among the SRI variables, addiction treatment outcome measures, and relapse predictor variables will be in accordance with the expected pattern of relationships. Table 2 lists the predicted relationships between the SRI variables and the other predictor variables. fl!22§h§§i§.21 The SRI variables, summary risk for relapse, exposure to drugs/alcohol, and summary reinforcing activities, will be retrospectively and prospectively related to relapse status. This concurrent and predictive validity relationship will be tested for data at both posttests. Hypetnesie . The SRI will be sensitive to the effects of an intervention by distinguishing a relapse prevention support group and a comparison group. The support group is expected to have higher scores on summary reinforcing activities and lower summary risk for relapse and exposure to drugs/alcohol than the comparison group at both posttests. Implied in this hypothesis in the notion that the intervention group will relapse less than the comparison group. This hypothesis will be tested in order to determine whether the intervention had any effect independent of the validity of the SRI. 29 A+. A.. A+. A-. A... A... A... A.. A.. ace-8A5 fan-3m A... A-. A-. A+. A+. A... A+. A+. Ao. c.5396 35v. A.. A-. A... A+. A+. A+. A... A+. A+. no; {85m .39... 833...; 83.30 toga 3.3. 9:38 .395 .39... .39... .28... :8 .33 to .33 .83 .5: 93-3.. «cg—Eu .30.. 2.33:3... R. .5. 030: .2385... eon-.2 segue 8 83:...» til: :3 to 8:58:32 8.2%.... .~ .3: 30 Hypoehesis 5. The pattern of relationships suggested in the theoretical model described earlier (in Figure 1) will be supported indicating construct validity. CHAPTER 2 Method P ot' r cedu es All measures, including the setting relapse indicator, were piloted among fifteen substance abuSers for clarity, item content, and item distribution across response choices. Pilot subjects were recruited from a local Alano Club using snowball sampling techniques. W This study reports on the first 85 participants that were recruited from a parent study, the Addiction Relapse Prevention Project. Participants with a diagnosis of schizophrenia or treatment with methadone maintenance were excluded from participating since both of these conditions could affect treatment outcome. The majority of the participants in this sample were male and white. Table 3 provides a more detailed demographic profile of the participants. Information was also collected regarding the participants’ substance use histories. Participants’ lifetime regular substance use averaged fourteen years and ranged from one year to thirty-five years. Sixty-five percent of participants identified alcohol as their major problem substance; the remainder of the participants 31 32 Table 3. Demographic profile of study participants. Demographic Variables Participant Profile Age Gender Race Education Employment Monthly income Marital Status §=36, §Q=9.2, RangeIZO-éo 65% male 35% female 71.8% shite 21.2% African American 7.1: Hispanic 18.8% < 12 years 51.8% 12 years 29.6% >12 years 76% Full-time 16.5% part-time 16.3% unemployed 15.3% 80 48.2% <=S1000 27.1% 52000-3000 9.4% >33000 22.6% Married 43.5% Separated, Divorced 34.1% Mever married 33 identified themselves as having a problem with one or more other drugs. Most participants (55%) had been treated more than once for substance abuse with the number of treatment episodes ranging from two to twenty. Participants' abstinence periods varied: 29% had a month or less; 41% had two to six months; 25% had seven to twelve months; and 6% had thirteen to sixteen months of sobriety. Witness; Most participants were recruited to this study after successfully completing a substance abuse treatment program. Programs varied and participants came from inpatient (28%), outpatient (25%), traffic (16%), and residential (25%) centers. A few participants (6%) were also recruited from Alcoholics Anonymous. Recruitment strategies used were: flyers posted at treatment centers and Alano clubs, referrals from addiction counselors, and group presentations made at treatment centers. At the time of recruitment, a "consent to be contacted" form was signed. Potential participants were contacted after completing treatment by an interviewer to arrange the pretest interview. Before beginning the pretest, an informed consent was presented explaining the study protocol, confidentiality, and all possible risks and benefits of participation. At that time, if the person agreed to participate, the first interview took place. 34 Dfiisn This study employed a longitudinal quasi-experimental design. The first 40 participants were recruited between February and March of 1992 and assigned to participate in a ten week skills building social support intervention. The remaining participants were recruited in two waves (the first from March to June of 1992, the second from September to January of 1993) and were assigned to the comparison no- additional treatment condition. Attempts were made to interview participants over the course of six months (see procedures). EIQEEQHIEE All participants received an initial pretest interview which was followed by participation in the support groups if so assigned. Three months and six months after the pretest, attempts were made to re-interview participants. Each interview lasted approximately two hours. Participants were paid a base rate ($5.00 per hour) for each interview, plus a bonus of five dollars for each consecutive completed interview: the total received upon completion of all interviews was forty-five dollars. All interviews were administered either at the treatment center, the research office, or other location of mutual convenience. 21g§§_t. Participants were paid ten dollars for this interview. Measures included at this assessment were: addiction severity, coping, social network, self-efficacy, social ecology involvement. 35 Interim. At each group meeting, attendance was taken in addition to other process information not reported here. Attendance ranged from 0-10 meetings with 45% of the participants attending at least half of the meetings. ggsttgst 1. This interview occurred three months after the pretest. Participants received a fifteen dollar compensation for this appointment. Variables measured were identical to the pretest with an addition of a relapse measure, the Timeline Calendar. Egsttest 2. Subjects were interviewed again six months after the pretest. (The rationale for this time period is that research shows the majority of relapses occur by six months post-discharge.) They were paid twenty dollars for completion of this interview. Variables measured were identical to the previous interview. Tracking. Names and addresses of three significant others was requested at the pretest. In addition, participants were contacted half way between the posttest (month two) and follow-up (month five) via a letter. Additional efforts to increase compliance included: beverages provided at the interviews, postage paid change of address cards, acceptance of collect calls, and business cards stating the project phone number and the date of the next appointment. These efforts resulted in a successful interview completion rate of 92% at the three month interview and 98% at the six month interview. 36 Intervention Description fispeert_£res£am dezyigg. The groups met on a weekly basis for two hours for ten weeks. Groups were led by a trained paraprofessional seeking their addiction counselor certification. ngeging_. Each meeting began with the reading of a confidentiality pledge. This was followed by pairs completing forms asking questions about risky situations encountered during the pgggigng week and coping methods that were used. The pairs then shared with the total group. This was followed by an educational exercise (e.g., reading) about addiction recovery selected either by the group leaders or by the participants. Participants then broke up into pairs again and completed anticipated risky situations for the gnggming week: pairs also discussed potential coping mechanisms. This was followed by reporting to the groups. Members next exchanged sobriety support cards containing their name and telephone number. In this manner, it was hoped that the group social support would be carried back into daily living situations. The group closed with general group concerns, the confidentiality pledge, and refreshments. (A video of the group process can be obtained from Dr. Reischl at Michigan State University.) MW Participants assigned to this group received any aftercare treatment that was part of the treatment program 37 from which they were recruited but did not receive any additional treatment from the project. eas s For copies of all measures see Appendices. es i Demographic information obtained included age, race, education, employment, monthly income, and marital status. WW §Q§i§i_nggggrk§. Social support was measured through a social network analysis similar to that of Norbeck, Lindsey, and Carrieri (1981). Social networks were delineated according to four areas: partner, family, friends, professionals, others. (The number of names in the network was not limited.) Questions were then asked of each member’s frequency of contact, closeness, extent of drug or alcohol use, and support for sobriety. From these questions, two variables were computed: (1) ratio of the number of non-users to the total number of network members and, (2) mean perceived support for sobriety from network members. aning. Participants responded to a twenty-two item scale according to the frequency of the use of coping strategies in response to a high risk situation. This inventory was based on the constructs in the COPE inventory (Carver, Scheler, & Weintraub, 1989). The reliability and validity of the original scale has been demonstrated using test-retest/internal consistency and convergent/divergent 38 assessments respectively. The revised scale was examined for internal consistency via a factor analysis which yielded a four factor solution: behavioral disengagement, cognitive coping, help seeking, and active-expressive coping (Reischl, Ramanathan, & Nguyen, 1993). The behavioral disengagement factor was used in this study as it best represented "negative" or ineffective coping strategies. Items were removed, however, that included use of drugs or alcohol. The remaining four items (g=.62) referred to isolation (e.g., be alone for a period of time) and withdrawal from others (e.g., give up trying to reach your goals in the situation). §gi§;gffiigggy. Self-efficacy was measured using a shortened version (twenty-four items) of the 100 item Situational Confidence Questionnaire (Annis, 1982a). The original items were developed from Marlatt's situation categories and were reliably coded in these categories (Annis, 1982b). In order to best represent the original questionnaire, three items were retained from each of the original subscales. The original questionnaire asked clients to rate, in percentages, how confident they are they could resist the urge to use drugs/drink in various situations. For this study the response choices were simplified to a five point Likert scale ranging from "not at all" to "extremely." The revised scale was shown to be internally consistent, a=.91. 39 Sev 't e . This instrument is used to provide composite scores for six problem areas: medical, employment, family/social relationships, drug/alcohol use, legal status, and psychiatric status (McLellan et al., 1985). These composite scores have been shown to be reliable and valid using test-retest and convergent and discriminant methods (McLellan, Luborsky, Cacciola, Griffith, Evans, Barr, & O’Brien, 1985). '- t ' s ' o 5. Social environment was assessed for the past month for three settings: home, work, and community. Within each setting, three variable domains produced were: exposure to substances in the setting, setting risk for relapse, and involvement in reinforcing activities in the setting. Thus, nine specific setting variables were produced (see Table 1). In addition, three summary domains were computed across the settings (see Table 1). Calculation of the specific setting and summary variables in described in detail below. aQmgL_EQIKL-QDQ-QQEEQDiEX—Bifik° Participants were asked two questions about how many days in the last month they: (1) had urges to use drugs or alcohol at (home, work, and community); and (2) felt that they were at risk for relapse when at (home, work, and community). The home risk variable was created by choosing the larger of the two numbers. (Community and work risk were also created this way.) The reason for not summing the two questions was that it was not clear if the urge or risk occurred on the same 40 day or different days. For this reason a conservative approach was taken by assuming the urge or risk occurred on the same day: thus, the largest number of days was used. For these risk variables, however, response distributions were highly skewed. Thus, home risk, community risk, and work risk were transformed to reduce skewness using an empirical and theoretical rationale. Three risk categories were produced: "0" was coded if no risk days occurred in the last month; "1" was coded if a week (one to seven days) or less of risk occurred: "2" was coded if more than one week (seven days) of risk occurred. MW. Participants were asked two questions about how many days in the last month they: (1) had been offered drugs or alcohol at (home, work, and community): and (2) had someone use drugs or alcohol in front of them at (home, work, and community). The home exposure variable was created by choosing the larger of the two numbers for the identical reasons as described for risk. (Community and work exposure were also created this way.) For these exposure variables, however, response distributions were also highly skewed. As with the risk variables, home exposure, community exposure, and work exposure were transformed to three categories: "0" was coded if no exposure days occurred in the last month: "1" was coded if a week (seven days) or less of exposure occurred: "2" was coded if more than one week (seven days) of exposure occurred. 