Little is known about the demographic and clinical characteristics of people with intellectual disabilities and substance abuse problems. Drawing on health care billing claims for people with Medicaid coverage aged 12–99 years, the characteristics of people with intellectual disability and a history of substance abuse (N = 9,484) were explored and compared with people with intellectual disability but without substance abuse. Age- and/or gender-adjusted odds ratios were derived from logistic regression analyses to consider differences in demographic and clinical diagnoses. People with intellectual disability and substance abuse constituted 2.6% of all people with intellectual disability, most of whom had a diagnosis of mild or moderate intellectual disability. People with intellectual disability and substance abuse problems were, on average, 2 years older than the comparison group and less likely to be White. The sample was more likely than the comparison group to have serious mental illness or depression and substance abuse–related disorders were not prevalent. These data provide a comparison point for existing studies of mental health diagnoses as well as new information about substance abuse disorders. Implications relate to the identification of substance abuse among people with intellectual disabilities as well as the establishment of demographic and clinical correlates.
Substance abuse is a noted concern among people with intellectual disabilities (Didden, Embregts, van der Toorn, & Laarhoven, 2009; Edgerton, 1986; Slayter, 2008, 2007; Slayter & Steenrod, 2009). Although much is known about the co-occurrence of mental health conditions among people with intellectual disability, little is known about the demographic or clinical characteristics or clinical diagnoses of people with intellectual disabilities and substance abuse, in part due to the potential for diagnostic overshadowing (Borthwick-Duffy, 1990; Reiss, Levitan, & Szyszko, 1982; Strain, Buccino, Brooner, Schmidt, & Bigelow, 1993). Few existing studies provide this information, most of which are based on small, localized studies with limited external validity (Christian & Poling, 1997; DiNitto & Krishef, 1983; Krishef, 1986; McGillicuddy, 2006; Walkup, Sambamoorthi, & Crystal, 1999; Westermeyer, Kemp, & Nugent, 1996; Westermeyer, Phaobtang, & Neider, 1988).
A first wave of studies addressing the prevalence of substance abuse among the general population of people with intellectual disabilities was driven by post-deinstitutionalization concerns. A 1980 survey of 314 community-based people with intellectual disabilities in Florida found that between 26% and 34% “had an alcohol problem,” with higher rates reported among males (DiNitto & Krishef, 1983/1984). In one of the best known follow-up studies on deinstitutionalization, a qualitative, longitudinal study tracked 181 Californians with intellectual disability. The presence of substance abuse conditions and substance abuse–related consequences were noted in several case vignettes (Edgerton, 1986a) from sample members in four distinct groupings (a “normalized” group, an independent-living group, a deinstitutionalized group, and a group categorized as “Black, inner-city”). Depending on the group, between 5% and 14% of the people in each setting were reported to have alcohol- or drug (AOD)-related problems (Edgerton, 1986a). Males appeared to be more likely to be involved with AOD-related problems, and no differences by race or ethnicity were noted. The presence of serious mental illness was not reported.
Despite little available knowledge about the prevalence of this problem or approaches to treating it, during the 1980s and 1990s, social service agencies working with people with intellectual disabilities began to address substance abuse among this population by experimenting with the provision of prevention and education-oriented programming (McCarver & Craig, 1974; Meyers, Branch, & Lederman 1988; Rivinus, 1988; Selan, 1976; Sengstock, Vergason, & Sullivan, 1975; Small, 1980; Wenc, 1980). One study conducted in the late 1990s is representative of this time, presenting carefully collected evidence regarding the prevalence of substance abuse among people with intellectual disability. Drawing on a sample of 122 community-based people with intellectual disabilities living in Buffalo, New York (McGillicuddy & Blane, 1999), who were categorized by the authors of that study into groups of nonusers, AOD users, and misusers, 18% of the group surveyed were found to be alcohol misusers, and 4% of the population reporting lifetime use of illicit drugs (McGillicuddy & Blane, 1999). Notably, marijuana use was more common among people with intellectual disabilities who were misusers of alcohol compared with those who only used or abstained from drinking alcohol. Demographic and clinical data about the sample who reported use or misuse of alcohol or drugs were not reported. Another study based on chart-review data collection methodologies conducted in a psychiatric day hospital program in Maryland between 1989 and 1990 determined that 26% of a sample of 145 patients with intellectual disabilities also had diagnoses of substance abuse and mental illness (Strain, Buccino, Brooner, Schmidt, & Bigelow, 1993).
