Skip To Content

Go to the Table Of Contents

Click for DHHS Home Page
Click for the SAMHSA Home Page
Click for the OAS Drug Abuse Statistics Home Page
Click for What's New
Click for Recent Reports and HighlightsClick for Information by Topic Click for OAS Data Systems and more Pubs Click for Data on Specific Drugs of Use Click for Short Reports and Facts Click for Frequently Asked Questions Click for Publications Click to send OAS Comments, Questions and Requests Click for OAS Home Page Click for Substance Abuse and Mental Health Services Administration Home Page Click to Search Our Site

 Health Services Utilization by Individuals with Substance Abuse and Mental Disorders

Chapter 4. Do Client Characteristics Affect Admission to Inpatient Versus Outpatient Alcohol Treatment in Publicly Monitored Programs?

Sarah Q. Duffy, Ph.D.
Gary A. Zarkin, Ph.D.
Laura J. Dunlap, M.A.

Introduction

Alcohol use disorders cost the United States some 100,000 lives and $184.6 billion annually, and 14 million people meet the diagnostic criteria for alcohol abuse and alcoholism (National Institute on Alcohol Abuse and Alcoholism [NIAAA], 2000; Subcommittee on Health Services Research, 1997). The Nation spends approximately $6.1 billion per year on treatment for those with alcohol use disorders, 63 percent of which is funded by Federal, State, and local governments (Mark et al., 1999). Of the more than 1.5 million admissions annually to substance abuse treatment facilities in the United States, almost 50 percent list alcohol as the primary substance of abuse (Office of Applied Studies [OAS], 1999).

Substance abuse treatment policy is largely a State responsibility, especially since the establishment of the Federal Substance Abuse Prevention and Treatment (SAPT) block grant program in 1981 (Jacobsen & McGuire, 1996). States undertake treatment facility credentialing and licensing, and by 1997, through either their own funding or Federal funding that they managed, States and local governments managed more than 47 percent of all substance abuse funding and 74 percent of all public funding (Coffey et al., 2001).

Descriptive evidence suggests that substantial variations in treatment systems may exist across States. For example, in 1989, per capita alcohol treatment funding varied from $5.85 in Mississippi to $51.76 in Alaska (Dayhoff, Pope, & Huber, 1994). In 1998, the proportion of clients admitted to inpatient treatment varied from 3 percent of all substance abuse treatment clients in Vermont to 30 percent in North Dakota (OAS, 2000).

Still, one aspect of the publicly funded treatment system shared by many States is insufficient publicly funded treatment capacity (e.g., see New Jersey Substance Abuse Prevention and Treatment Advisory Task Force, 2001). One way in which many States attempt to improve care and make the best use of their limited resources is by implementing guidelines to help place clients receiving publicly funded treatment in different levels of care, including whether they are treated as inpatients or outpatients (Gastfriend & McLellan, 1997; Mattson, 2003). According to these guidelines, clients with more severe substance use, emotional, and behavioral disorders are candidates for inpatient care. The purpose of this study, using the administrative data that States use to monitor their treatment systems, is to estimate the effect of disorder severity on the odds of inpatient admission and to explore how that effect varies across States.

We extend the analysis of treatment admission to the publicly monitored treatment systems in several States. In doing so, we include either those who pay for treatment out-of-pocket or those who receive publicly funded treatment to explore the extent to which results for this augmented population are consistent with the findings reported in earlier research on privately insured individuals (Goodman, Holder, Nishiura, & Hankin, 1992; Goodman, Nishiura, & Hankin, 1998). Further, we examine whether variables, such as age of first intoxication, employment, and housing status, which research has found are associated with referral to inpatient treatment (Gregoire, 2000) but that are unavailable in insurance claims data, are correlated with inpatient admission. Finally, we examine the extent to which the estimated relationships vary across States. Our findings suggest that, although there are differences across States in client characteristics and in the effect of these characteristics on admission, clients with more severe substance use disorders generally are more likely to receive inpatient treatment. These results suggest that admission decisions in the State-monitored substance abuse treatment system conform, at least to some extent, to available placement criteria. Given the considerable barriers that can exist in implementing these criteria (Gastfriend, Lu, & Sharon, 2000; Kosanke, Magura, Staines, Foote, & DeLuca, 2002), our results suggest that States' attempts to manage their substance abuse treatment resources effectively are meeting with some success.

Background

Alcohol rehabilitation treatment is aimed at changing drinking behavior and often consists of psychotherapy and sometimes pharmacotherapy. It may take place in a number of settings, including outpatient and residential specialty substance abuse treatment facilities (including some in hospitals) or the offices of private practitioners. In our work, we examine data on clients in the specialty substance abuse treatment system that is monitored by State substance abuse treatment agencies. We exclude data on those who obtain care from private practitioners and those involved only in self-help groups, such as Alcoholics Anonymous, because data on admissions to such programs are not systematically collected.

On a per-episode basis, outpatient substance abuse rehabilitation treatment costs less than inpatient rehabilitation treatment. Costs for residential programs average between $4,000 and $6,800, depending on their length, or more than twice the $1,800 cost of the average outpatient program (U.S. General Accounting Office [GAO], 1998). Inpatient treatment also may be more disruptive and costly for clients than outpatient treatment. For example, employed clients who enter inpatient treatment must miss work, either losing pay or using sick leave. If both types of treatment were equally effective for all clients, providing treatment only in outpatient settings would be most efficient. However, if the two types of treatment are not equally effective, providing solely outpatient treatment may not be cost-effective.

That these treatment options may not be equally effective for all clients has been recognized in guidelines, such as the American Society of Addiction Medicine Patient Placement Criteria (ASAM-PPC) (ASAM, 1996). The ASAM-PPCs consider indicators across several psychosocial dimensions to determine optimal client placement and suggest that clients with emotional or behavioral disorders and complications, high risk for relapse, or a poor recovery environment may benefit from inpatient treatment (ASAM, 1996; McKay et al., 1997). Although not universally accepted, the ASAM-PPCs are the most widely distributed, implemented, discussed, and reviewed criteria available (Gartner & Mee-Lee, 1995; Mattson, 2003). Several States, such as Iowa, Colorado, Massachusetts, and New Jersey, either use the ASAM-PPCs or other similar criteria as guidelines for client placement. State modifications generally include adding treatment settings, such as halfway houses and longer-term residential treatment, not recognized in the original ASAM-PPCs (Gartner & Mee-Lee, 1995).

The results of several empirical studies conducted in the 1970s and 1980s, however, suggested that inpatient treatment may not have been worth the extra cost (Annis, 1985–1986; Miller & Hester, 1986). Miller and Hester (1986), for example, reviewed several controlled studies and concluded that few differences in outcomes arose between more intensive and less intensive programs, except in some cases where the less intensive programs produced superior outcomes. Such findings, coupled with the growth of managed behavioral health care, led to a decline in the number of inpatient admissions for substance abuse treatment throughout the early 1990s (Subcommittee on Health Services Research, 1997).

However, none of the controlled studies that Miller and Hester (1986) reviewed included individuals with a co-occurring mental disorder, an important clinical indicator for inpatient treatment (Pettinati, Meyers, Jensen, Kaplan, & Evans, 1993). Although one small randomized study suggested that inpatient treatment is not more effective for those who are appropriately matched to it (McKay et al., 1997), other observational studies and more recent reviews of the earlier controlled studies suggest that inpatient programs benefit those with more severe disorders (Finney, Hahn, & Moos, 1996; Finney & Moos, 1996; Gastfriend et al., 2000; Harrison & Asche, 1999; Hartmann, Sullivan, & Wolk, 1993; Mattson, 2003; Pettinatti et al., 1999; Simpson, Joe, Fletcher, Hubbard, & Anglin, 1999). The authors of these studies concluded that inpatient substance abuse treatment should remain an option.

