Differences Between U.S. Substance Abuse Treatment Facilities That Do and Do Not Offer Domestic Violence Services

Domestic violence is as a major public health issue. Research consistently links substance use and abuse to violence perpetration and victimization (1–8). Given the high prevalence of substance use among victims and perpetrators, integrating domestic violence services into programs that address substance abuse has become increasingly important, but most programs experience challenges when treating the array of problems associated with the co-occurrence of substance abuse and domestic violence (9–14).

Attention has turned to the link between co-occurring disorders and domestic violence perpetration and victimization (15–19). The Substance Abuse and Mental Health Services Administration (SAMHSA) has called for greater integration of screening and treatment for mental and substance use disorders among individuals with co-occurring disorders. However, there is a paucity of such integrated programs for specialized populations, such as those needing domestic violence services (20–22).

Addressing substance abuse and mental health needs in the screening and treatment of persons affected by domestic violence is complicated. First, many states forbid or limit discussion of alcohol or drug use in programs for domestic violence perpetrators (23,24), focusing instead on holding the perpetrator solely accountable for violence perpetration and thus disregarding the role substance abuse issues may play. Second, subtypes of persons involved in domestic violence display varying degrees of substance abuse comorbidities (2–5,25), and interventions are rarely tailored to meet clients' multiple and unique needs (10,11,23). Finally, individuals who are mandated by the courts to enter substance abuse treatment may be unmotivated to discuss or acknowledge the extent to which problem behaviors, such as domestic violence, are related to their substance use.

Given the overlap between substance abuse and domestic violence issues, substance abuse treatment providers have been encouraged to screen clients for domestic violence (perpetration or victimization) and to provide treatments tailored to addressing the co-occurrence of these behaviors. Substance abuse treatment programs may be the first point of entry to intervene and prevent further violence. However, it is unknown whether substance abuse treatment facilities are moving toward this integrated model. Using a national survey of more than 13,000 U.S. substance abuse treatment facilities, we determined the percentages of facilities that do and do not offer domestic violence services, whether the two groups of facilities differ in the frequency of use of common therapeutic approaches for treating substance use disorders and whether the proportion of clients with co-occurring disorders at a facility is related to the degree to which facilities in the two groups differed in the types of therapeutic approaches offered.

Methods

Data source

We used data from the 13,342 facilities participating in the 2009 National Survey on Substance Abuse Treatment Services (N-SSATS) that reported information on the provision of domestic violence services. This is a publicly available data set of public and private substance abuse treatment facilities in the United States surveyed annually by SAMHSA. Of the 15,213 eligible institutions, 93% (N=14,209) completed the survey; 696 facilities were excluded because of reporting inaccuracies, and 171 facilities had missing data on the provision of domestic violence services, leaving a final sample of 13,342.

Independent variables

Our independent variable captured the presence or absence of domestic violence services at each facility. The question for this item was worded: “Which of the following services are provided by this facility at this location?” with a response option of “Domestic violence—family or partner violence services (physical, sexual, and emotional abuse)?” The survey item does not measure how many clients access this specific service. An item also asked about the percentage of clients being treated for co-occurring mental and substance use disorders at each facility. Response options range from 0% to 100%.

Dependent variables

The dependent variable was the frequency with which the facilities used eight commonly employed types of treatment approaches. Respondents rated how often the facility used each of the following services (1, never, to 4, always or often): substance abuse counseling, 12-step counseling, brief intervention, cognitive-behavioral therapy (CBT), motivational interviewing, trauma-related counseling, anger management, and relapse prevention.

