CHAPTER 7.  THE STUDY’S IMPLICATIONS FOR LOUISIANA’S

PUBLICLY FUNDED ALCOHOL AND DRUG ABUSE TREATMENT SYSTEM

7.1    Introduction

      This chapter explores the implications of this study for Louisiana’s publicly funded Alcohol and Drug Abuse Treatment System.  The question of interest is would it be a good investment for the state to make new investments in this system?  Would it get value for its money?  To answer this question, we first have to examine the structure, capacity, and utilization in the current system.  This will be addressed in the following section.  In Section 7.3, we will review the relevant findings from the current study.  The concluding section relates the relevant findings to the current situation to answer the question of interest posed above and to explain the qualifications that must be made.

7.2    Structure, Capacity, and Utilization in Louisiana’s Treatment System

      It should be noted at the outset, this section draws heavily on the recent study by Rachal and Ducharme (1999).  When one is discussing the desirability of expanding treatment services to new clients, two relevant questions come to mind.  Is there a need for such services? 

Is there excess capacity in the existing system?  If so, services can be extended to new clients relatively cheaply.  If not, and new facilities and new staff have to be provided, then the cost of providing such services will be more expensive.

7.2.1        Current Treatment System

      The Office of Alcohol and Drug Abuse (OADA), within the Louisiana Department of Health and Hospitals (DHH) is the single state authority for substance abuse.  The OADA substance abuse service system is a mixture of state-operated and contracted community-based programs.  During State FY 1998, the state directly operated 27 full-time outpatient clinics, 21 outreach/satellite outpatient clinics, seven inpatient programs, one medical detoxification program, and two social detoxification programs.  The state also managed a 140-bed, pre-release adult criminal justice program for adult incarcerated males.  In addition, OADA contracted with private providers to offer 15 full-time outpatient clinics; one adult and one adolescent inpatient program, two adult and one adolescent residential programs; one medical and six social detoxification programs; and one methadone program.  The state also contracted for the delivery of community-based programs which included 12 adult and two adolescent halfway houses; two three-quarter-way houses, and two therapeutic community programs (for a discussion of the nature and types of services offered by these modalities see Rachal and Ducharme, 1999, pp. 15-18).

7.2.2        The Need For Treatment

      A 1996 Louisiana household telephone survey revealed that approximately 292,000 persons (i.e., 9.4% of the state’s household population) were in need of treatment for alcohol or illicit drug abuse or dependence (Rachal and Ducharme 1999, p. 6).  According to the survey, only about 11,000 persons had received treatment from a detoxification, residential, or outpatient treatment program in the preceding 12 months.  Therefore, only 3.8% (= 11,000 ¸ 292.000) of those determined to be in need of treatment actually sought and received treatment.  About 35,000 persons expressed an unmet demand for treatment in the previous year; that is, these individuals felt a need for treatment but received either no treatment or less than they desired.  A distinction must be made between the need and the demand for treatment.  In Louisiana, as in other states, many persons in need of treatment will not demand treatment.  It is important to know how many persons needing treatment chose not to seek treatment and how many persons sought treatment and could not find it.  This raises the issue of the system’s capacity to provide treatments.

7.2.3        System Capacity

      Rachal and Ducharme (1999) develop two measures of capacity – static capacity and annual capacity.  Static capacity is a snapshot measure.  Static capacity refers to the number of treatment slots for each modality that could have been filled at each treatment program on a single day (i.e., June 30, 1997).  For all inpatient services, static capacity refers to a count of the number of beds.  Outpatient capacity varies with the number of patients who can be accommodated in a treatment group and the number of group and individual sessions that can be offered at any given time or over a period of time.  Both the number of sessions and the session capacity are determined by the number of counselors a program has on staff and the patient/counselor ratio.

      Annual capacity refers to the capacity of the system to provide services over the year.  It can be measured by the number of patients who could be treated over the course of one full year.  The flow of clients through the treatment slots depends on a number of factors.  First is the length of stay or retention rates in the treatment regimes.  Some modalities require longer periods of treatment than others.  Second, the intensity of treatments vary across clients, often because of differences in the severity of substance abuse or dependence.  Third, retention and turnover rates vary across clients and programs; clients may or may not complete a treatment episode, and some clients will re-enter the system multiple times.

      Rachal and Ducharme (1999) estimated the average length of stay by treatment modality, and treatment admissions for individual providers and for the treatment system as a whole.  They used this information to estimate the system’s annual capacity.  Total annual capacity for Inpatient Services was calculated as follows:

365 days x number of beds

average length of stay

Outpatient annual capacity is more difficult to estimate because treatment is typically offered in a group setting and because outpatient care is offered at different levels of intensity.  The annual capacity was estimated as equal to the static capacity multiplied by a factor determined by dividing 365 by the average length of stay (ALOS).  Table 7.1 presents the estimated values for the system’s static and annual capacity by treatment modality for 1998.