41 figme, Wgzk, gng Community Beinfgzcing Activities. For the work setting, participants were asked to describe the activities they did at work in the last month. For the home and community setting, participants responded to a predetermined list of activities in the last month. For each home, work, and community activity they were involved in, participants then rated: (1) how much they enjoyed the activity on a five point Likert scale ranging from "not at all" to "extremely" and (2) how often they did that activity in the last month. Each activity was weighted by multiplying its value by the proportion of time they were involved in that activity in the last month. These weighted activity ratings were then averaged across all the activities in each setting. Thus, the home, work, and community involvement in reinforcing activities variables could range from one to five. §BEEQI¥_Bi§B_§nQ_EKEQ§EL_- Summary risk and summary exposure were created by summing the specific setting variables (home, work, and community). Since the specific variables were transformed to range from zero to two, summary risk and summary exposure ranged from zero to six. WW- Summary reinforcing activities was created by averaging the specific setting variables (home, work, community). Thus, this variable ranged from one to five. Reliability of the SRI measure was assessed several ways (see Table 4). While using test-retest methods with 42 brief assessment intervals to assess the reliability of the SRI would have been optimal: this was not done due to practical constraints. Table 4 shows two sets of reliability indicators for the SRI variables. For the specific setting risk and exposure variables, alpha was not computed since these variables only contained one item. Instead, correlations between the two specific setting risk or exposure items were computed. For the reinforcing activities domain, alphas were not computed because each activity was not expected to make up an internally consistent scale. Instead, reinforcement was supposed to vary by activity. Finally, reliability of the risk, exposure, and reinforcing activities variables was also assessed by correlating the pretest and posttest one data (as an approximation of test-retest methods). See results section for reliability and validity data. We; §3n§§§n9e_g§g. The timeline calendar protocol was used to examine daily alcohol and drug consumption over the follow-up period using monthly calendars (Sobell, Maisto, Sobell, 8 Cooper, 1979). Several studies have demonstrated the reliability and validity of this method of assessing drinking behavior using test-retest and convergent methodologies (Maisto, Sobell, Cooper, 8 Sobell, 1979: Sobell, et al., 1980: Sobell et al., 1979: Maisto, Sobell, 8 Sobell, 1982: Cooper, Sobell, Sobell, 8 Maisto, 1981; Sobell 8 Sobell, 1980). For this study, 72% were abstinent at 43 Table 4. Correlations: reliability of the SRI variables. Variable Internal Test- Consistency retest Summary Risk 0.66 0.40** Summary Exposure 0.49 0.56** Summary Reinforcing 0.24 ' 0.45** Activities Work Exposure 0.68** .35* Home Exposure 0.57** .56** Community Exposure 0.81** .44** Work Risk 0.38** .45** Home Risk 0.27* .33** Community Risk 0.27* .16 Work Reinforcing Activities N/A .27 Home Reinforcing Activities N/A .42** Community Reinforcing N/A .41** Activities * p<.os, Hr p<.01 Note that internal consistency for summary variables is alpha, for the specific setting variables are correlations. 44 posttest one and 28% relapsed: at posttest two 64% were abstinent and 36% relapsed. Thus, the variable ”number of days of substance use" was high skewed. For this reason a dichotomous abstinence/reuse variable was created and used for all analyses. In addition, the number of days of drugs or alcohol use in each setting was obtained from the multi-setting relapse risk indicator measure. More participants reported using substances at home (17% and 19% at posttests one and two) and in the community (15% and 16% respectively) then at work (5% and 2% respectively). Due to the small number of participants who relapsed in any setting, further analyses examining factors related to reuse in a specific setting (home, work, community) could not be conducted. W fiynggne§i§_ii The hypothesized convergent and discriminant validity of the setting relapse indicators measure was examined by generating a Pearson's r correlation matrix. A count was made of the number of times the matrix supported the hypothesized pattern of relationships. flynggn§§i§_zi Because of the dichotomous dependent variable reuse, discriminant analyses were used to determine the predictive and concurrent validity of the SRI measure summary domains of risk, exposure, and enjoyment. Discriminant analyses were also computed for the setting variables by each of the domains. Concurrent validity was tested at two time points: posttest 1 and posttest two. 45 Predictive validity was tested two ways: (1) pretest SRI was used to predict posttest 1 reuse: (2) posttest 1 SRI was used to predict posttest two reuse. gypgthesis 3. In order to determine whether the SRI was sensitive in change due to either receiving or not receiving the support groups, a repeated measures Multiple Analysis of Variance (MANOVA) was computed for the summary domains (risk, exposure, reinforcing activities) and for the specific setting variables over the three assessments (pretest, posttest one, posttest two). To assess the effectiveness of the support groups on preventing relapse, a Chi-square was computed for participation in the support or comparison group and abstinence or relapse at posttest one and posttest two. flynggng§i§_gi Due to the dichotomous nature of the dependent variable, discriminant analyses were conducted for overall construct validity of the model shown in Figure 1 with setting risk, setting exposure, setting reinforcing activities, coping, and self-efficacy serving as the independent variables and reuse serving as the dependent variable. This analysis was done concurrently at both posttests and prospectively for the two time intervals described in hypothesis two. W Power was calculated three ways using analysis of variance tables based on sample size and number of independent variables included in the analysis (Cohen, 46 1992). For most of the discriminant analyses, the sample size was 80 and three independent variables were included; assuming a moderate effect size of .25 (Cohen, 1992), and a=.05, power was 0.56. (Since a meta-analysis had not been published on the substance abuse treatment outcome literature, the effect size was estimated.) For analyses in which specific work setting variables were included, the sample size was about 60 since some participants were unemployed. For this sample size, three independent variables, a moderate effect size, and §=.05, power was 0.43. Finally, for the construct validity analyses testing of the conceptual model, the sample size was restricted to about 60 since many participants did not experience a risky situation in which they could describe their coping strategies. For these analyses, five independent variables were entered: assuming a moderate effect size and standard alpha level as above, power was 0.31 for these analyses. CHAPTER 3 Results W Overall reliability results were fair (Table 4). For summary risk and exposure computed alphas were respectable since these scales were based on the three items (specific home, work, and community risk or exposure). Similarly, the test-retest correlations were moderate and significant for summary risk and summary exposure. For summary reinforcing activities, the alpha was low even with the fact that only three items made up the scale. The test-retest correlations, however, were moderate and significant for summary reinforcing activities. Reliability analyses for the specific settings mirrored those of the summary domains. At the pretest, the two exposure items were highly and significantly correlated for the work, home, and community settings. The test-retest correlations between the pretest and posttest one were moderate and significant for the specific exposure settings. The pretest correlations between the two risk items were lower but significant for the home, work, and community settings. Test-retest correlations for risk were moderate 47 48 and significant for the work and home settings but low and non-significant for the community setting. Test-retest correlations for the specific setting reinforcing activities variables were moderate and significant for the home and community setting but were low and non-significant for the work setting. '5 ° s Convergent and discriminant validity of the setting relapse indicator variables was assessed by correlating the setting indicators with other relapse indicators. Table 5 shows correlations from data obtained at the pretest between the summary SRI domains (risk, exposure, and reinforcing activities) and other psychological and social indicators of relapse. Greater summary exposure to drugs or alcohol was significantly related to less perceived support for sobriety from network members and to a smaller proportion of non- users in the network. Greater summary risk was significantly related to: relapse, less support for sobriety from network members, and lower self-efficacy. A count was made of the number of times the correlations were in the expected direction. Overall, 5 of 27 (19%) were significant and in the expected direction. Regardless of the significance of the correlation, 17 of 27 (63%) of the correlations were in the expected direction. The correLations among the summary setting risk was significant for threee of nine (33%) and in the expected direction for 49 003.303 2.0.33.2. 20... fl. 3.. «~.- 9. 3. S. .2. 3.. Elite. ~.. - 2.3.. L»... ~.. - mo. 8. 3.. B. 2.... 0.5085 {Em 2.3.- 005.- o..- 2. n... - 8.- 9... no. 39. x0... flea... 3039.4 0030...; tag at... 8.68 0.0.3.... 0.032.. 2038.. .38” .3 03.30.30» 30.320. 503-5.. 3330.. «£38.95 .30.. u...uo....u>0a \3 .m... 030.. 3.3.00.2. 030.0,. 5056 .80 0030.52, >550 .cm c0030.. / 1.3.3.0530 . m 03. h 50 seven out of nine (78%). For the exposure domain, correlations with the other variables was significant for two out of nine (22%) and as expected for seven out of nine (78%). The summary reinforcing activities domain was not significantly correlated with any of the other variables and was correlated in the expected direction for only three out of nine (33%) of the correlations. In order to determine whether the specific settings (e.g., home) showed that different patterns of relationships with the other relapse indicators, exploratory correlations were generated between the SRI variable domains separated by settings (home, work, community) and the social and psychological variables (Table 6). Overall, the pattern of relationships showed if the summary setting domain was significantly correlated with a relapse indicator, the specific settings also tended to be significantly correlated. Some of the specific settings, however, were uniquely correlated with other relapse indicators. For example, while the summary reinforcing activities domain was not significantly correlated with any of the other variables, home reinforcing activity and work reinforcing activities were significantly correlated with some of the other relapse indicators. A larger home reinforcing activities rating was associated with more family problems and more psychiatric problems. Conversely, a larger work 51 Table 6. Correlations among SR1 settings by domains and other relapse indicators. Domains by Sobriety Mon-user Self- Megative Reuse Settings Smport Ratio Efficacy Coping Reinforcing Activities Maine .18 -.01 .09 -.12 -.02 Work -.02 -.13 .17 -.43*" .01 Conn .07 .01 .26' -.07 -.06 Risk Home -.14 -.05 -.38** .12 .36“ Work - .33* - . 13 - .49" .43" .41" Comm -.28** -.20 -.29** .03 .25“ Exposure Home -.36“ -.23* -.22* .00 .13 Bork -31* -.19 -.13 -.04 .19 Co. -.30“ -.32" -.11 -.17 .09 * p<.05, ** p<.01 table continues Table 6 (cont.). 52 Domains by Family Legal Psychiatric Medical Employment settings Problems Problems Problems Problems Problems Reinforcing Activities Home .26* -.10 .23* -.03 .05 Work -.26 .15 -.27* -.23 -.14 Comm .17 .08 .11 .12 .18 Risk Home -.01 -.04 .15 .19 -.11 Hork .14 -.14 .34* .38** -.02 Comm .08 .05 .17 .15 -.02 Exposure Home .10 -.00 -.03 .11 .07 Work -.07 -.10 -.12 -.13 .14 Comm .13 .05 -.03 .01 -.000 * p<.05, ** p<.01 53 reinforcing activities rating was significantly correlated with less use of negative coping strategies and fewer psychiatric problems. The reason for the differing direction of the relationship between the reinforcing activities variable and psychiatric problem variable according to home or work setting is not clear. A greater perception of risk for relapse at work was associated with more medical problems and psychiatric problems. Greater work risk was also associated with less use of negative coping strategies. A greater perception of work, home, and community risk for relapse was associated with lower self- efficacy and relapse. Only community and work risk were associated with less support for sobriety from network members. High exposure to substances at work, home, and in the community was related to less sobriety support among network members. A low proportion of non-users in the network was related to home and community exposure to substances. A high self-efficacy was related to low home exposure to substances but not to work or community exposure. Construct validity was also determined by intercorrelating the specific setting risk indicators. Table 7 shows the correlation matrix between the SRI settings and domains. The triangles show the correlations between the home, work, and community settings for each 54 Table 7. Correlations among home, work, and community settings by variable domains. geinfgrcing Risk x ur Agtivitieg ii if C ii if C H U Reinforcing Activities Home __ com 037** -007 Risk Home -000 -011 -001 _ Work -.17 -.11 -.28* .39** com -003 -004 -005 037** 040** '- Exposure Home -.08 -.01 .01 .26* .16 .16 work -007 023 -011 004 029* 005 024 com 021* 021 001 -007 015 040* 014 033* *IFJB,**E&£1 Mote: Menome, Work, CsCommnity 55 domain. The correlations among the settings for the reinforcing activities domain were low for work and home, and, work and community. Home reinforcement and community reinforcement were highly correlated. For the risk domain, all three settings (home, work, community) were highly intercorrelated. For the exposure domain, community and work exposure were significantly correlated; home and work exposure, and, home and community exposure were not significantly correlated. The diagonals in the table show the correlations between similar settings across variable domains (e.g., home risk with home exposure) which can be contrasted with the correlations for different settings on either side of the diagonals (e.g., home risk with community exposure). For the risk and exposure diagonal, each of the settings was significantly correlated with the identical setting (e.g., home risk with home exposure) as opposed to different settings (e.g., home risk with work exposure). For the diagonals correlating reinforcing activities with risk or exposure, the expected pattern of higher correlations among similar settings as opposed to different settings was not supported. For example, home reinforcing activities was more likely to be correlated with community exposure than home exposure. As a final test of construct validity, the SRI summary variable were intercorrelated. Risk and exposure were significantly correlated (£3.34, p<.01). Reinforcing activities was not significantly correlated with risk 56 (rs-.19, n.s.) although a trend was observed in the expected direction. Reinforcing activities was not significantly correlated with exposure (;=.07, n.s.). WW Because of the dichotomous dependent variable of reuse status (abstinent or relapsed), a series of discriminant analyses were used to test the concurrent and predictive validity of the SRI variables in determining reuse. In all analyses, the 2 statistic based on Hotellings t-test was used to test the significance