In this same era, two studies examined the prevalence of intellectual disabilities among clients in substance abuse treatment programs (Westermeyer, Kemp, & Nugent, 1996; Westermeyer, Phaobtong, & Neider, 1988). In the earlier study detailing substance abuse treatment use by people with intellectual disability, a survey was conducted in two treatment settings (a hospital-based program and an Alcoholics Anonymous group designed for people with intellectual disability; Westermeyer, Phaobtong, et al.). In non–intellectual disability-specific service settings, IQ testing was conducted to identify the presence of this condition. Collectively, substantial percentages of people with intellectual disabilities from all of these settings had, in their lifetime, used marijuana (82.5%), amphetamines (32.5%), and cocaine (0.2%). Although the authors provided significant detail about types of substances used, for example, little is reported about the demographic or diagnostic clinical characteristics of the population. Males were more prevalent than females (by a 2∶1 ratio) in the sample, although the authors pointed out that this mirrors the demographics of the population with intellectual disability, suggesting equal risk between genders (Westermeyer, Phaobtong, & Neider, 1988). Notably, 40% of people with intellectual disabilities who had a substance abuse problem in this study had experienced an inpatient psychiatric admission in their lifetime. In addition, 25% had been admitted to a state psychiatric hospital for mental illness and 62.5% had received outpatient psychiatric treatment (Westermeyer, Phaobtong, & Neider, 1988). These results supported the concern that the problem of co-occurring mental illness and substance abuse conditions—a well-documented concern for the general population—may also impact people with intellectual disabilities (Westermeyer, Phaobtong, & Neider, 1988).
In a later survey, 642 patients sequentially admitted to an alcoholism/addictions program at two university-hospital–based programs in Minnesota and Oklahoma were studied (Westermeyer, Kemp, & Nugent, 1996). It is unclear whether this program was for detoxification alone or was a longer term, hospital-based treatment program. Of the programs' combined 642 patients, 40 people with intellectual disabilities (6.2% of the patient group) were identified, with a slight majority of males.
During the 1990s, the issue of substance abuse became a better recognized problem among practitioners as it pertained to people with disabilities of all types, including those with intellectual disability, likely spurred by continued interest in the impact of deinstitutionalization and community inclusion. A number of studies identified people with disabilities as being at increased risk of developing substance abuse problems but did not present new empirical data on the topic (Campbell, Essex, & Held, 1994; Cocco & Harper, 2002; Simpson, 1998). Drawing on this momentum, the National Association of Alcohol, Drugs and Disability (NAADD) conducted a largely qualitative study in a number of different U.S. states to assess levels of access to substance abuse treatment for all with disabilities (NAADD, 1999). One part of this study, conducted in a large Arizona county, found that 2.2% of all clients in substance abuse treatment had “developmental disabilities,” presumably including intellectual disabilities but using the former as an umbrella term. No data related to gender, age, race, ethnicity, or co-occurring mental illness were reported. This finding was echoed by a single statistic derived from the National Health Interview Survey's Disability Panel in 1995, which suggested that 2.2% of the community-based population with intellectual disabilities or developmental disabilities had received substance abuse treatment in the previous year, although again, no demographic or clinical characteristics were reported (Larson, Lakin, Anderson, & Kwak, 1999).