Two studies by Goodman et al. (1992, 1998) used private insurance claims data to examine factors affecting the choice between inpatient and outpatient substance abuse treatment. The first study examined data on 879 individuals with employer-sponsored, fee-for-service health insurance with comprehensive alcoholism coverage. The authors found that admission to short-term inpatient treatment was more likely for those with a diagnosis of alcohol dependence (vs. abuse) and a co-occurring substance use and mental disorder. The second study analyzed the relative contributions of client- and employer-level factors to treatment choice by examining claims submitted on behalf of 9,878 individuals who received their health insurance through 10 large self-insured firms from 1989 to 1991. The authors found that clients were more likely to receive inpatient treatment if they had a diagnosis of dependence (vs. abuse) or a psychosis, used drugs other than opiates, were younger or male, and received hourly wages as opposed to salaries. However, a large part of the observed variation occurred at the employer level. The authors concluded that treatment choice was driven mainly by firm or health insurance administrator policy (they could not distinguish between the two) and that treatment expenditures at some firms could be reduced by shifting treatment to outpatient settings. Although these results are suggestive, they are not generalizable to those who receive treatment in the publicly monitored system, many of whom not only have no employer-sponsored health insurance, but also often are unemployed and/or homeless.

A study by Gregoire (2000) provides clues about this population. The study examined referrals to inpatient versus outpatient substance abuse treatment in a study of 3,093 individuals diagnosed as drug dependent who sought admission to publicly funded treatment in Wichita, Kansas. Although the study revealed that referrals generally were consistent with clinical criteria, the two variables that were the strongest predictors of referrals to inpatient treatment were housing and employment status. Those who were homeless and unemployed were more likely to be referred to inpatient treatment than those with stable housing and those who were employed. Although these results are suggestive, the study has some limitations. First, it examined only drug-dependent clients in a single city. Second, it concerned referrals, not admissions. Because research suggests that substantial numbers of clients fail to attend treatment as referred (e.g., see Donovan, Rosengren, Downey, Cox, & Sloan, 2001), and many clients self-refer into treatment, it is of interest to see whether these same client characteristics affect treatment choice in a broader admissions sample.

Data

We used data from the 1996 Treatment Episode Data Set (TEDS), maintained by the Office of Applied Studies (OAS) of the Substance Abuse and Mental Health Services Administration (SAMHSA) (OAS, 1999). TEDS contains admissions data routinely collected by treatment providers at client admission and sent to State agencies, which use them to monitor their substance abuse treatment systems. These State data systems, which were enacted to satisfy the mandate to collect client data in the Comprehensive Alcohol Abuse, Drug Abuse, and Mental Health Amendments (1988), were designed with input from each State's treatment providers and with input and funding from SAMHSA. The data are submitted at regular intervals by the States in a common format to SAMHSA. The data include disorder severity information important in determining clients' treatment needs, as well as socioeconomic measures. Although these data have been used by the States and the Federal Government to generate descriptive reports, they have seldom been used for health services research (McCarty, McGuire, Harwood, & Field, 1998).

Our analysis focuses on adult males with alcohol as their primary substance of abuse. We did not include women because a variable that might be relevant to their treatment setting choice, whether or not they are pregnant, is not well reported, and another variable, whether or not they have dependent children, is not collected. We examine only alcohol clients because a relevant measure of the disorder severity for most other drugs, route of administration, also is not consistently reported.

Unfortunately, States vary in their ability to report all variables or collect data from all substance abuse treatment facilities in the TEDS universe (those receiving public substance abuse treatment funding). Therefore, we focus on nine States (Colorado, Iowa, Maine, Massachusetts, Nevada, New Jersey, New York, North Dakota, and Rhode Island) that provided data covering 90 percent or more of their estimated substance abuse treatment clients in programs receiving public funds in 1996 and that collected variables hypothesized to affect substance abuse treatment admission.

Although these data are fairly consistent across States, our review of information from each State, such as the instruction manuals that States give to providers, data collection forms, and the crosswalk between the State data systems and SAMHSA's common format, reveals that there are some differences. One important difference is the universe of reporting facilities. In some States, such as New York, all alcohol treatment facilities are required to report admissions data, regardless of whether they receive public funding or not. In North Dakota, in contrast, only the programs at the State's eight regional services centers and the State hospital report these data. In other States, such as New Jersey, only facilities receiving public funds are required to report these data, but many more do so voluntarily.

Another important difference among States is the definition of an admission. Although SAMHSA requests that States report only the initial admission to a treatment episode as an admission (OAS, 1999), the nature of the help-seeking behavior of those with substance use disorders can make it difficult for States to comply. Fewer than 53 percent of substance abuse treatment clients nationwide complete their planned treatment, and it is common for those with substance use disorders to make more than one attempt at treatment (OAS, in press). The question then becomes, when is an admission a new admission, as opposed to a continuation of the same treatment episode? States do not provide uniform instructions to providers. Iowa, for example, instructs providers to report an admission as an initial admission only if 2 months or more have passed since the individual's last discharge; in Nevada, the relevant time period is 30 days. And although SAMHSA requests that states report changes of service (e.g., from detoxification to rehabilitation) within an episode as a transfer rather than a new admission, four of the nine States we include in our analysis do not (Maine, Massachusetts, New York, and Rhode Island). As discussed below, these differences among States are one reason we chose to estimate the model separately for each State.

Empirical Framework

Rather than rely on a standard model of health care demand, such as the Health Capital Model (Grossman, 1972; Muurinen, 1982) or the Behavioral Healthcare Model (Andersen & Newman, 1973), which have each been used in studies of the demand for behavioral health care (Haas-Wilson, Cheadle, & Scheffler, 1989; Pottick, Hansell, Gutterman, & White, 1995), we combine elements of both approaches with unique characteristics of the substance abuse treatment system and its clients to inform our empirical specification. This exploratory approach is consistent with recent calls to integrate both behavioral and economic variables in empirical behavioral health services research (Brito & Strain, 1996; Green & Kagel, 1996; Montoya, Atkinson, & Trevino, 2000).

Baseline Model

We model desired alcohol treatment intensity as an underlying, unobserved, continuous dependent variable y* for which we have a discrete realization, yi, that equals 1 for inpatient admission and 0 for outpatient admission. The individual's observed treatment setting is a function of his demand for alcohol treatment, which is a function of his disorder severity and other characteristics, and the availability of treatment options, which is a function of State treatment policy. Given this, and the differences in the data systems described above, we estimate the model separately for each State. Importantly, we rejected the null hypothesis of a single pooled model based on the results of a chi-square test.

The probability that client i in State j is admitted to inpatient treatment is

equation ,     D  (1)

where yij equals 1 if individual i in State j is admitted to inpatient care and equals 0 otherwise; f(•) is the logistic function; image representing alphaj is the intercept for State j; Xik is a vector of k demand variables and client characteristics; Riimage representing script l is a vector of image representing script l referral source indicators; and the image representing betas are parameters to be estimated.

Virtually all health care demand models indicate that problem severity affects the intensity of care demanded. We use several variables to measure the severity of the client's substance use disorder (hypothesized effect in parentheses). Frequency of use (+): We include four dichotomous variables that reflect how this information is coded on each client's record at admission: daily use, use three to six times in the past week, use one to two times in the past week, and use one to three times in the past month. No use in the past month is the reference category. Although other studies have used International Classification of Diseases (ICD) or Diagnostic and Statistical Manual of Mental Disorders (DSM) code information to measure severity of substance use disorders, we believe the frequency of use variable, in combination with other variables in our model, is more appropriate for our purpose, especially given that it is better reported. Only 18 of the more than 50 States and jurisdictions that report to TEDS collect ICD or DSM data (OAS, 1999). Furthermore, of those 18 States, only 3 obtain valid values on 99 percent or more of their admissions. Frequency of use, in contrast, is collected in 47 States and jurisdictions, some 32 of which obtain it on 99 percent or more of their admissions. The fact that frequency of use is much better reported may mean that it is easier for treatment personnel to collect, and, given that they had input into the data elements that would be collected by States, perhaps more useful to them in their treatment planning decisions than DSM or ICD criteria.

We include several other variables to measure client severity as well. Intoxication before age 15 (+): Research suggests that individuals who first use alcohol before the age of 15 are more likely to become alcohol dependent (Grant & Dawson, 1997). We include a dichotomous variable indicating first alcohol intoxication before age 15. Secondary substance (+): Having a second substance of abuse can indicate a more severe disorder. We include dichotomous variables indicating marijuana/hashish, cocaine, and other secondary substance use, with no secondary substance use as the reference category. Number of prior treatment episodes (?): This variable is used in many treatment studies as an indicator of disorder severity (e.g., Etheridge, Craddock, Hubbard, & Rounds-Bryant, 1999; McLellan et al., 1999). We enter this as a categorical variable to allow the estimated relationship to be something other than linear. We include indicator variables for one prior treatment and two or more prior treatments, using no prior treatment as the reference category. Co-occurring mental disorder (+), Homeless (+): Based on the ASAM-PPCs, the literature on treatment effectiveness, and the results of prior research on referrals (Gregoire, 2000) and treatment matching (Kosanke et al., 2002), we expect that those with a co-occurring mental disorder and those who are homeless may have a higher likelihood of admission to inpatient treatment. To capture this, we enter two indicator variables, one for co-occurring mental disorder and the other for homelessness.

