26
CHAPTER OUTLINE
This chapter examines research evidence on the effects of substance use disorder (SUD) treatment settings, duration, and amount, drawing heavily on research syntheses. Most of the research focuses on treatment for alcohol use disorders, but research on treatment of drug use disorders, other than nicotine dependence, also is considered. We argue that, to some extent, research findings (e.g., those on the relative effects of inpatient/residential vs. outpatient treatment) have been extrapolated to populations beyond those involved in the studies (e.g., those with disorders complicated by serious psychiatric conditions, homelessness) and that key issues remain unaddressed. These issues include determining the specific types of persons who benefit more from initial treatment in inpatient/residential than in outpatient settings, whether certain types of individuals benefit from longer or more intensive treatment, and whether treatment should be spread out over longer periods for some patients.
Although inpatient treatment is more prevalent in some other countries (e.g., Germany), only about 10% of SUD patients in the United States receive residential treatment, and only 1% receive inpatient treatment where presumably medical or psychiatric care also is readily available (1). Considerable research has focused on whether inpatient/residential or outpatient treatment is more effective overall, but the more pressing issue is whether certain types of patients benefit more from an initial phase of inpatient/residential treatment before continuing outpatient care than from outpatient treatment alone.
Rationales for Inpatient/Residential and Outpatient Treatment
At least five rationales have been put forward for the superiority of an initial phase of treatment in inpatient/ residential SUD treatment settings (2). One is that such settings provide a respite for patients, removing them from unstructured and unsupportive environments that perpetuate their addiction, thereby allowing their efforts toward abstinence to be consolidated. Second, inpatient/residential settings may allow patients to receive more treatment because treatment is more intensive and patients may be less likely to drop out of treatment (3). A third rationale is that inpatient/ residential settings provide medical/psychiatric care (inpatient settings) and other comprehensive services to patients who otherwise would not have access to such care or support, services seen as crucial to achieving optimal substance use outcomes (4). Fourth, inpatient/residential treatment prepares a patient better to engage in continuing outpatient treatment (e.g., by stressing the need for continuing care) (5). Finally, inpatient/residential treatment may suggest to patients that their problems are more severe and that resolving them is more paramount than would be the case if treatment were offered in an outpatient setting (6).
Arguments in favor of outpatient treatment also focus on the patient’s usual life situation, but stress the advantages of leaving the patient in, rather than removing him or her from, that context (7,8). Proponents have suggested that outpatient treatment provides an opportunity for more accurate assessments of the antecedents of substance use and for testing coping skills in real-life situations while the patient remains in a supportive therapeutic relationship. Accordingly, greater generalization of learning should take place than would be the case in the atypical environment of an inpatient/residential treatment program (9). In addition, outpatient treatment may mobilize help in the patient’s natural environment (e.g., from a family physician or self-help groups), to a greater extent than does inpatient or residential treatment. Finally, it has been argued that outpatient treatment results in a more successful transition to continuing care when, for example, a patient begins to attend self-help group meetings near his or her home while still in treatment.
Relative Effectiveness of Inpatient and Outpatient Settings
Several early research reviews examined the relative effectiveness of alcohol treatment in inpatient and outpatient settings and concluded inpatient/residential treatment was no more effective than outpatient treatment (9–11). A later review (2) found that 7 of 14 relevant studies had significant setting effects on one or more drinking-related outcome variables at one or more follow-up points. In five studies, the outcome difference favored inpatient/residential (sometimes followed by continuing outpatient treatment) over outpatient treatment; in the other two, the outcome difference favored day hospital over inpatient treatment. Patients in the “superior” setting usually received more treatment.
This “box-score” approach to synthesizing the research literature has serious limitations. Nonsignificant differences between treatment groups may simply reflect lack of statistical power; significant findings may emerge by chance when multiple tests for treatment effects are conducted and not adjusted for “experimentwise” error. Indeed, Finney et al. (2) found that the seven studies yielding significant setting effects had greater statistical power and conducted more treatment contrasts, on average, than the studies with no difference in outcome. The shortcomings of box-score reviews prompted the development of meta-analytic techniques (12) that use an “effect size” to gauge treatment efficacy. A common effect size in this context is the difference in the average posttreatment functioning of two groups divided by the pooled standard deviation of outcome scores for the two groups. A between-group, standardized effect size allows one to determine by how many standard deviation units or by what proportion of a standard deviation unit the functioning of one group is superior to that of another. When Finney and Moos (13) calculated average, cross-study effect sizes on drinking-related outcome variables, only the effect size of 0.22 at 3-month follow-ups was significant and favored inpatient/residential over outpatient treatment (the effect sizes at 6- and 12- to 14-month followups were not significant).
Although some of the more recent studies of mixed SUD treatment have found small or scattered effects that favored inpatient or residential treatment on a few of many outcome variables (5,14–20), others focusing on treatment for alcohol use disorders (21–23) and cocaine abuse (24) have not. Likewise, reviews of research on outpatient methadone maintenance by Anglin and Hser (25) and outpatient drug-free treatment by Crits-Christoph and Siqueland (26) reported few differences in outcomes in comparison with residential therapeutic community programs.