To build on these findings, this study draws on Medicaid claims (e.g., billing) data from 49 states and the District of Columbia for people with and without intellectual disabilities to facilitate a comparison of clinical and demographic characteristics. I hypothesized that people in this sample would be more likely to be male, to have a mental illness, and to have a what is characterized in clinical settings as a more severe serious mental illness and less likely to have a chronic substance abuse–related disorder. Hypotheses related to age differences and racial/ethnic differences were exploratory.
Using a retrospective, cross-sectional design, research questions were answered through the analysis of claims data for the population of Medicaid beneficiaries with intellectual disabilities aged 12–99 years who received Medicaid-covered care in 1999. Specific statistical analyses conducted included the derivation of age- and gender-adjusted odds ratios by logistic regression analyses and independent samples t tests. To facilitate easy interpretation of the findings, results are reported as adjusted odds ratios. In this study, adjusted odds ratios are transformations of coefficient estimates obtained from logistic regression analyses, which provide easily interpretable data on the likelihood or odds of the sample having a mental illness diagnosis versus the comparison group, while controlling for age and gender. For example, an adjusted odds ratio that is below 1 (e.g., 0.70) is interpreted to mean that the sample is 30% less likely to have such a condition than is the comparison group while controlling for age and gender. However, an adjusted odds ratio of greater than one (e.g., 1.50) means that the sample is 1.5 times more likely to have such a condition than is the comparison group while controlling for age and gender. The use of adjusted odds ratio testing is appropriate to the analyses herein, because the primary focus of these analyses is the comparison of populations.
Data included Medicaid beneficiaries in all U.S. states and the District of Columbia (with the exception of Hawai'i, because no data were submitted by that state) and were derived from Medicaid Statistical Information System (MSIS) eligibility and claims files. Medicaid claims data are widely used for research focused on service use and expenditures as well as the identification of rare and elusive populations (Iezzoni, 2002). Although a full discussion of the positives and negatives of Medicaid claims data is beyond the scope of this article, a brief overview of major points is presented here (see, e.g., Iezzoni, 2002; Quam et al., 1993). This data source is often critiqued with respect to its accuracy, reliability, and validity: These data are still characterized as “a potentially rich source of clinical” data (Walkup & Boyer, 2000, p. 136), especially when triangulated with other types of data, such as case record review.
A number of studies have conducted reliability assessments of Medicaid claims, concluding an average to high rate of concordance depending on the condition or service use studied (see, e.g., Jacobson Vann, Feaganes, & Wegner, 2007; Steinwachs et al., 1998b). Although claims data are the mechanism by which service settings get paid, there is some evidence that visits are underreported for low-cost services or low-use patients, suggesting the potential for an undercount in the present study. Although triangulation with other data sources was not a possibility in this large study, the data source does provide an opportunity to look at a large number of people with intellectual disability and substance abuse diagnoses, a sample hard to find by other means.
Assessments of Data Availability in Medicaid Claims
Data availability assessments were conducted due to the increased prevalence of capitated health insurance plans (in which claims data can be limited in scope). In Medicaid plans with a managed care orientation, claims may not include data about subcontracting providers in behavioral health care “carve-outs,” because in such a system, capitated or lump-sum payments are provided to the entities providing such care to cover all costs, without individual claims reported at the state Medicaid program level (Garnick et al., 1996). A four-part data availability check was conducted to assess the validity of the claims. First, it was determined that Hawai'i alone had not submitted any data to MSIS in 1999. Second, the study could not include older adults enrolled in Medicaid plan types that did not report claims due to the use of capitated payment systems. Third, certain states did not report data during specific months, so annualized prevalence figures for individuals with substance abuse treatment claims were examined by state and plan type. Each beneficiary was associated with one primary Medicaid plan type. Any cell (e.g., state and plan type) with fewer than 100 persons who were elderly or with more than the expected number of these individuals (compared with calculations conducted with a 5% sample of the entire MSIS file for 1999) was eliminated.