The research mentioned earlier and other economic research (Becker & Murphy, 1988) conceptualize the behavior of those with substance use disorders as consistent with choice theory, suggesting that socioeconomic and other client characteristics should be included in models that predict their behavior. We include the following variables and note that potentially offsetting effects render the predicted direction of many of the effects uncertain. Employment status (?): Employed individuals have higher time cost associated with inpatient treatment and may be less likely to engage in inpatient treatment, other things equal (Kosanke et al., 2002). Furthermore, the fact that they are employed suggests that they may have a less severe disorder, and treatment providers may believe that those who are unemployed are more likely to benefit from inpatient treatment (Gregoire, 2000). On the other hand, employed individuals may be better able to pay for more intensive treatment. To examine which of these hypotheses the data support, we include a dichotomous variable indicating whether the client was employed at admission, either full- or part-time, with those who are unemployed or not members of the labor force comprising the reference category. Education level (?): According to Muurinen (1982), the relationship between years of education and the demand for medical care should be negative because the rate of depreciation of the health stock should be lower for better-educated individuals. At the same time, education may proxy higher income, which may suggest a more intensive treatment choice. We included two dichotomous variables to measure education: less than high school graduate and high school graduate. Some postsecondary education is the reference category. Age (?): Human capital theory suggests that age has a positive effect on treatment intensity because the rate of depreciation of the health stock is a positive function of age (Muurinen, 1982). On the other hand, one version of the model by Suranovic, Goldfarb, and Leonard (1999) suggests that those who are older are more motivated to quit their substance use, perhaps making it less likely that they would need intensive treatment. Because it is unclear which effect may dominate, we enter age and age-squared to allow for a nonlinear relationship. Race/ethnicity (?): According to the 4th edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV), people of different races and ethnicities tend to have different cultural attitudes about and physiological responses to alcohol (American Psychiatric Association [APA], 1994). However, this variable also may capture placement in a treatment setting that was not clinically indicated because of the lack of culturally competent alternatives in the individual's area (Gartner & Mee-Lee, 1995). We include three indicator variables to capture the client's race/ethnicity (Hispanic, non-Hispanic white, and non-Hispanic black) with "other" as the reference category. This was the only way to code this information uniformly across the States included in our sample. Marital status (-): Although imperfect, marital status may proxy for the ASAM-PPCs' emotional and behavioral disorders criteria. Currently married clients may have less severe emotional and behavioral disorders than those who are single, divorced, separated, or widowed, all of whom comprise the reference cell.

Another variable that may affect client placement is season of admission (?): The time of year the client is admitted also may affect the odds of inpatient treatment. For example, those who are seasonally employed, such as teachers and college professors, may be more likely to accept assignment to inpatient treatment in the summer. We include indicator variables for summer, fall, and winter, with spring as the reference category.

Finally, we include indicator variables for referral source (?). Although the process by which clients obtain referrals, and the interplay between the various referral sources and the treatment system is admittedly complex (e.g., see Kosanke et al., 2002), we believe it is important to include referral source as a control variable. For example, although the criminal justice system is a frequent source of referral into substance abuse treatment, and referral through that system may affect client placement, we make no a priori judgment about whether criminal justice referrals are more or less likely to be admitted to inpatient alcohol treatment. The effect likely depends on the referral practices of criminal justice systems, and the availability of different types of treatment, which vary across States. For example, in some States, such as New Jersey, the State substance abuse treatment agency is actively involved in assessing prerelease inmate and parolee needs and referring clients to treatment (New Jersey Substance Abuse Prevention and Treatment Advisory Task Force, 2001). In others, the court may mandate both treatment and the modality, as the Massachusetts court does for second-time drunk driving offenders (Bureau of Substance Abuse Services, 2001). We include indicator variables for self-referral, referral by an alcohol or drug treatment provider, other medical provider referrals, and community (employer, school, etc.) referral. Referral by the criminal justice system serves as the reference cell.

Expected Payer Model

In addition to the baseline model, we present coefficients from a model including expected pay source for this admission and estimate it for each of the seven States that collected these data. Economic theory suggests that individuals who pay out of pocket for their own treatment may demand less costly treatment than those who do not. However, we only observe the expected payer for this particular admission and do not know, for example, whether the client's insurance covered both inpatient and outpatient treatment. In some cases, expected payer and treatment setting may be jointly determined if, for example, an indigent client is placed in an inpatient treatment program because no publicly funded outpatient slots are available (Gartner & Mee-Lee, 1995). Therefore, the direction of the expected payer effects is unclear a priori, and the results of the analysis should be considered tentative. We include the following categories: self-pay, Medicare, Medicaid, private, other (e.g., worker's comp), and other government funding/no charge (the reference category). The reference category includes clients whose treatment is funded by State agency funds, including those received through the Federal SAPT block grant program.

Results

Table 4.1 displays the means and standard deviations for the combined sample and by State. It shows that States varied substantially in the proportion of adult males in treatment for alcohol use disorders admitted to inpatient care, ranging from about 13 percent in Iowa to 32 percent in New York. Statistically significant differences exist across States for all variables, except for no secondary substance use and high school education.

