Community Mental Health Journal

, Volume 42, Issue 4, pp 363–375

Availability of Integrated Care for Co-occurring Substance Abuse and Psychiatric Conditions

  • Lori J. Ducharme
  • Hannah K. Knudsen
  • Paul M. Roman

DOI: 10.1007/s10597-005-9030-7

Cite this article as:
Ducharme, L.J., Knudsen, H.K. & Roman, P.M. Community Ment Health J (2006) 42: 363. doi:10.1007/s10597-005-9030-7


The co-occurrence of psychiatric conditions and substance abuse presents significant challenges for behavioral healthcare providers. The need for integrated care has received substantial recent attention from clinical, research, and funding entities. However, the availability of integrated care has been low, carrying potential adverse implications for quality of care and treatment outcomes. This article describes the prevalence and key correlates of the availability of integrated care for co-occurring conditions within public and private-sector addiction treatment programs. Several organizational attributes, caseload characteristics, and service provision patterns were associated with the availability of integrated care.


co-occurring conditions substance abuse treatment integrated care. 


The co-occurrence of psychiatric and addictive disorders represents a significant challenge for behavioral healthcare providers, both in terms of the complexity and number of admissions to treatment. Recent Federal data indicate that between 7 and 10 million persons in the U.S. have co-occurring mental and substance use disorders (SUD) (DHHS, 1999; SAMHSA National Advisory Council, 1998). National epidemiological data estimate that roughly 20% of persons with SUD have co-occurring mood or anxiety disorders, and similar percentages of those with mood/anxiety disorders have an SUD (Grant et al., 2004; SAMHSA, 2004a). Other studies suggest that at least half of all persons with a lifetime substance abuse diagnosis also have at least one lifetime psychiatric diagnosis (Kessler et al., 1997). Substantial progress has been made in the science of treating psychiatric conditions as well as SUD, and it is well documented that both conditions are responsive to treatment. However, the treatment of these conditions has been highly segregated within the U.S. behavioral healthcare system.

Patients with co-occurring addiction and psychiatric disorders may present for treatment in clinical settings that are variously prepared to address their healthcare needs. Much of the nation’s behavioral healthcare system remains bifurcated along traditional service delivery and funding lines, with substance abuse treatment services occurring in one setting, and mental health care occurring in another. Often, neither setting has the capability to effectively address the condition that is outside its specialty area. Moreover, patients with co-occurring addiction and psychiatric conditions are likely to have significant difficulty negotiating the receipt of care in two separate service delivery systems.

A growing body of clinical research demonstrates that improvements in treatment utilization, retention, and outcomes are realized when psychiatric and addiction treatment services are integrated within the same service delivery setting (see review in Drake, Mueser, Brunette, & McHugo, 2004). Given these findings and growing recognition that the structure and funding of the service delivery system is itself impeding patient access and outcomes, the Substance Abuse and Mental Health Services Administration has recently initiated activities to enhance the capacity of behavioral healthcare providers to effectively address co-occurring psychiatric and substance abuse disorders via integrated services (SAMHSA, 2002). In order to provide a baseline against which the progress of these new initiatives might be monitored in this segment of the behavioral healthcare field, this article presents data on the current availability of integrated care for co-occurring psychiatric and substance abuse disorders among two large national samples of addiction treatment providers.

Integrated Care in Addiction Treatment Settings

Although there is near consensus about the importance of integrated service delivery, it is less clear how to facilitate the offering of such services within substance abuse treatment facilities (Brems, Johnson, & Namyniuk, 2002). A recent study of public-sector programs found that less than half of California counties provide integrated mental health and substance abuse treatment services (Young & Grella, 1998). Data from the private substance abuse treatment sector also shows a low rate of integrated care for clients with co-occurring serious mental illness, with only one-third of programs reporting this capability as recently as 2001 (Knudsen, Roman, & Ducharme, 2004). The majority of private centers referred these dually diagnosed clients to external providers.

