Iowa Case Management for Rural Drug Abuse

Abstract

Objective

The purpose of this research was to evaluate the effectiveness of a comprehensive, strengths-based model of case management for clients in drug abuse treatment.

Method

503 volunteers from residential or intensive outpatient treatment were randomly assigned to one of three conditions of Iowa Case Management (ICM) plus treatment as usual (TAU), or to a fourth condition of TAU only. All were assessed at intake and followed at 3, 6, and 12 months.

Results

Clients in all four conditions significantly decreased substance use by 3 months after intake and maintained most gains over time. However, the addition of ICM to TAU did not improve substance use outcomes.

Conclusion

Overall, the addition of case management did not significantly improve drug treatment as hypothesized by both researchers and clinicians. Some results were mixed, possibly due to the heterogeneous sample, wide range of case management activities, or difficulty retaining participants over time.

Because of the complexity of drug abuse and dependence, some contend that case management should be provided as a supplemental service with traditional drug treatment services (Cellini, 2003; Holloway, McLean, & Robertson, 1991; Sullivan, Wolk, & Hartmann, 1992; Vanderplasschen, Rapp, Wolf, & Broekaert, 2004). Furthermore, clients from rural areas may need assistance finding other services such as mental health counseling and employment services because most of these services are not found in rural settings. In most drug abuse treatment settings, case management has been one of the responsibilities of the primary addiction counselor. In actual practice, however, these drug counselors have little time beyond their primary drug counseling responsibilities to provide case management services, and usually they cannot receive reimbursement for these services (Thilges, Carswell, Vaughan, & Hall, 1996). By understanding whether case management helps drug abuse clients maintain healthier lifestyles or has other important effects, drug treatment agencies can determine whether resources should be devoted to this type of intervention.

The purpose of this research was to develop and evaluate a case management model that would have positive effects with clients, including those from rural areas, who were receiving substance abuse treatment services. Participants who received case management in addition to standard drug treatment were hypothesized to report better outcomes than those clients receiving substance abuse treatment only—especially in the long run (i.e., 12 months). Second, participants receiving case management were hypothesized to report higher levels of service utilization, which would contribute to better outcomes overall. Finally, participants who were from rural areas and who received case management were hypothesized to have better outcomes than urban participants because of outreach activities.

Background

Rural Substance Misuse

Rural communities may be heterogeneous when compared to one another, but within each community, populations may be homogenous and highly influenced by local culture (Murray & Keller, 1991). Rural communities may also be more closely socially integrated than urban areas (Dewey, 1960; Photiadis & Simoni, 1983), making it more difficult for rural people to seek help for personal problems with the confidence in anonymity afforded urban dwellers (Bischoff, Hollist, Smith, & Flack, 2004). Rural dwellers tend to be more politically conservative (Bushy, 1994; Melton, 1983), work-oriented (Bushy, 1994; Wooster, 1972), family-oriented (Flax et al., 1979), self-reliant (Bushy, 1994), religious (Wallace et al., 2003), traditional, provincial, and slower to change.

Rural society is sometimes viewed as simple and immune from big city problems, but when it comes to substance misuse, this stereotype breaks down. The economic systems in rural areas have been increasingly disrupted, giving rise to higher crime rates and substance abuse, domestic violence, and a large proportion of single-parent households (Conger, 1997). Owing to economic problems, rural areas may face increased risk for emotional, behavioral, and substance use disorders (Conger & Elder, 1994). Some reports have found that rates of rural drug use and misuse in the United States are very similar to those in urban environments (General Accounting Office Report, 1990). However, rural substance users may misuse different drugs than urban users, such as alcohol and methamphetamine (Schoeneberger, 2006). In studies completed outside the United States, problematic rates of substance abuse in rural areas have been documented (Chaturvedi, Phukan, & Mahanta, 2003; Furr-Holden & Anthony, 2003; Forsyth & Barnard, 1999; Hapke et al., 1998; Paykel et al., 2000; Peltzer, Malaka, & Phaswana, 2002).

Studies of adult substance users in rural areas suggest that these individuals have higher rates of comorbidity (Simmons & Havens, 2007) and lower rates of treatment utilization (Warner & Leukefeld, 2001). Rural adult populations have unique barriers to treatment including healthcare personnel shortages (American Academy of Family Physicians, 2002; Meyer & Morrissey, 2007), lack of rural clinics (Hauenstein, Petterson, Rovnyak, Merwin, Heise, & Wagner, 2007; Robertson & Donnermeyer, 1997), and transportation issues (Sloboda, 2002). Furthermore, the rural adult population is more likely to be poor and less likely to have health insurance (Reschovasky & Staiti, 2005).

Models of Case Management

Case management models are usually described by the methods employed or by the philosophy behind the model. Ross (1980) distinguishes case management models based on levels of comprehensiveness—minimal, coordinated (i.e., brokerage), and comprehensive. Minimal models of case management involve minimal supervision and referral. Brokerage models of case management attempt to match resources to client needs and are characterized by more office-based work, telephone contact, and larger caseloads. Comprehensive models of case management (e.g., Iowa Case Management (ICM); Hall, Vaughan, Vaughn, Block, & Schut, 1999) are characterized by greater intensity of services, including therapeutic services and smaller caseloads.

Comprehensive models of case management have been shown to be an effective intervention with several vulnerable client populations: (a) older adults (Hyduk, 2002; Kemper, 1990; Sullivan & Fisher, 1994; Yang, Garis, & McClure, 2005), (b) persons with AIDS (Gasiorowicz et al., 2005; Sowell & Grier, 1995), (c) mental health clients (Burns & Santos, 1995; Hopkins & Ramsundar, 2006; Rapp & Chamberlain, 1985; Solomon & Draine, 1995; Stein & Test, 1980), (d) children with severe emotional disturbance (Burns & Santos, 1995; Dwyer, 2002), and (e) people with substance use disorders (Rapp, 1997; Rapp, Kelliher, Fisher, & Hall, 1994; Sullivan et al., 1992; Vanderplasschen et al., 2004). Furthermore, the addition of a strengths-based approach to comprehensive case management is especially effective for clients with little motivation (i.e., traditionally called denial) for change (Rapp et al., 1994; Siegal, Rapp, Li, & Saha, 2001).