Extracting appropriate policy implications from this research literature requires consideration of the representativeness and types of patients who have been included in the studies. In some studies, relatively low percentages of patients in treatment have actually participated in the research. For example, 6 of the 14 studies reviewed by Finney et al. (2) noted the percentage of patients in treatment that participated in the research; in 4 of those studies, the percentage was 25% or lower. Most studies, especially randomized trials, have examined a restricted set of patients. Ethical concerns have prevented random assignment to outpatient treatment of highly impaired patients who on clinical grounds were candidates for inpatient/ residential treatment. Accordingly, studies often have excluded patients with major medical or psychiatric disorders or insufficient resources, such as an inability to commute to treatment, homelessness, or a lack of a telephone. Thus, the findings may not generalize well to more impaired individuals or those with fewer social resources for whom inpatient/residential treatment might provide more benefit. It perhaps is no coincidence that many of the investigations indicating superiority of inpatient/residential treatment have come from naturalistic studies of more impaired patients receiving treatment in publicly funded programs (5,19).
Who Benefits from Inpatient/Residential Treatment?
Even though some degree of patient homogeneity resulting from ethical concerns in randomized trials or from admission criteria for public and private programs operates against its emergence, considerable evidence indicates that more impaired patients benefit more from an initial episode of inpatient or residential treatment than from outpatient treatment alone. In other words, patient impairment has been found to interact with treatment setting in affecting patient outcomes.
A diagnosis of a serious psychiatric disorder often has been an exclusion criterion in studies of inpatient versus outpatient alcohol treatment (2), precluding its broad examination as a matching variable. Nevertheless, Ritson (27) found that patients who had personality disorders tended to have poor outcomes in outpatient treatment, though no relationship was found between personality disorders and outcome among inpatients. Likewise, research by Moos et al. (28) found that, for patients with psychiatric disorders, an episode of inpatient treatment before transfer to a community residential facility was associated with better outcomes than direct placement in a community residential facility. With respect to social resources, Kissin et al. (29) reported that more socially stable alcohol-dependent patients experienced better outcomes in outpatient treatment, whereas socially unstable patients had better outcomes after inpatient treatment. Among both alcohol and drug use disorder patients with middle-level psychiatric severity (defined by scores from one standard deviation above to one standard deviation below the mean Addiction Severity Index [ASI] psychiatric severity rating), McLellan et al. (30) observed that those who had more serious family, legal, or employment problems experienced poorer outcomes after receiving outpatient versus inpatient treatment.
Several studies (5,22,31–35) have found that patients with greater alcohol or drug use severity at treatment intake who receive an initial episode of inpatient or residential treatment experience better outcomes than those receiving only outpatient treatment (cf. (20,36)). For example, De Leon et al. (35) examined persons who were “undertreated” (i.e., treated in outpatient treatment even though their severity indicated long-term residential treatment) versus those who were “appropriately matched” to long-term residential care. For persons who remained in treatment for 90 days or more, those who had been in residential care were more likely than those who had been in outpatient care to report no alcohol use at a 1-year follow-up. Although a significant difference emerged on only this one of the six outcome variables assessed, it did so even though the individuals who were undertreated in outpatient care had less baseline severity than the “matched” individuals treated in residential care.
Matching Patients to Treatment Settings
The evidence summarized here provides general support for matching patients to different treatment settings. However, the strength of the interactions, which statistically indicate how strongly different treatment settings are linked to different outcomes for different types of patients, is usually difficult to determine from research reports. In one study providing this information (34), the two interaction effects found, although significant, were weak. A highly impaired sample from the Department of Veterans Affairs facilities may have constrained the strength of these interactions, because stronger interactions are more likely to emerge with greater variability among the patients studied. Likewise, we know little about the precise levels of severity at which more impaired patients do better in inpatient/residential versus outpatient treatment. In the study by Tiet et al. (34), only patients who had very extreme scores on ASI alcohol and drug composites did better when they had an initial episode of inpatient/residential treatment versus only outpatient treatment.
Overall, the general concept of the American Society of Addiction Medicine’s (ASAM) Patient Placement Criteria (37) is supported by much of the research reviewed here. The criteria attempt to match patients to five levels of care: (a) early intervention, (b) outpatient treatment, (c) intensive outpatient/partial hospitalization treatment, (d) residential/inpatient treatment, and (e) medically managed intensive inpatient treatment. Placement decisions are based on a patient’s standing on six dimensions: (a) acute intoxication and/or withdrawal potential, (b) biomedical conditions and complications, (c) emotional/behavioral/cognitive conditions or complications, (d) readiness to change, (e) relapse, continued use, or continued problem potential, and (f) recovery/living environment. However, research is still needed to validate the specific placement assessments and algorithms used in the ASAM system (38,39), especially because the use of the ASAM criteria is required in at least some programs in over 30 states in the United States (40). The fact that Witbrodt et al. (20) found no difference in abstinence between residential and outpatient treatment among patients who met all but the recovery/living environmental ASAM criterion for residential care gives one pause, but it is only a single study, and not all the criteria for residential treatment were met in its participants. Overall, we still do not have precise, empirically supported guidelines for allocating patients to different levels of care. This state of affairs is unfortunate in light of the concern that “[w] ithout well-validated placement criteria that justify intensive treatments on the basis of their greater effectiveness, the pressures of managed care to reduce costs will continue to threaten addiction treatment quality” (41).