of the independent variables in determining relapse status. In addition, standardized discriminant function coefficients and structure coefficients were produced to determine the relative influence of individual variables on relapse status. Because equality of variances is assumed in discriminant analysis, Box M’s test was calculated for each analyses to verify that this assumption was met. If this assumption was violated, follow-up analyses were conducted using a procedure that accommodates a dichotomous dependent variable but does not require equal variances: logistic regression. Concurrent validity was assessed for two time periods. For the first three months, the SRI variables at posttest one were used to distinguish reuse status (abstinent or relapsed) at posttest one. Concurrent validity was also assessed for the second three months as posttest two SRI variables were used to distinguish posttest two reuse status. Predictive validity was assessed by using the SRI 57 variables to differentiate reuse status prospectively: pretest SRI was used to classify posttest one reuse status (first three months); posttest one SRI was used to classify posttest two reuse status (second three months). Qgpgndgnt_yg;i§blg§. Reuse at posttest one was simply a dichotomous coding of whether the participant relapsed during the first three month follow-up. Reuse at posttest two was defined as those who used a substance during the second three month period and was independent of relapse status during the first three months. Two participants relapsed during the first three months but were abstinent during the second three months and were classified as abstainers for the posttest two reuse variable; all other abstainers were abstinent for the entire six months. Participants classified as relapsers at posttest two consisted of those who relapsed during the first and the second follow-up periods and those who only relapsed during the second follow-up period. Inggpgggggt_yari§blg_. SRI summary domains (risk, exposure, and reinforcing activities) were first used to determine reuse in the concurrent and predictive analyses. In order to determine the impact of the setting specific SRI variables, discriminant analyses were also performed for the specific setting variables. Due to power limitations, several sets of analyses were done to limit the number of independent variables included in the analyses. Three analyses were conducted to determine which setting was the 58 largest contributor of the domains. Independent variables were: (1) home risk, work risk, community risk: (2) home exposure, work exposure, community exposure: (3) home reinforcing activities, work reinforcing activities, and community reinforcing activities. Finally, three analyses were done to determine which domain was the largest contributor to the setting's impact. Independent variables were: (1) home risk, home expousre, home reinforcing activities: (2) work risk, work exposure, work reinforcing activities: (3) community risk, community exposure, community reinforcing activities. §g§12131g_. Table 8 shows means, standard deviations, and function and structure coefficients for the discriminant analyses testing the concurrent validity of the summary SRI domains (risk, exposure, reinforcing activities) at both posttests. At posttest two, the summary domains significantly distinguished abstainers from relapsers as indicated by a significant E test: at posttest one, the 2 test statistic approached significance. In both examinations of concurrent validity, the risk variable had the largest discrimination coefficients. Table 9 shows results from discriminant analyses testing predictive validity of the SRI summary domains over the first three month time interval. The pretest summary domains significantly explained reuse status at posttest one (first three months). The summary risk domain variable had the Table 8. Discriminant analyses: 59 concurrent validity of SR1 and reuse. Domain Status 5 Q Standard Coefficent Structure Coefficent Posttest 03' Summary Risk '90 '96 Abstinent 1.21 1.6!. Relapsed 2.1!. 1.1.9 Sumary Reinforcing -.27 -.39 Activities Abstinent 3.89 0.50 Relapsed 3.76 0.52 Sumry Exposure .07 .46 Abstth 1.30 1.1.3 Relapsed 1.73 1.26 Posttast THO“ Squary Risk .76 .93 Abstinent 0.85 1.18 Relapsed 1.73 1.1.1 Smary Reinforcing -.31 -.42 Activities Abstinent 3.83 0.50 Relapsed 3.65 0.6!. Sunnary Exposure .31 .52 Abstinent 1.06 1.12 Relapsed 1.53 1.61 * Note: ns78: 56 abstinent, 22 relapsed. llotellings F(7l.,3)=2.25, p<.10. "Note: MB: 53 abstinent, 30 relapsed. Hotel l ings man-3.1.9. mos. 60 Table 9. Discriminant analyses: predictive validity of pretest SR! and posttest one reuse. Domain Status Standard Structure 1 52 Coefficent Coefficent Summary Risk -.96 -.99 Abstinent 1.60 1.65 Relapsed 3.00 1.71 Summary Reinforcing -.03 .11 Activities Abstinent 3.89 0.50 Relapsed 3.86 0.61 Summary Exposure -.17 -.66 Abstinent 1.20 1.23 Relapsed 1.83 1.61 Note: n383: 60 abstinent, 23 relapsed. Hotellings F(79,3)-6.12, p<.01. 61 largest discrimination coefficient. E tests for posttest one summary domains predicting posttest two reuse (second three months) was not significant (£(73,3)=1.26: n.s.). om ° 8: w o i 's . Concurrent validity was assessed at posttest one and posttest two for each of the settings in the risk domains (home risk, community risk, and work risk) in classifying reuse status (Table 10). At posttest one, the E test was significant and at posttest two the E test approached significance- In both cases, W W For posttest two, the Box M’s test for homogeneity of the multivariate variance matrices was significant (Box H=25-29. E=3.93, p<.01) indicating inequality of the variance-covariance and pooled variance matrices. Thus, a follow-up analyses was conducted for the risk setting variables, logistic regression, that does not require equal variances. Results from the logistic regression analysis were identical to the discriminant analysis as the overall equation approached significance (Chi-square=6-27. p.<-10) and W; or 'a . Prospective analyses, showed that pretest home, work, and community risk settings significantly differentiated posttest one reuse status (Table 11). The work risk variable had the largest discrimination coefficient. Posttest one risk settings did not significantly explain posttest two reuse status (£(54,3)=1.54, n.s.). 62 table 10. Discriminant analyses: concurrent validity of risk settings and reuse. Domain Status X 52 Standard Coefficent Structure Coefficent Posttest One”r Bork Risk 0.56 0.08 Abstinent 0.56 0.77 Relapsed 0.67 0.76 Home Risk -0.99 -0.87 Abstinent 0.60 0.70 Relapsed 1.13 0.83 Community Risk -0.26 -0.39 Abstinent 0.60 0.58 Relapsed 0.67 0.62 Posttest Tud" work Risk 0.07 -0.36 Abstinent 0.31 0.63 Relapsed 0.67 0.70 Home Risk -0.93 -0.95 Abstinent 0.36 0.61 Relapsed 0.86 0.90 Community Risk -0.32 -0.62 Abstinent 0.26 0.53 Relapsed 0.62 0.69 * Rote: n858: 63 abstinent, 15 relapsed. Hotellings F(56,3)-6.86, p<.01. **Rote: n-66: 65 abstinent, 19 relapsed. Hotellings F(60,3)82.26, p<.10. table 11. Discriminant analyses: 63 pretest risk settings predicting posttest one reuse. Domain Status Standard Structure 5 so Coefficent Coefficent Hork Risk -0.76 -0.91 Abstinent 0.36 0.63 Relapsed 1.06 0.90 Home Risk -0.39 -0.66 Abstinent 0.72 0.79 Relapsed 1.29 0.85 Community Risk -0.16 -0.52 Abstinent 0.56 0.68 Relapsed 0.96 0.83 Note: n-56: 39 abstinent, 17 relapsed. uotellings F(52,3)=6.39, p<.01. 64 ' ' o 't osu e. Home, work, and community exposure did not significantly distinguish abstainers and relapsers in concurrent analyses (for posttest one E(54,3)=0.67, n.s.: for posttest two £(60,3)=1.17, n.s.). Non-significant results were also found for prospective discriminant analyses using pretest SRI to predict posttest one relapse (£(52,3)=1.45, n.s.) and using posttest one SRI to predict posttest two relapse (£(54,3)=0.47, n.s.). 9- ' ssh: , ° £098 .- . an! 0m".!. -' . . QQEiXiSigé- Home, work, and community reinforcing activities did not significantly distinguish reuse status in concurrent analyses (for posttest one £(S4,3)=0.41, n.s.: for posttest two £(60,3)=0.70, n.s.). Non-significant results were also found for prospective discriminant analyses using pretest SRI to predict posttest one relapse (E(51,3)=0.19, n.s.) and using posttest one SRI to predict posttest two relapse (2(5o,3)=1.os, n.s.). - e ' ' ' 8' ° e su {ginfiggging_g§tiyitig_. Home risk, exposure, and reinforcing activities significantly identified reuse status concurrently at both posttest one and posttest two (Table 12)- In both cases. h2me_risk_nad_the_large§f_di§2riminanf gggfifiigigngg. Prospectively, pretest home risk, exposure, and reinforcing activities significantly predicted posttest one reuse (Table 13). As in the concurrent assessments, home risk had the largest discrimination coefficients. 65 table 12. Discriminant analyses: concurrent validity of home setting and reuse. Domain Status 5 52 Standard Coefficent Structure Coefficent Posttest One. Home Risk 0.95 0.98 Abstinent 0.61 0.65 Relapsed 1.16 0.83 Home Reinforcing -0.16 -0.17 Activities Abstinent 3.99 0.53 Relapsed 3.89 0.55 Home Exposure 0.15 0.33 Abstinent 0.30 0.66 Relapsed 0.55 0.80 Posttest Tud" Home Risk 1.06 0.98 Abstinent 0.36 0.59 Relapsed 1.03 0.81 Home Reinforcing 0.00 -0.28 Activities Abstinent 3.90 0.61 Relapsed 3.71 0.67 Home Exposure -0.23 0.16 Abstinent 0.36 0.65 Relapsed 0.67 0.68 * Note: nr78: 56 abstinent, 22 relapsed. notellings F(76,3)-5.66, ps.01. **Note: n-83: 53 abstinent, 30 relapsed. Hotellings F(79,3)=6.90, p<.01. 66 1able 13. Discriminant analyses: pretest home setting predicting posttest one reuse. Domain Status Standard Structure 5 fig Coefficent Coefficent Home Risk 0.97 0.99 Abstinent 0.68 0.75 Relapsed 1.35 0.83 Home Reinforcing -0.05 -0.05 Activities Abstinent 3.96 0.61 Relapsed 3.93 0.56 Home Exposure 0.09 0.33 Abstinent 0.37 0.66 Relapsed 0.57 0.86 Note: na83: 60 abstinent, 23 relapsed. Hotellings F(79,3)86.06, p<.05. 67 Similar trends were observed in prospective analyses where posttest one home variables were used to identify posttest two reuse status: E tests approached significance (§(73,3=2.42, p<.10). e ' ' s ' S' ' c un' e osur cgnnunity reinforcing activities. Results from concurrent analyses of community variables (risk, exposure, and reinforcing activities) in classifying participants' relapse status were significant at posttest two (Table 14). The community reinforcing activities variable had the largest discrimination coefficients. At posttest one, the E test of community variables and relapse status approached significance (Table 14). Community reinforcing activities had the largest standardized function coefficient but did not have the largest structure coefficient: instead, all three community variables, risk, reinforcing activities, and exposure, were equally large. Non-significant results were also found for prospective discriminant analyses using pretest community variables to predict posttest one relapse (£(79,3)=2.15, n.s.) and using posttest one community variables to predict posttest two relapse (fi(72,3)=0.81, n.s.). c' ' set ' s: wo 's we e osu e wor reinforcing ncgiyitig . Work risk, work exposure, and work reinforcing activities did not significantly differentiate abstainers and relapsers concurrently at either posttest (£(53,3)=0.14, n.s.: E(60,3)=0.86, n.s.). Since Box M's 68 Table 16. Discriminant analyses: concurrent validity'of community setting and reuse. Domain Status 5 §Q Standard Coefficent Structure Coefficent Posttest Dne‘ Community Risk 0.58 0.66 Abstinent 0.62 0.57 Relapsed 0.68 0.65 Comm. Reinforcing -0.66 -0.55 Activities Abstinent 6.11 0.58 Relapsed 3.85 0.85 Community Exposure 0.62 0.63 Abstinent 0.66 0.70 Relapsed 0.96 0.79 Posttest 1uo** Community Risk 0.27 0.61 Abstinent 0.25 0.52 Relapsed 0.61 0.63 Comm. Reinforcing -0.78 -0.70 Activities Abstinent 6.08 0.66 Relapsed 3.72 0.88 Community Exposure 0.59 0.58 Abstinent 0.62 0.61 Relapsed 0.69 0.71 * Note: n877: 55 abstinent, 22 relapsed. Hotellings F(73,3)=2.66, p<.10 **Note: n381: 52 abstinent, 29 relapsed. Hotellings F(77,3)s3.11, p<.05. 69 test for the posttest two analysis was significant, violating the assumption of equal variances, a follow-up logistic regression was conducted. Results were identical (Chi-square=4.75, n.s.). Prospectively, pretest work variables significantly categorized participants’ reuse status at posttest one (Table 15). Work risk had the largest discriminant coefficients. The E test was not significant however for posttest one work variables determining posttest two reuse status (£(50,3)=1.78, n.s.). 0 5'5 ° ens' ' ' rv Before testing the effects of participation in the support groups on the SRI variables, analyses were conducted to determine the impact of the groups on relapse (e.g., to test the validity of the intervention). A Chi-square was computed to determine the effects of participation in the support groups on relapse. At posttest one, the support group tended to have fewer persons relapse than the comparison group (Chi-square=3.50, p<.10). At posttest two, the two groups were not significantly different in respect to the number of persons who relapsed (Chi-square=2.01, n.s.). Thus, the groups did not substantially influence relapse. Despite the apparent weakness of the support groups, analyses were still conducted determine if the groups effected the SRI variables. Repeated measures analysis of variance was used to determine group, time, and interaction effects of group assignment on the SRI variables over the three assessments. 70 1able 15. Discriminant analyses: work setting predicting posttest one reuse. Domain Status Standard Structure 5 §Q Coefficent Coefficent work Risk -1.03 -0.98 Abstinent 0.37 0.63 Relapsed 1.06 0.90 Bork Reinforcing -0.18 -0.03 Activities Abstinent 3.51 1.87 Relapsed 3.56 0.92 Uork Exposure 0.11 -0.19 Abstinent 0.60 0.68 Relapsed 0.53 0.80 Note: ns83: 38 abstinent, 17 relapsed. Hotellings F(51,3)-3.56, p<.05. 71 Table 16 shows the means, standard deviations, and E tests associated with the repeated measures analysis of variance for the SRI summary domains. For summary risk domain, both the support and comparison groups’ showed a significant decline in perceived risk over time as indicated by the significant E test for the time effect. For the summary reinforcing activity domain, both groups tended to decline over time but the E test associated with this effect was not significant. No other significant group, time, or interaction effects were observed for the three summary domains. Similarly, repeated measures analysis of variance analyses were conducted for the separate settings for the risk, exposure, and reinforcing activities domains. The significant time effect for the summary risk domain described above resulted from declines in risk in the home and work settings and not the community setting (Table 17). In addition, when separating out the different settings, the comparison group had significantly greater home risk than the support group. A similar trend was observed for community risk however this was not significant. Table 18 shows similar home, work, and community exposure setting data for the two groups. Although none of the summary exposure 3 tests were significant, a trend towards a time effect for the work setting was evident with both groups showing a slight decline over time. A trend towards an interaction effect was also observed for exposure at home as 5.6.... .36». 