SAMHSA’s National Survey of Substance Abuse Treatment Services (N-SSATS, formerly the Uniform Facility Data Set) provides limited information about service availability on a national scale. Of facilities responding to the 2003 survey, only 35.1% reported offering “programs or groups for clients with co-occurring conditions” (SAMHSA, 2004b). The percentage was even lower (26.9%) among facilities with a primary focus on substance abuse services. Further, only 43.6% of treatment facilities reported providing comprehensive mental health assessments as part of their client intake process.

While these data provide some indication of service availability, N-SSATS inquires only about “programs or groups” and the availability of mental health assessments. These data cannot identify the actual proportion of patients receiving comprehensive psychiatric assessments in addiction treatment settings, nor specify the clinical content of the “programs or groups” available for this population. Moreover, N-SSATS appears to undercount programs in the private sector, where the availability of psychiatric services may in fact be greater.

The need for specificity in measures of service availability is critical for understanding whether these service providers are offering integrated care. Substance abuse treatment programs admit many clients with co-occurring psychiatric disorders, but may not adequately screen for such conditions, nor have the capacity to treat clients’ psychiatric symptoms. Depending on the severity of their psychiatric needs, substance abuse treatment centers may refer severely mentally ill clients to other service providers for their care; in such circumstances, substance abuse services may be provided concurrent with, or subsequent to, mental health care delivery. In either case, when substance abuse treatment centers refer dually diagnosed clients to other agencies for some or all of their needed services, they are not providing integrated care for this population, and may in fact be increasing the risks of client drop-out. Although referrals to other agencies may represent an improvement over receipt of no mental health care, they are not consistent with an integrated service delivery model under which substance abuse treatment and psychiatric care are provided by a single treatment team within a single setting (Drake et al., 2001; Steele et al., 2002; Watkins, Burnam, Kung, & Paddock, 2001).

In the analyses that follow, we present data obtained from nationally representative samples of substance abuse treatment facilities in the public and private sectors in 2002–2003. We present both bivariate and multivariate analyses to identify organizational, funding, and other factors that are associated with the provision of integrated care. These data shed additional light on the availability of integrated services for clients with co-occurring conditions who seek care in addiction treatment programs.



Data are drawn from the 2002–2003 wave of the National Treatment Center Study (NTCS), conducted by the University of Georgia under grant support from the National Institute on Drug Abuse. The NTCS focuses on the long-term adaptation of drug treatment organizations and the environmental contexts in which they operate. The NTCS has several components, including nationally representative samples of specialty addiction treatment programs in the public and private sectors.

The NTCS uses a two-stage random sample of treatment programs, stratifying first on geographic location (county) and then sampling treatment facilities within geographic strata. Separate samples were drawn for the public and private treatment center components. In both samples, eligible facilities are organizations offering treatment for alcohol and drug problems, providing a level of care at least equivalent to structured outpatient programming as defined by ASAM patient placement criteria (Mee-Lee, Gartner, Miller, & Shulman, 1996). Excluded from the study are counselors in private practice, halfway houses and residential facilities, DUI or driver education programs, and units offering exclusively methadone maintenance services.

The sample of private sector treatment programs for this study was originally drawn in 1994. Unique to this study, “private sector” programs are defined as those receiving less than 50% of their annual operating revenues from Federal, state, or local grants or contracts, including criminal justice funds. In practice, these centers receive an average of 10.4% of their revenues from such sources. They include both for-profit and non-profit programs. Based on a review of SAMHSA’s national treatment facility inventory, we confirmed that 26% of the private centers in our sample are not included in the federal directory; thus, this sample appears to tap a segment of the service delivery system that is not reflected in available national data. To compensate for sample attrition due to program closure over time, replacement centers are randomly selected within the appropriate geographic strata to maintain a target sample size of 400 eligible units. This paper reports data from the 401 privately funded programs that were interviewed in 2002–2003. The response rate for this wave of interviews was 87% of those centers that were sampled and eligible.