Table 4.1 Analysis Sample Means and Standard Deviations of the Model Variables, by State and for All Nine States Pooled
  All States Colorado Iowa Maine Massachusetts Nevada New Jersey New York North Dakota Rhode Island
N 113,948 4,654 17,591 3,762 10,797 1,615 10,426 62,093 1,028 1,982
Inpatient Treatment 0.255* 0.300 0.126 0.150 0.162 0.305 0.212 0.320 0.252 0.161
  (0.436) (0.458) (0.332) (0.357) (0.368) (0.461) (0.409) (0.467) (0.434) (0.368)
Frequency of Useimage representing a dagger denoting a footnote
No use in the past month 0.324* 0.358 0.389 0.532 0.055 0.363 0.218 0.353 0.332 0.324
(0.468) (0.480) (0.488) (0.499) (0.228) (0.481) (0.413) (0.478) (0.471) (0.468)
1 to 3 times in the past month 0.145* 0.169 0.226 0.112 0.110 0.172 0.121 0.130 0.245 0.155
(0.352) (0.375) (0.418) (0.315) (0.313) (0.377) (0.326) (0.336) (0.430) (0.362)
1 to 2 times in the past week 0.125* 0.130 0.125 0.042 0.239 0.107 0.186 0.099 0.121 0.192
(0.331) (0.337) (0.331) (0.200) (0.427) (0.309) (0.389) (0.298) (0.326) (0.394)
3 to 6 times in the past week 0.118* 0.137 0.109 0.157 0.193 0.102 0.138 0.100 0.128 0.121
(0.322) (0.344) (0.312) (0.364) (0.395) (0.302) (0.345) (0.300) (0.335) (0.326)
Daily 0.288* 0.206 0.151 0.157 0.402 0.257 0.337 0.318 0.174 0.208
(0.453) (0.404) (0.358) (0.364) (0.490) (0.437) (0.473) (0.466) (0.379) (0.406)
Age of First Intoxication Less Than 15 Years 0.387* 0.394 0.345 0.443 0.415 0.359 0.339 0.396 0.517 0.426
(0.487) (0.489) (0.475) (0.497) (0.493) (0.480) (0.473) (0.489) (0.500) (0.495)
Secondary Drugimage representing a dagger denoting a footnote
None 0.484 0.597 0.592 0.624 0.510 0.661 0.587 0.409 0.569 0.458
(0.500) (0.490) (0.491) (0.484) (0.500) (0.473) (0.492) (0.492) (0.495) (0.498)
Marijuana/hashish 0.229* 0.249 0.302 0.308 0.259 0.160 0.171 0.206 0.357 0.259
(0.420) (0.432) (0.459) (0.462) (0.438) (0.367) (0.376) (0.404) (0.479) (0.438)
Cocaine/crack 0.229* 0.102 0.036 0.036 0.182 0.070 0.184 0.328 0.018 0.211
(0.420) (0.302) (0.187) (0.186) (0.386) (0.255) (0.388) (0.470) (0.135) (0.408)
Other 0.058* 0.052 0.070 0.031 0.049 0.108 0.058 0.057 0.055 0.072
(0.235) (0.222) (0.255) (0.174) (0.215) (0.311) (0.234) (0.233) (0.229) (0.258)
Prior Treatment Episodesimage representing a dagger denoting a footnote
None 0.325* 0.307 0.415 0.312 0.287 0.547 0.502 0.270 0.304 0.401
(0.468) (0.461) (0.493) (0.463) (0.453) (0.498) (0.500) (0.444) (0.495) (0.490)
1 episode 0.249* 0.218 0.287 0.285 0.241 0.289 0.251 0.239 0.215 0.258
(0.432) (0.413) (0.452) (0.452) (0.427) (0.453) (0.434) (0.426) (0.411) (0.438)
2 episodes or more 0.426* 0.475 0.299 0.403 0.472 0.164 0.246 0.491 0.482 0.341
(0.495) (0.499) (0.458) (0.491) (0.499) (0.370) (0.431) (0.500) (0.500) (0.474)
Demographics
Age 35.770* 34.165 34.293 35.401 35.948 36.692 36.152 36.237 34.993 35.391
(10.359) (10.218) (10.675) (10.308) (10.764) (9.923) (10.461) (10.153) (11.464) (9.638)
Employed 0.447* 0.526 0.651 0.455 0.476 0.490 0.562 0.357 0.399 0.457
(0.497) (0.499) (0.477) (0.498) (0.499) (0.500) (0.496) (0.479) (0.489) (0.498)
Homeless 0.124* 0.109 0.014 0.065 0.057 0.209 0.045 0.185 0.093 0.080
(0.330) (0.312) (0.116) (0.247) (0.231) (0.407) (0.208) (0.388) (0.291) (0.272)
Mental disorders 0.174* 0.170 0.158 0.201 0.265 0.051 0.085 0.179 0.390 0.099
(0.379) (0.376) (0.365) (0.401) (0.441) (0.221) (0.279) (0.383) (0.488) (0.299)
Married 0.246* 0.244 0.338 0.206 0.191 0.302 0.253 0.231 0.208 0.221
(0.431) (0.429) (0.473) (0.404) (0.393) (0.459) (0.435) (0.421) (0.406) (0.415)
Education Levelimage representing a dagger denoting a footnote
No high school 0.318* 0.309 0.204 0.302 0.303 0.291 0.265 0.362 0.244 0.395
(0.466) (0.462) (0.403) (0.459) (0.460) (0.454) (0.442) (0.481) (0.430) (0.489)
High school 0.466 0.448 0.565 0.511 0.478 0.473 0.502 0.430 0.471 0.413
(0.499) (0.497) (0.496) (0.500) (0.500) (0.499) (0.500) (0.495) (0.499) (0.492)
Post high school 0.216* 0.243 0.230 0.188 0.219 0.236 0.233 0.208 0.281 0.192
(0.412) (0.429) (0.421) (0.390) (0.413) (0.425) (0.423) (0.406) (0.450) (0.394)
Race/Ethnicityimage representing a dagger denoting a footnote
Non-Hispanic white 0.674* 0.602 0.904 0.957 0.778 0.638 0.687 0.573 0.729 0.819
(0.469) (0.490) (0.295) (0.203) (0.416) (0.481) (0.464) (0.495) (0.445) (0.385)
Non-Hispanic black 0.191* 0.063 0.044 0.012 0.093 0.071 0.196 0.279 0.007 0.093
(0.393) (0.242) (0.205) (0.109) (0.291) (0.256) (0.397) (0.448) (0.082) (0.290)
Hispanic 0.106* 0.293 0.034 0.007 0.099 0.064 0.103 0.126 0.011 0.063
(0.309) (0.455) (0.182) (0.081) (0.298) (0.246) (0.304) (0.331) (0.103) (0.242)
Other 0.027* 0.043 0.018 0.024 0.029 0.225 0.014 0.023 0.254 0.026
(0.164) (0.202) (0.133) (0.154) (0.167) (0.418) (0.118) (0.149) (0.439) (0.158)
Primary Source of Referralimage representing a dagger denoting a footnote
Individual 0.176* 0.177 0.176 0.173 0.201 0.146 0.193 0.168 0.196 0.215
(0.381) (0.382) (0.381) (0.378) (0.400) (0.353) (0.394) (0.374) (0.397) (0.411)
Alcohol/drug treatment provider 0.239* 0.150 0.072 0.150 0.177 0.063 0.099 0.343 0.094 0.109
(0.426) (0.358) (0.259) (0.357) (0.381) (0.243) (0.299) (0.475) (0.292) (0.312)
Other health care provider 0.078* 0.052 0.068 0.073 0.083 0.050 0.101 0.080 0.094 0.060
(0.269) (0.222) (0.251) (0.260) (0.276) (0.218) (0.301) (0.272) (0.292) (0.237)
School, employer, community 0.086* 0.082 0.046 0.052 0.069 0.110 0.075 0.101 0.205 0.098
(0.280) (0.275) (0.210) (0.222) (0.253) (0.313) (0.263) (0.301) (0.404) (0.297)
Criminal justice 0.422* 0.538 0.637 0.552 0.471 0.631 0.532 0.308 0.410 0.518
(0.494) (0.499) (0.481) (0.497) (0.499) (0.483) (0.499) (0.462) (0.492) (0.500)
Season Entering Treatmentimage representing a dagger denoting a footnote
Spring 0.265* 0.259 0.322 0.278 0.257 0.221 0.259 0.251 0.264 0.305
(0.441) (0.438) (0.467) (0.448) (0.437) (0.415) (0.438) (0.433) (0.440) (0.460)
Summer 0.238* 0.257 0.190 0.227 0.249 0.285 0.255 0.246 0.240 0.228
(0.426) (0.437) (0.392) (0.419) (0.432) (0.452) (0.436) (0.431) (0.427) (0.420)
Fall 0.239* 0.239 0.188 0.234 0.244 0.323 0.230 0.253 0.240 0.229
(0.427) (0.427) (0.391) (0.423) (0.430) (0.468) (0.421) (0.435) (0.427) (0.420)
Winter 0.257* 0.244 0.300 0.261 0.250 0.171 0.256 0.250 0.255 0.239
(0.437) (0.430) (0.458) (0.439) (0.433) (0.377) (0.437) (0.433) (0.437) (0.426)
Expected Source of Paymentimage representing a dagger denoting a footnote
N 40,824 4,654 17,591 3,762 NA 1,615 10,414 NA 1,028 1,760
Self-pay 0.200* 0.402 0.065 0.175 NA 0.515 0.322 NA 0.121 0.099
(0.400) (0.490) (0.247) (0.326) NA (0.500) (0.467) NA (0.326) (0.300)
Private 0.141* 0.028 0.159 0.100 NA 0.057 0.185 NA 0.109 0.183
(0.348) (0.165) (0.366) (0.301) NA (0.233) (0.389) NA (0.312) (0.387)
Medicare 0.018* 0.007 0.016 0.012 NA 0.001 0.030 NA 0.032 0.003
(0.132) (0.083) (0.127) (0.108) NA (0.024) (0.172) NA (0.176) (0.058)
Medicaid 0.050* 0.005 0.053 0.125 NA 0.005 0.033 NA 0.051 0.135
(0.219) (0.069) (0.223) (0.331) NA (0.070) (0.179) NA (0.219) (0.342)
Other 0.044* 0.111 0.003 0.183 NA 0.032 0.042 NA 0.042 0.003
(0.206) (0.314) (0.059) (0.387) NA (0.175) (0.201) NA (0.200) (0.053)
Other Government/
No Charge
0.541* 0.437 0.703 0.405 NA 0.396 0.387 NA 0.478 0.577
(0.498) (0.496) (0.457) (0.491) NA (0.489) (0.487) NA (0.500) (0.494)
Unknown 0.005* 0.010 0.000 0.000 NA 0.001 0.000 NA 0.168 0.000
(0.074) (0.101) 0.000 0.000 NA (0.025) 0.000 NA (0.374) 0.000
* Statistically significant differences in means among States at the 0.01 level.
image representing a dagger denoting a footnote Statistically significant differences in distributions among States at the 0.01 level using Pearson's chi square.
NA = not available (States do not collect the variable).
Note: Standard deviations are included in parentheses.
Source: SAMHSA, Office of Applied Studies, Treatment Episode Data Set, 1996.