72 2.6 . hm... E.n on... 8.n .m.o oo.n 3 contoeou 33.3.3 3... aid 3.: ~m.o 8.n 36 3..” .36 5..” sn toga .50.. f!!!» 3.. ow... or... 3.. on. an. 3 coat-089 0.5085 3.~ 8.. o... o... -.. 3.. 3.. 8.. 8.. kn toga fol!» 3.. De... 3.. 3.. «3.. o~.~ 3 coat-9.8 x»; K; 2.5.x. mo.~ mm... c... 3.. «~.. hr... 3.. sn toga fag—5m you w I 1.. I t a x mm m 8 x as: 30.3 x 99.6 tutu 2: 99.6 9-3:... 58... o 9.33.8 58... n :39: 33. u. votes «55332 a 033...; £833 6.50338 sec... 39.8 2395 coat-93 pc- acoonas to. 2.8.8 :6 Elisa to. 983-38 2353» use aces: .0. 63a. 8.6.: .36.... 2.6 . 73 3.9 5.9 3.9 3.9 3.9 99.9 3 contain". as... 99.. 2.99.0 .36 3.9 2.19 99.9 3.9 3.9 3.9 km «.8999 {63 no.9 and 05.9 3.9 3.9 2.9 3 coat-389 so... 99.~ no.9 99.9 mud 3.9 3.9 3.9 95.9 3.9 km toga >239 05.9 3.9 3.9 3.9 3.9 99.. 3 coatoouou so... 3.. com... L96 Q39 .96 3.9 3.9 99.9 3.9 pm «conga use: .8... m a m m m m al.. .023 .00»; x 99.9 as: 99.9 9.336. 5.8! 0 9.3623 5:... n «630.8 33. u votes .5332! 9. Boat... guano 6.50303. 09.... 39.6. 099.9 coat-go pc- ucooga .6. «83.0. 6.3.. .3. 9.3.3.6.. Hence: pro aces: .s. 03-. .o.vmucc Kevan. aria . 74 .5.9 99.9 95.9 99.9 ~5.9 ~5.9 99 89.339 0.5896 99.9 ..9.~ 9~.~ and 99.9 95.9 99.9 3.9 5m.9 5m .3999 .20: 3.9 99.9 99.9 99.9 99.9 99.9 99 89.39.89 0,589.9 99.9 8.9 99.9 .9.9 99.9 .9.9 99.9 3.9 9~.9 5m .8959. 5.53.89 9m.9 3.9 95.9 99.9 95.9 99.9 99 coat-9.59 0.5896 .8... :6 t... n... 3... 8... 3.... on... a... R :89... 82. .8 I I .3 aw. x w x fl 9 as... .839 .829 x 96.9 on: 39.9 9.3020. 9.8-. 9 9.333. 5...! n .83.... 9...». u 90...: .5633: m 03...... 8.3.3 6.539093 8.2. 39.0- 9995 coat-.33 n..- .3999 .8. 995...... 0.5896 .8. 9.3.9.3.. 9.35.. 9.... 9.90.. .9. 03-. 75 the support groups’ exposure increased over time and the comparison groups’ exposure at home decreased over time. Finally, work reinforcing activities showed a trend toward a time effect with both groups scores declining over the follow-up period (Table 19). WW Because of the dichotomous dependent variable of reuse, a series of discriminant analyses were used to test the overall construct validity of the conceptual model presented in Figure 1. No attempts were made to test the indirect and direct effects hypothesized in the conceptual model since this was not the central focus of this study. Additionally, the small sample size and dichotomous dependent variable of relapse would have made this difficult. Variables that were included in the discriminant analyses were the: (l) summary SRI variable domains (risk, exposure, reinforcing activities), (2) self-efficacy, (3) negative coping. Social network variables were not included in these analyses since these network variables were not hypothesized to mediate the effects of SRI on reuse. The model was tested concurrently by using the SRI variables at posttest one to determine reuse status at posttest one and by using posttest two SRI variables to determine posttest two reuse status. The model was tested prospectively by using the: pretest SRI variable to predict posttest one reuse status (first three months); posttest one SRI to predict posttest two reuse status (second three months). 76 3%.: 549.2% . no > 8... a: R... 8..” on... 8d 3 5.28.3 2. :2 8.9.35... 3.0 .3.~ 3.: 3.0 86 and 36 on... 86 um £806 in: no u > no on... 3:... ~56 33 $6 54 3 coat-“loo _ _ _ < 059.352. 3... and 3.: on... SJ 3.: SJ and :3 kn ton—9m 525.38 .3 > u 3 ~59 3:... no... 3...... S... $.n 3 52898“. . _ . 8.9.852. no... 36 mo.~ 3.. 3.n no.9 2...n 3.9 swim um found «12. «co m I I z w m a x a x «I: booty 333 x 90.5 at: 995 9.3033 5.8.. 0 3.33.3 5:9- n 330.... 33.. m .822. 58.333 a 030...; «833 .3533:- 0055 :98- !303 53.398 u:- utonii .3. acct—«o. 3.233. 959.352 .3. acct-:60 utopia. u:- Ilos .o— 03-» 77 Tables 20 and 21 show the concurrent discriminant analyses for posttest one and posttest two model variables distinguishing relapse status. In both analyses, 3 tests were significant. For posttest one, negative coping had the largest discrimination coefficients followed by self- efficacy. However, Box M's test was significant for posttest one (Chi-square=23.32, p<.05). Thus, a logistic regression was performed indicating similar overall significant results (Chi-square=14.73, p<.05). Posttest one coping and self-efficacy were the two significant variables as indicated by the Wald statistic. For posttest two, self- efficacy had the largest discriminant coefficients followed by summary risk. Prospectively, pretest model variables significantly predicted posttest one relapse status; self- efficacy and summary risk had the largest discriminant coefficients respectively (Table 22). Posttest one model variables did not significantly predict posttest two relapse status (£(41,5)=0.89, n.s.). 78 Table 20. Discriminant analyses: concurrent validity of conceptual model and reuse at posttest one. Domain Status Standard Structure 5 & Coefficent Coefficent Self-efficacy -0.61 -0.63 Abstinent 3.03 0.79 Relamed 2.58 0.78 Negative Coping 0.86 0.72 Abstinent 0.56 0.66 Relapsed 1.28 0.92 Risk 0.19 0.69 Abstinent 1.55 1.56 Relapsed 2.53 1.36 Reinforcing Activities -0.23 -0.34 Abetimnt 3.94 0.56 Relapsed 3.70 0.52 Exposure ~0.20 0.16 Abstinent 1.36 1.45 Relapsed 1.67 1.29 Note: n48: 33 abstinent, 15 relapsed. Hotellinga F(62,5)-3.63, r.05 79 table 21. Discriminant analyses: concurrent validity of conceptual model and reuse at posttest two. Domain Status Standard Structure 5 fig Coefficent Coefficent Self-efficacy 0.74 0.85 Abstinent 3.30 0.76 Relapsed 2.62 0.91 Negative Coping 0.14 -0.17 Abstinent 0.75 0.84 Relapsed .0.89 0.85 Risk -0.46 -0.70 Abstinent 0.97 1.24 Relapsed 1.88 1.45 Reinforcing Activities 0.05 0.31 Abstinent 3.85 0.53 Relapsed 3.68 0.59 Exposure -0.17 -0.36 Abstinent 1.09 1.24 Relapsed 1.56 1.47 Note: n-59: 34 abstinent, 25 relapsed. flotellings F(53,5)-2.54, p<.05 80 Table 22. Discriminant analyses: pretest conceptual model predicting postest one reuse. Domain Status Standard Structure 1 §Q Coefficent Coefficent Self-efficacy 0.76 0.76 Abstinent 3.06 0.72 Relapsed 3.37 0.66 Negative Coping -0.10 -0.18 Abstinent 0.73 0.85 Relapsed 0.93 0.88 Risk -0.51 -0.67 Abstinent 1.93 1.44 Relapsed 3.17 1.43 Reinforcing Activities -0.25 0.21 Abstinent 3.92 0.47 Relapsed 3.80 0.35 Exposure -0.33 -0.37 Abstinent 1.34 1.26 Relapsed 2.00 1.57 Note: n-59: 41 abstinent, 18 relapsed. Notellings F(53,5)83.93, p<.01 CHAPTER 4 Discussion Evidence for the validity of using social setting influences to understand relapse was mixed. This may indicate that the measurement approach taken in this study requires refinement, that social setting indicators are not consistent determinants of relapse, or that design imperfections prevented consistent detection of setting impacts on relapse. A key finding, however, was that a recovering person’s perception of risk for relapse in a setting is an important determinant of abstinence or reuse of alcohol or drugs. Additionally, negative aspects of settings, such as perceived risk for relapse and exposure to substances, were more frequently related to relapse than positive setting attributes, such as providing access to reinforcing activities. Finally, home, work, and community settings were often differentially related to other relapse indicators. This finding is important in that it may suggest that settings have distinct characteristics that uniquely effect relapse. Examination of specific setting factors, therefore, could serve as a useful technique for targeting relapse prevention interventions. The following discussion of specific construct, concurrent, and predictive 81 82 validity tests provides justification for sanctioning further investigation into setting based determinants of relapse. E 1. 1.1.! Reliability of the SRI measure could not be precisely assessed in this study due to practical constraints and the types of questions asked. Some evidence exists, however, for the accuracy of the setting indicator variables. Internal consistency for setting risk and setting exposure across all settings was promising considering that these domains only contained three items and internal consistency analyses are somewhat dependent on the number of items in the scale. Results of internal consistency analyses for setting reinforcing activities was not favorable. Results from analyses for specific setting risk and exposure indicators, however, provided some evidence for their internal consistency. Applying test-retest procedures to assess the reliability of the setting indicators became convoluted when the setting indicators may change across time in relation to relapse. This problem could have been resolved by using brief measurement intervals. Since this was not possible, an attempt was made to approximate test-retest reliability by comparing the pretest setting indicators to the first posttest. Especially due to the length of the assessment interval, three months, test-retest comparisons provided some confidence in the dependability of the specific setting 83 indicators. Except for risk for relapse in the community setting and involvement in reinforcing activities at work, test-retest comparisons for the remaining specific risk, exposure, and reinforcing activities setting indicators were encouraging. Future studies should provide more conclusive evidence for the reliability of the setting relapse indicators. s Val Results of correlational analyses provided some evidence supporting the discriminant and convergent validity of SRI. When comparing the summary risk, exposure, and reinforcing activities, to other known psychological and social predictors, some evidence was found for the construct validity of setting risk and setting exposure. To illustrate, involvement in settings that were risky or that provided exposure to drugs and alcohol were related to relapse. These settings were also associated with having fewer proportions of people that did not use substances and less social support for sobriety. Similarly, involvement in risky settings or those in which substances were available was related to lower self-confidence in remaining sober. Not astonishingly, setting risk and exposure were also related. Involvement with reinforcing activities in a setting was not a robust predictor of relapse or most other indicators of relapse. Thus, reinforcing activities in settings did not tend to demonstrate construct validity. It is not surprising then that setting risk and setting 84 exposure to drugs or alcohol were not associated with involvement in reinforcing activities. While exploratory, findings for specific setting attributes and other psychological and social relapse indicators and intercorrelations among setting indicators have several interesting implications. First, some indication was detected for the stability of a recovering person’s lifestyle regardless of setting. For example, if a relationship was observed between setting risk or exposure across all settings and the relapse indicator (as described above), all three setting specific variables (home, work, community) and the relapse indicator tended also to be associated. Upon examination of the pattern of relationships across home, work, and community settings for risk for relapse and exposure to substances, a similar conclusion was reached as they were consistently related. Thus, the newly recovered person's settings may be risky or alcohol and drug laden as a result of their former "addictive" lifestyles (Gorski, 1986). This finding may imply that specific settings are not important to examine in lieu of more general lifestyle factors. This conclusion is contraindicated however based on the finding that some of the specific settings factors (i.e., involvement in reinforcing activities at work) were associated with relapse indicators (i.e., less use of negative coping strategies), while the same factor in a different setting was not substantial (e.g., involvement in 85 reinforcing activities in the community and the use of negative coping strategies). If only general lifestyle factors were examined, this information would have been diluted and perhaps lost when setting factors are averaged across settings. In a similar fashion, involvement with reinforcing activities in one setting was not necessarily related to involvement with reinforcing activities in another setting. For example, while involvement in reinforcing activities at home was related to involvement with reinforcing activities in the community, participation in reinforcing activities at work was not related to reinforcement in the home or community settings. Again, these data may offer support for the inclusion of setting specific data collection methods as a home lifestyle may be different than a work lifestyle. A final implication discerned from the interrelationships among settings is that individual settings have comparable attributes. This conclusion is based on the finding that relationships between constructs within setting (e.g., home risk with home exposure, work risk with work exposure, and community risk with community exposure), tended to be more consistent than those across settings (e.g., home risk with work exposure). This result was only true for risk and exposure in a setting: it was not seen for involvement with reinforcing activities in a setting. This finding may substantiate the validity of the 86 measurement approach as participants appeared to be able to distinguish settings. Wham Similar to results regarding the construct validity of the setting relapse indicators, evidence for concurrent validity was also mixed depending on the specific setting construct examined. Irrespective of the assessment (the three month or six month posttest), results for concurrent validity tests were nearly identical. For all concurrent analyses, hypotheses were supported in that the abstainers had lower risk and exposure scores and higher reinforcing activities scores. More specifically, setting risk for relapse, exposure to drugs or alcohol, and involvement in reinforcing activities across settings, were highly indicative of relapse for parallel time intervals. Setting risk was the most powerful determinant of abstinence or relapse. Upon inspection of follow-up analyses for specific setting risk, specific setting exposure, and specific setting involvement in reinforcing activities, it was apparent that only setting risk was a consistent relapse determinant. Home risk best distinguished abstainers and relapsers (as compared to work risk and community risk). An analogous theme emerged from follow-up analyses by setting. When examining setting indicators for each setting, home and community setting characteristics demonstrated concurrent validity by distinguishing relapse: 87 the work setting did not. In the home setting, risk was the most influential variable. In the community setting, risk, exposure, and reinforcing