Covariates

Covariates were selected on the basis of factors used in previous studies of the N-SSATS and factors that have been empirically shown to affect substance abuse treatment services (26–28). They were primary treatment focus (substance abuse, mental health, mix of mental health and substance abuse, general health care, or other); private versus government ownership; facility region (Northeast, Midwest, South, West, or U.S. jurisdiction or territory); hospital location; provision of standard case review with a quality review committee; treatment in language other than English; provision of nonhospital residential substance abuse treatment; provision of regular outpatient treatment; use of a sliding fee scale or treatment free of charge; Medicare reimbursement accepted; Medicaid reimbursement accepted; receipt of federal, state, county, or local funds (yes or no); and licensed or certified by a state substance abuse agency, mental health department, state health department, or hospital authority (one response selected).

Analytic plan

Prevalence rates and differences between facilities.

We calculated the percentages of facilities that offered and that did not offer domestic violence services. We used a one-way analysis of variance test to examine differences between the two groups of facilities on the basis of the percentage of clients with co-occurring disorders. We then used chi square tests to compare differences between facilities that offered domestic violence services and those that did not on key organizational characteristics and frequency of use of the eight different treatment approaches.

Principal-components analysis.

To decrease the probability of type I error that would result from using each of the eight different treatment approaches as separate outcomes and to condense the eight approaches into a smaller number of underlying categories, we conducted a principal-components analysis (PCA). Factors with eigenvalues >1 were retained. Direct oblimin oblique rotation was used, and factor loadings >.4 were interpreted. Factor scores were saved by using the regression method and were used as our dependent variables (29).

Hierarchical regression analyses.

Hierarchical regression analyses that controlled for organizational covariates were performed to examine whether facilities that did or did not offer domestic violence services differed in their frequency of using each PCA-derived factor. Analyses also examined whether the association between offering domestic violence services and the frequency of using different treatment approaches was influenced (moderated) across facilities by the percentage of clients with co-occurring disorders. To evaluate main and interaction effects, we created a meaningful “zero point” by standardizing scores for the percentage of clients with co-occurring disorders.

We first examined an overall unadjusted model, which included all potential organizational covariates that differed at the bivariate level between facilities that offered domestic violence services and those that did not. The second step of the model included the variable “offered domestic violence services” versus “did not offer such services” (yes or no) and the percentage of clients with co-occurring disorders. The interaction (domestic violence × percentages with co-occurring disorders) was added in the final step of this model.

To reduce the set of parameters, regression models were rerun (adjusted model) by using a backward selection procedure, in which only organizational variables with p<.15 were included. Final adjusted models included the most statistically robust correlates of the PCA-derived dependent variables.

Results

Prevalence and differences between facilities

Of the 13,342 facilities, 36% (N=4,779) offered domestic violence services; whereas 64% (N=8,563) did not. Facilities offering domestic violence services had a higher percentage of clients with co-occurring disorders (42%±32%) than facilities not offering these services (39%±32%) (F=33.79, df=1 and 13,341, p<001).

As shown in Table 1, the two groups of facilities differed on a variety of organizational factors, including primary treatment focus, location, funding stream, insurance and reimbursement type, and licensing credentials, among others. As indicated in Table 2, facilities not offering domestic violence services were more likely than facilities offering domestic violence services to offer most treatment approaches (more likely to respond “always” or “often”).

PCA results

The PCA showed a two-factor solution, accounting for 44% of the total variance (Table 3). The first factor, psychosocial services, included brief intervention, CBT, trauma-related counseling, motivational interviewing, and anger management. The second factor, traditional substance abuse services, included substance abuse counseling, 12-step counseling, and relapse prevention. The two factors were moderately correlated (r=.31, p<.001), suggesting that they were distinct.