Using similar stratification and identification procedures, a separate sample of public-sector treatment programs was drawn in early 2002. Eligibility rules were the same as for the private sample, except that public sector programs were defined as receiving at least 51% of their annual operating revenues from government grants or contracts. In practice, sampled programs received an average of 85.6% of their annual operating revenues from government funds. Because the NTCS focuses on treatment programs that are available to the general public, facilities operated by the Veteran’s Administration and those based in correctional facilities were excluded from the sample. The response rate for the 362 public-sector facilities interviewed in 2002–2003 was 80%.

Face-to-face interviews were conducted with the administrators of all participating programs. Interviews gathered information on the structure, staffing, and services of each facility, including the availability of integrated care for co-occurring substance abuse and psychiatric disorders. Pooled, unweighted data from both samples (total N=763) are reported in this paper. We include principal funding source as a control variable in our analyses to account for any systematic differences between the two samples, and to examine its relationship to the availability of integrated care.


Respondents indicated whether the treatment center accepts clients with co-occurring psychiatric conditions and, if so, whether the center treats both the substance abuse and psychiatric conditions of dually diagnosed clients. A number of centers indicated that they could treat some, but not all, psychiatric conditions. Only centers that could treat the psychiatric conditions of all clients (including those with active psychosis and other severe mental illness) were coded as providing integrated care. In the analyses presented below, a dichotomous dependent variable is used, with centers coded ‘1’ if they treat both the addiction and psychiatric conditions of dually diagnosed clients, and those that refer or do not admit some or all such clients coded ‘0.’

It should be noted that this measure represents a conservative definition of “integrated” care. SAMHSA’s recently published Treatment Improvement Protocol on co-occurring conditions (SAMHSA, 2005) employs a broader definition of integrated care, recognizing that there may be a variety of mechanisms by which clients’ needs are met, including formalizing arrangements with other providers to deliver services that cannot be met by the treatment facility to which the client has sought admission. The broad objective is to encourage a seamless system of care, in which there is “no wrong door” for patients needing treatment. However, the data obtained in this study are insufficient to address the extent to which these broader, system-level goals are being achieved, nor the extent to which client referrals to other providers are clinically appropriate. Thus, the analyses presented below use a fairly strict definition of service integration.

Several organizational variables were included in the analysis. First, because larger treatment programs are likely to have more resources available and may therefore be better positioned to provide a diverse array of services, treatment center size (measured in the number of full-time equivalent employees [FTEs]) is included. To facilitate interpretation, size is reported as measured in the bivariate analyses; to compensate for skewness, the natural log transformation of this variable is used in the multivariate regression analysis. Next, the age of the center (in years) is included, as newer centers may be more likely to be offering integrated services in response to recent funding and research priorities. Third, based on prior research that found for-profit units to be less likely to offer integrated care for co-occurring conditions (Knudsen et al., 2004), treatment centers’ profit status is also included as a potential predictor variable.

In terms of ownership, government-owned programs may be more responsive to recent Federal initiatives to improve the availability of integrated services; on the other hand, these programs may have established greater linkages to community-based service providers, making them more likely to refer clients with severe mental illness elsewhere for treatment of their psychiatric conditions. For this reason, government ownership is examined as a potential predictor variable. Relatedly, treatment centers’ reliance on public grant funds may render them less able to provide integrated services to the extent that those funds are earmarked for specific addiction treatment or ancillary supportive services. These analyses include a dichotomous measure of the proportion of the center’s past-year annual operating revenues obtained from public grants and contracts (centers with majority public funds are coded 1, 0 otherwise).

Hospital-based treatment centers may be more likely to offer integrated care, because of access to medical staff and related services, including inpatient beds and psychiatric medications. Within hospitals as well as other behavioral healthcare settings, accreditation is a hallmark of program quality, and has been associated with the availability of mental health care services in addiction treatment facilities (Friedmann, Alexander, & D’Aunno, 1999). Programs in the NTCS indicated whether they were accredited by the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) or the Commission on Accreditation of Rehabilitation Facilities (CARF). We include both hospital location and accreditation status as potential predictors in these analyses.