Case management in various formats has been linked to improved retention in substance abuse treatment (Laken & Ager, 1996; Mejta, Bokos, Mickenberg, Maslar, & Senay, 1997; Rapp, Siegal, Li, & Saha, 1998; Siegal, Rapp, Li, Saha, & Kirk, 1997), greater use of primary care and other medical services (Knowlton et al., 2001; McCoy, Dodds, Rivers, & McCoy, 1992; Schlenger, Kroutil, & Roland, 1992), and fewer employment problems (McLellan et al., 2003; Siegal et al., 1996). Additionally, case management has been found to improve family functioning (Loudenburg & Leonardson, 2003; McLellan et al., 2003; Sharlin & Shamai, 1995) and to reduce substance use among parents, which in turn reduces associated individual and family risk factors brought into the home by the substance abuser (Kerson, 1990; Lanehart, Clark, Dratochvil, Rollings, & Fidora, 1994; McLellan et al., 2003).

Iowa Case Management Project

The Iowa Case Management Project (ICMP) for Rural Drug Abuse was funded by a grant from the National Institute on Drug Abuse (NIDA). From 1993 until 1999, a total of 1,434 clients who were referred to either residential (RES) or intensive outpatient (IOP) treatment programs in a community-based agency were assessed at intake and then invited to participate in this study. A total of 909 clients agreed to participate and they were randomly assigned to one of four research conditions. In one condition that served as the control group, participants received standard substance abuse treatment only. In the other three conditions, participants received services from case managers who used the Iowa Case Management (ICM) model. All participants (and nonparticipants) were offered the usual treatment provided by a drug treatment agency. The ICM model addressed issues that are not typically part of traditional drug treatment by incorporating home visits, transportation, and other out-reach activities. The study presented in this article includes participants for whom baseline and follow-up assessments were attempted.

Preliminary Results

Preliminary results from ICMP suggested that ICM significantly improved client psychological functioning and employment status (Hall et al., 1999). At follow-up sessions, residential clients, but not intensive outpatient clients, receiving ICM reported fewer days of psychological distress and more days paid for working than did clients in the control condition. Clients in all conditions, including those in the control group, experienced significant reductions in drug and alcohol use during the 12 months following treatment, with the largest reduction from intake to the 3-month follow-up session. Other reports from the ICMP study showed that ICM may play a role in improving family functioning and parental attitudes (Vaughan-Sarrazin, Huber, & Hall, 2002) and increasing utilization of medical and drug treatment services (Vaughan-Sarrazin, Hall, & Rick, 2000). More specifically, ICM residential clients reported improved family functioning on the Index of Family Relations (IFR; Hudson, 2000) and for those with children, improved parenting attitudes on the Index of Parental Attitudes (IPA; Hudson, 2000). Additionally, residential clients with a case manager whose office was located at the drug treatment agency received more drug treatment and more medical services compared with clients in other conditions over the first 6 months following intake.

Other analyses with these data had differing results. Focusing only on composite scores from the Addiction Severity Index (see below) and on substance-use-free days, Saleh et al. (2002) found only modest support for ICM in the legal, employment, and psychiatric domains. In a subsequent study, Saleh et al. (2003) found that clients receiving ICM decreased their use of mental health services but increased their use of inpatient care, access to physicians, and visits to the emergency room—thus increasing overall costs.

Impact of Rural Status

Using the Rural–Urban Commuting Area (RUCA) coding system (Economic Research Service, 2005), characteristics of rural clients (small town) in our ICMP study were compared with urban and large town clients. Clients from rural areas were more likely to have automobile transportation available as well as a valid driver’s license than clients from these other two areas. Clients from large towns outside urban centers reported more employment problems and lower monthly income. Clients from large and small towns were more likely to report the use of amphetamines (most likely, methamphetamine) than did clients from urban centers. Finally, several other differences were found that are discussed in another manuscript (see Hall, Sarrazin, Huber, Smith, & Jang, 2007).

Method

All study procedures were approved by the University’s institutional review board (IRB) before any data were collected, and the IRB reviewed the project on an annual basis. No adverse effects were reported.

Community-Based Setting

The ICMP study was conducted in collaboration with the Mid-Eastern Council on Chemical Abuse (MECCA), a community-based non-profit substance abuse treatment agency with multiple programs. The main MECCA office was located in Johnson County, Iowa, which had a population of 98,000 and was considered Metro by the U.S. Department of Commerce Bureau of Economic Analysis (BEA) at the time of this study. The three additional counties that comprise MECCA’s catchment area were classified as rural counties adjacent to a metro county by the USBEA and had a total population of 42,000. Eighty-five percent of MECCA’s clients originated from within the four county catchment areas.

Clients are referred to MECCA from a variety of sources, including hospitals (7%), other health and social service agencies (7%), the legal system (30%), family and friends (6%), committals (10%), and self-referrals (40%). Initially, potential clients were screened by MECCA staff for substance use severity using the guidelines of the American Society of Addiction Medicine (ASAM, 1996) and, based on results, assigned to type of treatment (e.g., residential, outpatient). For clients assigned to residential, intensive outpatient or extended outpatient treatment programs, the MECCA staff person scheduled a time for the Clinical Intake Assessment (CIA). By agreement, a study research assistant conducted the CIA for MECCA using the Addiction Severity Index (ASI) for clients assigned to the residential (RES) or intensive outpatient (IOP) programs.

At the end of each CIA interview, the research assistant attempted to recruit the client into our study. By conducting these comprehensive assessments for MECCA, the overall burden was reduced for the clients, who could have been assessed by MECCA for clinical purposes and again by research assistants for study purposes—primarily to gather the same data. Research assistants were explicit from the beginning of the interview session that they were employees of the University of Iowa and that they were conducting these initial interviews for MECCA purposes.

Study Participants

From October 1995 through June 1998, 862 clients were admitted to the MECCA residential treatment program and 517 were admitted to the MECCA intensive outpatient program. Of the residential clients, 719 (83%) were eligible for the ICMP and 539 (75%) of those agreed to participate. Of the intensive outpatient clients, 512 (99%) were eligible for the ICMP and 207 (41%) of eligible clients agreed to participate. Residential clients were less likely to participate if they did not have a significant other or were referred by the criminal justice system, although the likelihood of participation increased if the client was confident in stopping drug use (Vaughn, Vaughan-Sarrazin, Saleh, Huber & Hall, 2002). For clients in the intensive outpatient program, the likelihood of participation in the research project increased if the client was female or had a major abuse problem with alcohol, cocaine, or multiple substances. From our study on participation (Vaughn, Sarrazin, Saleh, Huber, & Hall, 2002), a convenience sample provided some information about not wanting to participate. Potential clients’ reasons were as follows: not interested (26%), too busy (21%), expected to move soon (13%), too much time required (10%), and not comfortable sharing sensitive information (3%). In addition, 20% of those approached about participation said they would get back to us about their decision but could not be located after they left drug treatment.