Hierarchical regression models

Table 4 presents results of the final adjusted regression models. The interaction between domestic violence and percentage of patients with co-occurring disorders was significant for psychosocial services as the dependent variable (b=–.08, t=−3.78, df=8,113; p<.001). Explication revealed that the association between the percentage of patients with co-occurring disorders and the frequency of offering psychosocial services was stronger for facilities not offering domestic violence services (b=.25, t=18.36, df=8,113, p<.001) than for facilities offering domestic violence services (b=.17, t=9.82, df=8,113, p<.001). Provision of domestic violence services and the percentage of patients with co-occurring disorders emerged as significant main effects of offering traditional substance abuse services, but they did not interact. In other words, the relationship between provision of domestic violence services and offering traditional substance abuse treatment services was not significantly influenced by whether or not a facility had a greater percentage of clients with a co-occurring disorder. Facilities providing domestic violence services offered traditional substance abuse treatment services significantly more often than facilities that did not provide domestic violence services (b=.05, t=2.39; df=8,133, p=.019), but facilities with a higher percentage of clients with co-occurring disorders offered traditional substance abuse services significantly less often than facilities with a lower percentage of clients with co-occurring disorders (b=−.02, t=22.13, df=8,133, p=.033).

Discussion

This study determined the percentages of U.S. substance abuse treatment facilities that did and did not offer domestic violence services, how these two facility groups differed in the types of treatment approaches they offered, and whether the percentage of clients being treated for co-occurring disorders moderated the degree to which the two facility groups differed in the types of treatment approaches offered.

Only a little over one-third (36%) of the facilities surveyed offered domestic violence services. It is important to uncover barriers in the adoption and implementation of domestic violence services by substance abuse treatment programs. Identifying barriers likely requires consideration of other factors associated with adoption and implementation of such services, such as counselor burden and attitudes toward domestic violence, staff training to address comorbid problems, client motivation and degree of psychiatric severity, and organizational factors. One reason for the small proportion of facilities offering domestic violence services may be that clients are receiving specialty domestic violence services elsewhere. However, even if that is the case, this finding highlights a problem in that substance abuse treatment programs are not offering concurrent domestic violence services to the extent that they are needed (25).

A variety of organizational factors were differentially associated with whether or not facilities offered domestic violence services. Notably, among facilities offering domestic violence services, 40% offered a mix of mental health and substance abuse treatment. In contrast, only 26% of facilities that did not offer domestic violence services provided this type of comprehensive focus. When the mean percentage of clients with co-occurring disorders was examined across facilities, the facilities that offered domestic violence services had a larger percentage of such clients (42%) than facilities that did not offer domestic violence services (39%). Even though these differences may appear small, they are encouraging in light of research indicating that interventions can be more effective at reducing the occurrence and consequences of domestic violence if they take into the account both mental health and substance abuse issues affecting individuals (6,20,25). Certain types of substance abuse treatment facilities should be the focus of enhanced resources that allow for the provision of domestic violence screening and services in their programs. Such facilities should include those located in the West, those that do receive funding earmarked for special projects, those that do not offer quality case review, and those that do not accept Medicaid or Medicare, among others.

Bivariate analyses also showed that the primary differences between facilities that did and did not offer domestic violence services appear to be within the treatments categorized as psychosocial, rather than traditional substance abuse treatments. Although both types of facilities appeared equally likely to offer traditional substance abuse treatment services, those that offered ancillary domestic violence programs were more likely to offer brief intervention, CBT, motivational interviewing, trauma-related counseling, and anger management. This distribution of services is consistent with the data in Table 1, indicating that 47% of the domestic violence programs reported that their primary treatment focus was either on mental health treatment (7%) or a mix of mental health and substance abuse treatment (40%).

Results from a PCA of eight different treatment approaches suggested the presence of two separate dimensions—psychosocial services and traditional substance abuse services—which were used as dependent variables in regression models. Final adjusted models showed that facilities not offering domestic violence services were more likely than facilities offering domestic violence services to offer psychosocial services when those facilities also treated a higher percentage of clients with co-occurring disorders. This finding indicates that the delivery of general mental health treatments to clients in substance abuse treatment facilities was largely contingent on the number of clients being assessed and treated for co-occurring disorders, as well as whether or not that facility offered ancillary domestic violence services. Facilities that offer domestic violence services and that treat clients with co-occurring disorders may lack the resources to offer general psychosocial services as frequently as facilities not offering domestic violence services. Perhaps facilities not offering domestic violence services focus more on rehabilitating the symptoms associated with the underlying mental illnesses of the clients, whereas those offering domestic violence services spend more time and resources delivering interventions targeted toward reducing stressors in the social environment (violence) or problems specific to substance abuse.