Almost 30 percent of clients in our sample reported using alcohol daily at admission. The percentage ranged from 15 percent in Iowa to 40 percent in Massachusetts. Another third of clients reported no use in the past month, ranging from 6 percent in Massachusetts to 53 percent in Maine. The percentage of clients who reported having been first intoxicated before age 15 was somewhat less variable across States, averaging about 39 percent and ranging from 34 percent in New Jersey to 52 percent in North Dakota. Almost half of the clients did not report any secondary substance use. Marijuana was the most common secondary substance of abuse in most States, ranging from 16 percent of clients in Nevada to 36 percent in North Dakota. However, in New Jersey and New York, cocaine (including crack) was the most commonly reported secondary substance. In all States but Nevada and New Jersey, most clients had at least one prior treatment episode.

Only 17 percent of clients across States indicated a mental disorder; however, this percentage ranged from 5 percent of clients in Nevada to 39 percent in North Dakota. Homelessness also varied across States, ranging from 1.4 percent in Iowa to almost 21 percent in Nevada.

The average client in the dataset was almost 36 years old, and the majority of clients were non-Hispanic white (67 percent overall). Although age did not vary substantially across States, race/ethnicity did: The percentage of non-Hispanic white clients ranged from 57 percent in New York to 96 percent in Maine. Almost 70 percent of clients in our sample had at least a high school education, and almost 45 percent were employed. Fewer than 25 percent were currently married.

The criminal justice system was the most common route of treatment referral for clients in all States except New York, ranging from 31 percent in New York to almost 64 percent in Iowa. The most common route of referral in New York was an alcohol or drug treatment provider (34 percent). In all States but New York and North Dakota, self-referral was the second most common referral route.

Baseline Models

Table 4.2 shows the coefficients from our baseline logit models. Clients with more severe alcohol disorders, as measured by TEDS data, generally were more likely to receive inpatient alcohol treatment. However, there were differences across the States in the estimated magnitudes. Daily alcohol users were significantly more likely to receive inpatient treatment than clients who did not use in the 30 days prior to admission (omitted category). In Colorado, clients who reported daily alcohol use in the past 30 days were twice as likely to enter inpatient treatment as clients who reported no use in the past 30 days (based on eimage representing beta, which gives the effect of a one-unit change in the independent variable on the odds of inpatient treatment, where image representing beta is the estimated coefficient). In the other States exhibiting this relationship, the increases in the odds due to daily use were much larger (e.g., 4.5 in New Jersey, 7.2 in New York, 11.1 in Iowa). The exception is Massachusetts, where the frequency of use variables were not statistically significant. Reporting cocaine as a secondary substance of abuse increased the odds of inpatient admission in all States except North Dakota (which, as shown in Table 4.1, also has the lowest proportion of clients, 2 percent) and Massachusetts. Effects ranged from 1.44 in Iowa to 4 in Rhode Island. On the other hand, marijuana/hashish as a secondary substance significantly increased the odds of inpatient admission in three of the nine States (Colorado, Maine, and New York) and decreased it in two others (Iowa and North Dakota). Reporting other drugs as a secondary substance increased the odds of inpatient admission in four of the nine States (Colorado, New Jersey, New York, and Rhode Island) and was insignificant in the other States. Another severity measure, age of first intoxication younger than 15, significantly increased the odds of inpatient admission in three of nine States (Maine, Massachusetts, and New York) and was insignificant in all others. Having one prior treatment episode increased the odds of admission to inpatient treatment compared with having no prior treatment in the six States for which it was significant (Iowa, Massachusetts, Nevada, New Jersey, New York, and Rhode Island). Effects ranged from 1.2 in New Jersey to 6.06 in Massachusetts. Having two or more prior treatment episodes also increased the odds of inpatient treatment for clients in Colorado, Iowa, Massachusetts, New York, and Rhode Island. Effects ranged from 1.3 in Colorado to 2.8 in Massachusetts. Having two or more prior treatment episodes was not significant in the other four States.

Table 4.2 Probability of Seeking Inpatient Alcohol Abuse Treatment: Coefficients and Standard Errors from Baseline Logit Models
  Colorado Iowa Maine Massachusetts Nevada New Jersey New York North Dakota Rhode Island
N 4,654 17,591 3,762 10,797 1,615 10,426 62,093 1,028 1,982
Frequency of Use
1 to 3 times in the past month -0.358*** 0.403*** -0.311 0.109 0.537** -0.605*** 0.005 1.922*** -0.012
(0.127) (0.090) (0.245) (0.149) (0.236) (0.169) (0.043) -0.537 (0.284)
1 to 2 times in the past week 0.152 0.926*** -1.017* 0.239* 0.701** -0.249* 0.670*** 2.853*** -0.198
(0.127) (0.097) (0.572) (0.136) (0.271) (0.133) (0.044) (0.662) (0.320)
3 to 6 times in the past week 0.334*** 1.703*** -0.375* 0.018 1.183*** 0.686*** 1.258*** 2.760*** 0.551**
(0.119) (0.088) (0.202) (0.142) (0.248) (0.112) (0.038) (0.635) (0.274)
Daily 0.767*** 2.403*** 0.368** -0.043 2.095*** 1.507*** 1.980*** 3.263*** 1.570***
(0.108) (0.078) (0.151) (0.140) (0.194) (0.093) (0.028) (0.600) (0.212)
Age of First Intoxication
Less Than 15 Years
0.136* 0.107* 0.307** 0.167** 0.032 0.050 0.064*** 0.266 0.269
(0.081) (0.057) (0.135) (0.066) (0.154) (0.063) (0.023) (0.376) (0.171)
Secondary Drug
Marijuana/hashish 0.253*** -0.278*** 0.350** -0.136* -0.112 -0.113 0.273*** -0.824** 0.169
(0.094) (0.065) (0.148) (0.075) (0.209) (0.091) (0.034) (0.389) (0.249)
Cocaine/crack 0.651*** 0.369*** 0.991*** 0.162* 0.855*** 0.452*** 0.636*** -0.031 1.407***
(0.123) (0.118) (0.278) (0.089) (0.291) (0.079) (0.029) (1.108) (0.213)
Other 0.511*** 0.170* 0.059 -0.166 0.289 0.616*** 0.489*** -1.419 1.732***
(0.162) (0.096) (0.299) (0.163) (0.227) (0.113) (0.047) (0.917) (0.269)
Prior Treatment Episodes
1 episode 0.094 0.230*** 0.032 1.803*** 0.545*** 0.175** 0.733*** -0.360 0.481**
(0.114) (0.068) (0.215) (0.092) (0.169) (0.076) (0.037) (0.588) (0.240)
2 episodes or more 0.236** 0.428*** 0.336* 1.028*** 0.306 0.001 0.829*** -0.814 0.972***
(-0.101) (0.066) (0.189) (0.096) (0.208) (0.076) (0.034) (0.514) (0.209)
Demographics
Age 0.153*** 0.037*** 0.037 0.091*** 0.058 -0.027* -0.040*** -0.074 -0.005
(0.022) (0.013) (0.033) (0.017) (0.046) (0.016) (0.006) (0.091) (0.060)
Age squared -0.002*** -0.001*** 0.000 -0.001*** -0.001 0.000 0.000*** 0.001 0.000
(0.000) (0.000) (0.000) (0.000) (0.001) (0.000) (0.000) (0.001) (0.001)
No high school education -0.020 0.221*** -0.554*** -0.547*** 0.047 0.239*** -0.079** 0.125 -1.088***
(0.112) (0.082) (0.181) (0.087) (0.214) (0.088) (0.031) (0.506) (0.000)
High school education 0.149 0.112* -0.527*** -0.256*** 0.128 0.086 0.005 0.351 -0.309
(0.097) (0.067) (0.165) (0.074) (0.181) (0.078) (0.030) (0.455) (0.212)
Employed -1.608*** -0.995*** -1.870*** -0.043 -2.085*** -0.923*** -0.856*** -0.645 -0.970***
(0.082) (0.056) (0.212) (0.065) (0.166) (0.066) (0.029) (0.411) (0.205)
Non-Hispanic white 0.145 -0.830*** -0.609* 0.433** 0.205 -0.266 0.024 -0.632 0.555
(0.189) (0.153) (0.322) (0.194) (-0.192) (0.270) (0.077) (0.397) (0.540)
Non-Hispanic black -0.091 -0.673*** 0.074 -0.473** -0.003 -0.500* -0.043 -1.449 0.246
(0.239) (0.188) (0.635) (0.232) (-0.323) (0.277) (0.079) (2.112) (0.583)
Hispanic -0.032 -0.811*** -3.073*** 0.291 -2.334*** -0.167 -0.347*** 0.384 0.243
(0.196) (0.215) (1.158) (0.220) (0.589) (0.286) (0.083) (2.867) (0.646)
Homeless 0.508*** 1.748*** 3.346*** 1.188*** 0.354* 1.699*** 0.256*** 1.919*** 2.073***
(0.124) (0.163) (0.254) (0.143) (0.187) (0.138) (0.027) (0.589) (0.245)
Mental disorders -0.497*** 0.141** -0.044 -0.415*** -0.058 1.172*** -0.156*** 7.375*** -0.071
(0.107) (0.065) (0.142) (0.082) (0.334) (0.090) (0.028) (0.866) (0.247)
Married -0.106 -0.456*** -0.130 0.093 -0.182 -0.080 -0.025 0.483 -0.205
(0.096) (0.061) (0.182) (0.076) (0.174) (0.076) (0.029) (0.452) (0.223)
Primary Source of Referral
Individual -0.264** 0.206*** 1.567*** -2.641*** 0.525** 1.190*** 0.170*** -0.189 0.640***
(0.110) (0.070) (0.203) (0.136) (0.216) (0.083) (0.038) (0.446) (0.212)
Alcohol/drug treatment provider 1.274*** 1.475*** 2.606*** -1.491*** 1.415*** 1.561*** 1.846*** ne 1.406***
(0.119) (0.080) (0.199) (0.091) (0.307) (0.099) (0.032)   (0.239)
Other health care provider -0.284* 0.447*** 1.857*** -4.360*** 0.764** 1.085*** 0.319*** -3.462*** 0.800**
(0.169) (0.089) (0.246) (0.453) (0.311) (0.099) (0.047) (0.873) (0.328)
School, employer, community -1.477*** -0.464*** 0.824** -3.844*** -0.944*** 0.316** 0.083* 3.594*** -0.389
(0.196) (0.147) (0.337) (0.344) (0.273) (0.126) (0.045) (0.504) (0.348)
Season Entering Treatment
Summer 0.090 0.206*** 0.034 -0.257*** -1.486*** 0.088 0.036 0.524 0.287
(0.105) (0.075) (0.177) (0.084) (0.209) (0.082) (0.031) (0.527) (0.223)
Fall -0.056 0.067 -0.130 -0.216*** -1.846*** 0.007 -0.032 -0.285 0.291
(0.109) (0.078) (0.180) (0.083) (0.208) (0.089) (0.031) (0.496) (0.233)
Winter 0.037 0.247*** -0.167 -0.150* -0.247 0.043 0.000 0.278 0.067
(0.107) (0.068) (0.176) (0.082) (0.222) (0.082) (0.031) (0.513) (0.223)
Intercept -3.642*** -2.918*** -3.171*** -3.972*** -1.370 -1.909*** -2.187*** -7.936*** -3.911***
(0.460) (0.301) (0.779) (0.399) (0.924) (0.421) (0.148) (1.958) (1.257)
Pseudo R2 0.2446 0.2528 0.4650 0.2451 0.3913 0.3290 0.3340 0.7900 0.4082
ne = not estimable (see text).
* Significant at 0.10 level. ** Significant at 0.05 level. *** Significant at 0.01 level.
Note: Standard errors in parentheses.
Source: SAMHSA, Office of Applied Studies, Treatment Episode Data Set, 1996.