7.2.4        System Utilization

      Rachal and Ducharme (1999) estimated capacity utilization ratios for the system by developing a ratio of the active static cases and annual admissions by the static and annual capacity estimates, respectively.  The estimated capacity utilization ratios are reported in Table 7.2.  Proportions less than one indicate some slack in the system and proportions greater than one indicate utilization greater than capacity. 

      In the present study, we are primarily interested in the annual capacity utilization ratios.  The overall annual capacity utilization rate is 0.94, which does not suggest a lot of slack in the system.  It appears that services to adolescents could be expanded without creating new facilities, but the absolute number of adolescents that can be treated is quite small.  The Intensive Outpatient, Inpatient Adult, Residential Adult, Halfway House Adult and Three-Fourths Way House are clearly being fully utilized.  All in all, it appears

Table 7.1

Estimates of Static and Annual Capacity of

Louisiana’s Treatment Programs, 1998

Source:  Rachal and Ducharme (1999, p. 28 and p. 30).

Table 7.2

Capacity Utilization Ratios by Treatment Modality, 1998

Source:  Rachal and Ducharme (1999, p. 35 and p. 36).

that any expansion in AOD services offered to new clients will require an expansion in treatment modalities and facilities.  This means the benefits flowing from those facilities will be postponed for a period of time.

      Rachal and Ducharme (1999) did not survey methadone maintenance programs so they could not estimate capacity utilization ratios for this treatment modality.  They rationalized this omission stating, “methadone maintenance represents a fundamentally different treatment approach than drug-free modalities and the characteristics of methadone patients, the prevalence of opiate abuse, and treatment utilization rates have been remarkably stable over time, lessening the need for current estimates of capacity and utilization” (p. 24).  The authors did note that on October 1, 1999, 6.3% of all clients enrolled in treatment in the State of Louisiana (in either public or private programs) were receiving opioid substitution therapy (methadone or levo-alpha-acetyle-methadol [LAAM]).  But the state only funded about 2.5% of those receiving treatments.  To date, the state has made only a limited investment in methadone maintenance treatment (p. 38).

7.3    Client Characteristics

      As shown above, client characteristics can have an important effect on treatment outcomes.  Therefore, we need to compare Louisiana’s treatment population with that used in other studies.  Table 7.3 presents the available information on client’s characteristics at admission to Louisiana’s OADA substance abuse treatment system compared to those at CALDATA, TOPS, and DATOS.  Unfortunately, not all characteristics are reported for all systems or are not reported the same way.

      Louisiana has a higher percent of male clients than the other systems.  Over 75 percent of Louisiana clients are male.  The figure for TOPS clients is not much smaller at

Table 7.3  Client Characteristics At Admission

^aRachal and Ducharme (1999) excludes methadone maintenance clients.

^bGerstein et al. (1984) includes Residential, Social Model, and Outpatients only.

^cHubbard et al. (1989) includes data for residential and outpatient drug free.

^dSimpson et al. (1999) includes cocaine clients only.

72.3 percent, but CALDATA (57.7) and DATOS (65.0) have a much smaller percent of males.  These differences are probably not important because gender has not been shown to have a significant effect on either alcohol or drug abuse treatment outcomes.

      Louisiana has 45.1 percent white clients and CALDATA has 49.5 percent.  These figures are influenced by the relative large black population in Louisiana and the large Hispanic population in California.  The low percentage figure (32.0%) for                         DATOS probably reflects the high use of cocaine among minorities.  In any case, race has not been shown to have a significant effect on treatment outcomes.

      Age is reported differently in the samples, so it is difficult to make inferences.  In general, age does not have an effect on treatment outcomes.  But a few studies reported that outcomes are less favorable for clients age 55 and older.  Louisiana only has 8.3 percent of its clients age 45 or older.  Based on age, there is no reason to believe that treatments in Louisiana’s system would be less effective than elsewhere.

      Only two systems reported education levels for their clients.  The percent of clients with 12 or more years of education was 52.6 percent in Louisiana and 52.7 percent in TOPS.

      Only two systems provided information on marital status.  The percent of married clients was a little lower in Louisiana (15.5%) than in TOPS (18.7%).  Stable marriage is a characteristic that is positively correlated with favorable treatment outcomes.  But the difference in percent married is small, so this should not have much effect on predicted treatment outcomes in Louisiana.