Characteristics of the 539 residential and 207 outpatient clients who agreed to participate are displayed in Table 1. The average age of both residential and outpatient participants was about 33 years. Roughly one third (36%, n = 73) of outpatients were female, while 40% (n = 209) of residential participants were female. Most residential clients (n = 440; 82%) were White, while 14% (n = 73) were African American, and 4% (n = 25) were other races. Outpatient clients were also predominantly White (85%, n = 174), while 8% (n = 16) were African American, and 7% (n = 15) were other races. The main drugs of abuse were alcohol, marijuana, methamphetamine, cocaine, and multiple substances.

Table 1.

Client Demographic Characteristics and Baseline Status in ASI Domains by Drug Treatment Status (Residential Vs. Intensive Outpatient)

Demographic Characteristics	Outpatient (n = 207)	Residential (n = 539)
Mean age (SD)	31.9 (9.1)	33.5 (8.8)
Percentage White (n)	85% (174)	82% (440)
Percentage African American (n)	8% (16)	14% (73)
Percentage Other (n)	7% (15)	4% (25)
Percentage female (n)	36% (73)	40% (209)
Percentage from urban centers	80% (187)	60% (363)**
Percentage from large town	3% (7)	19% (115)
Percentage from small town	17% (40)	20% (122)
Days of the last 30
Bothered by medical problems (SD)	4.3 (9.3)	6.5 (10.9)**
Bothered by psychological problems	10.0 (10.8)	14.8 (11.9)**
Paid for working (SD)	11.2 (10.6)	6.6 (9.4)**
Troubled by family relationships (SD)	1.5 (5.4)	3.7(8.2)*
Troubled by social relationships (SD)	2.7 (6.6)	3.6 (7.6)
Days of the last 30
Using alcohol (SD)	9.5 (10.1)	13.3 (12.0)**
Using any nonalcohol drug (SD)	6.9 (10.9)	14.3 (12.7)**
Using marijuana (SD)	5.5 (10.2)	8.2 (11.6)
Using cocaine (SD)	1.3 (4.6)	5.5 (9.9)**
Using amphetamines (SD)	0.8 (3.7)	5.2 (10.0)**
Using multiple substances (SD)	3.5 (7.0)	8.3 (10.9)**
Engaging in illegal activity (SD)	0.5 (3.0)	2.3 (6.6) **

^*Note: Differences between residential and intensive outpatient programs: p < .05.

^**p < .01.

As expected, residential clients tended to have more severe characteristics at intake than outpatient clients. Residential participants reported more days troubled by medical problems during the 30 days prior to recruitment (p < .001), more days bothered by psychological or emotional problems (p < .001), fewer days paid for working (p < .001), more days bothered by family problems (p < .01), and more illegal activity (p < .001) than did outpatient clients. Residential clients also reported significantly more days using alcohol (p < .001), marijuana (p = .03), cocaine (p < .001), and amphetamines (p < .001) than did outpatient clients. The most frequently used substances by both residential and outpatient clients were alcohol, marijuana, cocaine, and amphetamines.

Iowa Case Management Procedures

Iowa Case Management had five broad, integrated clinical guidelines, which included contracting and negotiating, assessment and monitoring, brief solution-based counseling, planning and referral, and evaluation of process and outcomes (Hall et al., 1999). Additionally, the case-worker focused on client strengths and resources, rather than weaknesses, in order to achieve treatment goals.

The comprehensive model of Iowa Case Management was designed to supplement the interventions provided by a drug abuse treatment agency (Hall, Carswell, Walsh, Huber, & Jampoler, 2002). Five ICM functions were identified as broad clinical activity categories: (a) contracting and negotiating, (b) assessment and monitoring, (c) brief solution-based counseling, (d) planning and referral, and (e) evaluation of process and outcomes. Although these functions are described separately, they are fully interconnected when applied in case management activity sessions. The comprehensive model includes outreach activities such as visiting clients in their homes, assisting with transportation to and from client services, and providing emergency funds on a limited basis. This model has been described in greater detail elsewhere (Hall et al., 2002).

Goals of Iowa Case Management

ICM was designed to be a supplemental program to the interventions provided by drug treatment staff with rural clients. More specifically, this model has six program goals to help clients to (a) decrease or cease drug use; (b) improve their health, mental health, legal status, occupational status, and other key areas of their lives; (c) develop new areas of interest in their lives and develop better relationships with others in their social network; (d) lead more fulfilling lives, learning to see themselves as competent and having strengths to build on; (e) develop positively oriented goals and work with their case managers to operationalize solution plans to achieve these goals; and (f) learn problem solving, strengths and resources, and solution-planning techniques in order to undertake future work on their own.

Strengths and resources assessment

The first clinical contact between Iowa Case Managers and their clients included signing of the Client–Case Management agreement form, a review of basic needs (e.g., shelter, food, safety), and the Strengths & Resources Assessment (SRA; Rapp et al., 1994). The SRA consisted of four key steps. First, the client conducted a self-evaluation of strengths and resources using the Performance Evaluation Scale (PES) over nine life domains (e.g., social support, employment, financial). Second, the case manager reviewed each of these domains with the client and attempted to identify personal strengths and environmental resources using the principles and skills of ICM. After the completion of the SRA, the case manager and client independently completed post-SRA Personal Evaluation Scales. Finally, the case manager compared his or her PES ratings with the client, and final scores for each of the domains were arrived at by consensus. This process was considered the first intervention provided by the strengths-oriented case managers.

Case manager training and supervision

Case manager training included experiential training, regular supervisory sessions, and guest lectures covering a variety of associated topics (e.g., communications skills, substance abuse client needs, psychiatric issues, HIV risk reduction, and issues pertaining to ethnic minorities and women). Training included discussion and review of audiotaped sessions and role-play of treatment intervention techniques and problematic situations. Direct supervision included biweekly face-to-face meetings between the case management supervisor and ICM case managers, weekly case manager team meetings, and telephone consultation as needed. In addition, ICM case managers received treatment manuals that described the theory underlying the ICM model and detailed the specific functions and intervention techniques.

Documentation of ICM activities

Case managers documented all activities on a computerized case management information system (CMIS). Details regarding activity descriptions included the time of the activity, type of activity (i.e., referral, counseling, transporting, paperwork, etc.), mode of activity (phone call, meeting, etc.), activity duration, and persons involved in the activity. Most case management sessions were audio-recorded for supervision and fidelity evaluation.

Engagement in ICM

Engagement in case management was an indicator of whether the client ever substantially participated in case management. A portion of clients who agreed to participate in the ICMP rarely or never met with their assigned case manager. For clients assigned to receive ICM, case management engagement means that the client completed at least 15 min of the initial strengths and resources assessment that reflected the initial function of formal case management. Clients assigned to the control group, by definition, did not engage in ICM. This criterion might seem too small to be meaningful, but this level was established through examination of our case management data and through a survey of case managers that asked them to estimate the minimum time to establish a therapeutic alliance with the clients.