The presence of a co-occurring mental disorder increased the odds of inpatient admission in only three States (Iowa, New Jersey, and North Dakota). Contrary to expectations based on clinical criteria and previous studies of insured individuals, it significantly decreased the odds in Colorado, Massachusetts, and New York and it was insignificant in the other three States. Homelessness, on the other hand, significantly increased the odds of inpatient admission in eight of the nine States examined, and the effect was generally large.

Several socioeconomic variables were significantly associated with the probability of inpatient treatment, and, again, the results varied across States. Age was significant in four of the nine States, in all cases in a nonlinear form. Employed clients were significantly less likely to be admitted to inpatient treatment in seven of the nine States (Colorado, Iowa, Maine, Nevada, New Jersey, New York, and Rhode Island). The education variables were significant in six of the nine States but revealed inconsistent effects. Race/ethnicity variables were significant in five of the nine States, and in four States (Iowa, Maine, Nevada, and New York), Hispanic clients were less likely to have been admitted to inpatient treatment, other things equal. Marital status was significant in only one State, and season of admission was significant in three States but with no discernible pattern.

In all States except Massachusetts and North Dakota, individuals referred by an alcohol or drug treatment provider had greater odds of entering inpatient treatment than those referred by the criminal justice system. In Massachusetts, being referred by an alcohol or drug treatment provider decreased the odds of entering inpatient treatment. In North Dakota, all those referred by an alcohol or drug treatment provider entered outpatient treatment, which is why we do not report parameter estimates for that cell. In six States (Iowa, Maine, Nevada, New Jersey, New York, Rhode Island), individuals who self-referred into treatment had higher odds of being admitted to inpatient treatment than those referred by the criminal justice system.

Expected Payer Model

Descriptive statistics on the expected payer analysis file variables are displayed at the end of Table 4.1. In most States, the primary expected payer was other government funding/no charge. The proportion of clients in this category, which includes those whose care is funded by the SAPT block grant, ranged from 39 percent in New Jersey to 70 percent in Iowa. The second most frequent expected payer in most States was the individual, ranging from 6.5 percent in Iowa to over 51 percent in Nevada. Taken together, these two categories of expected payer account for 74 percent of the clients in the sample, revealing that this is a very different population than that studied in the research mentioned earlier (Goodman et al., 1992, 1998).

Table 4.3 displays the coefficients and standard errors from our expected payer model. Coefficients and standard errors on the additional variables, which did not change much from those in the baseline model, are available from the lead author upon request. Massachusetts and New York are excluded from Table 4.3 because they did not collect data on expected payer. For Nevada, we do not report estimated parameters for Medicaid and Medicare as expected payer because all of those with Medicaid or Medicare entered outpatient treatment. Similarly, we do not report parameter estimates for other insurance in Rhode Island because all clients with that payer were admitted to outpatient treatment, or unknown insurance in North Dakota, as all clients with that payer were admitted into inpatient treatment. In Nevada and Rhode Island, the number of clients with these payers was fewer than 10, rendering any inference inconclusive at best. In North Dakota, however, the payer for almost 17 percent of the clients was unknown at the time of admission, providing better evidence of an association.

As Table 4.3 reveals, in five of the seven States, individuals who were expected to pay for care themselves had significantly lower odds of entering inpatient treatment than those in the omitted category (i.e., other government funding/no charge). Individuals with private health insurance, Medicare, or Medicaid also in many cases had lower odds of entering inpatient treatment facilities than individuals with other government funding/no charge. A notable exception was North Dakota, in which individuals with private health insurance, Medicare, or Medicaid had much higher odds of entering inpatient treatment and in which all of those with unknown insurance entered outpatient treatment.

Discussion

This study examined the effects of the severity of the alcohol use disorder, as measured by routinely collected administrative data, on the odds of admission to alcohol treatment programs in nine States. In contrast to previous studies, which used claims data from clients with employer-sponsored health insurance, our data include clients who were unemployed, uninsured, and homeless. Our results reveal that having a more severe disorder generally increased the odds of inpatient treatment, but the magnitude, and sometimes the direction, of the estimated effects vary across the nine States considered here.

Individuals with more severe substance use disorders (as measured by more frequent alcohol use, use of cocaine as a secondary substance, and a prior treatment episode), as well as those who were homeless, generally had higher odds of admission to inpatient treatment. Those who were employed had consistently lower odds of inpatient admission. Thus, Gregoire's (2000) finding extends to admission, at least in these nine States. Whether employed persons really have less severe disorders, or choose outpatient treatment due to time constraints or some other reason, is a question for future research. Taken together, our findings suggest that client placement in these State substance abuse treatment systems appear, at least to some extent, to conform with the current thinking on client placement.