      Only two systems provided employment information, but not in the same way.  Louisiana reported that 59.5 percent of its clients were not employed.  DATOS reported that 47.0 percent of its clients are unemployed.  It is not clear if these figures are comparable.  If they are, treatment outcomes might be less favorable in Louisiana because it has been shown that employment is positively correlated with favorable treatment outcomes.

      Finally, Louisiana reported that 52.0 percent of its clients were multiple drug users, whereas only 27.0 percent were in the DATOS system.  Multiple drug use has been shown to have a negative effect on favorable treatment outcomes.

      Altogether, the marriage, employment status, and multiple drug use characteristics suggest that treatments might be slightly less effective in Louisiana than in other systems.  We say slightly because while these characteristics have been shown to have a statistically significant effect on treatment outcomes, the magnitude of that effect is quite small.

7.4    Interpreting the Study’s Findings and Their Implications for the Louisiana OADA System

      This study has reviewed a massive literature on alcoholism and drug abuse treatments in the United States.  The estimated social costs of drug abuse ($143.4 billion) and  of alcoholism ($184.6 billion) in 1998 are staggering.  Despite a number of methodological problems, a consensus has been reached that both alcoholism and drug abuse treatments are effective at least in the short-term.  Treatments appear to reduce the consumption of alcohol and drugs, the amount of criminal activity, and the level of medical care utilization and spending, especially with respect to inpatient hospital care.  The magnitude of these effects is still under debate.

      The cost-effectiveness of alcoholism treatments has been rarely studied.  There is some evidence indicating that outpatient alcoholism treatment is more cost-effective than inpatient treatment.  There have been no full-blown cost-benefit studies of alcoholism treatments.  There have been some “cost-offset” studies that indicate that such treatments more than pay for themselves in terms of reductions in future medical care costs.  This is important, but it must be remembered that medical care costs represent only approximately 4% of the estimated social cost of drug abuse.  If treatments are going to significantly lower the social cost of alcoholism, they need to reduce morbidity and mortality rates because the lost productivity attributed to these two factors account for over 67 percent of the social costs of alcoholism.  It should be noted that lost productivity is important from the client and society’s point of view; not necessarily from the state government’s point of view.

      Cost-benefit analysis from the state’s viewpoint would focus on the benefits that would yield the state cost-savings.  Of these, medical care cost-savings (to the state) would be important as would cost-savings in the criminal justice system and public assistance programs.  As we shall review below, a few studies considered the simultaneous effects of state alcoholism and drug abuse treatments on these benefits.

      There have been more benefit-cost studies of drug abuse treatments than alcoholism treatments.  The findings from the private sector benefit-cost studies are summarized in Table 7.4 .  The natural experiment studies by McGlothlin and Anglin (1981) and Anglin et al. (1987) only considered the cost-savings associated with legal, arrest, and incarceration costs.  These are well designed studies and provide strong evidence that drug  abuse  treatments  yield  significant  cost-savings  in  these  areas.   The  studies  by

Table 7.4

Summary of Private Sector Cost-Benefit

Studies of Drug Abuse Treatments

Rajkumar and French (1997) and Flynn also indicate that drug abuse treatment yield significant reductions in crime costs.  This finding is very important because crime related costs accounted for 62.0% of the total social cost of drug abuse.  The estimated values of Rajkumar and French tend to be higher than those of Flynn because Rajkumar and French included the intangible cost of crime in their analysis.

      Hubbard et al. (1989) considered three types of crime cost-savings in their analysis:  victim cost, criminal justice cost, and criminal-career productivity costs.  Their estimated value of the B/C ratio is somewhat higher than those of McGlothlin and Anglin (1981), Anglin et al. (1989), and Flynn (1997).  This is in part because they included a greater array of crime costs than the other studies.  It is also true that the Hubbard et al. (1989) study was less well designed than the McGlothlin and Anglin (1981) and Anglin et al. (1989) studies.  Hubbard et al. (1989) had no control group and the pre-and-post treatment periods were short (1 year).  Their estimates of cost-savings and B/C ratios are upwardly biased due to the “ramp-up” and “regression-to-the-mean” effect and the “spontaneous recovery” effect.  The Rajkumar and French (1997) and Flynn (1997) studies suffer from the same problems.  Tabbush’s (1986) estimated values for the B/C ratio are extraordinarily high and we tend to discount them for the reasons provided by French (1995).

      Despite their individual weaknesses, these studies collectively provide strong evidence that drug abuse treatments yield significant crime-related cost-savings to society and to state governments.