activities were comparably influential. In these analyses, the home and community settings emerged as the most important settings for determining relapse. It was not surprising that the home and community setting were most predictive of relapse since more people reported using substances in the home or community settings than at work. Combined, these data suggest that the perception of risk for relapse in the home setting is the most valid predictor of abstinence or relapse during similar time periods. The only concurrent evidence for considering exposure and reinforcing activities as important for understanding relapse is suggested by the parallel impact of these constructs and risk in the community setting. Ezegigtive Validity Predictive validity was differentially supported by the various setting influences depending on the variable considered, setting it occurred in, and time interval over which relapse was predicted. Overall, pretest setting indicators typically determined reuse prospectively for the first three months of the study. The setting indicators, however, did not determine relapse status for the second three months of the study. It could be that setting attributes are more important determinants of relapse early in recovery as the recovering person adjusts to old 88 environments with a new sober persona. Later in recovery setting attributes may not influence relapse. As the person accommodates to the surroundings, setting variables such as risk and exposure may no longer be influential: instead, other factors may become important in determining relapse. Since the sobriety time of the participants varied, this post-hoc explanation requires replication. For the initial three month interval, the question remains as to which settings and indicators were effective in predicting relapse. When considering setting risk, exposure, and involvement in reinforcing activities simultaneously across settings, evidence for their predictive validity was observed as they dependably anticipated relapse. Once again, setting risk was the most potent determinant of relapse. Results for the analysis including pretest home risk, work risk, and community risk were also useful in classifying abstainers and relapsers. The relative importance of specific indicators was unexpected in that work risk was the most powerful discriminating variable, followed by home risk. This finding is in contrast to the concurrent analyses in which home risk was the most powerful variable. In follow-up analyses by setting, the social indicators measured in the work setting appropriately determined reuse at three months: work risk was clearly the most important variable. Identical results were found for setting 89 indicators in the home setting as they reliably determined reuse with home risk being the largest contributor. The distinction of whether home or work risk influenced relapse status may be a result of other factors not examined in these analyses. For example, research has found that an alcoholic’s perception of the work environment only influenced functioning for those who were not married (Moos & Ingel, 1974: Bromet & Moos, 1977). Thus, whether a recovering person's risk perception is greater at home or at work, and whether this determines relapse, may be buffered by other influences. The community setting attributes did not appear to prospectively determine relapse. A final implication is apparent from many of the results of the previous construct, concurrent, and predictive validity tests showing the differential importance of home, work, and community settings. Some settings may be riskier than others in precipitating relapse. This finding is consistent with Marlatt and Gordon's belief in high risk situations (Marlatt & Gordon, 1980). While situational data show that most relapses occur under negative emotional states or during social pressure (Marlatt & Gordon, 1985), in this study setting data established that overall relapse is best explained by examining qualities of the home setting. Before examining whether participation in support groups impacted social settings, the validity of the 90 intervention in effecting relapse was assessed. It is not clear if the support groups influenced relapse since the number of people who relapsed was small, random assignment was not used, and attendance to the support group was varied. Despite these limitations, those who attended the support groups tended to relapse less than the comparison groups. However, this may have been due to individual differences (such as sobriety time) rather than impacts of the support group itself. For the purposes of this study, the answer to the question of whether those who participated in the support group experienced changes in their social environments is "no". Some initial differences between the support and comparison groups’ settings were found. The comparison group experienced greater risk for relapse at home than the support group. Both groups’ risk perceptions tended to decline over time. This may suggest that risk perception declines as sobriety time increases. This could result either from increased confidence resulting in lower perceptions of risk or of fewer social situations occurring that were risky. It is possible that more robust relapse prevention interventions could impact social settings. When assessing the sensitivity to interventions, the setting relapse indicator measure might be more effectively examined with interventions that target changing the substance users’ social environments. 91 Weds]. Results testing the conceptual model provided some final evidence for the construct validity of perceived setting risk in understanding relapse. Combined, setting risk, setting exposure, setting involvement in reinforcing activities, coping, and self-efficacy significantly explained relapse concurrently at both posttests. As hypothesized, relapsers had lower self-efficacy, lower reinforcing activities, higher use of negative coping, and greater risk and exposure ratings. Interestingly though, different constructs emerged as the largest contributors to relapse at the two posttests. At the first posttest, negative coping (e.g., isolation) was the strongest variable. This was followed by summary risk and self- efficacy as the next largest contributors. At the second posttest, self-efficacy was the largest contributor. This was followed by summary risk, and then exposure and reinforcing activities. Negative coping had very low discrimination coefficients. It may be that relapse determinants changes as sobriety time increases. The model was also validated prospectively, as pretest model variables significantly predicted posttest one relapse status. Self-efficacy was the greatest contributor to distinguishing relapse followed by setting risk, setting exposure, setting reinforcing activities, and coping respectively. As before, other prospective analyses for the 92 entire six month interval or, for the second three month interval, were not consequential. These findings partially support the cognitive behavioral model of relapse and the empirical literature showing self-efficacy was an important determinant of relapse (Marlatt & Gordon, 1985: Condiotte 8 Lichtenstein, 1981: Yates & Thain, 1985). However, coping’s effect on relapse is not clear from this data. Previous studies have not examined the effects of negative coping strategies but assume all coping to have some benefit and that no response is superior to another (Shiffman, 1982: Curry & Marlatt, 1985: Billings & Moos, 1983). The finding that setting risk was a powerful determinant of relapse may justify measuring setting variables in addition to psychological variables to better understand relapse. WW Results of this study are limited by two types of methodological concerns: design problems and measurement problems. Diversity in the sample on sobriety time, addiction treatment history, drug of choice, etc., made interpretation of the these results more difficult. Future studies are warranted that examine the psychometric properties of this measure while better controlling these within group variances. Power was low in this study possibly resulting in Type II error (missing an effect). The fact that some of the results were significant, despite 93 this, may indicate that the true impact of setting indicators was under-estimated. Generalization of these findings to the population of addicted persons in recovery is compromised by the fact that the majority of persons in this sample remained abstinent. This is in direct contrast to the relapse literature which reports relapse rates of seventy-five percent (Brownell, Marlatt, Lichenstein, 8 Wilson, 1986; Hunt, Barnett, 8 Branch, 1971). The sample in this study appeared to be made up of persons having extensive psychological or environmental resources. This is further evidenced by the distribution of the setting relapse indicators as many participants had no risk or exposure situations in their environment: participants also had appeared to have access to reinforcing activities in their settings as on average they "very much enjoyed" these activities. Again, the significant results found even with the restriction in variance, increases the confidence in the these results. 5 o s While these results are considered a preliminary attempt at examining the influence of addiction specific setting factors on relapse, results clearly substantiate further research into setting based determinants of relapse. Unfortunately, for this study reuse in specific settings could not be examined in relation to relapse indicators. Future studies with larger samples of persons who relapse 94 could examine the link between a setting indicator and relapse in the setting. While some evidence for the validity of the measurement approach used and for the validity of the risk and exposure constructs, several revisions are suggested. For the risk construct, it might be useful to ask an open-ended question about on what participants are basing their risk perceptions. For the exposure construct, questions regarding whether the persons using in front of the participant was using the participant’s drug of choice are recommended. Also, whether other persons present in the setting were supportive of substance use or of abstinence should also be included. While involvement in reinforcing activities should conceptually buffer the effects of other variables on relapse (as proposed in the lifestyle and behavioral choice theories), this construct lacked validity experimentally. This may be due to the way this construct was operationalized or it may be that positive effects do not influence relapse as strongly as negative effects such as perceived risk or exposure. These two explanations should be investigated by refining the definition of involvement in reinforcing activities in other ways to better capture how this variable might influence relapse. Revised questions might focus on changes in reinforcing activities to those that are substance free. These questions might only be asked for activities that are really important, reinforcing, 95 and those in which the person is invested as opposed to all the activities that take place in a setting. Perceived risk for relapse was consistently a valid predictor of relapse both prospectively and concurrently. What is not clear is if peoples’ perceptions of setting risk reflects actual setting attributes or if perception of risk becomes a self—fulfilling prophecy indicating where or why a person might relapse. Finally, specific settings were often differentially related to relapse with the home setting showing the most consistent impact. These findings for the influence of social settings on relapse are promising and warrant continued research in order to better understand and eventually prevent relapse. APPENDICES 96 Consent to Contact Form an Opportunity To Participate in a Research Study on Folloonp Treatment With the support of this treatment agency and Michigan Department of Public Health. Office of Substance Abuse Services. a Michigan State University program and research team is offering you an opportunity to participate in a research project that is designed to increase the knowledge about preventing addiction relapse. Persons who volunteer to participate in this project will attend weekly Addiction Recovery Groups. a support group program. for ten weelc. In addition. all participants will be interviewed three times. The participants wil receive at least $10 for each 2-hour interview and everyone will receive ash bonuses for consecutive interviews. If you corrplete all three interviews. you will receive $45. Providing Permission to Contact You By providing my name. address. and phone nurrber and by signing my name. I understand that I am p'oviding permission for a Michigan State University project staff member to contact me to arrange an initial appointment. l understand that at this initial appointment. the project staff member will first inform me of (a) all procedures that involve collecting information about me. (b) my rights to refuse participation or to withdraw from the project without penalty, and (c) all procedures to protect my identity and to keep information about myself confidential. I will then indium my voluntary decision to participate further in the project. Please provide your social security number below. so we can give you a $10 check at the time of the first interview. By signing below. i indicate only my understanding that (a) project staff at Michigan State University will identify themselves as representatives of the MSU Health Stud_y, (b) project staff will not reveal my identity to anyone outside the project staff and (c) this form (with my name. address and phone number) will be destroyed after the initial appointment. Print Your Name: Your Signature: Print Your Address: Today's Date: Your Social Security Number. Your Phone Number: Best times to reach you at this phone number: When are good times for you to attend a two-hour Addiction Recovery Group? Please write all possible two-hour times you could meet. For example. you should write under Mondays: 9-11 am. 3-5 pm. 7-9 pm (if you are available at those n'mes). Mondays Tuesdays Wednesdays Thursdan Fridays Saturdays List all possible times here: if you have questions about this project. please call the Addiction Relapse Prevention Project office at Michigan State University at (517)453-9936. You may reach the answering machine. but please leave your name and phone nurrber. The project directors are Prof. Flam Ramanathan (Social Work) and Prof. Tom Fleischl (Psychology). The project research coordinator is Maureen Walton. 