A particular strength of this study is the fact that it did not rely on one location or one health care setting, instead drawing on claims from outpatient, inpatient, and long-term care settings across the United States. Medicaid claims are widely used in health services research to examine a range of conditions, including substance abuse (Sambamoorthi, Warner, Crystal, & Walkup, 2000). Claims data are especially helpful in locating hard-to-study populations (Quam et al., 1993). Although they are useful in examining proxy estimates of the prevalence and/or incidence of various conditions, it is important to remember that codes are “not meant to tell stories, (but) rather to generate reimbursement” (Horner, Paris, Purvis, & Lawler, 1991; Iezzoni, 2002, p. 348). Diagnostic undercounts may occur due to stigma-related underreporting of substance abuse, perceived irrelevance of ancillary diagnoses, lack of space on reporting forms, or a focus on another health condition during a visit (Garnick, Hendricks, & Comstock, 1996). For example, a person with intellectual disabilities might receive health care for a broken leg and contusions in relationship to alcohol consumption, but the claim submitted to Medicaid might only include International Classification of Diseases, Ninth Revision Clinical Modification (ICD-9-CM; Health and Human Services Department, Centers for Disease Control and Prevention; and Centers for Medicare and Medicaid Services, 2005) codes for the broken leg and contusions.
Despite the potential problems associated with the use of Medicaid claims, the ability to study the proxy prevalence of demographic and clinical characteristics in a large sample of people with intellectual disabilities across the United States can provide clinicians with helpful epidemiological data. Last, Medicaid claims are particularly useful for studying people with intellectual disability, because Medicaid is a primary source of health coverage for this population, although findings from this study are not strictly generalizable to the entire population of people with intellectual disabilities in the United States.
Through the use of two sets of diagnostic codes (for intellectual disability and substance abuse), the population of Medicaid beneficiaries with intellectual disability and substance abuse (n = 9,484) was identified. Characteristics were compared with Medicaid beneficiaries with intellectual disability but without substance abuse diagnoses (n = 366,606). Intellectual disabilities was identified by ICD-9-CM diagnostic codes for “mental retardation” (“MR”), including 317 (“mild MR”), 318 (“moderate MR”), 318.1 (“severe MR”), 318.2 (“profound MR”), and 319 (“MR not otherwise specified”). Medicaid claims files were searched for the first instance of one of the five ICD-9-CM diagnosis codes for “MR,” looking at both primary and secondary code fields.
Similarly, substance abuse was identified through the examination of 1999 Medicaid claims histories to identify those who had at least one diagnosis of substance abuse at any time during the year. A set of ICD-9-CM diagnosis codes designed by the Washington Circle (“a group of national experts in substance abuse policy, research and performance management who seek to improve the quality and effectiveness of prevention and treatment services through the use of performance measurement systems”; Washington, Circle, 2010), were used for this purpose (Garnick, Lee, et al., 2002; McCory, Garnick, Bartlett, Cotter, & Chalk, 2000). Demographic variables extracted from Medicaid eligibility files included gender, age, and a combined race and ethnicity variable (e.g., White and non-White, the latter including “Hispanic ethnicity”). Although the clinical validity of race and ethnicity markers are limited, the power of these socially constructed categories renders them important to consider in light of the presence of racial and ethnic disparities in the United States (Moscou, 2008).
Four types of clinical factors were examined. Clinical factors included a diagnosis of developmental delay, mental health diagnoses, a set of diagnoses categorized as serious mental illness, and a set of chronic, substance abuse–related disorders. First, the ICD-9-CM code 305 for “specific delays in development” was included as a measure of developmental delay. This variable includes subcategories, such as 315.5 (mixed developmental disorder), 315.8 (specific delays in development), and 315.9 (unspecified delays in development). Health care providers are cued to only use this code if the patient in question meets Federal guidelines, specifically “disabilities originating before age 18 that constitute substantial barriers to normal functioning. Use a more specific term if possible.”