      More germane to the present study are the findings of the recent cost-benefit studies of  state substance abuse treatments, which are summarized in Table 7.5.  The Gerstein et

Table 7.5

Summary of Cost-Benefit Studies of State

Substance Abuse Treatments

al. (1994) study considered an array of benefits and potential cost-savings from drug abuse treatments.  This included:  crime, medical, productivity, theft, and government transfers.  The citizen taxpayer B/C ratios are larger than the society’s B/C ratios.  This is because the former includes theft and government transfer cost-savings, which are quite large, and the latter does not.  We believe both sets of estimated B/C ratios are upwardly biased.  Gerstein et al. (1994) did not include a control group and their pre- and post-treatment periods were short (1 year).  Therefore, their estimated treatment effects include “ramp-up” and “regression-to-the-mean” effects and “spontaneous recovery” effects.  The precise magnitude of the bias is unknown, but the “ramp-up” “regression-to-the-mean” effect could be as large as 30 to 50% and the “spontaneous recovery” effect could be 10% (Cartwright 1998) to 30% (Saxe 1983).

      The Mauser et al. (1994) study only considered the effects of drug abuse treatments on criminal justice costs for the total sample and criminal justice, medical, and employment benefits for the small (n=25) sample in the follow up.  This study finds significant criminal justice cost savings in both samples, but the medical and employment effects were negative for the smaller group.  We discount the estimate B/C ratio for the smaller sample because the sample is too small.  This study confirms that drug abuse treatments have a significant negative effect on criminal justice costs.

      This can clearly be seen in the Wickizer and Longhi (1997) study which ignored the effect of alcohol and drug abuse treatments on criminal costs.  They only considered medical, public assistance, and treatment reentry costs in their calculation of the B/C ratio.  Their estimated value B/C = 0.38 is the lowest of any study (private or public) reviewed.  Still, they conclude that alcohol and drug abuse treatments do produce significant positive cost-savings. 

      The most methodologically sound study was conducted by Finigan (1996).  He used a matched control group of early program dropouts and relatively long pre (2 years)-and-post (3 years)-treatment periods.  Also, he used multiple regression analysis to control for the influence of confounding factors.  His citizen taxpayer B/C of 5.60 was based on consideration of criminal justice, public assistance, victim, and theft costs.  Finigan’s government B/C of 1.64 was based only on criminal justice and public assistance cost-savings.  Conspicuously absent is the effect of alcohol and drug abuse treatments on medical care cost-savings reported in so many other studies.

      To fill in this omission, we turned to the study by Estee and Nordlund (2001).  This study also used a good control group of individuals who needed treatment, but did not get it, and multivariate analysis that controlled for the influence of confounding factors.  Based only on medical care cost-savings, Estee and Nordlund estimate the alcohol and drug abuse treatment to have a B/C of 2.19.  We have no way of knowing how much of the medical costs are paid by government.  If we assume 100%, then we can add Estee and Nordlund’s estimated B/C of 2.19 to Finigan’s (1996) estimated B/C of 1.64 to obtain an overall estimate of state alcohol and abuse treatment programs B/C of 3.83.  This indicates that each dollar the state spends on alcohol and drug abuse treatment programs leads to a $3.83 reduction in state spending on criminal justice, medical care, and public assistance programs.  This is our best estimate of the value of these programs to state government.

      Also of interest is the benefit-cost ratio of these programs from society’s point of view because this is important for the optimal allocation of society’s scarce resources.  We cannot simply add Estee and Nordlunds’ (2001) B/C of 2.19 based only n medical care cost savings to Finigan’s (1996) citizen taxpayer B/C of 5.60 ratio.  The latter ratio includes theft and public assistance cost-savings which are not relevant from the society point of view.  Based on other studies (see Table 7.4 and Table 7.5) we know the crime cost savings from alcohol and drug treatment programs are significant and that the benefit-cost ratios based on crime cost-savings generally exceeds the costs of treatment.  We estimate the average value of crime cost-savings based B/C ratios adjusted for bias conservatively to be in the range of 1.5 to 3.0.  If we add these values to Estee and Nordlund’s (2001) B/C of 2.19 which is based only on medical care cost-savings, we estimate the benefit-cost ratio for state alcohol and drug abuse treatments from a society point of view to range from 3.69 to 5.19.  This is an underestimate because it only considers crime and medical cost-savings.  It ignores employment and earnings gains (believed to be small) and the intangible benefits to the client and his family (Quality of Life Effects thought to be large).  Based exclusively on crime and medical care cost-savings, we calculate that society gains in reductions in medical care and crime costs between $3.69 to $5.19 for each dollar spent on alcohol and drug abuse treatments.

References