To evaluate the impact of engagement, the rates of case management engagement were compared across condition and among levels of treatment (Table 2). Engagement in case management did not differ significantly by case management condition, although it did differ by level of treatment. Engagement rates were lower for outpatient clients than for residential clients (63% vs. 81%). Overall, case managers found that outpatient clients thought that case management sounded helpful in the beginning of treatment, but as the clients returned to their regular routines, they did not actually need or chose not to pursue case management services.

Table 2.

Engagement in ICM by Case Management Condition

Percentage Engaged in ICM	Outpatient (n = 139)	Residential (n = 364)
ICM-MECCA (Inside)	64%	84%
ICM-LSS (Outside)	58%	79%
ICM-Telecommunication	68%	79%
ICM-Overall	63%	81%
χ 2	0.93 (p = .63)	0.86 (p = .65)

Note: ICM: Iowa Case Management; MECCA: Mid-Eastern Council on Chemical Abuse; LSS: Logical Structure of Sentence.

Research Design

Clients who were admitted to intensive outpatient or residential substance abuse treatment at a local community-based drug treatment agency were invited to participate in this study. Participants were randomly assigned to one of four treatment conditions, which included three variations of ICM and a standard drug-treatment-only group. These conditions were designed to evaluate the impact of case management on outcomes as well as to determine whether the proximity of the case manager to the substance abuse treatment facility or use of a telecommunication system enhanced case management effectiveness (Hall et al., 1999). All participants in the study received drug treatment services as usual from MECCA, which meant that participants in the fourth condition received only drug treatment (i.e., the control condition). Clients assigned to Iowa Case Managers were eligible for services for up to 12 months, although most activities occurred within the first 3 months after intake.

At the end of the CIA, eligible clients were invited to participate in our research project. Clients were deemed eligible if they were admitted to MECCA’s intensive outpatient or residential treatment programs and lived in a county within 60 miles of MECCA. After clients signed IRB-approved consent documents, they were randomly assigned to a condition and then completed additional instruments within a few days of recruitment. Participants were followed over a 12-month period and assessed at three follow-up points: 3, 6, and 12 months following intake using the follow-up version of the ASI. Clients were paid $20 for each completed follow-up research assessment.

Research assistants who conducted the CIA at intake attempted to recruit participants into our study immediately after the CIA and subsequently conducted the follow-up sessions at 3, 6, and 12 months. Case managers were university employees and provided case management services, documented specific details of service provision and kept in contact with local agencies. Research assistants and case managers were aware of each other but only shared client location information at specific times under supervision of the project director and principal investigator.

Case management conditions

The four case management conditions were (a) inside case management, (b) outside case management, (c) telecommunications case management, and (d) standard services only (control group). These conditions were developed to address policy and structure issues in the drug treatment system. Directors of drug treatment agencies wanted case managers to be housed in the agency under their control. The state funding agency wanted case managers to be housed in an outside agency to provide more objectivity in decision making for services. Federal officials wanted to know if a telecommunication approach would be cost-effective.

In the first condition, ICM was delivered by two case managers who were housed in the drug treatment agency and who had an average active caseload of 40 clients each. In the second condition, ICM was delivered by two case managers who were located at a social service agency that was not directly connected with the drug treatment agency and who also had an active caseload of 40 clients. In the third condition, ICM was delivered by one telecommunications case manager who was located at the ICMP research office, who used the Iowa Telecommunication System (ITS) to facilitate communications with clients (Hall & Huber, 2000) and who had an active caseload of 80 clients, which was double the caseload of the other case managers. To improve client engagement with case management, the telecommunication case manager completed the Strengths and Resources Assessment in face-to-face sessions with clients. Following these initial two to three sessions, the vast majority of services were provided through the ITS. Although the ideology, function, and technique of the telecommunications case manager was the same as for the other ICM conditions, greater efficiency was anticipated in this condition because of less wasted time on missed communications and travel. Finally, clients in the control condition received standard treatment services from the drug abuse treatment agency, which included standard individual counseling and group treatment from primary drug counselors.

Assessment Procedures

The main instrument used for the Clinical Intake Assessment was the Addiction Severity Index (ASI; McLellan et al., 1992). The ASI is a widely used, reliable, valid, and comprehensive interview that collects information on seven domains: general demographic information, medical status, employment and education, family and social background, psychiatric status, history of drug and alcohol use, and legal status. The instrument inquires about the client’s status during the 30 days preceding treatment as well as lifetime characteristics. Although the ASI was the primary data collection instrument, other instruments were administered to collect data on each of the conceptual domains in the ASI. However, this study focuses only on ASI composite scores.

Data Analysis Procedures

Data analysis focused on evaluating (a) differences in client outcomes across the four case management conditions (controlling for the client’s status at treatment intake), (b) changes in client outcomes over time by condition and differences between conditions in the patterns of change over time, (c) the relationship between client characteristics (e.g., rural status) at baseline (treatment intake) and the effectiveness of case management, and (d) the impact of engagement in case management on outcomes. The extent to which case manager location and use of a telecommunication system affected clients’ outcomes was determined by differences across the conditions. The effect of ICM on outcomes and differences across case management conditions were investigated separately for clients admitted to the outpatient and residential programs. To be considered significant, the results had to have a probability of less than .05.

Client characteristics were described using data self-reported by clients during the ASI interview at intake and subsequent assessments. Selected variables reflect client characteristics in the 30 days prior to the interview. These data were obtained during the interview by asking the client to report the number of days that he or she experienced a specific outcome indicator. For example, alcohol use was assessed by asking the client, “How many days of the past 30 did you use alcohol?” Variables used as outcomes are days of the last 30 (a) using alcohol, (b) using drugs, (c) experiencing medical problems, (d) experiencing psychological problems, (e) experiencing problems with family relationships, (f) experiencing problems with social relationships, (g) paid for working, and (h) engaged in illegal activity. Although ASI composite scores were created from several variables, both composite scores and specific variables were included in our analyses.

The variables examined for impact on client outcomes were client characteristics at baseline, the admission environment, months since recruitment, case management condition, rural status, and whether the client engaged in case management. Then the same measures were repeated over time and became the dependent variables in analyses. For example, one outcome is the number of days of the last 30 that the client has used drugs, measured at each follow-up point. The baseline measurement for that model is the number of days of the 30 days prior to treatment that the client used drugs, assessed at intake. Analyses controlled for baseline measures over time.

In many cases the scheduled interview did not occur in the exact month in which it was planned because of scheduling difficulties with clients. Therefore, the number of months since recruitment was calculated for each interview and months since recruitment was included in the model as a continuous variable to indicate time.