These findings suggest that substance abuse treatment facilities are not entirely moving toward targeting the unique psychological needs of persons involved in domestic violence. More focus should be given to providing additional training and resources to counselors at facilities that serve clients involved in domestic violence but that do not focus on clients with co-occurring disorders. Resources should include providing training to screen for co-occurring problems; enhancing treatment planning to address relationships among mental health problems, substance abuse, and violence; and teaching counselors to adapt interventions to handle a variety of co-occurring problems.

Results from final regression models showed that a facility was less likely to provide traditional substance abuse services if that facility reported a higher percentage of clients with a co-occurring axis I disorder and a substance use disorder (regardless of whether that facility offered ancillary domestic violence services). No interaction between domestic violence and percentage of clients being treated for co-occurring disorders was found. This finding further underlines the above-mentioned findings that facilities with large percentages of clients with co-occurring disorders may not be focusing as much on clients’ substance abuse issues as on their mental health problems. Further, regression models showed that a facility offering domestic violence services was significantly more likely than one not offering these services to provide traditional substance abuse services; but again, this relationship was not contingent on the percentage of clients being treated for co-occurring disorders at that facility. We would expect that facilities offering domestic violence services would be more likely than those not offering these services to provide specific substance abuse treatments, given the encouragement by SAMHSA to address the high correlation between these factors. This finding suggests some promising movement toward an integrated approach, at least in terms of the focus on substance abuse and violence-related issues. However, given that relatively few of the more than 13,000 facilities surveyed offered domestic violence services, more attention and resources should be given to help facilities identify and target concurrent domestic violence, substance abuse problems, and mental illness.

This study had several limitations. First, because we used a preexisting data set, our research questions and interpretations of findings were limited to the information available. Second, we recognize that domestic violence services were broadly defined as those for both victims and perpetrators, and services focused on perpetrators differ substantially from services for victims. Third, we were unable to discern in the survey data how many clients actually accessed domestic violence services. Without more information about services provided and services accessed, causal interpretations from the analyses are limited. Fourth, even though we found that compared with facilities not offering domestic violence services, those offering domestic violence services treated a significantly higher percentage of clients with co-occurring disorders, the discrepancy was small given the large sample. In addition, it should be noted that the large sample affected significance levels in contingency tables. Thus, although differences between facilities that did and did not offer domestic violence services may be significant, the clinical meaningfulness should be interpreted with caution. Finally, we were not able to assess client-level characteristics that may affect associations between domestic violence services and treatment for substance use problems.

Conclusions

The low rate of provision of domestic violence services in substance abuse treatment facilities suggests that efforts should be improved to tailor treatments to address the links between domestic violence, mental health problems, and substance abuse. Our findings are not entirely consistent with the suggestion by SAMHSA, which calls for a multifaceted approach to reducing co-occurring problems among persons with substance use disorders. Substance abuse counselors may lack expertise to focus on issues related to partner violence, or their high caseloads may prevent them from devoting time to effective screening for domestic violence. Leadership issues, high clinician turnover, lack of administrative support and resources for staff training, and issues with the funding stream may also be barriers to concurrently treating domestic violence and substance abuse (14,26).

Although adding domestic violence services to substance abuse treatment programs will prove incrementally helpful to individuals and families in need, offering a truly integrated approach to mental health care that addresses a variety of issues is likely to be most effective. Identification of factors that are correlated with whether or not facilities provide integrated domestic violence services may lead to improved screening, referral, and implementation of coordinated care and may inform restructuring of organizational weak points and barriers to implementation and adoption of domestic violence services in these settings.