9'7 SETTING RISK INDICATOR MEASURE Sarmrdfiemaskywabatyurdaflywehmmwsdmgsaphcesm Mascfioolmornaandccnmfiry. MqueadomrdermdiepestMOND-l. WorkiSchoolSettlnge ’ 4 ' ‘ 1 Do you consider yourself rriosdy as a: 1- currently working worker (continue below) 2-temporarily non-working worker (go to pg. 3) a- homemaker (go to pg. 3) «student (continue below) s-unemployed (go to pg. 3) In the last MONTH: - WORKER/STUDENT 2 you said you wenttcworkorschool days(iromEconomic Opportunities). ' (days) 3 Onthetypicalday.howmanyhoursperDAYdldyouspendat work /school? (hours) WORKER/STUDENT lnthelastMONTl-lyouhavebeenatworkorschool‘ 'days. 4 On how many DAYS did you use drugs or alcohol whie at work/school? (include lunch tlme) (days) 5 OnhwmanyDAYSwereyoudlrectlyoffereddmgsoralcoholat work/school? (days) 6 On how many DAYS did someone at work/school use drugs or alcohol in front of you? (days) 7 Oh how many DAYS did your work /school present a situation that put you at risk for using drugs or alcohol? (days) 8 OnhowmanyDAYSdidyouhaveanurgetousedrugsordrink alcohol at work/school? (days) 0 Copyright We lion 1 98 WORK/SCHOOL SETTING In the last MONTH: ‘9. Please tell me all the major 10. For each of these 11. On the AVERAGE. how much actMtles that take place as part activities. l'm going to ask time in hours is spent doing each of your wont/school setting ‘ . you how much you enjoy activity (Lfstathlesbydayandby dolngthatactMty: waltdonothdudedrug'orfi 0- rnotatall' H H H alcohol use): ' " t - a little W" 9" t W" 9" m 9" . 2- somewhat DAY was: MONTH 3- very much 4- extremely A1 I El IL :2 12 A3 1'3 4; a 14 _A5 ES IS A! as m a n 48 EB re 49 ‘ as re A10 E10 T10 Note: Provide other examples such as lunch. smoozing with co—workers. breaks. weekly staff meetings. etc. 9 Copyright Walton 2 959 lD-___. HOME SETTING ln the last MONTH: HOME 12 On how many DAYS did were you at home? (not on vacation or out of town) (days) 13 If a Worker/student: A On the typical day you work or go to school. how many hours per DAY did you spend at home? (do not include sleeping time) (hours) B. On the typical day-off. how many hours per DAY did you spend at home? (do nor include sleeping time) (hours) 14 If a Homemaker or not currently working or unemployed: A On the typical day. how many hours per DAY did you spend at (hours) home? (do not include sleeping time) HOME In the last MONTH you have been at HOME days. 15 On how many DAYS did you use drugs or alcohol while at HOME (days; 16 Oh how many DAYS were you directly offered drugs or alcohol at HOME? (days) 17 Oh how many DAYS did someone at home use drugs or alcohol in nontolyou? (days) 18 On how many DAYS did your home present a situation that put you at risk for using drugs or alcohol? (days) 19 On how many DAYS did you have an urge to use drugs or drink alcohol at home? (days) 0 Copyright Walton 3 100 10: HOME SETTING In the last MONTH: 20. Please tell me 20A. 21. For each of these 22. On the AVERAGE. how whether you do Circle activities. I'm going to ask much time In hours is spent these activities yes or you how much you enjoy doing each activity. when you are at no doing that activity: home. please 0- not at all H H H answer yes or no- 1. a "we 00': 9" 0U“ 9" DU" 9" 2_ somewhat DAY WEEK MONTH 3- very much 4- extremely A1 WATCH TV 1-YES o . NO E1 71 A2 PRAY/ 1-‘YES MEDITATE/ o - NO RELAX :2 T; A3 HOBBIES 1=YES (PAINTING. o = No GARDENING) ea ,3 A4 RENT MOVIES 1-YES 0- NO E4 n A5 VISIT WITH 1-YES FAMILY OR 0 - NO FRIENDS ES 1'5 A6 READ 1-YES 0: NO 56 re A7 TALK ON THE 1 sYES PHONE o - NO E7 17 A8 PLAY WITH 1-YES CHILDREN OR 0: NO PETS E8 T8 A9 TAKE 1-YES ' WALKS /WALK 0- NO DOG E9 79 A10 1-YES COOK/EATING o-NO :10 T10 A11 CHORES/ 1-YES CLEANING/ O-NO RUNNING ERRANDS Ell T11 A12 OTHER: 1-YES E12 712 0- NO 0 Copyright Walton 101 10. Community Setting In the last MONTH: COMMUNITY 23 On how many DAYS did you go to a community or social event? (daysi 24 On the typical DAY. how many hours did you spend in the community (at a some kind of social event)? (hoursl COMMUNITY In the last MONTH. you went to a community or social event days. 25 On how many DAYS did you use drugs or alcohol while at a community or social event? (days; 26 On how many DAYS were you directly oIfered drugs or alcohol at a community or social event? (days; 27 On how many DAYS did someone at a community or social event use drugs or alcohol in front of you? (daySI 28 On how many DAYS did your community or social events present a situation that put you at risk for using drugs or alcohol? (daysi 29 On how many DAYS did you have an urge to use drugs or drink alcohol at a community or social event? (days) °CopvrtghtWalton COMMUNITY/LEISURE SETTING In the last MONTH: 102 0 Copyright WaltOn 30. Please tell me whether 30A 31. For each of these 32. Of those activities you do these activities when Circle yes activities. how much you participated In. you are in the community. or no did you enjoy doing how many hours did please answer yes or no. each activity? you spend doing each activity in the last 0 a: Not at all month? 1 - A little 2 1: Moderately 3 2 Very much 4 - Extremely A1 SELF-HELP MEETINGS I-YES (M) 0:: NO El 11 A2 COUSELING/ 1 -YES AFTERCARE 0- NO 52 T? A3 VOLUNTEER WORK 1=YES 0 - NO 53 13 AA SHOPPING (MALL OR 1-YES SWAPMEET) o . NO E4 1’4 A5 PLAYING SPORTS. GYM 1-YES O: NO ES 15 A6 OUTDOOR RECREATION 1=YES (BBO. HUNTING) o-NO E6 T6 A7 VISITING FAMILY OR 1aYES FRIENDS 0: NO ET T7 A8 EATING OUT 1-YES a. NO as T8 A9 MOVIES 1 sYES 0.140 59 19 A10 CONCERTS/ SPORTS 1 sYES EVENTS 0.. NO :10 110 A11 DANCES/ PARTIES/ I-YES BINGO 0.140 £11 111 A12 MEDICAL / LEGAL CARE I-YES 0 - NO 612 112 A13 RELIGION 1 sYES or. no E13 113 A14 OTHER: 18YES E14 114 0- NO 6 103 Participant Consent Form 1 Instructions to researcher: Read each section of the consent form aioud to the Indwtdual and ask if the individuai understood the section before reading the next section Do not read the next section untII the V indivrduaI Indicates a clear understanding of the section. g1}, _,3 .. 3 Procedures and Purposes of the Study You are invited to participate in a study that could improve addiction counselors’ knowledge and skills for helping addicted persons stop using addictive substances. About half of the persons who voluntarily agree to participate in this project will be randomly chosen to participate in a 10-week support group program. The support groups will meet once a week and the emphasis will be placed on helping group members cope with situations where the urge to use addictive substances is strong. A requirement of those in the support group is that they participate in the evaluation research for this program including three 2-hour interviews so that the researchers can assess the effectiveness of the support group program. The first interview will occur before the participant begins their involvement in the support group program. The second interview will occur alter the 1o-week program is over. And the third interview will occur about 3 months after the second intervrew. The other half of the participants will be invited to participate in the three interviews. but not in the support group program. In addition to the interviews. the participants will be asked to permit the research staff to interview their primary treatment counselors (or caseworkers) about the nature and success of the participants' most recent treatment program. The chief purposes of this study are to evaluate the effectiveness of a the support group program and to learn more about the circumstances in which some recovering addicts reuse addictive substances after completing a treatment program. This information could be used to improve the current approaches for helping persons recovery from their addictions. If you have any questions or concerns about your participation in this study. please contact either Professor Chathapuram S. Ramanathan (517-853-8616) or Professor Thomas M. Reischl (517-353-5015). Participant Interview Procedures The three interviews will each take about two hours. During the first interview. the interviewer will ask you about some background information such as your age, religion. marital status. and your occupation. After the background information is discussed. the interviewer will ask a series of questions about your health status. employment status. recent substance use, legal status. family’s problems. and your relationships with your family and friends. use of social services, stressful life events. your expectations of using substances. your confidence in your ability to control your use of substances. your recent coping strategies. and you recent sources of psychological distress. During the second and third interviews. most of these questions will be asked again to learn how much your life has changed. These interviews will occur at a place that is convenient. private. and safe. You will be paid for your participation in these three interviews. Since the interviews will take about 2 hours. you will receive $10 after the first interview. You will receive $10 plus a $5 bonus (515 total) for completing the second interview. If you complete the first two interviews, you will receive $10 plus a $10 bonus ($20) for completing the third interview. As you can see. you will receive a total of $45 if you complete all three interviews. 104 Counselor Interview Procedures In addition to interviewing you. the researchers are asking your permission to interview your primary counselor or caseworker from your last addiction treatment program to learn more about the types of services you received and the counselors perception of how well those services worked for you. This interview will last about 20 minutes. The researcher will only interview the agency counselor. The agency counselor will probably refer to your treatment file. but the researcher will not examine your private file. Procedures for Contacting You For Future Interviews Because we will want to interview two more times in the next 6 months. we are asking your to provide your address and a convenient telephone number. We are also asking for your permission to contact up to three persons who are most likely to know the best way to contact you during the next six months. The researchers will identify themselves as researchers from the MSU Health Study in order to prevent anyone besides yourself knowing that you are involved in a study about addiction recovery. 01 course. you may tell other people that you are involved in an addiction recovery study. but our research staff will keep that information confidential. ‘ Your Participation is Voluntary Your participation in this study is voluntary. You may choose to not participate in any part of this study. For instance. you may choose to not answer an interview question if you feel uncomfortable. You may also choose to not participate in this study at all. There are no penalties for choosing not to participate in the study or for withdrawing your consent to participate in the study. Your Participation and information About You Will Be Confidential The researchers will adopt procedures that will best ensure that any information about you collected in this study will never be identified with you or be used to hurt you. Here’s what the researchers will do to protect your identity and the confidentiality of the information about you: 1. All information about you will be linked only with a 5-digit secret code number. Your name will never appear on your interview forms. In fact. the only form that will have your name will be this consent form and a list that has both the secret code numbers and the names of the participants. And this consent form and the code number list will be kept in a locked file box in a secure place at Michigan State University. 2. All research staff and project staff will be required to sign a confidentiality pledge which requires them to never report any information about any client or participant involved in this study. 3. The researchers will have secured a Confidentiality Certificate from the US. Department of Health and Human Services for the period of your involvement in this study which authorizes the researchers to protect your identity from all persons outside of the research project. This certificate authorizes the researchers to protect you identity and information gathered in this study from any Federal. State. or local civil. criminal. administrative. legislative. or other proceedings. 4. After the third interview. all information that could be used to identify you or link you with any other information about you will be destroyed. 105 Potential Risks and Benefits There are very few risks to yourself for participating in this study. The interviews are two hours long, but you will be compensated for your time and if you do get tired. you can take a break. You might think that some of the questions in the inten/iew are too personal or upsetting to talk about. If this happens. please tell the interviewer and you can either stop the interview or go on to the next part of the interview. Despite the safeguards for confidentiality. there is a minimal risk that another person outside of the research staff could learn about your participation in this study. This information could cause you social embarrassment or could be used in legal proceedings. This risk. however. is minimized by the procedures to protect the confidentiality of information about you. The are several benefits to participating in this project. If you are randomly selected to receive one of the 10-week programs. you will have the opportunity to receive 10 weeks of additional follow-up programming at no cost to yourself. Participation in one of the programs may help the participants in their recovery from alcohol or drug addiction. There is no guarantee. however, that the programs will help in the recovery process. If you complete the research interviews. you will be paid for your time. You may also benefit from participating in the interviews because the interviews will give an opportunity to review the circumstances in your life that could help you in your recovery process. Consent to Participate in the Interviews By providing my name. address. telephone number. and signature below. I indicate (a) my complete understanding of the information in this consent form and (b) I am voluntarily choosing to participate in this research study. Print Your Name: Print Your Address: Best Times to Your Telephone Number. Contact You: Your Signature: Date: Consent to Interview Treatment Agency Counselor By providing my signature below, I indicate (a) my complete understanding of the information in this consent Iorrn and (b) I am voluntarily choosing to allow a member of the MSU research staff interview my primary treatment agency counselor from my last treatment program. Counselors Name: Agency: Your Signature: Date: 106 Consent to Contact Other Persons to Help Locate You By providing the names. addresses. and telephone numbers of three other persons. and my signature below. I indicate (a) my complete understanding of the information in this consent form. (b) I am voluntarily choosing to allow a member of the MSU research staff contact these persons in order to locate and contact me for future interviews. and (c) I am giving permission to these contact persons to tell the MSU researcher the best way to contact me. Contact Person A: I give my permission to the MSU research staff member to contact the following person and ask the best way to contact me for a research interview. I also give my permission to the following person to tell the MSU researcher the best way to contact me. Print Person’s Name: Person's Telephone Print Person's Address: Number. Person’s Relation to You: Contact Person B: I give my permission to the MSU research staff member to contact the following person and ask the best way to contact me for a research interview. I also give my permission to the following person to tell the MSU researcher the best way to contact me. Print Person's Name: Person’s Telephone Print Person's Address: Number: Person's Relation to You: Contact Person C: I give my permission to the MSU research staff member to contact the following person and ask the best way to contact me for a research interview. I also give my permission to the following person to tell the MSU researcher the best way to contact me. Print Person's Name: Person’s Telephone Print Person's Address: Number. Person's Relation to You: Print Your Name: Your Signature: Date: 107 “lime-line Calendar Protocol ID: mumwdmammmumwmm)? Yes No If yes. continue below. If no. skip to next Instrument. Wearelriarestedhmdmwngyoupamddmgwaicohdwa. Ushgthecalendvforthe mammrdllketohdpyoummcaflmdaflydmfldngmddmgusa. Wehwetound Wafsnaadflficukmsk.mciaflyumenywusemwauarfardaence. Ihavewrta‘anontha calerrtarmadateyoucaruetedthefirstlntarviewardtodarsdate. Also,standardholldaysare niariradonlhecafendarmhelpywrrecall. Bdorewebegfnlisdngyotralcoholmdmguan Mikemwwmdmmmmhmwmmwaum Multan-doom. Recordspecialdays. MandoyouMmebookuwumwrymywmmhmpu remamberyoudn‘nkingordrugusa? Ifyas sayflaasenkeloa Ywmmdflsbaflwudufinfifispafidfinm lino. continuebelow. mmwomummwwmmmmumaemwmm mWMw'thaMWadeMdfinfing/dmym. timers notwelltwasonthaflrorMJhatisokay. Tellmeyowbestquess Obviouslyathorsdils wedmmmryousaidMGerfinks. Againbebeasaccurateaspaasiblehowever tywm'fmflWmedM/dmmWaTdekwbast shot. Wedonotnaadtokrwthemtyouuaed. MIMweamrIadybbagln. Sanafimspeodehawcaflinmdambdwirdfinfingadmywauflflscanhdpdnmh fillingouthacalandar. Foretranrplellywusuallygoatwmmtdsonfiidayamsmway mmmimmcallfintyoumuHMVeIndacatainmnbadansmvnsemmgs.or ywnathamkad/hmkdaydwhyoudnhfimuwugwepufupsbasadmwk sdnddnumwmmripsdc. Canyouthlnkclawpattembyptrddnfdngor druguae? lfyes.firsttalk breiIlyofthe general pattern. Then. begin recordingthepattemofdaysthatiseasiest to resell. (Not easily remembered days can be completed later although It is best to complete a week.andthenmonthatatimebeforecontinulngontothenutone.) Ifno.Mnbeglnwlththodayofthepo&testInterview. Generalquestionstoaskclienhwywdrfnlrduaadrugson ’ auditing dramatic! ,ordutnglhemonhd 7 lfyes. ask: Mattypedafcomf/drugsddywcorm? (record responseoncdendar). If no. continue on to next day/week/month. When every day of the calendar Is completed. continue on to next section. 108 Timeline Code Sheet Date Alcohol Cocaine Amphat amine Heroin Mari- iuana Opiate Hallu- cinogon Hyp- notic Barbit- urata Other 0.; uni-J N... “a g.‘ (II—a a—O fi-a “II. .aN ON O-a MN TOTAL If DAYS: RELAPSE LEVEL: 109 Timeline Code Sheet ID: Date Alcohol Cocaine Amphet amine Heroin Mari- I'uana Opiate Hallu- cinogen Hyp- notic Berblt- mate Other flu OU ON “N N” COM 0!” AN UN db (3‘ CU CU NU mu 0‘“ DO U“ ”U ‘5 “5 Nb 110 Timeline Code Sheet ID: Date Alcohol Cocaine Amphet Heroin Meri- Opiate Hellu- Hyp- Beroit— Other amine I'uene cinogen notic urete a S a 6 e 7 4 8 4 9 5 0 5 ‘I 5 2 5 3 5 4 5 5 5 B 5 7 5 8 5 9 0 O 6 9 1 000 0'0 .0 0“ MG 111 Timeline Code Sheet ID:— Date Alcohol Cocaine Amphet Heroin Marl- Opiate Hallu- Hyp- Ier'bit- Other amine juana cinogen notic urate B 7 6 8 6 9 7 0 7 1 7 2 7 3 7 a 7 5 7 6 7 7 7 8 7 9 8 0 8 1 no N. O“ 0|. .1. NC NC 112 Coping with Urges l. Generate List of Recent Urge Situations Todayis: Somamorrttugowas: Ask:Canyoumflmabafldwdnnsdwhgmpeymmmewgemusadmgsaabdeasm scorned? Thesecwldbesihrab‘msadmyouacflraflydidusedmgsoralcohd. Quickly generate a list of urge situations Be sure to have the participant tell you when the event occurred-do not record the event if they cannot specify a time. Then ask the participant to rate each situation with the W Urge Rating gag. Risky Date Strength Situ. MM /DD/YY Brief Urge Situation Description of Urge M N1 2 _ _ / _ _ / - _ .32 uz ~1ch 8 _ _ / _ _ / __ __ BS U3 MON3 4 / _ _ / _ _ 34 U4 " ' MO 5 _ _ / _ _ / _ _ as us MONS 6 __ _ / _ _ / _ _ B6 us MONs ll. Description of STONGEST URGE Situation checklist to be sure all questions are answered. Ask: Could you tailrnernoreaboutthetimewhen. .(lnsert situation with highest urge strength)? Canyoulelfme mabommwashapperfingbdaechwnemthmmeugemsmeW Iwwfdafso Iiketoirnowwheredrishappened wttodsewasherearfludryyoutffinkttrissiunfionhappened. Write in situation number. urge strength rating of the event with the highest strength rating. Ask follow-up questions if necessary to obtain details of the event‘s history. Write a brief summary of the event and use the Sit. #: <1 Urge Strength: Circumstances: 151 Sit. code (office use only): 59591 Ehflkflst When? What? Where? Who? Why? Modified from Carver at al. 113 lIl. Coping Responses to STRONGEST URGE Situation lD:____ Asklwowdflkerokrmumatlypesdrespmsesyouaiedinmmfim. Faeadirespampleasatellme Inwaflmyouflgddnmsparswmver. orientwodmes. Weetofwrdmesmorethanfivetimes. In rem '0 this M0"- did W0 W to... 1-2 3.4 s T111108 Never Times TIrnes or More 1. Let your feelings out by crying or yelling? 0 1 2 3 2. Think about the situation in a more positive way. like 'It could be worse'? 0 1 2 3 3. Accept that this happened and that it can't be changed? 0 1 2 3 4. Find something funny about the situation? 0 1 2 3 5. Give up trying to reach your goals in the situation? 0 1 2 3 6. Hold back or restrain yourself untl the time was right to do something? ....... 0 1 2 3 7. Make a plan about the best way to deal with the situation? ....... . ................... o 1 2 3 8. Put aside other activities so you could deal with this situation? ..................... 0 1 2 3 9. Take action to get rid of the problems in the situation? 0 1 2 3 10. Seek spiritual comfort by praying or meditating? 0 1 2 3 11 Take your mind off the situation by doing other things? 0 1 2 3 12. Tell someone your feelings about the situation to get some support? ......... 0 1 2 3 13. Get some advice from someone about what to do? 0 1 2 3 14. Bealoneforaperiodoftime? 0 1 2 3 15. Help yourself feel better by using addictive drugs or alcohol? .......................... 0 1 2 3 16. Express your emotions by trying to destroy something or hurt someone? 0 1 2 3 17. Think about the situation as a chance to learn or grow as a person? .......... 0 1 2 3 18. Decide to learn to live with the situation? 0 1 2 3 19. Make jokes about the situation? 0 1 2 3 20. Stop your attempts to deal with the situation? 0 1 2 3 21. Avoid making matters worse by acting too soon? 0 1 2 3 22. Think hard to come up with a strategy for the situation? 0 1 2 3 23. Focus on the situation and let other things slide a little? 0 1 2 3 24. Take direct action to get around the situation? 0 1 2 3 25. Seek God's help or put your trust in God? 0 1 2 3 26. Think about other things so you could forget about the situation? ......... - ....... 0 1 2 3 27. Get some understanding or sympathy from someone? 0 1 2 3 28. Talk to someone who could do something to help you? 0 1 2 3 29. Get away from everything and everyone so you could deal with this alone? 0 1 2 3 30. Think about the situation less by drinking alcohol or taking drugs? .............. 0 1 2 3 Modified from Carveret al. 114 ID: Self-Efficacy Questionnaire Sariwfllraadawnbarolsrmanons' ‘ oreverrtsinwhr'chsomepeopfeexpenenc‘ eadrinkingordrug problem. Imagine youselfasyouare rightNOlll/in each ofthese situations. Irxticateon mascots provrd’ edhowconfidentyouaredtatyouwr'lfboabletoresrst' meta-gatedrinkaicoholand/orusedmgs in that sltzration. Ask: HOW sure am you M you would be we to (0981' me Not at A Little Moderately Very Extremely urge to dn'nk or use drugs: "' MW" 2 When your stomach felt like it was tied in knots. 0 1 2 3 4 3 When something good would happen and you would feel like 0 1 2 3 4 celebrating. 4 When you would start to think that just one drink or drug use 0 1 2 3 4 wouldn't hurt. 5 When you would suddenly have an urge to drink/use drugs. 0 1 2 3 4 6 When you had an argument with a friend or family member. 0 1 2 3 4 7 When you would be at a party and other people would be 0 1 2 3 4 drinking or using drugs. 8 When you wanted to heighten your sexual enjoyment. 0 1 2 3 4 9 When you felt you had let yourself down. 0 1 2 3 4 10 When you felt nauseous. O 1 2 3 4 12 When you wanted to prove to yourself you could control 0 1 2 3 4 your drinking or drug use. 14 When pressure would build up at work/school. 0 t 2 3 4 16 When you wanted to celebrate with friends. 0 1 2 3 4 17 When you were afraid that things weren't going to work out. 0 1 2 3 4 19 When you felt satisfied with something you had done. 0 1 2 3 4 21 When you would pass a liquor store or ran into your dealer. 0 1 2 3 4 23 When you would meet an old friend who suggested that you 0 1 2 3 4 drink or use drugs together. ModIerdfrom Annis 1982 115 Social Network ID: Nomi willaskyouahoutpeoplewhohavebaenimportanttoyouinyourlifedruingdielaatTHREE MONTHS. Thesearepeoplewhoyou haveheenincpr1ta¢wid1lbyphone.letter.orinpersonlindtelastTHR£E MONTHS. Today is: So. deeennorrdisaoo was: Relationshipto respondent First name. last initial How frequently do you have contact with this person (Visit. phone. letter) (Enter 2 digit 0 code) I yea rly few tr mes 0 YOU! 2 monthly 3 weekly 4 daily How long have you knownthis person? Years Mons. How close/special is this person to you? 0 Not at all 1 A little 2 Moderately 3 Very Much 4 Extremal)! How many days on a typical week does this person use drugs or drink alcohol? HOW MUCH 0003 this person suoDOrt your effortsto recover? 0 Not at all 1 A little 2 Moderately 3 Very much 4 Extremely Live-impouselpartnerttt monthsl I. I I Family/relative“ months) sues-Ito Friends (3 months) 1. 2. 3. 4. 5. Professionalslcgc doctor. socialwork'er.psychologist.clerey. and so on.) 3 months 1. 2. 3. Others (e.g.. member. co-worlterneighbor. etc.) 3 months ’ NOTE: You should not limit the number of importantpersonsthey list. If necessary.use additionalsheeta. 0 Copyright Ramanathan. Montcalm. ti Reischl REFERENCES 116 REFERENCES Alterman, A.I., Kushner, R., Holahan, J.N. (1990). Cognitive functioning and treatment outcome in alcoholics. Ihghlggpngl g; hehygus and ughthl Disgggg, 175(5), 494-499. Annis, H.M. & Davis, C.S. (1989). Relapse prevention training: A cognitive-behavioral approach based on self- efficacy theory. l2uraal_2f_Qhem12al_Dseenden§¥_TreaLaens. 81-103. Annis. H-M- (1982a). E11ua112aal_Qanfidenca_guessiennaire. Toronto: Addiction Research Foundation of Ontario. Annie, H.M. (1982b). A relapse prevention model for treatment of alcoholics. In Miller, W.R. & Heather, N. eds. Treating Addgctivg Behaviors, Ehocessgs gf ghahgg. New York: Plenum Press. Baer, J.S., Kamarack, R., Lichtenstein, E., & Ransom, C.C. (1989). Prediction of smoking relapse: Analysis of temptations and transgressions after initial cessation. ournal of ns a l ' PS 0 , 52(5), 623- 627. Bandura. A- (1977). §221al.learniae_1hecrx- Englewood Cliffs, NJ: Prentice-Hall. Barker, R.G. (1963). On the nature of the environment. Journal of Social Issues, 12, 17-38. Barrios, F.X. & Niehaus, J.C. (1985). The influence of smoker status, smoking history, sex, and situational variables on smokers' self-efficacy. hgdictive hehgvigrs, lg, 425-429. Billings, A.G., & Moos, R.H. (1983). Psychosocial processes of recovery among alcoholics and their families: Implications for clinicians and program evaluators. Addictive hehaviohg, 8. 205-215, Bliss, R. E., Garvey, A. J., & Heinold, J. W. (1989). The influence of situation and coping on relapse crisis outcomes after smoking cessation. ourna f Co 5 in d Esychology, 57, 443- 449. Brandon, R. R., Tiffany, & Baker, T.B. (1988). The process of smoking relapse. o s 0 us es Monograph §gzi fig, 1;, 104-117. Brandon, R.H., Tiffany, S.T., Obremski, K.M., & Baker, T.B. (1990). Postcessation cigarette use: The process of relapse. Addictlye Behavigns, 15, 105-114. 117 Bromet, E., & Moos, R. H. (1977). Environmental resources and the posttreatment functioning of alcoholic patients. Qghghgl 9: Health and Sggial Behgxigr 15. 326- 338. Bronfenbrenner, U- (1979). The_e22legx_gf_humen_de2elenment- Cambridge, MA: Harvard University Press. Brown, S.A., Vik, P.W., & Creamer, V.A. (1989). Characteristics of relapse following adolescent substance abuse treatment. Addigtixe_neh§xigr§. 11.291-300- Brownell, K.D., Marlatt, G.A., Lichtenstein, E., 8 Wilson, G.T. (1986). Understanding and Preventing Relapse. Ahgziggn Esycholggish, 41(2), 765-782. Burling, R.A., Reilly, P.M., Moltzen, J.C., & Ziff, D.C. (1989). Self-Efficacy and relapse among inpatient drug and alcohol abusers: A predictor of outcome. igurhal g: ghudies on Alcohol, 50(5), 354-360. Captain, C. (November, 1989). The impact of family stress- regulating variables on recovery from alcoholism. nilihgzy Medicing, 1511111, 539-546. Carver, C.S., Weintraub, J.K., & Scheier, M.F. (1989). Assessing coping strategies: A theory based approach. ourna of ersona 't nd So PS 0 , 55(2), 267- 283. Catalano, R.F., & Hawkins, J.D. (1985). Project SKILLS: Preliminary results from a theoretically based aftercare experiment. Progress in the development of cost-effective treatment for drug abusers, National Institute on Drug Abuse Research Monograph, No 58., pp. 157-181, Rockville, MD. Chaney, E.F., O’Leary, M.R., & Marlatt, G.A. (1978). Skill training with alcoholics. u o 'n n ’ ' Esycholggy, 55(5), 1092-1104. Childress, A.R., McLellan, A.T., & O’Brien, C.P. (1986). Role of conditioning factors in the development of drug dependence. Esyghiatrig Clihigg g: HQIEh Amghicg, 2(3), 413- 425. Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdal, New Jersey: Lawrence Erlbaum Associates. Condiotte, M.M., & Lichtenstein, E. (1981). Self-efficacy and relapse in smoking cessation programs. Jghzhgl_gfi onsul n a d nica s c o , 12, 648-658. 118 Congdon, D.C., & Holland, T.P. (1988). Measuring the effectiveness of substance abuse treatment: Toward a theory- based index. Journa1_2f_§2§ial_fierxiee_3e§eargh. 12. 23-48. Cooper, M.L., Russell, M., & George, W.H. (1988). Coping, expectancies, and alcohol abuse: A test of social learning formulations. l2urnal_of_Abnorma1_2§¥2hologx. 