Second, a grouping of diagnoses was used to create a composite variable for mental health diagnoses. These variables included anxiety, dissociative and somatoform disorders (300); personality disorders (310); sexual and gender disorders (302); physiological malfunction arising from mental disorder (306); special symptoms or syndromes not elsewhere classified (307); acute reactions to stress (308); adjustment reactions (309); specific nonpsychotic mental disorders due to brain damage (310); depressive disorder, not elsewhere classified (311); and disturbances of conduct, not elsewhere classified (312).
Third, a grouping of serious mental illness diagnoses was also created. Clinical research using serious mental illness constructs is not consistent regarding groupings of diagnostic codes (Ruggeri, Leese, Thornicroft, Bisoffi, & Tansella, 2000; Schinnar, Rothbard, Kanter, & Jung, 1990). Serious mental illness was originally defined by the Alcohol, Drug Abuse and Mental Health Administration (ADAMHA) in Public Law 102-321 and included any mental illness (other than substance abuse or “mental retardation”) lasting 12 months or more (Alcohol, Drug Abuse & Mental Health Administration [ADAMHA] Reorganization Bill, Public Law Number 102-321; Kessler et al., 1996). Some consensus exists about the inclusion of bipolar disorder, major depression, schizophrenia, and schizoaffective disorder in a serious mental illness grouping. There is less clarity about the inclusion of a wide spectrum of psychotic diagnoses (e.g., Codes 290, 293, 294, and 299; Kessler et al.). Although duration–persistence is a key component of serious mental illness, it is not measurable in a cross-sectional study (looking at only 12 months). Three articles used varied combinations of ICD-9 Codes 295–299 to define serious mental illness (Daumit, Pratt, Crum, Powe, & Ford, 2002; Ford, 1999; Fujii, Wylieb, & Nathan, 2004; Gianfrancesco Wang, & Yu 2005). Based on this work, a wide net was cast for serious mental illness identification in this study, including schizophrenia (295), affective psychoses (296), paranoid states (297), nonorganic psychoses (298), and psychoses with origin specific to childhood (299).
ICD-9-CM codes can be cross-referenced with the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV; American Psychiatric Association, 1994). Code 295 refers to various subtypes of schizophrenia, (e.g., schizophreniform, schizoaffective disorder). Code 296 includes diagnoses for various forms of what the DSM-IV refers to as Bipolar Disorders I and II and major depressive disorder. Code 297 includes diagnoses for the DSM-IV's psychotic states, delusional disorders and shared psychotic disorder. Code 298 includes diagnoses categorized in the DSM-IV as brief psychotic disorder and other psychotic disorder not otherwise specified. ICD-9-CM Code 299 includes diagnoses of what the DSM-IV refers to as autistic disorder, infantile autism, childhood disintegrative disorder, disintegrative psychosis, Asperger's disorder, other specified early childhood psychoses, pervasive developmental disorder not otherwise specified, and Rett's disorder.
Fourth, chronic substance abuse–related disorders were also extracted from Medicaid claims. These conditions are considered as a proxy indicator of longer term substance abuse. Based on a grouping of substance abuse–related diagnostic codes compiled by the Washington Circle, the following codes were used: alcoholic pellagra (265.2), alcohol polyneuropathy (357.5), polyneuropathy due to illicit drugs (357.6), alcoholic cardiomyopathy (425.5), and alcoholic fatty liver–cirrhosis of the liver or alcoholic liver damage (571.0).
Levels of Intellectual Disability
An examination of findings related to reported levels of intellectual disabilities in the present study's analytic file led to the decision against reporting findings by level of intellectual disability for reasons related to potential coding validity (Table 2). A majority of people in the intellectual disability–substance abuse group had mild or moderate intellectual disability, but, unexpectedly, 9.5% of this group were coded with severe or profound intellectual disability. Both diagnoses are suggestive of a low level of community involvement that would not be conducive to accessing alcohol or illicit drugs. These findings may stem from coding of intellectual disabilities by clinicians who were unfamiliar with how to differentiate between levels of intellectual disability. Only five beneficiaries were noted to have more than one “MR” diagnostic code in their claims files, in their case, both 317 and 319.