Treatment center staff credentials may also be a predictor of integrated service delivery. Workforce professionalism, as measured by educational attainment and/or credentials such as certification or licensure, are indicators of an organization’s capacity to access and implement knowledge about new techniques and practices (Knudsen & Roman, 2004). Thus, we examine whether treatment centers having a higher proportion of certified addictions counselors have greater capacity for offering integrated care.

Other factors that may influence the availability of integrated services relate to the treatment centers’ caseloads. Among persons with co-occurring conditions, women are more likely than men to receive treatment (SAMHSA, 2004c). In addition, adolescents entering substance abuse treatment have higher rates of depression, anxiety disorders (Hoffmann, Mee-Lee, & Arrowood, 1993) and affective disorders (Bergman, Smith, & Hoffmann, 1995). As a result, centers treating proportionately more women or more adolescents may be more likely to offer integrated care.

A treatment center’s location may also significantly impact the availability of integrated care. Treatment centers in rural locations may have fewer options for referring clients to other area agencies, and may therefore need to provide a more comprehensive array of services under one roof. In metropolitan areas with more dense networks of specialized service providers, agencies may be more likely to “buy” than “make” services for clients with complex treatment needs. Centers are coded as being in “rural” areas if the county in which they are located falls outside a metropolitan statistical area according to current Census Bureau definitions.

Finally, three measures of the psychiatric orientation of the treatment centers are also included as potential predictor variables. Administrators reported whether the center treats any of a variety of conditions other than alcohol and drug abuse. Although eligibility criteria required all centers to treat substance abuse conditions, a number of centers also offered specific treatment services for pathological gambling, eating disorders, and/or Internet addiction. Because they demonstrated an ability to provide treatment for these conditions, centers providing such services were hypothesized to be more likely to offer integrated care for clients with co-occurring conditions. A dichotomous measure is used, with 1=center offers any of these other behavioral health services and 0=center does not offer any of these services. Similarly, centers that reported offering a distinct inpatient psychiatric program for adults and/or adolescents were hypothesized to be most likely to treat substance abuse clients with co-occurring mental illness. Finally, administrators reported the proportion of incoming clients who receive a psychiatric assessment conducted by a psychiatrist or psychiatric nurse at intake. Because they are more likely to identify and diagnose co-occurring conditions, centers that are more active in conducting formal psychiatric assessments should be more likely to offer integrated care for those conditions.


In all, 57.5% of the sampled substance abuse treatment centers offered integrated (i.e., concurrent, co-located) care for co-occurring addictive and psychiatric disorders. The remaining centers either do not admit dually diagnosed clients, or refer clients elsewhere for the treatment of their psychiatric conditions. This is an encouraging finding when considering the conservative definition of “integrated care” used here. To understand which types of substance abuse treatment settings are more likely to meet this definition, we explore the organizational correlates of this service availability.

To compare centers that do and do not offer integrated care, we first examined bivariate associations between each of the organizational characteristics described above and the availability of integrated care, using t-tests and chi-square tests as appropriate. The first three columns of Table 1 show the frequency distribution for the variables of interest, as well as tests of significant differences in these characteristics between treatment centers offering integrated care and those that either refer or turn away clients with co-occurring psychiatric conditions. As shown, treatment centers providing integrated care appear to be substantially different in terms of their organizational characteristics. The provision of integrated care is more common among larger centers (as measured in FTEs), those based in hospitals, accredited facilities, and those treating more women and more adolescents. Notably, centers with a greater percentage of addictions-certified staff were significantly less likely to offer integrated care, as were centers with a greater dependence on public funding sources. As expected, substance abuse treatment centers that perform more psychiatric assessments at intake, those with available inpatient psychiatric services, and those treating various non-substance abuse conditions are significantly more likely to offer integrated care for clients with co-occurring psychiatric conditions. Rural location, government ownership, profit status, and age of the center (data not shown) were not significantly associated with provision of integrated care in these bivariate tests.
Table 1