The case management condition to which the client was assigned was identified in each model for the first, second, and third conditions (inside, outside, and telecommunications). The fourth condition was the control group and was the default in each analysis model. The client’s admission environment (i.e., outpatient or residential) was also included in each model. The interaction of the case management condition with the client’s admission environment was evaluated, and separate coefficients are reported for clients admitted to outpatient and residential programs.

Other data were analyzed, including (a) investigating the relationship between the total minutes of case management and outcomes, (b) evaluating case management effectiveness only for those clients who engaged in case management, and (c) evaluating case management only for those clients who reported problem days in their initial evaluation. Finally, similar analytic models were developed in which the outcomes were the ASI composite score from each domain rather than the number of days of reported problems. ASI composite scores and days of reported problems are generally highly correlated (r > .9; McLellan et al., 1992).

Data analysis procedures for repeated measures were employed (Laird & Ware, 1982; Liang & Zeger, 1986) using the MIXED procedure of SAS (The SAS Institute, 2003) to allow for the inclusion of time-specific and client-specific variables in the repeated measures analyses. A key component of this type of repeated measures analysis is determining the appropriate structure for modeling the variance–covariance structure of the repeated measures. For each analysis, the use of an unstructured covariance model was first considered, followed by tests of several other structures. A first-order auto-regressive structure provided a reasonable fit to the data and was preferred for all models, particularly in that this structure requires estimation of fewer parameters than did the other structures considered and therefore is more parsimonious.

Results

Evaluation of Client Outcomes

Results from the repeated measures statistical models are described in the narrative below. To facilitate interpretation of our results, we present means and standard deviations in Table 3, but model coefficients can be obtained from the authors. To evaluate changes over time, analysis coefficients were examined to determine whether the coefficient differed significantly from zero. The coefficient for each ICM condition within treatment subgroups (i.e., outpatient or residential) reflected the difference between that condition and the control group for residential or outpatient clients. The coefficient for engagement reflected whether engagement in ICM is related to outcomes. The interactions of case management condition with months since recruitment and client baseline status were not statistically significant and thus are not reported. Finally, the overall impact of ICM, regardless of location of the case manager, was tested using a χ² statistic to contrast the coefficients for the three ICM conditions combined and compared to the control group.

Table 3.

Comparison of Key Outcome Variables Over Time and Between Conditions: Combined Totals, Outpatients, Residential—Days of the Last 30