Taken in conjunction with the fact that 31% of people with intellectual disability–substance abuse had an intellectual disability diagnosis categorized as not otherwise specified, led to the decision to look at intellectual disabilities without consideration of diagnostic levels. However, researchers have posited that the identification of intellectual disability, regardless of the accuracy of level coding, is unlikely to represent a false positive. Other researchers, using Medicaid claims with ICD-9-CM codes for intellectual disabilities to examine the prevalence of HIV/AIDS have noted that “physician diagnosis (of intellectual disability ) may be more prone to false negatives than false positives” (Walkup, Sambamoorthi, & Crystal 1999, p. 357), especially for community-based Medicaid beneficiaries with intellectual disabilities.
People with intellectual disability and substance abuse constituted 2.6% of all people with intellectual disability (Table 3). On average, the sample was 2 years younger than the comparison group and 50% less likely to be categorized as White. Looking at each non-White racial and ethnic category individually, people with intellectual disability–substance abuse were equally likely to be Asian (0.4% vs. 0.5%) and more likely to be Black (28.5% vs. 17.1%, odds ratio [OR] = 1.9, p < .001), Latino/a (3.4% vs. 2.7%, OR = 1.2, p < .001), or Native American (1.6% vs. 0.6%, OR = 3.0, p < .001).
Clinical factors reported included a diagnosis of developmental delay, mental health diagnoses, a set of diagnoses categorized as serious mental illness, and a set of chronic, substance abuse–related disorders. First, regarding developmental delays, 4.1% (n = 389) people with intellectual disability and substance abuse had this diagnosis, at a rate that was slightly higher than their counterparts without substance abuse (n = 11,774 [3.3%], OR = 1.2, p < .001). However, people with intellectual disability–substance abuse were 1.3 times (p < .001) more likely to have a diagnosis of unspecified delays in development, a subtype (2.9%, n = 277 vs. 2.1%, n = 7,524). It is unclear whether health care providers in charge of coding procedures would distinguish between codes for “mental retardation” and the developmental delays listed here, given the common confusion about the diagnostic distinctions between the two as well as the vernacular usage of “developmental delay” to refer to intellectual disability.
Second, regarding mental health diagnoses, the hypothesis was supported, because the sample was four times more likely to have any mental health or serious mental illness diagnosis (52.9%, n = 5,014 vs. 19.1%, n = 68,061; OR = 4.7, p < .001; Table 4). People in the sample were more likely to have all types of mental health diagnoses measured here compared with their counterparts without substance abuse. Third, regarding serious mental illness diagnoses, the hypothesis was supported, as the sample was five times more likely to have any serious mental illness diagnosis (Table 5). Gender differences were noted in both groups without substance abuse diagnoses, the sample and comparison groups. Males in both groups, for example, were very much more likely to have been treated for psychoses with origin specific to childhood. Among people with intellectual disability–substance abuse, whereas females were more likely to have been treated for affective psychoses, males were more likely to have been treated for schizophrenia. Females with intellectual disability–without substance abuse diagnoses were slightly more likely to have all serious mental illness types. Fourth, with respect to diagnoses for chronic, substance abuse–related disorders, the hypothesis was supported, as the prevalence of such disorders was very low, with the largest number of people having a diagnosis of alcoholic fatty liver–alcoholic cirrhosis of the liver or alcoholic liver damage (Table 6).
A number of notable differences in the demographic and clinical characteristics of people with intellectual disability and substance abuse were noted compared with those with intellectual disability without substance abuse diagnoses. Overall, the signifcant findings suggest that compared with people with intellectual disability but without substance abuse diagnoses, people with intellectual disability and substance abuse are slightly younger, more likely to be male or non-White, and more likely to have either a mental illness or a serious mental illness. Results have particular relevance to mental health clinicians and case managers specializing in work with people with intellectual disabilities and mental illness or serious mental illness, vis-à-vis the identification of substance abuse and/or the prevention of substance abuse–related social problems.