Descriptive Statistics, Bivariate Associations, and Multivariate Logistic Regression Model of Availability of Integrated Services for Co-occurring Conditions on Organizational Characteristics


All Centers (N=763) Mean or %

Centers Offering Integrated Services (N=439 [57.5%]) Mean or %

Centers Without Integrated Services (N=324 [42.5%]) Mean or %

Logistic Regression: Odds Ratios for Variables in Model

% Clients receiving psychiatric assessments





Inpatient psychiatric unit





Treat other non-substance abuse conditions





Size (FTEs)a





Age of center (years)





For-profit status





Government owned





Majority funds from public sources





Hospital based program










Rural location





% Certified addictions counselors





% Female clients





% Adolescent clients





*p<.05; **p<.01.

aLogistic regression included log-transformed FTEs variable.

Next, we estimated a multivariate logistic regression model to identify significant predictors of the availability of integrated care. The coefficients and odds ratios resulting from this analysis are shown in column 4 of Table 1. As shown, a number of these variables are significant predictors of integrated care, net of the effect of other variables measured.

Considering first the availability of other behavioral health care services, treatment centers in which a greater proportion of patients receive formal psychiatric assessments are significantly more likely to offer integrated care (odds ratio [OR]=1.02, p<.01). Specifically, the likelihood of offering integrated care is 61.2% greater in centers conducting the mean percentage of psychiatric assessments, relative to those that do not perform any psychiatric assessments. Centers offering an inpatient psychiatric program in addition to their substance abuse treatment services were almost three times as likely to offer integrated care (OR=2.88, p<.01), while those also treating non-substance abuse conditions (such as pathological gambling and eating disorders) are more than twice as likely as other centers to offer integrated care (OR=2.13, p<.01).

Among the organizational characteristics included in this model, program size and staff credentials were significant predictors of integrated care delivery. Larger centers (measured by the natural log transformation of the total number of FTEs) were significantly more likely to offer integrated care, while those with a greater proportion of addiction-certified counselors were significantly less likely to do so. Neither accreditation status nor hospital location were significant predictors of integrated care. Because all hospital-based programs are among those that are accredited, we ran alternative models (not shown), to examine whether including both of these variables was suppressing their individual effects. However, removal of the hospital variable did not cause accreditation to become significant, and removing accreditation did not cause hospital location to become significant.

Centers receiving a majority of their funds from public sources (government grants and contracts) were no more likely to offer integrated care. In fact, the direction of the coefficient suggests a trend toward being less likely to offer such services. However, this measure of program funding was not a significant predictor in the model shown. Alternate models (not shown) examined different funding measures, including the absolute percentage of revenues received from public sources, as well as a dichotomous variable measuring the receipt of any public funds. There were no models in which funding emerged as a significant predictor of integrated care, net of ownership and other measures.

Finally, only one of the two measures of caseload characteristics was significant. Centers with caseloads including a higher percentage of adolescent clients were significant more likely to report offering integrated care for co-occurring substance abuse and mental health conditions (OR=1.01, p<.05). Relative to centers treating no adolescent clients, the likelihood of offering integrated care is 22.1% greater in centers where adolescents comprise one-quarter of the caseload. However, the proportion of female clients in a center’s caseload was not significantly associated with the provision of integrated care in this multivariate model.


Consistent with some of the findings from existing federal data sets, patients with co-occurring conditions who seek admission to substance abuse treatment facilities are likely to find that a substantial proportion of these facilities are unable to fully address their needs. Addiction treatment programs offering a more “mixed” set of behavioral healthcare services—including inpatient psychiatric services, psychiatric assessments, and treatment for other non-substance abuse problems—are better able to accommodate patients with co-occurring psychiatric conditions. Meanwhile, those with a more specific focus on substance abuse services—signaled in part by a higher percentage of certified addictions counselors among their staff—are less likely to be able to meet the needs of dually diagnosed patients. In short, while other national data suggest that there are significant differences in services between the mental health and substance abuse treatment sectors, our data indicate that there also appear to be two distinct systems of care for co-occurring conditions within the substance abuse treatment system itself.