A more fundamental issue not addressed in existing studies is the relative attractiveness of treatment in the two types of settings, that is, their ability to induce certain types of individuals to seek and enter treatment. In randomized trials, patients already have opted for treatment and are usually preselected for their willingness to accept either inpatient/ residential or outpatient treatment. Under normal conditions of treatment delivery, inpatient/residential programs may be more effective than outpatient programs in attracting individuals who have significant barriers to receiving treatment in other settings (e.g., homeless individuals and persons who lack transportation or who live some distance from a treatment facility) (42). The findings of Milby et al. (19) support the beneficial effects of providing homeless individuals a place to stay, especially in abstinence-oriented housing, while receiving SUD treatment. More broadly, if inpatient/residential programs are not available, administrators may inappropriately be able to point to “reduced treatment demand” as evidence to support cutbacks in SUD treatment services.
DURATION AND AMOUNT OF TREATMENT
Although the chronic, relapsing nature of many individuals’ SUDs suggests the need for extended treatment, the tendency in the United States has been toward shorter episodes of treatment, given reduced insurance coverage for SUD care (43). This section reviews evidence on the effectiveness of longer versus shorter stays in inpatient/residential treatment and the effects of participation in continuing outpatient care.
Because other chapters in this text review the evidence on screening and brief interventions, we only note here the need for more research to determine who is as likely to benefit from a brief intervention as from more extensive care. At present, low to moderate alcohol severity patients with positive life contexts and without severe skill deficits appear to be the best candidates for brief interventions (44–46). Also, Ashton (47) pointed to evidence suggesting that brief motivational interventions are best directed toward persons who are ambivalent about changing their substance use behavior. Persons already committed to reducing or eliminating their substance use may be “set back” by consideration of positive aspects of substance use that is a component of some brief interventions using motivational interviewing principles.
Length of Stay in Inpatient/Residential Treatment
Mattick and Jarvis (10) and Miller and Hester (11) provided early reviews of several randomized trials comparing different lengths of inpatient or residential treatment for alcohol abuse. The consistent finding was no difference in outcome. Several more recent randomized trials also have found no or only isolated (i.e., on a few outcomes) beneficial effects for longer inpatient/residential alcohol or substance abuse treatment (15,16,48,49). These findings suggest it is not useful to assign unselected clients to longer stays in residential or inpatient treatment. In contrast, many naturalistic studies of substance abuse treatment have found longer stays in treatment to be associated with better outcomes, even a reduction in premature mortality (50). For example, longer episodes of inpatient and residential care (51–53), extended care (28), community residential care (54–57), and care in therapeutic communities (58) have been associated with better substance use outcomes and psychosocial functioning, as well as lower readmission rates for subsequent inpatient care, including among SUD patients with co-occurring psychiatric disorders (59).
However, there appear to be limits to the positive effects of increasingly greater durations of residential care, even in observational studies. Harris et al. (60) examined the relationships between length of stay (LOS) and outcome for over 1,300 persons in 28 VA residential SUD treatment programs. In a mixed-effect analysis controlling for average program LOS, patient-level LOS was unrelated to substance use outcomes, but was related to greater improvement in the ASI employment composite. However, programs whose average LOS was greater than 90 days showed the least improvement in an average of 7.5 months after admission on the ASI alcohol composite score relative to programs whose average LOS was 15 to 30 or 31 to 45 days. No significant relationships were found on the ASI drug composite, and no significant interactions were found between symptom severity or prior treatment utilization and LOS in predicting patients’ outcomes. Not surprisingly, programs with average LOSs greater than 90 days had higher costs than those with shorter average LOSs. The authors concluded VA residential SUD treatment programs of longer than 90 days could not be justified and raised questions about treatment exceeding 60 days in length.
It may be that beneficial effects of longer stays in inpatient/ residential treatment apply only to more impaired patients with fewer social resources, although the research on this issue is not current. For example, Welte et al. (61) found no relationship between LOS and outcome of alcoholism treatment for higher social stability patients; in contrast, for patients with lower social stability, those with longer stays had better outcomes (62,63). The finding that clients in naturalistic studies have better outcomes with longer treatment than clients who stay in treatment for shorter periods suggests that many individuals may be able to determine whether or not they will benefit from longer treatment. Thus, within limits (60), having longer courses of treatment available can be important for clients who seek them.
Continuing Outpatient Care
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