Differences in Racial and Ethnic Categorizations
Findings related to racial and ethnic categorization must be interpreted with caution, as these results measure use of substance abuse treatment as opposed to overall rates of substance abuse. Therefore, two types of social factors could inform these results. First, these results could be indicative of the populations of people that are more likely to be referred to substance abuse treatment. Second, results specific to racial and/or ethnic status reflect the numbers of people living in urban areas, where substance abuse treatment programs are more plentiful, which could in turn interact with where the majority of Non-White Medicaid beneficiaries reside. Third, an argument could be made that increased rates of Non-White people who received substance abuse treatment could relate to disproportionate rates of criminal court involvement in these populations. A critical mass of U.S. criminal court cases have some alcohol or drug component, and, therefore, criminal defendants are often coursed through alternatives to incarceration and jail or prison diversion programs (CASA, 1998; Pettit & Western, 2004). Last, that there are higher proportions of non-White groups among the intellectual disability and substance abuse group is especially interesting, given that this population is noted to have more limited access to behavioral health treatment than the White population (Satcher, 2001).
Prevalence of Serious Mental Illness Among People With Intellectual Disability and Substance Abuse
Findings related to serious mental illness prevalence among members of the sample are consistent with a variety of existing higher bound prevalence estimates in the literature that suggest an overall higher rate of such disorders among people with intellectual disabilities in general. Estimates in the overall population with intellectual disabilities suggest prevalence rates between 7% and 26% (Horwitz, Kerker, Owens, & Zigler, 2000). There are no comparable studies of the prevalence of serious mental illnesses among people with intellectual disability and substance abuse compared with people with intellectual disability without substance abuse diagnoses, so comparisons were not possible. Whether populations of people with intellectual disabilities who access substance abuse treatment are demographically representative of all people with intellectual disabilities who may have substance abuse problems or be in need of substance abuse treatment is not known. People with an intellectual disability and substance abuse were more likely to be treated for schizophrenia, affective psychoses, paranoid states, and nonorganic psychoses. People with intellectual disability but without substance abuse diagnoses, however, were more likely to be treated for psychoses with origin specific to childhood, a category that includes disorders on the autism spectrum.
Gender differences were noted in both the sample and comparison groups. Males in both groups were very much more likely to have been treated for psychoses with origin specific to childhood. Among the sample, although females were more likely to have been treated for affective psychoses, males were more likely to have been treated for schizophrenia. Females with intellectual disability without substance abuse diagnoses were slightly more likely to have all serious mental illness types. This study's findings on serious mental illness among people with intellectual disability and substance abuse are consistent with a variety of existing prevalence estimates in the literature that suggest an overall higher rate of such disorders among people with intellectual disabilities.
Baseline Data on Chronic, Substance Abuse–Related Conditions
To my knowledge, the results in this study represent the first findings related to the prevalence of substance abuse–related conditions among people with intellectual disability and substance abuse. A review of the epidemiology of alcoholic liver diseases suggested that U.S. mortality from this disease generally occurs between ages 45 and 55 years (Mann, Smart, & Govoni, 2003). Although data used for the present study measure treated prevalence as opposed to mortality, an exploratory comparison was conducted. Although the ages of all people with this diagnosis in this data file fell within this range, people in the sample group were diagnosed with this condition in their mid-30s.
Two other factors might help to explain the earlier age onset of treatment for alcoholic liver disease. Research suggests that people with intellectual disabilities may be at increased risk for infection with hepatitis A and B, often in relation to poor hygiene. No identified research has examined hepatitis C in this population. although hepatitis A does not become chronic, hepatitis B does. Chronic hepatitis infection can increase the potential for liver damage, which might explain the rates noted here, although this theory could not be tested because information on hepatitis infection was not available in the data extract used for this file and administrative data are not most suitable for this type of analysis. Furthermore, given the higher rate of psychopathology among people with intellectual disability, there is a concurrently higher rate of prescribed psychotropic medication use in this population, such as antipsychotic neuroleptic medications, that can create acute liver damage, but not necessarily chronic liver damage. A cursory examination of this issue in the present study's data revealed that people with intellectual disability, substance abuse, and an alcoholic liver disease diagnosis were more likely to have a serious mental illness, for which they might be treated with such medications (45.9% vs. 17.8%; OR = 3.9, p < .01). This area of inquiry would be best examined in more detail through a study based on chart review.