	ICM-MECCA				ICM-LSS				ICM-Telecommunication				Control
	Baseline	3 Months	6 Months	12 Months	Baseline	3 Months	6 Months	12 Months	Baseline	3 Months	6 Months	12 Months	Baseline	3 Months	6 Months	12 Months
1. Total
Using Drugs	13.53 (12.75)	1.88 (5.78)	1.59 (3.85)	2.8 (6.71)	13.93 (12.79)	2.73 (7.02)	3.56 (7.83)	2.78 (6.98)	16.83 (13.13)	2.72 (7.11)	3.71 (8.24)	4.47 (9.1)	13.85 (12.71)	3.94 (8.42)	3.14 (7.81)	4.79 (9.67)
Using   Alcohol	13.77 (11.48)	3.19 (7.25)	4.44 (8.41)	5.23 (8.4)	13.19 (12.05)	2.89 (6.43)	3.52 (7.32)	2.74 (6.24)	11.85 (11.56)	4.47 (8.18)	4.41 (8.17)	4.85 (7.68)	13.82 (11.98)	3.9 (7.26)	4.05 (6.58)	6.14 (10.07)
Bothered by   Psychological   Problems	13.16 (12.53)	8.58 (10.25)	9.99 (11.89)	8.64 (10.68)	14.07 (12.19)	10.21 (11.63)	8.16 (10.59)	9.74 (11.37)	14.83 (13.03)	9.15 (12.06)	9.39 (11.95)	7.68 (9.3)	13.57 (12.68)	9.84 (11.72)	10.15 (11.49)	9.87 (11.2)
Bothered by   Medical   Problems	1.34 (1.51)	0.99 (1.33)	1.30 (1.48)	1.40 (1.58)	1.33 (1.48)	1.13 (1.24)	0.88 (1.17)	1.04 (1.38)	1.06 (1.39)	0.78 (1.21)	1.18 (1.43)	1.27 (1.38)	1.22 (1.53)	1.01 (1.4)	0.99 (1.28)	1.15 (1.39)
Bothered by   Family   Relationship   Problems	3.66 (7.76)	1.28 (3.92)	1.99 (5.61)	1.53 (5.5)	3.74 (8.40)	0.93 (2.57)	1.44 (4.63)	1.95 (5.65)	3.8 (7.53)	1.11 (3.53)	1.05 (2.83)	0.41 (1.00)	3.08 (7.48)	1.64 (5.71)	1.3 (4.65)	2.05 (6.80)
Bothered by   Social   Relationship   Problems	4.31 (8.57)	1.15 (2.84)	2.74 (7.03)	2.56 (6.67)	3.68 (7.5)	1.76 (4.41)	1.12 (2.57)	2.49 (6.94)	3.02 (6.36)	1.25 (4.48)	2.42 (6.71)	2.58 (7.36)	3.57 (7.59)	0.97 (3.47)	0.87 (3.77)	1.97 (6.19)
Paid for   Working	5.86 (9.01)	12.3 (10.95)	12.11 (9.81)	13.69 (10.03)	6.56 (9.21)	9.45 (10.1)	12.14 (10.62)	13.24 (10.64)	8.8 (10.11)	11.83 (10.16)	14.94 (9.65)	12.69 (11.07)	6.71 (9.89)	10.63 (9.98)	11.44 (10.48)	11.56 (10.29)
Engaged in   Illegal   Activity	1.83 (5.44)	0.15 (0.88)	0.54 (3.41)	0.36 (1.48)	1.56 (5.69)	0.21 (1.33)	0.05 (0.22)	0.58 (3.24)	1.54 (5.62)	0.73 (4.03)	1.38 (5.68)	1.03 (3.69)	2.98 (7.38)	0.01 (0.12)	0.38 (3.31)	0.06 (0.41)
2. Outpatient
Using Drugs	14.91 (13.15)	1.67 (5.77)	1.50 (5.33)	1.80 (3.36)	13.13 (14.17)	4.31 (9.2)	2.79 (7.45)	2.44 (7.31)	11.00 (14.08)	1.00 (3.16)	2.33 (4.95)	5.00 (9.11)	7.38 (11.21)	1.64 (3.27)	3.08 (8.39)	2.62 (6.85)
Using Alcohol	11.06 (9.1)	1.00 (1.86)	4.57 (8.34)	2.73 (4.43)	11.54 (12.5)	4.00 (8.97)	1.60 (2.95)	2.74 (4.85)	6.29 (8.18)	3.30 (7.75)	1.44 (2.4)	5.00 (6.66)	9.13 (10.54)	3.86 (6.15)	2.85 (3.29)	6.54 (10.63)
Bothered by   Psychological   Problems	10.06 (12.47)	5.82 (8.93)	5.58 (7.61)	3.82 (5.06)	10.09 (12.78)	7.46 (12.91)	5.22 (9.05)	6.63 (10.28)	11.36 (12.33)	4.67 (7.57)	4.78 (7.46)	4.00 (5.54)	7.88 (10.72)	7.69 (10.36)	8.58 (10.07)	5.08 (8.53)
Bothered by   Medical   Problems	0.56 (1.09)	0.33 (0.89)	0.43 (0.94)	0.36 (0.67)	1.25 (1.54)	1.31 (1.6)	0.70 (0.92)	1.21 (1.23)	0.86 (1.1)	0.50 (0.85)	1.10 (1.29)	1.27 (1.42)	0.50 (1.21)	0.57 (1.16)	1.08 (1.31)	1.00 (1.29)
Bothered by   Family   Relationship   Problems	2.44 (6.52)	0 (0) (1.34)	0.36 (0.9)	0.27 (1.02)	0.42 (3.33)	0.92 (1.61)	0.45 (7.18)	2.53 (7.88)	3.93 (0.63)	0.2 (0.67)	0.22 (1.51)	0.58 (7.46)	2.5 (7.99)	2.29 (4.16)	1.15	0 (0)
Bothered by   Social   Relationship   Problems	1.94 (5.28)	1.25 (4.03)	0.79 (2.67)	0 (0)	4.25 (9.16)	2.46 (8.29)	0.8 (1.47)	3.11 (7.36)	2.71 (4.98)	1.50 (3.72)	0.56 (1.33)	0.25 (0.62)	2.31 (5.00)	0.21 (0.8)	0.46 (1.13)	1.77 (4.42)
Paid for   Working	12.13 (10.05)	18 (8.93)	14.64 (9.64)	19.18 (8.26)	11.58 (10.75)	15 (8.47)	15.79 (10.63)	13.33 (10.26)	9.64 (9.77)	13.4 (10.17)	13.67 (7.48)	13.75 (11.39)	11 (12.87)	13.29 (9.51)	19.46 (9.3)	14.15 (10.81)
Engaged in   Illegal Activitya	0.44 (1.75)	0 (0)	0.07 (0.27)	0 (0)	1.3 (6.26)	0 (0)	0.05 (0.22)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
3. Residential
Using Drugs	13.37 (12.77)	1.93 (5.83)	1.61 (3.52)	2.97 (7.12)	14.15 (12.46)	2.37 (6.47)	3.79 (7.99)	2.89 (6.94)	17.77 (12.82)	3.09 (7.67)	3.94 (8.68)	4.35 (9.18)	15.06 (12.66)	4.53 (9.22)	3.15 (7.76)	5.24 (10.14)
Using Alcohol	14.23 (11.82)	3.67 (7.89)	4.41 (8.48)	5.68 (8.87)	13.64 (11.96)	2.64 (5.78)	4.13 (8.15)	2.74 (6.65)	12.81 (11.82)	4.7 (8.32)	4.91 (8.69)	4.82 (7.94)	14.66 (12.08)	3.91 (7.56)	4.28 (7.02)	6.06 (10.04)
Bothered by   Psychological   Problems	13.69 (12.53)	9.18 (10.49)	10.79 (12.39)	9.47 (11.19)	15.14 (11.88)	10.83 (11.36)	9.00 (10.91)	10.7 (11.59)	15.43 (13.13)	9.92 (12.57)	10.15 (12.42)	8.46 (9.78)	14.60 (12.78)	10.35 (12.06)	10.43 (11.77)	10.87 (11.48)
Bothered by   Medical   Problemsb	1.47 (1.54)	1.13 (1.38)	1.47 (1.51)	1.58 (1.62)	1.35 (1.48)	1.10 (1.17)	0.94 (1.24)	0.98 (1.43)	1.10 (1.44)	0.83 (1.27)	1.20 (1.47)	1.27 (1.38)	1.36 (1.56)	1.13 (1.44)	0.97 (1.28)	1.18 (1.41)
Bothered by   Family   Relationship   Problem	3.87 (7.96)	1.56 (4.28)	2.31 (6.08)	1.75 (5.92)	4.66 (9.28)	0.93 (2.41)	1.74 (5.18)	1.77 (5.14)	3.78 (7.52)	1.28 (3.82)	1.18 (3.02)	0.37 (0.88)	3.18 (7.52)	1.48 (5.06)	1.33 (4.76)	2.46 (7.39)
Bothered by   Social   Relationship   Problemsc	4.70 (8.95)	1.13 (2.56)	3.13 (7.56)	3.00 (7.14)	3.52 (7.02)	1.61 (3.13)	1.21 (2.83)	2.30 (6.85)	3.07 (6.59)	1.21 (4.63)	2.71 (7.17)	3.07 (8.02)	3.79 (7.96)	1.16 (3.85)	0.94 (4.08)	2.02 (6.52)
Paid for Working	4.81 (8.44)	11.05 (11.02)	11.59 (9.83)	12.75 (10.06)	5.16 (8.27)	8.36 (10.09)	11.09 (10.46)	13.22 (10.84)	8.66 (10.22)	11.53 (10.23)	15.14 (10)	12.46 (11.09)	5.94 (9.15)	9.96 (10.06)	9.91 (10.04)	11.05 (10.2)
Engaged in   Illegal Activityd	2.06 (5.81)	0.19 (0.97)	0.63 (3.75)	0.42 (1.6)	1.63 (5.57)	0.26 (1.47)	0.05 (0.21)	0.75 (3.66)	1.8 (6.04)	0.87 (4.4)	1.6 (6.1)	1.27 (4.07)	3.52 (7.91)	0.02 (0.14)	0.45 (3.61)	0.08 (0.44)

Note: ICM: Iowa Case Management; MECCA: Mid-Eastern Council on Chemical Abuse; LSS: Logical Structure of Sentence.

^aKey: Over time, clients receiving ICM from the MECCA case manager reported significantly more days of illegal activities (in the last 30 days).

^bKey: Over time, clients receiving ICM in any condition reported significantly more medical problems than those in the control.

^cKey: Over time, clients receiving ICM through the telecommunication system reported significantly more problems than those in the control group.

^dKey: Over time, clients receiving ICM through the telecommunication system reported significantly more days of illegal activities than those in control group.