211214 218- 230. Cronkite, R.C., & Moos, R.H. (1980). Determinants of posttreatment functioning of alcoholic patients: A conceptual framework. Jo2rnal_2f.92nanlting_and_glinigal Esycholggy, Ag, 305-316. Curry, S. & Marlatt, G.A. (1985). Unaided quitters' strategies for coping with temptations to smoke. In Wills, T.A. and Shiffman, S. (Eds.). n an ubs nce hhuse. New York: Academic Press, 243-265. Donovan, D.M., & Marlatt, G.A. (1988). Assgsshgnh g: gddictive behaviohs. New York: Guilford Press. Eriksen, L., Bjornstad, S., & Gotestam, K. G. (1986). Social skills training in groups for alcoholics: One-year treatment outcome for groups and individuals. agg1gh12g_fighgyigz§, 11, 309- 329. Fagan, R.W. & Mauss, A.L. (1986). Social margin and social reentry: An evaluation of a rehabilitation program for skid row alcoholics. gourngl of §§Qd1gs gn Algghgl, 47(5), 413- 425. Favazza, A.R., & Jackson-Thompson, J. (1984). Social networks of alcoholics: Some early findings. A co : Clinical and Experimenhgl Besgazgh, 5, 9-15. Finney, J. W., & Moos, R. H. (1984). Environmental assessment and evaluation research: Examples from mental health and substance abuse programs. Evalughigh ghd Prograh £1ghnihg, 1, 151- 167. Finney, J.W., Moos, R.H., & Mewborn, C.R. (1980). Posttreatment experiences and treatment outcome of alcoholic patients six months and two years after hospitalization. ourna of Consu t'n nd i i s , 55(1), 17-29. Fitzgerald, M.R., Davies, W.H., Zucker, R.A., & Klinger, M. (in press). Developmental systems theory and substance abuse: A conceptual and methodological framework for analyzing patterns of variation in families. In L.L'Abate (Ed.). fiahdbooh of developmehhgl fi§h11y psyghglggy gag pexghopathologx. New York: Wiley- 119 Fraser, M.W., & Hawkins, J.D. (1984). The social networks of opioid abusers - WW5. 19(3), 903-917. Galanter, M. (1987L Social network therapy for cocaine dependence. Adxan2e__in_Al_oh91_and_§ub_tan___abu§e 5. 159- 175. Gorski, T.T. (1986). Relapse prevention planning: A new recovery tool. A122h2l_Health.and_Besear2h_florld. 6-11. Havassy, B.E., Hall, S.H., & Tschann, J.M. (1986). Social support and relapse to tobacco, alcohol, and opiates: Preliminary findings. Proceedings from the 48th Annual Scientific Meeting of the Committee on Problems of Drug Dependence, pp. 207-213, Rockville, MD. Hawkins, J.D., & Fraser, M.W. (1987). The social networks of drug abusers before and after treatment. Ihg_lnhgnnnh1gnnl Journal 0: the Addictions, 22, 343-355. Hunt, W.A., Barnett, L.W., & Branch, L.G. (1971). Relapse rates in addiction programs. 0 r a o l'n'ca h o 21, 455-456. Insel, P.M., & Moos, R.H. (March, 1974). Psychological environments: Expanding the scope of human ecology. Angz1gnn Esygholgglgt, 179-188. Joe, G.W., & Simpson, D.D. (1983). Social factors related to the follow-up status of opioid addicts. Journal 9f Psychogctlvg hnngs, 15, 207-216. Kadden, R.M., Getter, H., Cooney, N.L., & Litt, M.D. (1989). Matching alcoholics to coping skills or interactional therapies: Posttreatment results. on Co 5 ' Clinical Esychology, 57(5), 698-704. Kosten, T.R., Rounsaville, B.J., & Kleber, H.D. (1987). Multidimensionality and prediction of treatment outcome in opioid addicts: 2.5 year follow-up. annnghgnglyg Psychiatry, 25, 3-13. Lazarus, R.S. (1966). Esychological stress and the gonlng process. New York: Appleton-Century-Crofts. Lazarus, R.S., & Folkman, S. (1984). ess a's d ggn1ng. New York: Springer Publishing. Lesieur, H.R., & Blume, 8.8. (1991). Evaluation of patients treated for pathological gambling in a combined alcohol, substance abuse, and pathological gambling treatment unit 120 using the Addiction Severity Index. B;1§1gh_59nznnl_gfi Addlcgign, 55, 1017-1028. Leukefeld, C.G., 8 Tims, F.M. (1989). Relapse and recovery in drug abuse: Research and practice. Ihg_1n§gzng§1gn§l Journal of hhg 5551951535, 25(5), 189-201. Levin. K. (1935). Erinei2les_2f_Tonological_£§xchologx. New York: McGraw-Hill. Litman, G.K., Eiser, J.R., Rawson, N.S., 8 Oppenheim, A.N. (1979). Differences in relapse precipitants and coping, behaviour between alcohol relapsers and survivors. fighnyign; Besehrgh 8 Ihgzgny, 11, 89-94. Litman, G.K., Stapleton, J., Oppenheim, A.N., Peleg, M., 8 Jackson, P. (1984). The relationship between coping behaviours, their effectiveness and alcoholism relapse and survival. Briti§h_gournal_of_AQ§iction. 12. 283-291. Living Sober (1975). h1y1ng_§ghg;. New York: Alcoholics Anonymous World Service Inc. Longabaugh, R. (1989). Recommended research priorities for the assessment of treatment outcomes. Egyghglggy_g£ Addictive 5ehavigr§, 2(5), 131-141. Lowe, W.C., 8 Thomas, S.D. (1976). Assessing alcoholism treatment effectiveness. ggurngl g: ghugies gn Algghgl, 57(7), 883-889. Ludwig, A.M., 8 Stark, L.H. (1974). Alcohol Craving. Quanhgrly igurnal g: Studigs gn Algghgl, 55, 899-905. MacDonald, J.G. (1987). Predictors of treatment outcome for alcoholic women. Ihe Inhernatigngl ggurngl of hhg Addictions, 22(3), 235-248. Maisto, S.A., 8 Connors, G.J. (1988). Assessment of treatment outcome. In D. Donovan 8 G.A. Marlatt's (Eds.) Assessment of Adglctlve fighavlgng (pp. 421-453). Maisto, S.A., Sobell, M.B., Cooper, A.M., 8 Sobell, L.C. (1979). Test-retest reliability of retrospective self- reports in three populations of alcohol abusers. ggnnnhl_g; 5ehavloral Assessmenh, 1, 315-326. Maisto, S.A., Sobell, M.B., 8 Sobell, L.C. (1982). Reliability of self-reports of low ethanol consumption by problem drinkers over 18 months of follow-up. nnng_gng Alcohol Denengengg, 2, 273-278. 121 Marlatt, G.A. (1977). The drinking profile: A questionnaire for the behavioral assessment of alcoholism. In E.J. Mash 8 L.G. Terdal (Eds.) e - ° ' B_haxi2r_thera2x_asses§ment1_diagnosi§1 gesign, and gvglngh1gn (pp. 121-137). New York: Springer Publishing. Marlatt, G. A. (1985L Lifestyle modification. In G. A. Marlatt & J. R. Gordon (Eds. ) Bela2se.2rexention__maintenanse §tra§eg1es in hhg hrgatmenh 9f aggigtixg_h£hg_121§ (PP. 280- 345). New York: Guilford Press. Marlatt, G.A. 8 Gordon, J.R. (1980). Determinants of Relapse: Implication for the Maintenance of Behavior change. In: P.O. Davidson 8 S.M. Davidson (Eds.). v' e - Qhanging_flealth_Lifestxles. (pp-410-452). Elmsford. NY: Pergamon. Marlatt, G.A., 8 Gordon, J.R. (Eds.). (1985). 351525; revent'o : ' anc e ' s ' ' 'v a ' ghangg. New York: Guilford. Martin, G.W., 8 D.A. Wilkinson. (1989). Methodological issues in the evaluation of treatment of drug dependence. Advances 1n 5ehav1grgl Besgangh gng Ihgrgny, 11, 133-150. McCrady, B.S. (1989). Extending relapse prevention models to couples. Addic§1vg 2§h§y1ons, 15, 69-74. McLatchie, B.H. 8 Lomp, R.G.E. (1988). Alcoholics anonymous affiliation and treatment outcome among a clinical sample of problem drinkers. er' Jon 0 Abusg, 15(5), 309-324. McLellan, et al. (1985). Gu'de e d ion Sev Index. DHHS Publication No. (ADM) 85-1419. Rockville, MD: National Institute on Drug Abuse. McLellan, A.T., Luborsky, L., Woody, G.E., O'Brien, C.P., 8 Kron, R. (1981). Are the "addiction-related" problems of substance abusers really related? Ihg_59n;ngl_gfi_flg;ygn§_nng Mental 2155555, 152(5), 232-239. McLellan, A.T., Luborsky, L., Woody, G.E., O’Brien, C.P., 8 Druley, K.A. (1983). Predicting response to alcohol and drug abuse treatments. Archivgs gf nggral Esygh1ghry, 55, 620- 625. McLellan, A.T., Luborsky, L., Cacciola, J., Griffith, J., Evans, F., Barr, H.L., 8 O’Brien, C.P.(1985). New data from the addiction severity index. Ihg ggnnngl 9f heryous gng Mengal Qiseasg, 175(7), 412-423. ‘ 122 McLellan, A.T., Luborsky, L., O'Brien, C.P. (1986). Alcohol and drug abuse treatment in three different populations: Is there improvement and is it predictable? Angniggn_Jgn:nnl_Qj Erug_and_Alsohol_Ahu§e. 1211821. 101-120. McLellan, A.T., O'Brien, C.P., Woody, G.E., Luborsky, L., 8 8 Druley, K.A. (1982). Is drug abuse treatment effective? National Institute on Drug Abuse: Research Monograph Series, 51, 223-229. McLellan, A.T., Woody, G.E., Luborsky, L., O'Brien, C.P., 8 Druley, K.A. (1983). Increased effectiveness of substance abuse treatment. Ihe1l92rnal_9f_Nerxou§_and_uental_ni§ea§e. 171(10), 597-605. Mermelstein, R., Lichtenstein, E., 8 McIntyre, K. (1983). Partner support and relapse in smoking-cessation programs. Journal of Consu1t1ng ang Ql1n19a l E_¥Qh_1QQY, 51111 465- 466. Moberg, D.P., Krause, W.K., 8 Klien, P.E. (1982). Posttreatment drinking behavior among inpatients from an industrial alcoholism program. e rn t'o a rn the Adgighi1gns, 17 (5), 549- -567. Moos, R.H. (August, 1973). Conceptualizations of human environments. An§r1can 23ygholgg1s§,652-665. Moos, R.H. (1974). Family Environment Scale Preliminary Manual. Palo Alto: Consulting Psychologists Press. Moos, R.H., 8 Bromet, E. (1977). Environmental resources and the posttreatment functioning of alcoholic patients. Jgnrnnl of Health and §oc1al 5ehav1or, 15, 326-338. Moos, R.H., Bromet, E., Tsu, V., 8 Moos, B. (1979). Family characteristics and the outcome of treatment for alcoholism. Journal 0: §tudies on Algghol, 55, 78-88. Moos, R.H., 8 Insel, P. (1974) Work Environment Scale Preliminary Manual. Palo Alto: Consulting Psychologists Press. Moos, R.H. 8 Finney, J.W. (1983). The expanding scope of alcoholism treatment evaluation. An§r1gnn_£§yghglgg1§§, 55. 1036-1044. Nathan, P.E., 8 Skinstad, A. (1987). Outcomes of treatment for alcohol problems: Current methods, problems, and results. 0 a of C nsu 'n a d C ' 'ca s , 55(3), 332-340. 123 Norbeck, J., Lindsey, A., 8 Carrieri, V. (1981). The development of an instrument to measure social support. nuns1ng Besggngh, 59(5), 264-269. O'Leary, D.E., O’Leary, M.R., 8 Donovan, D.M. (1976). Social skill acquisition and psychosocial development of alcoholics: A review. Agg1gh1yg_55hny1gz§, 1, 111-120. Page, R. D. 8 Badgett, S. (1984). Alcoholism treatment with environmental support contracting. Angn1gnn_Jgnznnl_gfi_nzng 5nd Alcohol Ahhsg, 19(5), 589-605. Pargament, K.I. (1986). Refining fit: Conceptual and methodological challenges. Amerisan_lournal_9f_29mmunitx Esycholggy, 15(5), 677-684. Perkins, D.V., Burns, T.P., Perry, J.C., 8 Nielson, K.P. (1988). Behavior setting theory and community psychology: An analysis and critique. I9urna1_of_§9mmunitx_zsxchologx. 15. 355-372. Reischl, T.M., Ramanathan, C.S., 8 Nguyen, H.X. (June, 1993).Psychological predictors of relapse following a relapse prevention intervention. Symposium presentation at the Biennial Community Research and Action Conference, Williamsburg, VA. Rosenberg, H. (1983). Relapsed versus non-relapsed alcohol abusers: Coping skills, life events, and social support. Addigtivg 5ghnv1ong, 5, 183-186. Sanchez-Carbonell, J., Cami, J., 8 Brigos, B. (1988). Follow-up of heroin addicts in Spain (EMETYSY Project): results lyear after treatment admission. 5:151§h_Jnn;nnl_gj Addictign, 5_, 1439- -1448. Sarason, 8.8. (1977). Wgrh, aging, and 599151 chnngg. New York: The Free Press. Schuckit, M.A., Schwei, M.G., 8 Gold, E. (1986). Prediction of outcome in inpatient alcoholics. Jgnnn51_gfi_5hng1g§_gn Alcohol, 57(2), 151-155. Schuckit, M.A., Morrison, C.R., 8 Gold, E.O. (1984). A pragmatic alcoholism treatment outcome scale. Angn1gnn Jounnal oz 555g 5nd Alcghgl Ahngg, 15111, 125-131. Sheeren, M. (1987). The relationship between relapse and involvement in alcoholics anonymous. on u Alcohol, 42(1), 104-106. 124 Shiffman, S. (1982). Relapse following smoking cessation: A situational analysis. J2urna1_2f.99nsulting_and_slinisal Esychology, 55(1), 71-86. Shiffman, S. (1989). Trans-Situational consistency in smoking relapse. Health_zsyshology. 9111. 471-481. Shiffman, S., 8 Wills, T.A. (1985). an1ng_5n5_5nh§hnngg use. Academic Press: London. Smith, J.W., Frawley, P.J., 8 Howard, M.O. (1991). Environmental risks to be considered in substance user treatment planning, implementation, and assessment. Th5 International Journgl 5: hhg 1551551555, 25(5), 371- 375. Sobell, M.B., Maisto, S.A., Sobell, L.C., Cooper, A.M., Cooper, T., 8 Sanders,B. (1981). Developing a prototype for evaluating alcohol treatment effectiveness. In: Sobell, L.C., Sobell, M.B., 8 Ward, E., (Eds.). Eynlnnh1ng_nlggh51 .nd d .- .ous- t -atm-n - - t ve.-_ : 1- - .2 .1 -_, (pp.129-150). New York: Pergamon Press. Sobell, L.C., Maisto, S.A., Sobell, M.B., 8 Cooper, A.M. (1979). Reliability of alcohol abusers' self-reports of drinking behavior. 5ehav1or Besennch 5nd Thgznny, 12, 157- 160. Sobell, L.C. 8 Sobell, M.B. (1980). Convergent validity: An approach to increasing confidence in treatment outcome conclusions with alcohol and drug abusers. In L.C. Sobell, M.B. Sobell and E. Ward (eds.), at'n c d Ahuse greatment Effect1veness: Regent Advanges, Pergamon Press, New York. Stewart, J., de Wit, H., 8 Eikelboom, R. (1984). Role of unconditioned and conditioned drug effects in the self- administration of opiates and stimulants. £§y5h515g1551 Eeview, 21(2), 251-268. Svanum, S. 8 McAdoo, W.G. (1989). Predicting rapid relapse following treatment for chemical dependence: A matched- subject design. Jou na 0 s a d C ' ' Esycholggy, 52(2), 222-226. Thurstin, A.H., Alfano, A.M., 8 Nerviano, V.J. (1987). The efficacy of AA attendance for aftercare of inpatient alcoholics: Some follow-up data. Ih5_1n55;n5515n51_Jnnnnnl 5f the A551c515ns, 22, 1083-1090. Timmer, S.G., Veroff, J., 8 Colten, M.B. (1985). Life stress, helplessness, and the use of alcohol and drugs to cope: An analysis of national survey data. In Shiffman, S. 125 and Wills. T.A. (Eds.). Q9ping_and_supstanee_apuse. New York, Academic Press, 171-198. Towle, L.H. (June, 1974). Alcoholism treatment outcomes in different populations. Proceedings from the fourth annual alcoholism conference of the National Institute on Alcohol Abuse and Alcoholism: research, treatment and prevention. DHEW # 76-284: Rockville, MD. Tucker, J.A., Vuchinich, R.E., 8 Gladsjo, J.A. (1991). Environmental influences on relapse in substance use disorders. Ihe_Internati9nal_lournal_9f_the_bddictions. 25(7A & 51), 1017-1050. Vaillant,G.E. (1988). What can long-term follow-up teach us about relapse and prevention of relapse in addiction? 5ritish Journal of Addigh1on, 55, 1147-1157. Vuchinich, R.E., Tucker, J.A. (1988). Contributions from behavioral theories of choice to an analysis of alcohol abuse. 19urnal_of_A2n2rmal_Esyshology. 21121. 181-195. Wallace, B.C. (1989). Psychological and environmental determinants of relapse in crack cocaine smokers. o a f 5ubstange Ahnsg Treatment, 5, 95-106. Walton, M.A. (1988). Lapse and relapse: A process study in methamphetamine and cocaine users. Unpublished masters thesis, San Diego State University: CA. Wanberg, R.W., Horn, J.L. (October, 1983). Assessment of alcohol use with multidimensional concepts and measures. Anericnn Esychglog1sh, 1055-1069. Wells, E.A., Hawkins, J.D., Catalano, R.C. (1988). Choosing drug use measures for treatment outcome studies. I. The influence of measurement approach on treatment results. Th5 international Journal of Lhe Add1gt1ons, 25(5), 851-873. Wicker, A.W. (1972). Processes which mediate behavior- environment congruence. 55hny15;51_5g15n55, 12, 265-277. Wicker, A.W. (1987). Behavior settings reconsidered: Temporal stages, resources, internal dynamics, context. In D. Stokols and I. Altman (Eds.) Handbook of environmental psychology. New York: John Wiley 8 Sons. Wilkenson, D.A., 8 LeBreten, S. (1986). Early indications of treatment outcome in multiple drug users. In W.E. Miller and N. Heather (Eds.) Treat1ng Addigt1ve 55h5y15rs, nnggesses 5: ghange (pp. 239-261). New York: Plenum Press. 126 Williams, J.S., Stout, J.R., 8 Erickson, L (1986). Comparison of the importance of alcoholics anonymous and outpatient counseling to maintenance of sobriety among alcohol abusers. Esycholog1gal 3550155, 55, 803-806. Woody, G.E., McLellan, A.T., Luborsky, L., 8 O'Brien, C.P. (1987). Twelve-month follow-up of psychotherapy for opiate dependence. WWW. 1515.1. 590-596. Yates, A.J. 8 Thain, J. (1985). Self-efficacy as a predictor of relapse following voluntary cessation of smoking. Addict1V5 5eh5v1ons, 15, 291-298.