Need for Identification of Potential Substance Abuse Problems
Among populations with intellectual disability, family, friends, and caregivers can use these data as a rough guide to considerations of the identification of substance abuse. Given concerns about early onset of substance abuse, teachers and ancillary workers in special education–related settings who work with youth with intellectual disabilities before they “age out” into adult systems of support and care will need to expand their capacity to screen for either substance use (to take advantage of “teachable moments”) or substance abuse in a population-specific manner. For adults, the daily staff who are the backbone of the intellectual disabilities system workforce (e.g., workers in shared living sites, supported work programs, or state intellectual disabilities agencies) should also develop basic competencies in screening and brief prevention intervention approaches for substance abuse in this population. Clinicians most commonly involved in the ongoing health care of people with intellectual disabilities who are living in community-based settings are also in a prime position to both screen as well as assess and refer to appropriate treatment agencies. Clinicians in this category include primary care physicians, nurse practitioners, psychologists, psychiatrists, and social workers. These clinicians should foster partnerships with state substance abuse agencies to learn more about this condition without assuming that it does not exist among people with intellectual disability.
Recommendations for Future Research
The baseline data presented here suggest the need for additional research in three specific areas. First, with respect to the use of claims data for tracking behavioral health conditions in this population, additional research using a triangulated data collection process involving both case record review and analyses of Medicaid claims data would be helpful in evaluating the feasibility and utility of this data source as an epidemiological surveillance tool. Although administrative data are not ideal for an epidemiological study on substance abuse among people with intellectual disability, as noted, given the exclusion of this population from most studies of substance abuse, these data are valuable for examining a hard-to-reach population and establishing baseline clinical and demographic data.
Second, although this study identifies clinical and demographic characteristics about people with intellectual disability and substance abuse diagnoses, evidence-based practices for screening and assessment of substance abuse are unavailable beyond a limited number of items on the Reiss Screen for Maladaptive Behavior (Reiss, 1997). This work could begin by conducting psychometric research on the modification of existing screening and assessment tools for people with intellectual disabilities. Third, given the dearth of evidence-based treatments for this population, process studies focused on substance abuse treatment could examine salient factors for clinicians to consider in matching people with intellectual disability to appropriate substance abuse treatment modalities, commonly referred to in the substance abuse treatment sector as treatment matching. Specifically, researchers need to study the applicability of treatment modalities or curricula for this population, in particular because of the prevalance of reported co-occurring diagnoses. Fourth, given the relatively small size of the population of people with intellectual disability and substance abuse diagnoses, arguments must be made for why such a small population should have priority. This could be achieved through the examination of costs of substance abuse–related care incurred by the sample compared with people with intellectual disabilities but without substance abuse diagnoses.
I thank Drs. Marty Wyngaarden Krauss, Constance Horgan, and Deborah Garnick at The Heller School for Social Policy and Management, Brandeis University as well as Dr. Richard Saitz (Boston University) and Dan Gilden (Jen Associates, Inc.) for support in designing and implementing this study at the dissertation stage. Additional thanks go to Rosemary Hakim (Centers for Medicare and Medicaid Services) for access to these data and general guidance. Support for this research included a predoctoral training grant from the National Institute on Alcohol Abuse and Alcoholism as well as an American Dissertation Fellow award from the American Association of University Women.
Editor-in-Charge: Steven J. Taylor
Elspeth Maclean Slayter, PhD (E-mail: email@example.com), Assistant Professor, Salem State University, School of Social Work, Salem, MA 01970.