Table 4.3 Probability of Seeking Inpatient Alcohol Abuse Treatment, by State: Coefficients and Standard Errors on the Expected Payer Variables
  Colorado Iowa Maine Nevada New Jersey North Dakota Rhode Island
N 4,654 17,591 3,762 1,615 10,414 1,028 1,760
Expected Source of Payment
Self-pay -1.239*** 0.182* -1.369*** -1.076*** -1.263*** -0.273 -1.209***
(0.101) (0.101) (0.384) (0.175) (0.098) (1.488) (0.320)
Private -0.708** -0.153* -1.501*** 0.242 -0.115 2.401*** -3.446***
(0.300) (0.081) (0.370) (0.333) (0.092) (0.849) (0.748)
Medicare -0.007 -0.004 -1.433** ne -0.354** 2.760** ne
(0.417) (0.156) (0.645)   (0.155) (1.095)  
Medicaid -0.964* -0.344*** -0.556*** ne -0.973*** 2.573*** -3.572***
(0.575) (0.109) (0.170)   (0.159) (0.928) (0.549)
Other 2.054*** -0.531 -1.155*** -1.899*** -0.021 4.323*** ne
(0.140) (0.476) (0.198) (0.501) (0.128) (1.274)  
Unknown -0.775* na na ne na ne na
(0.397)            
Pseudo R2 0.3509 0.2542 0.4850 0.4217 0.3496 0.8108 0.5011
na = not applicable (i.e., the State had no clients in that category).
ne = not estimable (see text).
* Significant at 0.10 level. ** Significant at 0.05 level. *** Significant at 0.01 level.
Notes: Standard errors in parentheses.
Source: SAMHSA, Office of Applied Studies, Treatment Episode Data Set, 1996.

However, co-occurring mental disorders did not play a consistent role across even most States, and it was negatively related to the likelihood of inpatient treatment in several. This is surprising given its importance as a clinical indicator for inpatient treatment and its significance in earlier studies (Goodman et al., 1992, 1998). We cannot tell whether we obtain our results because individuals with co-occurring mental disorders are less likely to choose inpatient treatment settings, are less likely to be referred to them, or are refused admission to them. However, of note is that two of the three States for which co-occurring mental disorders increased the odds of inpatient admission, New Jersey and North Dakota, had special programs for those with both mental and substance use disorders (New Jersey Substance Abuse Prevention and Treatment Advisory Task Force, 2001; North Dakota Department of Human Services, 2003). Iowa, the third State for which co-occurring mental disorders significantly increased the odds of inpatient admission, had two characteristics that might have worked together to promote appropriate placement. The first was that it promulgated its own set of client placement guidelines, which, similar to the ASAM-PPCs, consider emotional and behavioral factors in client placement (Chemical Dependency Treatment Programs of Iowa, Iowa Substance Abuse Program Directors Association, & Iowa Department of Public Health, 1991). The second is that it was one of the first public systems in the country to contract with a behavioral managed care company (Division of Criminal and Juvenile Justice Planning, 1998). Taken together, these results suggest that States may need to take steps in lieu of or in addition to promulgating guidelines to promote admission of clients with co-occurring mental disorders to inpatient treatment. Both Colorado and Massachusetts had guidelines at the time, but in both States we find that co-occurring mental disorders decreased the odds of inpatient admission (Gartner & Mee-Lee, 1995; O'Keefe & Fisher, 2001). Further research is needed to determine the effect of specific State policies on client placement.

Another finding meriting further examination is that, in the four States for which it was significant, Hispanic ethnicity reduced the odds of inpatient admission. Again, we cannot discern whether this is because Hispanic individuals chose not to enter inpatient treatment, were not screened as carefully, or because culturally competent inpatient treatment was unavailable. Although the latter might be understandable in Maine, which reported having only 26 Hispanics in treatment in 1996, it would be less understandable in New York and Nevada, which have substantial Hispanic populations.

Finally, of mention is the finding that, in seven of our nine States, referral by an alcohol or drug treatment provider is strongly and positively associated with inpatient treatment. One possible explanation for this finding may be that providers of inpatient detoxification treatment believe that clients should be stepped down gradually to lower levels of care, so they refer their clients to inpatient treatment programs, as found in a small study by Kosanke et al. (2002). Because we cannot link records to create an episode for a given client, we are unable to test this hypothesis empirically. However, it is a plausible explanation for our empirical finding.

In other chapters in this compendium, we extend the research on setting choice. Chapter 5 investigates the choice among types of inpatient treatment (hospital, long-term residential, and short-term residential) and outpatient treatment (standard outpatient and intensive outpatient). Chapter 8 investigates the effect of reporting requirements, one of the ways in which State data systems diverge, on estimates of the effects of client characteristics on treatment setting choice.

References

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (DSM-IV) (4th ed.). Washington, DC: Author.

American Society of Addiction Medicine. (1996). Patient placement criteria for the treatment of substance-related disorders (2nd ed., ASAM PPC-2). Chevy Chase, MD: Author.

Andersen, R., & Newman, J. F. (1973). Societal and individual determinants of medical care utilization in the United States. Milbank Memorial Fund Quarterly: Health and Society, 51(1), 95–124.

Annis, H. M. (1985–1986). Is inpatient rehabilitation of the alcoholic cost effective? Con position. Advances in Alcohol & Substance Abuse, 5, 175–190.

Becker, G. S., & Murphy, K. M. (1988). A theory of rational addiction. Journal of Political Economy, 96, 675–700.

Brito, D. L., & Strain, C. K. (1996). A model of the consumption of alcohol. In L. Green & J. H. Kagel (Eds.), Advances in behavioral economics: Volume 3. Substance use and abuse (pp. 163–182). Norwood, NJ : Ablex Publishing Corporation.

Bureau of Substance Abuse Services. (2001, January). Substance abuse prevention, treatment, and services directory 2001 revised. Boston, MA: Massachusetts Department of Public Health.

Chemical Dependency Treatment Programs of Iowa, Iowa Substance Abuse Program Directors Association, & Iowa Department of Public Health. (1991). Iowa client/placement criteria for the treatment of psychoactive substance use disorder. Des Moines, IA: Authors.

Coffey, R. M., Mark, T., King, E., Harwood, H., McKusick, D., Genuardi, J., Dilonardo, J., & Chalk, M. (2001). National estimates of expenditures for substance abuse treatment, 1997 (SAMHSA Publication No. SMA-01–3511). Rockville, MD: Substance Abuse and Mental Health Services Administration, Center for Substance Abuse Treatment and Center for Mental Health Services.

Comprehensive Alcohol Abuse, Drug Abuse, and Mental Health Amendments, Pub. L. No. 100–690, Title II, ch. 2, § 2052 (1988).

Dayhoff, D. A., Pope, G. C., & Huber, J. H. (1994). State variations in public and private alcoholism treatment at specialty substance abuse treatment facilities. Journal of Studies on Alcohol, 55, 549–560.

Division of Criminal and Juvenile Justice Planning. (1998). Substance abuse treatment. Excerpted from Iowa criminal and juvenile justice plan: 1998 update. Des Moines, IA: Iowa Department of Human Rights.

Donovan, D. M., Rosengren, D. B., Downey, L., Cox, G. B., & Sloan, K. L. (2001). Attrition prevention with individuals awaiting publicly funded drug treatment. Addiction, 96, 1149–1160.

Etheridge, R. M., Craddock, S. G., Hubbard, R. L., & Rounds-Bryant, J. L. (1999). The relationship of counseling and self-help participation to patient outcomes in DATOS. Drug and Alcohol Dependence, 57, 99–112.

Finney, J. W., Hahn, A. C., & Moos, R. H. (1996). The effectiveness of inpatient and outpatient treatment for alcohol abuse: The need to focus on mediators and moderators of setting effects. Addiction, 91, 1773–1796. Discussion in 91, 1803–1820.

Finney, J. W., & Moos, R. H. (1996). The effectiveness of inpatient and outpatient treatment for alcohol abuse: Effect sizes, research design issues, and explanatory mechanisms. Addiction, 91, 1803–1820.

Gartner, L., & Mee-Lee, D. (Consensus Panel Co-Chairs). (1995). The role and current status of patient placement criteria in the treatment of substance use disorders (Treatment Improvement Protocol [TIP] Series 13). Rockville, MD: U.S. Department of Health and Human Services, Public Health Service, Substance Abuse and Mental Health Services Administration, Center for Substance Abuse Treatment.

Gastfriend, D. R., Lu, S. H., & Sharon, E. (2000). Placement matching: Challenges and technical progress. Substance Use & Misuse, 35, 2191–2213.