In Table 4, the numbers of follow-up interviews completed by months since recruitment are displayed. Overall, 364 of the 539 residential clients (68%) completed at least one follow-up assessment, and 139 of the 207 outpatient clients (67%) completed at least one assessment. Overall, retention did not differ across the four treatment groups. In Table 5, the mean numbers of follow-up sessions that were completed are compared across conditions. In Table 6, the percentages of follow-up sessions completed are compared across condition and over time. More detailed information about participation and retention with a preliminary sample is described in a previous article (Vaughn et al., 2002).

Table 4.

Retention of Participants at Follow-Up Sessions Completed by Months Since Recruitment

Months Since Recruitment	Outpatient (n = 207)	Residential (n = 539)
2-4 Months	99 (48%)	237 (44%)
5-9 Months	102 (49%)	253 (47%)
10-14 Months	111 (53%)	249 (46%)
No. of clients  with at least  one follow-up  assessment	139 (66%)	364 (68%)

Table 5.

Retention of Participants at Follow-Up Sessions Completed by Condition^*

	n	M	SD
ICMP-Inside	229	1.34	1.14
ICMP-Outside	232	1.45	1.23
ICMP-Telecommunication	230	1.40	1.13
ICMP-Control	218	1.47	1.21

Note: ICMP: Iowa Case Management Project.

^*p = 0.56 using ANOVA.

Table 6.

Retention of Participants at Follow-Up Sessions Completed by Condition and by Follow-Up Time

	3-Month Follow-Ups			6-Month Follow-Ups			12-Month Follow-Ups
	n		%	n		%	n		%
ICMP-Inside	97		42	115		50	95		42
ICMP-Outside	114		49	120		52	102		44
ICMP-Telecommunication	106		46	112		49	103		45
ICMP-Control	105		48	115		53	100		46
Chi-square		2.49			0.85			0.96
p		0.48			0.84			0.96

Note: ICMP: Iowa Case Management Project.

Results for All Participants: Changes Over Time

Baseline variables

As expected, baseline variables were significant predictors of outcomes for all eight outcomes (p < .001). In all cases the coefficient was positive, indicating that previous behavior predicted behavior at follow-up time points.

Months since baseline

Overall, for all participants, months since baseline was significantly related to client status at follow-up for days of medical problems (p < .01) and days paid for working (p < .01). The beta coefficients were positive and significant for both outcomes, indicating that participants reported more medical problems and days of paid work over time. Furthermore, each additional passing month since recruitment was associated with 0.21 additional days worked.

The month-by-month coefficients for the other six outcomes were not significant. Initially, the use of alcohol decreased significantly from an average 12.5 days (over the last 30) at intake to 3.2 days at the 3-month follow-up, and the use of drugs decreased significantly from an average 13.3 days at intake to 3.0 days at the 3-month follow-up. Although these initial decreases were significant, a client’s drug and alcohol use tended to remain stable from 3 months through the end of the follow-up period in this study.

Admission status

Clients assigned to the outpatient and residential programs were compared at baseline. In contrast to the outpatient clients, residential clients tended to come from large and small towns, and reported more days bothered by medical problems, psychological problems, or family relationships. Although outpatient clients reported significantly more days of paid work in the last 30 days than the residential clients, the residential clients reported significantly more substance use (six of seven categories) than outpatient clients. See Table 1 for more details.

Case Management Condition and Key Outcome Variables

Engagement in case management

The relationship between case management engagement and outcomes for residential clients were not significant. For outpatient clients, engagement in case management was significantly related to a decrease in days bothered by social problems (β = −1.70, p = .03). Engagement in case management was not significantly associated with any other outcomes for outpatient clients.

Outpatient ICM conditions vs. control

In general, for outpatient clients, differences between ICM conditions and the control group were not significant. The exception was that clients in the MECCA case management condition reported significantly more days of illegal activity (β = 1.43, p < .05) over time.

Residential ICM conditions vs. control

Upon inspection, the coefficients for residential clients in all three case management conditions were positive, indicating that they reported more days of medical problems over time than those in the control group.

Although the coefficients for all three residential ICM conditions and social problems (client reported significant problems with individuals in his or her social network by roles) were positive, the coefficient was significant only for clients who received ICM through the telecommunication system (β = 1.76, p = .02). Additionally, clients who received ICM through the telecommunication system reported significantly more days of illegal activity than those in the control group (β = 0.99, p < .05).

Combined ICM

Data from all follow-up points were considered together to evaluate the amount of time (in days) between intake and each follow-up point. When data for each ICM conditions were combined for clients in outpatient treatment, case management was not significantly related to any outcome variables. For residential clients, results for the three combined case management conditions were significantly related to outcomes in two areas: medical and social relationship problems. Overall, residential clients who received ICM reported significantly more days of medical problems at follow-up than did clients in the control group (χ² = 5.53, p = .02) and more days of problems with social relationships (χ² = 4.08, p = .04). The coefficients for all three case management conditions were positive, suggesting more days of medical and social problems reported at follow-ups by residential clients. The evidence did not indicate that ICM had a direct, statistically significant impact on the quality of the client’s family relationships, psychological status, employment status, illegal activity, or drug and alcohol use for either residential or outpatient clients.

Discussion and Applications to Practice

The overall purpose of the Iowa Case Management Project was to evaluate the impact of comprehensive case management with rural clients in drug abuse treatment. Based on these final results, the ICM model does not appear to improve outcomes consistently for clients admitted to residential or intensive outpatient treatment. Thus, our first hypothesis was not supported (i.e., ICM would lead to better outcomes). Second, we hypothesized that participants receiving case management would report higher levels of service utilization and that would contribute to better outcomes overall—but this hypothesis was also not supported. Finally, we hypothesized that participants who were from rural areas and who received case management would have better outcomes than urban participants because of outreach activities—but this hypothesis was not supported. Although differences were found at baseline between clients from small towns, large towns, and urban centers, these differences did not affect outcomes over time nor were there significant differences in outcomes between these three groups.

The findings indicated that residential clients who received ICM actually reported significantly more days of medical problems and social relationship problems than did clients who did not receive ICM. These results are contrary to what was anticipated. However, clients may have become more aware of their status in each life domain through the case management process. This increased awareness may create a sort of Hawthorne effect and result in greater reported days of problems. Furthermore, clients in drug abuse treatment typically have not utilized health and social services while actively using substances and may have used drugs to cope with their problems—albeit inappropriately.