These data illustrate that a substance abuse client who also has psychiatric treatment needs cannot assume that those needs will be fully met within a given addiction treatment program. While adolescents appear to have better odds of receiving integrated care within addiction treatment settings, other patients appear to be particularly at risk of not having access to integrated services, and are likely to be referred to other treatment facilities for some or all of their care. Although these external referral patterns may mean that clients with dual diagnoses are linked to services that they might not otherwise receive, the data nevertheless indicate that many addiction treatment centers are not providing the integrated treatment that has been identified as most effective (Drake et al., 2001, 2004). Furthermore, although referrals to other providers may ensure that the client’s mental health needs are met, such referrals are no assurance that the needed services are received. While it is unknown whether the external providers are able to fully meet clients’ substance abuse treatment needs, there is also the potential for delayed or denied admission. This highlights the unrealistic expectation that clients will be able to successfully obtain the services they need across multiple treatment systems. Thus, inadequate care continues to threaten the recovery of a significant population of persons with co-occurring conditions who seek care in addiction treatment programs.

Several limitations of these data should be noted. First, the sample selection criteria limited participation to specialty substance abuse treatment facilities. Organizations whose primary focus was on the provision of psychiatric services—including psychiatric hospitals—were not included in the study design. Such organizations are likely to face parallel challenges in terms of providing integrated care for clients with co-occurring SUD, and may utilize different service delivery models to more effectively accomplish this goal. Second, these analyses employ a conservative definition of “integrated care.” These data do not speak to the presence of formal or well-established referral networks that may be in place in these facilities. As a result, these data cannot address other activities in which these treatment centers engage to obtain needed services for clients with co-occurring psychiatric conditions. Finally, the data cannot provide an assessment of staff competencies in identifying and working with clients who have co-occurring conditions in the context of delivering substance abuse treatment services. These areas warrant attention in future research.


Despite the known high rates of co-occurring psychiatric conditions among patients entering substance abuse treatment, just over half (57.5%) of the treatment providers surveyed offer integrated care for dually diagnosed clients. A number of organizational, staffing, and case mix factors differentiate among integrated and non-integrated treatment settings in both bivariate and multivariate analyses. These systematic variations among treatment providers raise important concerns about the quality of care received by substance abuse clients with co-occurring psychiatric disorders. Patients with severe and persistent mental illness are particularly at risk of being referred to other settings for treatment of their psychiatric condition. As a result, these patients must negotiate separate systems of care when they themselves are most significantly compromised. This fragmentation of service delivery has been identified as disadvantageous in terms of treatment retention, patient outcomes, and long-term cost effectiveness (Hoff & Rosenheck, 1999). The significance of this continued bifurcation of service delivery for the welfare of clients with dual diagnoses warrants continued monitoring of trends within the behavioral healthcare field. In particular, companion analyses of the availability of integrated care within mental health settings are especially needed.


The authors thank the journal’s reviewers for their helpful comments, and gratefully acknowledge funding support from the National Institute on Drug Abuse (research grants R01DA14482 and R01DA13110). The opinions expressed are those of the authors and do not reflect the official position of NIDA.

Copyright information

© Springer Science+Business Media, Inc. 2006

Authors and Affiliations

  • Lori J. Ducharme
    • 1
  • Hannah K. Knudsen
    • 1
  • Paul M. Roman
    • 1
    • 2
  1. 1.Institute for Behavioral ResearchUniversity of GeorgiaAthensUSA
  2. 2.Department of SociologyUniversity of GeorgiaAthensUSA

Personalised recommendations