Gastfriend, D. R., & McLellan, A. T. (1997). Treatment matching: Theoretic basis and practical implications. Medical Clinics of North America, 81, 945–966.

Goodman, A. C., Holder, H. D., Nishiura, E., & Hankin, J. R. (1992). A discrete choice model of alcoholism treatment location. Medical Care, 30, 1097–1110.

Goodman, A. C., Nishiura, E., & Hankin, J. R. (1998). A discrete choice model of drug abuse treatment location. Health Services Research, 33, 125–145.

Grant, B. F., & Dawson, D. A. (1997). Age at onset of alcohol use and its association with DSM-IV alcohol abuse and dependence: Results from the National Longitudinal Alcohol Epidemiologic Survey. Journal of Substance Abuse, 9, 103–110.

Green, L., & Kagel, J. H. (Eds.). (1996). Advances in behavioral economics: Volume 3. Substance use and abuse. Norwood, NJ: Ablex Publishing Corporation.

Gregoire, T. K. (2000). Factors associated with level of care assignment in substance abuse treatment. Journal of Substance Abuse Treatment, 18, 241–248.

Grossman, M. (1972). On the concept of health capital and the demand for health. Journal of Political Economy, 80, 223–255. [Available at http://ideas.repec.org/a/ucp/jpolec/v80y1972i2p223–55.html]

Haas-Wilson, D., Cheadle, A., & Scheffler, R. M. (1989). Demand for mental health services: An episode of treatment approach. Southern Economic Journal, 56, 219–232.

Harrison, P. A., & Asche, S. E. (1999). Comparison of substance abuse treatment outcomes for inpatients and outpatients. Journal of Substance Abuse Treatment, 17, 207–220.

Hartmann, D. J., Sullivan, W. P., & Wolk, J. L. (1993). Inpatient and outpatient outcomes in Missouri's alcohol and drug treatment programs. Journal of Health & Social Policy, 5(2), 67–76.

Jacobsen, K., & McGuire, T. G. (1996). Federal block grants and state spending: The Alcohol, Drug Abuse, and Mental Health block grant and state agency behavior. Journal of Health Politics, Policy and Law, 21, 753–770.

Kosanke, N., Magura, S., Staines, G., Foote, J., & DeLuca, A. (2002). Feasibility of matching alcohol patients to ASAM levels of care. American Journal on Addictions, 11, 124–134.

Mark, T., Coffey, R. M., King, E., Harwood, H., McKusick, D., Genuardi, J., & Das, M. (1999). National spending estimates for mental health, alcohol and other drug abuse treatment, 1987–1997: Draft round-two report (Contract deliverable for Contract No. 270–96–007, Project 1: Task 19). Rockville, MD: U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration.

Mattson, M. E. (2003). The search for a rational basis for treatment selection. In M. Galanter (Ed.), Recent developments in alcoholism: Volume 16. Research on alcoholism treatment (pp. 97–113). New York: Kluwer Academic/Plenum Publishers.

McCarty, D., McGuire, T. G., Harwood, H. J., & Field, T. (1998). Using state information systems for drug abuse services research. American Behavioral Scientist, 41, 1090­1106.

McKay, J. R., Cacciola, J. S., McLellan, A. T., Alterman, A. I., & Wirtz, P. W. (1997). An initial evaluation of the psychosocial dimensions of the American Society of Addiction Medicine criteria for inpatient versus intensive outpatient substance abuse rehabilitation. Journal of Studies on Alcohol, 58, 239–252.

McLellan, A. T., Hagan, T. A., Levine, M., Meyers, K., Gould, F., Bencivengo, M., Durell, J., & Jaffe, J. (1999). Does clinical case management improve outpatient addiction treatment. Drug and Alcohol Dependence, 55(1–2), 91–103.

Miller, W. R., & Hester, R. K. (1986). Inpatient alcoholism treatment: Who benefits? American Psychologist, 41, 794–805.

Montoya, I. D., Atkinson, J. S., & Trevino, R. A. (2000). Economics as a factor in models of behavioral motivation and change. Substance Use & Misuse, 35, 329–346.

Muurinen, J. M. (1982). Demand for health: A generalised Grossman model. Journal of Health Economics, 1(1), 5–28.

National Institute on Alcohol Abuse and Alcoholism. (2000, June). 10th special report to the U.S. Congress on alcohol and health (NIH Publication No. 00–1583). Rockville, MD: U.S. Department of Health and Human Services, National Institutes of Health. [Available at http://www.niaaa.nih.gov/publications/10report/intro.pdf & http://www.niaaa.nih.gov/publications/publications.htm]

National Institute on Alcohol Abuse and Alcoholism. (2001). Alcoholism: Getting the facts (NIH Publication No. 96–4153). Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health. [Available at http://www.niaaa.nih.gov/publications/booklet.htm]

New Jersey Substance Abuse Prevention and Treatment Advisory Task Force. (2001). Improving substance abuse treatment in New Jersey. Trenton, NJ: New Jersey Department of Health and Senior Services. [Available as a PDF download at http://www.ncaddnj.org/PublicAffairs/taskforce.htm]

North Dakota Department of Human Services. (2003). Home page of the North Dakota Department of Human Services. Retrieved October 16, 2003, from http://lnotes.state.nd.us/dhs/dhsweb.nsf

Office of Applied Studies. (1999). Treatment Episode Data Set (TEDS): 1992–1997: National admissions to substance abuse treatment services (DHHS Publication No. SMA 99–3324, Drug and Alcohol Services Information System Series S-7). Rockville, MD: Substance Abuse and Mental Health Services Administration. [Available at http://www.oas.samhsa.gov/dasis.htm#teds2]

Office of Applied Studies. (2000). Uniform Facility Data Set (UFDS): 1998 data on substance abuse treatment facilities (DHHS Publication No. SMA 00–3463, Drug and Alcohol Services Information System Series S-10). Rockville, MD: Substance Abuse and Mental Health Services Administration. [Available at http://www.oas.samhsa.gov/dasis.htm#nssats2]

Office of Applied Studies. (in press). Alcohol and Drug Services Study (ADSS): Phase II: Client record abstract report. Rockville, MD: Substance Abuse and Mental Health Services Administration. [To be available as a PDF at http://www.oas.samhsa.gov/adss.htm]

O'Keefe, M. L., & Fisher, E. N. (2001). Overview of substance abuse treatment services: Fiscal year 2001. Colorado Springs, CO: Colorado Department of Corrections. [Available as a PDF at http://www.doc.state.co.us/AlcoholDrug/pdfs/annual01.pdf]

Pettinati, H. M., Meyers, K., Evans, B. D., Ruetsch, C. R., Kaplan, F. N., Jensen, J. M., & Hadley, T. R. (1999). Inpatient alcohol treatment in a private healthcare setting: Which patients benefit and at what cost? American Journal on Addictions, 8, 220–233.

Pettinati, H. M., Meyers, K., Jensen, J. M., Kaplan, F., & Evans, B. D. (1993). Inpatient vs outpatient treatment for substance dependence revisited. Psychiatric Quarterly, 64, 173–182.

Pottick, K., Hansell, S., Gutterman, E., & White, H. R. (1995). Factors associated with inpatient and outpatient treatment for children and adolescents with serious mental illness. Journal of the American Academy of Child and Adolescent Psychiatry, 34, 425–433.

Simpson, D. D., Joe, G. W., Fletcher, B. W., Hubbard, R. L., & Anglin, M. D. (1999). A national evaluation of treatment outcomes for cocaine dependence. Archives of General Psychiatry, 56, 507–514.

Subcommittee on Health Services Research, National Advisory Council on Alcohol Abuse and Alcoholism. (1997). Improving the delivery of alcohol treatment and prevention services: A national plan for alcohol health services research: Executive summary (NIH Publication No. 4224). Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism. [Available at http://www.niaaa.nih.gov/publications/exsum.htm]

Suranovic, S. M., Goldfarb, R. S., & Leonard, T. C. (1999). An economic theory of cigarette addiction. Journal of Health Economics, 18(1), 1–29.

U.S. General Accounting Office. (1998, March 27). Drug abuse: Research shows treatment is effective, but benefits may be overstated (GAO/HEHS-98–72). Washington, DC: U.S. Government Printing Office. [Available as a PDF at http://www.gao.gov/]

Go to the Table of Contents

This page was last updated on June 03, 2008.