Clients also reported an increase in problems in several areas. Residential clients reported an increase in the number of days they were bothered by medical problems. In fact, the numbers of days bothered by medical problems increased by 0.19 every month following recruitment, which may be attributed to increased awareness of and less tolerance for enduring medical problems as the clients remain sober and can therefore focus on other pressing needs. Additionally, residential clients reported an increase in the number of days they were bothered by social relationship problems. The finding that ICM is associated with increases in reported social problem days and medical problems at first appears puzzling. However, by addressing these issues with a case manager, clients could have become more aware of their status and circumstances. As the craving for substances is reduced, other life issues become more important and are thus reported to our research team. The result is more reported problem days for clients who received ICM who may have the same number of problems but not be actively addressing problems in these areas as compared to clients in the control group.

Another unexpected outcome was that outpatient and residential clients reported an increase in the number of days they engaged in illegal activity. Although an explanation for these results is difficult, clients with case managers may, over time and involvement, learn to trust ICMP staff more and, therefore, provide more honest reports of illegal activity than do clients in the control group. A further consideration is that clients whose admission was prompted by the legal system tended not to participate in the ICMP (Vaughan-Sarrazin, Hall, & Rick, 2000). These research data, therefore, may not represent the true impact of ICM on the general MECCA population, especially as related to clients in substance abuse treatment because of prompting by legal difficulties.

When the results from all conditions of ICM were viewed together, ICM did not appear to have a direct, statistically significant impact on the quality of the client’s family relationships, psychological status, employment status, illegal activity, or drug and alcohol use for either residential or outpatient clients. Possibly, ASI scores were not sufficiently sensitive to adequately reflect changes in these areas or the actual activities of the ICM case managers were not focused on ASI domain topics. These findings demonstrate that ICM did not improve outcomes compared with the control group, but data do demonstrate that all participants significantly decreased their drug use and problems due to drug use—which we interpret to mean that the drug treatment program, and not the case management services, actually had the main effect.

Other explanations may address why case management did not have the intended impact with these substance abuse clients. Based on the results that were observed, substance abuse clients may carry severe latent problems in the medical and social life domains that emerge in importance after cravings are reduced and when clients are in a controlled environment. These latent problems require additional services and extended surveillance through case management services to be effectively addressed. Thus, participants in this study may not have resolved or addressed key problems in their lives by the conclusion of case management services.

Second, our case management program might have been focused on too many issues across the broad range of life domains. Rather than focus merely on drug use, our case managers customized and focused on client strengths as determined by their clients. For example, one client could be working on employment issues, while another client would work on family problems. Then, a third client could be working a little in four or five life domains. The ICM model had 10 goals upon which the client and case manager could focus. This variability in case management focus and dosage will be discussed later as a limitation for our overall study and has been addressed elsewhere (Huber, Hall, & Vaughn, 2001; Huber, Sarrazin, Vaughn, & Hall, 2003).

Limitations

Our results must be interpreted further by considering important limitations that pertain to (a) missing data and sample attrition, (b) case manager characteristics, and (c) case management dosage. First, some clients are not reflected in the outcome data either because they did not recruit for the study, or they recruited and subsequently dropped out, or were otherwise nonresponsive. The issue of missing data is not terribly disturbing if it can be assumed that the data are missing at random. The data were examined for evidence of bias and, overall, no indication was found that the follow-up of active clients was biased according to client baseline severity and demographics. The exceptions were that female clients tended to complete follow-up interviews more often than did male clients, and clients whose admission was prompted by the legal system tended not to complete follow-ups. The low participation by persons with legal involvement is particularly relevant, given the lack of results in that area.

A second limit to the generalizability of these data pertains to the specific style and characteristics of individual case managers. With only one or two case managers in each condition, the impact of the personal style and motivation of the case manager on outcomes cannot be ignored. Requiring all case managers to use a manual to direct their activities minimized the potential for personal style to affect results. Furthermore, all case managers, regardless of where they were housed, had the same supervisor with whom they regularly met.

Iowa Case Management was designed to supplement the primary drug treatment being provided to clients in both residential and intensive outpatient programs. Thus, the ratio of drug treatment to case management could have been an important factor—especially for those assigned to an Iowa Case Manager. Theoretically, some clients who were assigned to an Iowa Case Manager could have (a) dropped out of treatment and never met with their case manager, (b) stayed in drug treatment but never met with their case manager, (c) dropped out of drug treatment and met with their case manager, and (d) stayed in drug treatment and met with their case manager. If the case management program were truly supplemental, we would expect minimal clients in the first three options and the vast majority in the fourth option.

Also, all five of the ICM case managers were employees of our research project and were supervised by a clinician who was independent from the treatment agency. Even the case managers located at MECCA maintained some clinical independence from MECCA, although they did attend regular MECCA staff meetings. This arrangement made it possible for the ICM case managers to be rigorous in their application of the strengths-based approach prescribed by the ICM model, which is somewhat different from the traditional language of substance abuse treatment. Thus, it is not known if the ICM type of case management would be feasible if the case managers were employees of the substance abuse treatment agency and supervised directly by treatment agency personnel.

Additionally, depending on the magnitude of the impact of case management, the ASI measures may not be sensitive enough to identify positive changes in client functioning. For example, it is possible that the finding that case management did not affect drug and alcohol use is related to the sensitivity of our measure of drug and alcohol use and its ability to reflect improvements due to case management. Since many ways can be used to measure use of drugs and alcohol (i.e., frequency, quantity, duration), a more sensitive measure may be needed for longitudinal tracking of outcomes.

Finally, it may be that the variability of case management activities across ICM clients meant that we were comparing different subinterventions within each ICM condition. It is also possible that 12 months is too short a time to follow these clients as outcomes may manifest at a further time trajectory. Additional research could compare the goals and activities of the case manager and clients to determine if certain goals and/or activities were related to better outcomes than others.

In previous work (Hall, Carswell, Walsh, Huber, & Jampoler, 2002), Iowa Case Management was presented as a variation of social casework and thus a possible evidence-based model that could be used by practitioners in substance abuse settings and possibly with other populations. Based on the results reported here, statistical support for the effectiveness of our model with clients in drug abuse settings was not found. However, participants in all four conditions significantly decreased substance use by the 3-month follow-up session and for the most part these gains were maintained to the 12-month session. Anecdotal data about the positive impact of ICM was given by staff persons from the drug treatment agency, staff persons from other local agencies, and many of our clients, but these anecdotes do not translate into evidence. For social work practice, additional research should evaluate modifications of ICM and other social casework models as the theoretical support seems strong that clients in drug treatment need more than just drug treatment. This assumption has been supported with evidence from our project and others, but the evidence for the efficacy or effectiveness of these models has not been reported. Feedback from drug counselors and clients in drug treatment might provide clinical wisdom and other interpretations of our results that can be integrated with empirical results to improve the ICM model leading to a new evaluation with improved odds of success.